OT: MOSFET question
Detector195
class-A tube and JFET gain stages both used in front-end circuits, but
never a class-A MOSFETs stage. Yet, according to spending 5 minutes
with LTSpice, a MOSFET seems to make a nicely linear class-A
amplifier, with behavior similar to a tube. (At last we are back on
topic!).
Comments? Is there anything particularly unpleasant about a MOSFET
that prevents this application from being practical?
Patrick Turner
Detector195 wrote:
I have built a pair of class A mosfet amps
which employ transformer coupling to the speaker, see the schematic at
http://www.turneraudio.com.au/htmlwebpgs02/schema5050mosfet.htm
Patrick Turner.
Denis
Considering linearity, very probably nothing can beat a class A
complementary emitter follower made of old good bipolar transistors.
The two-stage Darlington follower can work well with a driver stage
with Ri < 5kOhm. If we'll use the robust output devices in the TO-3
cases like MJ15023/24, we'll get much wider safe operating area and
far better reliability than with any MOSFETs.
Denis Afanassyev
Robert Casey
> Yet, according to spending 5 minutes
>with LTSpice, a MOSFET seems to make a nicely linear class-A
>amplifier, with behavior similar to a tube. (At last we are back on
>topic!).
>
>Comments? Is there anything particularly unpleasant about a MOSFET
>that prevents this application from being practical?
>
>
MOSFETs tend to have a lot of capacitance compared to tubes. When you
run a simulation,
be sure to insert between the signal source and the circuit input a
resistor to simulate the
impedance of a real signal source. Otherwise you won't see any effects
the FET's capacitance
might have in the real world.
Patrick Turner
Denis wrote:
Gees Denis, I thought you'd frozen to death in a russian winter!
Nice to see you back here.
I agree with low voltage high current class A BJTs.
But I have my mosfets running with very little stress,
and nowhere near their limits, with only about
25 watts maximum dissipation.
I am about to change the circuit of mine to incorporate the use of
high Gm j-fets for the front end, which will have the same format of the
3 transistor input driver amp.
The 2SK369 are to be used.
If a BJT output is chosen, then they are much harder to drive since
the base current is quite high, so darlington pair drivers are needed,
and still its much lower impedance than the gates of a mosfet.
There is however much more gain in common emitter mode than with mosfets,
so more NFB can be applied if a circuit like mine is employed.
The amp of mine is set up like a tube amp, with the mosfets in
common source mode, and a balanced 1.6 vrms vrms is all that is needed to drive the
gates for full power.
Quite nice class A mosfet amps can be made using a 3 NPN and 3 PNP devices,
in a complementary series pair arrangement, from about +/- 25v, and
each device with 1 amp of I flow.
The drive voltage is then an SE drive, and a tube could even be used
if suitably buffered from the highish Cinput of the mosfet gates.
My own amp with an OPT has RC coupling between the driver amp and
output mosfets, and there is zero sign of any instability, and like a guitar amp,
it can be driven way into over drive and all that happens is that the output mosfets are biased
off like in a guitar amp, and so nothing bad happens to the mosfets.
Patrick Turner.
Denis
> Denis wrote:
>
> > Patrick Turner <in...@turneraudio.com.au> wrote in message news:<40B4626F...@turneraudio.com.au>...
> > > Detector195 wrote:
> > >
> > > > Oh, I know this is off-topic, but maybe not completely. I have seen
> > > > class-A tube and JFET gain stages both used in front-end circuits, but
> > > > never a class-A MOSFETs stage. Yet, according to spending 5 minutes
> > > > with LTSpice, a MOSFET seems to make a nicely linear class-A
> > > > amplifier, with behavior similar to a tube. (At last we are back on
> > > > topic!).
> > > >
> > > > Comments? Is there anything particularly unpleasant about a MOSFET
> > > > that prevents this application from being practical?
> > >
> > > I have built a pair of class A mosfet amps
> > > which employ transformer coupling to the speaker, see the schematic at
> > >
> > > http://www.turneraudio.com.au/htmlwebpgs02/schema5050mosfet.htm
> > >
> > > Patrick Turner.
> >
> > Considering linearity, very probably nothing can beat a class A
> > complementary emitter follower made of old good bipolar transistors.
> > The two-stage Darlington follower can work well with a driver stage
> > with Ri < 5kOhm. If we'll use the robust output devices in the TO-3
> > cases like MJ15023/24, we'll get much wider safe operating area and
> > far better reliability than with any MOSFETs.
> >
> > Denis Afanassyev
>
> Gees Denis, I thought you'd frozen to death in a russian winter!
> Nice to see you back here.
>
Who is frozen? God save the beer!
> I agree with low voltage high current class A BJTs.
> But I have my mosfets running with very little stress,
> and nowhere near their limits, with only about
> 25 watts maximum dissipation.
>
Right now I'm experimenting with moderate to high voltage BJT follower
with output transformer. The results are very promising.
The transformer with turn ratio of just 2-3 makes the output
transistors operating within few amperes of the collector current
where the current gain is almost absolutely constant.
> I am about to change the circuit of mine to incorporate the use of
> high Gm j-fets for the front end, which will have the same format of the
> 3 transistor input driver amp.
> The 2SK369 are to be used.
>
> If a BJT output is chosen, then they are much harder to drive since
> the base current is quite high, so darlington pair drivers are needed,
> and still its much lower impedance than the gates of a mosfet.
>
Indeed the input impedance of a BJT follower can be considered active
well over 20 kHz. When a BJT operates well within class A limits, the
AC component of base-emitter voltage measures a few millivolts and the
large input capacitance can be ignored. The transfer capacitance of
smaller front-end transistors of a Darlington stage has a little
effect too. What about a full power bandwidth of more than 200 kHz and
negligible distortion without global NFB?
The behavior at clipping is nice too.
> There is however much more gain in common emitter mode than with mosfets,
> so more NFB can be applied if a circuit like mine is employed.
>
> The amp of mine is set up like a tube amp, with the mosfets in
> common source mode, and a balanced 1.6 vrms vrms is all that is needed to drive the
> gates for full power.
>
> Quite nice class A mosfet amps can be made using a 3 NPN and 3 PNP devices,
> in a complementary series pair arrangement, from about +/- 25v, and
> each device with 1 amp of I flow.
> The drive voltage is then an SE drive, and a tube could even be used
> if suitably buffered from the highish Cinput of the mosfet gates.
>
So the buffer stage is indeed needed and we return to some kind of
Darlington.
Another weak side of the MOSFETs is their inferior reliability in
linear applications as compared to industry standard BJTs. The
generally faster thermal response of the solid-state devices makes
them inferior to tubes in respect of handling extreme peaks of power
dissipation. This means that the choice of output transistors in the
audio amplifier is dictated not only by the mean dissipation, but the
most severe situation while driving the reactive loads. The linear
BJTs with high thermal inertia and wide SOA are superior to all other
solid-state devices. I saw a lot of various MOSFET amplifiers with
output devices dead even if sophisticated protection circuits are
employed.
Denis Afanassyev.
Sander deWaal
>So the buffer stage is indeed needed and we return to some kind of
>Darlington.
>Another weak side of the MOSFETs is their inferior reliability in
>linear applications as compared to industry standard BJTs. The
>generally faster thermal response of the solid-state devices makes
>them inferior to tubes in respect of handling extreme peaks of power
>dissipation. This means that the choice of output transistors in the
>audio amplifier is dictated not only by the mean dissipation, but the
>most severe situation while driving the reactive loads. The linear
>BJTs with high thermal inertia and wide SOA are superior to all other
>solid-state devices. I saw a lot of various MOSFET amplifiers with
>output devices dead even if sophisticated protection circuits are
>employed.
Well Denis, I don't quite agree with you.
I have built a hybrid tube/MOSFET amplifier with SK135/SJ50, each 4
pieces, driven by a SRPP consisting of 2 x E288CC in parallel.
No output- or driver transformers.
The F response is ruler flat up to 200 kHz, and limited only by an
anti-slewing network at the input.
Capable of over 50 watts in 8 ohms, it roughly doubles up power up to
1 ohm.
THD is under 0,2 % without NFB, current capability over 20 amps.
Power supply is an 800 VA toroid/channel, and a grand total of 200.000
uF per channel.
Each MOSFET draws an idle current of ca. 750 mA, and they're mounted
on huge heatsinks.
No drift in idle current, no avalanche effects, nothing.
In fact, it runs cooler when playing :-)
This is one of the best amps I've ever built.
The same MOSFETS are used in the Acoustat Transnova 200, and I've seen
many of these on my bench.
Rarely, if ever, one of the output transistors failed, rather the
voltage regulator MOSFETS died.
--
Sander deWaal
Vacuum Audio Consultancy
Denis
Nothing unusual for such deep class A operation. On can suppose this
amp can give its full output in class A below 4 Ohms load. The
described SRPP driver stage has the output impedance lower than 1kOhm.
This can overcome the input capacitance problem to some extent.
Indeed, single MJ15024 or15025 BJT has roughly threefold permissible
power dissipation against 2SK135/SJ50 at the realistic operating
temperatures. And I'm using two pairs of such mighty BJTs in similarly
rated amplifier. This is a realistic limit for safe operation with
hard to drive loudspeakers. In contrary,I'm not quite sure about the
long-term reliability of your hybrid amplifier.
The abovementioned emitter follower accompanied with a pair of
suitable driver BJTs, say MJE15032/33 can be driven by a simple
common cathode gain stage built on a half of any low- or medium-MU
twin triode, and will provide essentially resistive plate load for
this driver stage.
Another drawback of any source (and cathode!) follower is its
excessive local NFB. This results in very harsh distortion spectrum
and smearing of the low-level spectral components of the amplified
signal. These effects especially manifest themselves on complex
waveforms and cannot be fully studied with steady-state sine waves.
The finite current gain of a BJT effectively eliminates this problem.
Moreover, the complementary emitter follower does not necessarily
require full power class A bias for linear operation without global
NFB.
When 2SK135/SJ50 MOSFETS were manufactured, their operating
limitations, especially the importance of thermal timeconstant for
linear audio applications were not clearly understood by their
designers (and this situation seems to be continued now with currently
available Hitachi audio MOSFETs), so there were massive failures.
Actually almost all series manufactured amplifiers with these MOSFETs
have failed. The replacement is difficult because these MOSFETs are
obsolete and cannot be substituted by other ones without major
alteration of the circuit.
Of course, your amplifier provides easier operating environment for
these MOSFETs, but the doubts remain.
Phil Allison
"Denis" <denis_af...@mail.ru>
>
> When 2SK135/SJ50 MOSFETS were manufactured, their operating
> limitations, especially the importance of thermal timeconstant for
> linear audio applications were not clearly understood by their
> designers (and this situation seems to be continued now with currently
> available Hitachi audio MOSFETs), so there were massive failures.
** That is an absolute pile of rot !!!!!!!
> Actually almost all series manufactured amplifiers with these MOSFETs
> have failed.
** Ever bigger lie than the first one.
> The replacement is difficult because these MOSFETs are
> obsolete and cannot be substituted by other ones without major
> alteration of the circuit.
** Giant lie number three !!!
Ever heard of Semelab or Magnatec mosfets ??
> Of course, your amplifier provides easier operating environment for
> these MOSFETs, but the doubts remain.
** I have no idea where this **fool** got this utter drivel from - but
anti Hitachi mosfet *racism* is a new one.
........... Phil
Phil Allison
"Sander deWaal" <nonon...@citroen.demon.ln> wrote in message
>
** That goes double for me !
> I have built a hybrid tube/MOSFET amplifier with SK135/SJ50, each 4
> pieces, driven by a SRPP consisting of 2 x E288CC in parallel.
> No output- or driver transformers.
> The F response is ruler flat up to 200 kHz, and limited only by an
> anti-slewing network at the input.
> Capable of over 50 watts in 8 ohms, it roughly doubles up power up to
> 1 ohm.
> THD is under 0,2 % without NFB, current capability over 20 amps.
> Power supply is an 800 VA toroid/channel, and a grand total of 200.000
> uF per channel.
> Each MOSFET draws an idle current of ca. 750 mA, and they're mounted
> on huge heatsinks.
> No drift in idle current, no avalanche effects, nothing.
> In fact, it runs cooler when playing :-)
>
> This is one of the best amps I've ever built.
>
> The same MOSFETS are used in the Acoustat Transnova 200, and I've seen
> many of these on my bench.
> Rarely, if ever, one of the output transistors failed,
** Hitachi TO3 lateral mosfets ( 2SJ50 / 2SK 135 and the similar 2SJ56 /
2SK176 ) are the most rugged and reliable output devices ever made for audio
amp output. They need only a simple zener diode gate clamp to limit output
current into a short or low Z plus normal fast acting fuses in the DC rails.
Their ability to drive the most severely reactive loads without problem is
legendary. They also need no ballast resistors for current sharing and do
**NOT ** fail if the chip temp becomes excessive for *whatever* reason !!
By comparison, BJT power amps always need a complex VI limiting systems to
allow very low impedance, shorted or highly reactive loads to be driven
without device failure. They also need a thermal shutdown system to protect
against excess chip temperature if the heatsink gets too hot. This is not
to mention the Vbe/Ic matching problem for parallel operation and the need
for emitter ballast resistors for current sharing, the need for driver
transistors and the fact that failure of one device will cause a chain
reaction and produce many other failures in a multi-device output stage.
BJTs have a thermal cycle (fatigue) lifespan as well - while TO3 pack
mosfets do not.
The only genuine plus for BJTs is cost per device - but when a fully
protected output stage is considered there is no cost advantage over using
mosfets.
.......... Phil
Patrick Turner
Phil Allison wrote:
I have to agree with Phil about the Hitachi mosfets being reliable,
but I am not so sure they are the most reliable.
But I have repaired quite a few mosfet amps which failed from one reason or
other,
and to suggest they don't fail like other devices is sheer BS.
>
>
> By comparison, BJT power amps always need a complex VI limiting systems to
> allow very low impedance, shorted or highly reactive loads to be driven
> without device failure. They also need a thermal shutdown system to protect
> against excess chip temperature if the heatsink gets too hot. This is not
> to mention the Vbe/Ic matching problem for parallel operation and the need
> for emitter ballast resistors for current sharing, the need for driver
> transistors and the fact that failure of one device will cause a chain
> reaction and produce many other failures in a multi-device output stage.
>
> BJTs have a thermal cycle (fatigue) lifespan as well - while TO3 pack
> mosfets do not.
>
> The only genuine plus for BJTs is cost per device - but when a fully
> protected output stage is considered there is no cost advantage over using
> mosfets.
>
> .......... Phil
I have to say the vast majority of repairs to SS amps is on those with bjts.
But then the % of amps using mosfets is low.
If the mosfet costs were lower, we'd see more of them in mainstream gear.
Mosfets are the obvious choice of the low volume maker and DIYer,
and experimenter, they are like pentodes which have sucked on
bottles and bottles of steroids.
It shrinks their voltage abilities, but increases their current strength.
A pair of mosfets in class A such as Phil nominates above in a complementary
source follower pair give about 25 relaible class A watts at about 1 ohm output
impedance,
and less than 1% of thd 1 dB below clipping, and at a watt or two, no
otyher NFB is really required.
An 8 ohms load is seen by each device as 16 ohms, and each device has a
voltage gain of around Gm x RL = 0.8 amps/V x 16 = 12.8, so the series voltage
NFB
involved with the follower action means the open loop gain is reduced from
12.8 to just under 1, so that's about 22 dB of NFB.
The problem is then to make sure the driver amp is linear, since the source
follower needs
a voltage slightly greater than the output voltage to the speaker.
And the drive voltage is a single ended one to the mosfet gates.
Some folks have realised this means a tube as a driver would be good,
and Micrex of WA make such an amp but with a low bias class aB mosfet output
stage,
and an EL34 in triode operating off 110v to drive the fet gates, and an opamp
as a differential input pair.
Its a showy smarmy way to build a tube amp without actually doing so.
I have listened, and heard no benefits.
Better would have been to be brave, and run a class A output stage, and be
braver,
run the EL34 with 250v, and a choke load, and use a tube for V1 input.
Less global FB would then be required. But EL34 in triode
has Ra about 1.5k, and easily drives into the 150 pF of the follower
input capacitance of a comp pair class A output stage.
An EL86 in triode has Ra = 1.4 k at 20 mA of idle current.
6EM7 is another twin triode with dissimilar units which would drive
a mosfet output stage well.
With source follower, the high gate to source capacitance is reduced,
due to follwer action.
But if a capacitor load is used at the output, the HF voltage gain is low.
But so what, if a driver circuit has low impedance, and some global FB,
the voltage at the gates can be increased easily, and in phase fast enough
to boost the current flow into a cap load to maintain output signal linearity.
Its easy to build mosfet driver circuits so the relationship between
driver impedance and capacitance is the same as or better than that required
with an
all triode amp.
I'd like to be sociable and agree with Denis that BJTs are better than mosfets,
but my experience tells me my disagreement is valid, but without angst.
Patrick Turner.
Phil Allison
"Patrick Turner" <in...@turneraudio.com.au>
>
> I have to agree with Phil about the Hitachi mosfets being reliable,
> but I am not so sure they are the most reliable.
>
> But I have repaired quite a few mosfet amps which failed from one reason
or
> other, and to suggest they don't fail like other devices is sheer BS.
** The Turneroid Parrot will find no mention of Hitachi mosfets in his
copy of the RDH4 - so he as to consider them as heathens to his fanatical
religion.
If the Turneroid Parrot had even *slight acquaintance* with Hitachi
lateral mosfets he would be aware that when the case and hence chip reaches
a critical temperature (about 165 C chip temp ) the device "throttles back
" and refuses to pass more current than is needed to maintain that chip
temperature.
In other words the devices thermally self protect - a characteristic of
lateral mosfets.
I have seen many commercial power amps using Hitachi TO3 mosfets where
this overtemp condition has existed for hours and hours on end because ( for
example) the mounting bolts in the cases had all worked loose - and
**no** devices failed !!!!
Once the bolts in each TO 3 package were re-tightened the amps worked again
exactly as specified.
Try loosening all the mounting bolts with *any* BJT amp - if you dare
!!!!!
( Better have new ones on hand to replace the whole lot when the chain
reaction starts.)
.............. Phil
Patrick Turner
Phil Allison wrote:
> "Patrick Turner" <in...@turneraudio.com.au>
>
> >
> > I have to agree with Phil about the Hitachi mosfets being reliable,
> > but I am not so sure they are the most reliable.
> >
> > But I have repaired quite a few mosfet amps which failed from one reason
> or
> > other, and to suggest they don't fail like other devices is sheer BS.
>
> ** The Turneroid Parrot will find no mention of Hitachi mosfets in his
> copy of the RDH4 - so he as to consider them as heathens to his fanatical
> religion.
Phil, you really are a stupid twit.
WTF are you on, LSD?
Forget your medication again?
By what absurd reasoning do you come this idiotic above conclusion?
Now do you see why the world thinks you are a joke?
>
> If the Turneroid Parrot had even *slight acquaintance* with Hitachi
> lateral mosfets he would be aware that when the case and hence chip reaches
> a critical temperature (about 165 C chip temp ) the device "throttles back
> " and refuses to pass more current than is needed to maintain that chip
> temperature.
>
> In other words the devices thermally self protect - a characteristic of
> lateral mosfets.
So why did the devices you say are so bullet proof suffer damage, and refuse to
work?
I have even blown up a couple of Hitachi mosfets myself, during my learning
experiences with mosfets. It was dead easy to do that, just an accidently
grounded rail, and the transient stuffed them. A tube is tickled by such
treatment.
All SS circuitry is sudden death, and difficult to work with unless
you know what you are doing. The circuits are always complex, and nearly always
arranged on a board so there is little you can do to follow the circuit unless
you spend all this time with circuit and layout sheets.
When the bean counter driven designers draw up plans for SS gear,
the last consideration is easy serviceing.
Repairing SS amps etc is like cleaning toilets.
>
>
> I have seen many commercial power amps using Hitachi TO3 mosfets where
> this overtemp condition has existed for hours and hours on end because ( for
> example) the mounting bolts in the cases had all worked loose - and
> **no** devices failed !!!!
>
> Once the bolts in each TO 3 package were re-tightened the amps worked again
> exactly as specified.
>
> Try loosening all the mounting bolts with *any* BJT amp - if you dare
> !!!!!
>
Ha, really you make nonsense, why youre'e as bad as Trevor
Wilson trying to say that BJTs will last forever. I tried in vain to convince
him that they
don't, and quit the argument.
Now if you are to say that mosfets last so much better,
then I also refuse to argue further.
Nothing lasts, not even a bloomin mosfet!!!!!
ALL electronics has a lifespan, no matter what it is.
> ( Better have new ones on hand to replace the whole lot when the chain
> reaction starts.)
>
> .............. Phil
I have rebuilt a few mosfet amps, and I am aware of the thermal
character, which means you don't need a VBE section to control the bias if
the devices heat up, but heat isn't always their killer.
I have repaired enough mosfet amps to know that like many of the BJT amps that
land on my repair bench,
they go phut pretty damn easy sometimes.
I have experienced random out of the blue failures where they have simply
fused into a short circuit.
If mosfets were far more reliable than BJTs, methinks the mainstream industry
would have adopted them more fully, that has never happened.
And for regulator pass elements in tube amp supplies, I much prefer
a BU208A compared to any mosfet, after having shorted quite a few.
I still like mosfets, and they sound OK to me when used with
care. I also like lots of fuses everywhere and active protection which
*turns the mains off* when a fault occurs.
In my own class A mosfet amps, there is RC coupling, and OPT,
and fixed bias, and they do throttle back beautifully when the amp
is driven hard into a load lower than recommended.
The OPT stops the DC ruining a speaker should
the devices short to one of the rails.
Unlike Nelson Pass Zen amps where he designs the mosfet SE amp
to idle at about 51 watts, I run mine at no more than an EL34,
ie, 25 watts max.
My friend who tried to DIY a Zen amp quickly went through 3 expensive IRF
mosfets before he just gave up.
People should remember that the actual chip which the current and voltage
happens is
only a few mm square, and its very easy to damage such a square of silicon
material.
Patrick Turner.
Sander deWaal
>Indeed, single MJ15024 or15025 BJT has roughly threefold permissible
>power dissipation against 2SK135/SJ50 at the realistic operating
>temperatures. And I'm using two pairs of such mighty BJTs in similarly
>rated amplifier. This is a realistic limit for safe operation with
>hard to drive loudspeakers. In contrary,I'm not quite sure about the
>long-term reliability of your hybrid amplifier.
In the late '80s and early '90s, the Dutch magazine "Audio & Techniek"
published several class A MOSFET designs with the SJ50/SK135 combo.
Thousands of them were built, of which I have seen many do their duty
every day without complaints up until this day.
Also, the beforementioned Acoustat amplifier stems from the '80s.
The SJ50/SK135 combo is one of the most robust devices I've ever
encountered.
My current amp (pun intended) incubated on a bread bord, and many
mistakes were made when trying to get to the final version.
These MOSFETS can withstand an enormous amount of abuse.
I've never had one single transistor fail on me.
Apart from high gate voltages, I wouldn't know an easy way to destroy
them.
Note, I'm NOT talking about IRF series MOSFETS, as one encounters in
e.g. Arcam amplifiers. They are well known for their firecracker
nature.
Compared to BJT amps, of which I've seen and still see many on my
bench, Hitachi MOSFET amps are the utmost of reliability.
I've repaired many BJT amps like ML, Classe, Krell etc., they all fail
at some point in time, mostly due to heat problems.
We'll probably speak eachother in another 10 years or so, when I'm
still enjoying my amp.............and to think I'm currently driving 2
pair of Maggies in parallel with it :-)
That situation will alter when my KT88 monoblocks will be ready from
tinkering status to final neat project, then I'm going to bi-amp.
BTW while the SK135/SJ50 may be obsolete, they're still available in
some places.
There also are Taiwanese or Chinese versions about, don't make the
mistake of putting them in place of the Hitachis! They're indeed
likely to fail on you while pulling the ON-switch for the first time!
Phil Allison
"Patrick Turner" <in...@turneraudio.com.au>
>
> Phil Allison wrote:
> >
> > ** The Turneroid Parrot will find no mention of Hitachi mosfets in his
> > copy of the RDH4 - so he as to consider them as heathens to his
fanatical
> > religion.
>
> Phil, you really are a stupid twit.
** The Turneroid Parrot is getting cranky - polly want a cracker ??
> > In other words the devices thermally self protect - a characteristic
of
> > lateral mosfets.
>
> So why did the devices you say are so bullet proof suffer damage, and
refuse to
> work?
>
** The Turneroid Parrot is not asking about anything in my post.
>
> Repairing SS amps etc is like cleaning toilets.
>
** Correcting the Turneroid Parrot's errors is like cleaning out an putrid
sewer.
> >
> > I have seen many commercial power amps using Hitachi TO3 mosfets where
> > this overtemp condition has existed for hours and hours on end because
( for
> > example) the mounting bolts in the cases had all worked loose - and
> > **no** devices failed !!!!
> >
> > Once the bolts in each TO 3 package were re-tightened the amps worked
again
> > exactly as specified.
> >
> > Try loosening all the mounting bolts with *any* BJT amp - if you
dare
> > !!!!!
> >
>
> Ha, really you make nonsense, ....
** Snip squawking, naked, parrot droppings.
>
> If mosfets were far more reliable than BJTs, methinks the mainstream
industry
> would have adopted them more fully, that has never happened.
** For 15 years or more high powered amplifiers made with Hitachi mosfets
dominated the market in Australia, NZ the UK and elsewhere. The major
brands were Perreaux (NZ) , ZPE ( NZ), Australian Monitor, H-H ( UK),
Harrisson (UK), ARX (Aust), Jands (Aust), Hertz ( Germany) and a plethora
of smaller players.
Circa 1996 Hitachi withdrew the TO3 pack devices.
However, near identical devices are available from Magnatec of the UK under
the brand Semelab - these are made in Scotland. The numbers are BUZ901 and
BUZ 906 plus other similar numbers, some with dual chips inside giving
double the current and power ratings of the Hitachi ones. Australian Monitor
still make top class mosfet power amplifiers using the above mentioned
Semelab devices - now assembled in India !
In the world of hi-fi amps Haffler of the USA made and sold massive numbers
of power amps made with Hitachi mosfets from the late 1970s onwards. The
world of DIY kit amplifiers was completely dominated in Australia by amps
using Hitachi mosfets until very recently.
How much evidence do you want ?
........... Phil
mike diack
<phila...@tpg.com.au> wrote:
> However, near identical devices are available from Magnatec of the UK under
>the brand Semelab - these are made in Scotland. The numbers are BUZ901 and
>BUZ 906 plus other similar numbers, some with dual chips inside giving
>double the current and power ratings of the Hitachi ones.
Q: Do Semelab brew their own silicon or do they package die purchased
from Hitachi ?
M
Phil Allison
"mike diack" <
"Phil Allison"
>
>
> > However, near identical devices are available from Magnatec of the UK
under
> >the brand Semelab - these are made in Scotland. The numbers are BUZ901
and
> >BUZ 906 plus other similar numbers, some with dual chips inside giving
> >double the current and power ratings of the Hitachi ones.
>
> Q: Do Semelab brew their own silicon or do they package die purchased
> from Hitachi ?
** They definitely brew their own - right in sunny Scotland. The chip
making facility is called Semefab.
It appears they were making lateral power mosfets for a long while prior to
Hitachi's withdrawal of their TO3 types. They were supplying their devices
mainly for industrial and military applications ( SONAR amplifiers ?) until
then.
The folk at Australian Monitor did a world search for replacements for the
Hitachi parts and were *delighted* to find Semelab existed and could supply
them with equivalent devices at decent prices - since it meant they could
stay in business !!!
........... Phil
Patrick Turner
> >
> > If mosfets were far more reliable than BJTs, methinks the mainstream
> industry
> > would have adopted them more fully, that has never happened.
>
> ** For 15 years or more high powered amplifiers made with Hitachi mosfets
> dominated the market in Australia, NZ the UK and elsewhere. The major
> brands were Perreaux (NZ) , ZPE ( NZ), Australian Monitor, H-H ( UK),
> Harrisson (UK), ARX (Aust), Jands (Aust), Hertz ( Germany) and a plethora
> of smaller players.
>
> Circa 1996 Hitachi withdrew the TO3 pack devices.
>
> However, near identical devices are available from Magnatec of the UK under
> the brand Semelab - these are made in Scotland. The numbers are BUZ901 and
> BUZ 906 plus other similar numbers, some with dual chips inside giving
> double the current and power ratings of the Hitachi ones. Australian Monitor
> still make top class mosfet power amplifiers using the above mentioned
> Semelab devices - now assembled in India !
>
> In the world of hi-fi amps Haffler of the USA made and sold massive numbers
> of power amps made with Hitachi mosfets from the late 1970s onwards. The
> world of DIY kit amplifiers was completely dominated in Australia by amps
> using Hitachi mosfets until very recently.
>
> How much evidence do you want ?
>
> ........... Phil
Sure, but the majority of all amps made are bjt based.
If you start talking about specific segments of the market, it may not be the
case that BJTs are the major choice of output devices.
Ben Duncan's 1996 book backs up what I am saying.
Care to write a book to refute this?
Anyway, I quite like mosfets, but I remain wary about their longevity, and
remain fixed in my belief that mosfet amps can and do fail just like their
BJT brothers.
But its rare that a mosfet based amp ever comes into my shed for a repair,
but then the majority of all domestic home hi-fi amps are BJT based.
Halcro use mosfets, and sell quite well despite the absurd prices.
And I once used mosfets and my own much simpler aB circuit to replace all the
guts of
a very poor condition Quad 405 amp, with good sonic results.
I have the task ahead of me to completely gut and rebuild a
Phase Linear amp to my own design this winter.
The owner and I were able to source some special flat pack BJTs at a
small fraction of the mosfet price.
I would have preferred mosfets.
Its for a sub woofer, and it will do the job OK, and be reliable, which is all
that really matters.
Patrick Turner.
Phil Allison
"Patrick Turner" <in...@turneraudio.com.au>
> > Snip all the references to parrots, etc,
** Squawk, squawk squawk ........
** Sure, because BJTs in flat packs cost only cents each in 100,000 off
quantities.
Hitachi TO3 lateral power mosfets were made by just one company -
Hitachi.
> If you start talking about specific segments of the market,
** Yep - the segment where reliability and ruggedness is paramount.
> Ben Duncan's 1996 book backs up what I am saying.
** Yawn.....
>
> Care to write a book to refute this?
>
** There is never any need to refute Ben Duncan - he does that al the
time himself.
> Anyway, I quite like mosfets, but I remain wary about their longevity, and
> remain fixed in my belief that mosfet amps can and do fail just like their
> BJT brothers.
** There is nothing quite as obnoxious as a true bigot.
............. Phil
Denis
> People should remember that the actual chip which the current and voltage
> happens is
> only a few mm square, and its very easy to damage such a square of silicon
> material.
>
>
> Patrick Turner.
Now what plain figures say: TO-3 BJTs like my favourite MJ15024/25
permit chip temperature of 200C. MOSFETS can withstand no more than
150C (some vertical switching types can 175C). The permissible heat
dissipation at 25C is 250W for those MJ and only 100W for 2SK135/SJ50.
At a realistic case temperature of 100C, which is actually a design
value for class A amplifiers, the mentioned BJT can dissipate 140W,
and that MOSFET – only 40W.
Now, which amplifier can we build with a single pair of respective
devices? I mention, of course, a class A design with wide range of
load impedances and a possibility to avoid any kind of protection
circuits. The output power of such amplifier should continuously
increase with reduction of load impedance, and the SOA of the output
devices will permit the continuous short circuit at the output with
full uncompressed audio signal applied. Very much like a good tube
amp.
The answer is simple: 20W at 8Ohm with a pair of MJ15024/25 and just
7W with a pair of 2SK135/SJ50.
Yes, even MJ15025 is roughly a half of Her Majesty EL509!
Denis Afanassyev
Phil Allison
"Denis" <denis_af...@mail.ru>
>
> Now what plain figures say: TO-3 BJTs like my favourite MJ15024/25
> permit chip temperature of 200C.
** 200 C is the absolute max, peak chip temp - never to be exceeded or the
chip will fail by going dead short and taking its associates with it in
spectacular fashion. Any sane design keeps the devices well away from that
catastrophe by a wide margin.
> MOSFETS can withstand no more than 150C (some vertical switching types can
175C).
** Wrong - 150 C is the maximum chip temp that Hitachi guarantee the
specs for. Above that temp device parameters shift so that the device loses
its effective power gain ( yfs to be precise) and goes into self protective
limiting. This is a huge difference in favour of the mosfet compared to a
BJT since no thermal protection is needed and devices can be allowed to run
as close to or even over the design temp limit without fear of amp failure.
> The permissible heat dissipation at 25C is 250W for those MJ and only 100W
for 2SK135/SJ50.
> At a realistic case temperature of 100C, which is actually a design
> value for class A amplifiers, the mentioned BJT can dissipate 140W,
> and that MOSFET - only 40W.
** Wrong.
The MJs have a 0.7 degrees per watt thermal resistance from junction to
case plus another 0.7 for the mica and grease insulation giving a total of
1.4 degrees C per watt from chip to heatsink mounting surface. At 140 watts
dissipation the chip would be at the **unsafe** temp of 200 C while the
heatsink would have to be kept at a chilly 4 degrees C ( 200 - 1.4 x 140)
and that 4 degrees C must exist right under the transistor !!! Anyone
for dry ice ????
In practice, with similar sized heatsinks used, the safe dissipation of
Hitachi TO3 mosfets is very similar to that of the MJs. The best evidence
of this is to be found by examining commercial amplifiers and seeing how
many devices are used for a given power output.
The Perreaux 6000B is a commercial amplifier that uses the Hitachi 2SK125
and 2SJ50 devices - a total of **six per channel** mounted on a fan
cooled heatsink. The measured output power is 600 watts sine wave into a 4
ohm load - per channel. Another model the Perreaux 8000B uses 10 per
channel of the 2SK176 and 2SJ56 devices for 1000 watts into 4 ohms per
channel - this model is regarded widely as the most reliable 1000 wpc amp
ever made.
BJT amps using the MJ15024/25s can barely equal this result using the
same number of devices per channel even with the benefit of elaborate SOA
limiting circuitry and thermal protection systems - and are still far less
reliable in service than the mosfets amps have proved to be.
> Now, which amplifier can we build with a single pair of respective
> devices? I mention, of course, a class A design with wide range of
> load impedances and a possibility to avoid any kind of protection
> circuits. The output power of such amplifier should continuously
> increase with reduction of load impedance,
** With a class A design that will never happen.
> The answer is simple: 20W at 8Ohm with a pair of MJ15024/25 and just
> 7W with a pair of 2SK135/SJ50.
** The answer is simply more of this Denis whacko's lies.
........... Phil
Patrick Turner
Denis wrote:
> Patrick Turner <in...@turneraudio.com.au> wrote in message news:<40B75824...@turneraudio.com.au>...
> > People should remember that the actual chip which the current and voltage
> > happens is
> > only a few mm square, and its very easy to damage such a square of silicon
> > material.
> >
> >
> > Patrick Turner.
>
> Now what plain figures say: TO-3 BJTs like my favourite MJ15024/25
> permit chip temperature of 200C. MOSFETS can withstand no more than
> 150C (some vertical switching types can 175C).
Who in their right mind would design SS to ever get near that these Ts?
I don't know of too many devices surviving happily....
I believe the reliability of SS electronics reduces 10% for every +10degrees C rise in temp,
so that when 100C is reached, the amp is 100% unreliable.
> The permissible heat
> dissipation at 25C is 250W for those MJ and only 100W for 2SK135/SJ50.
> At a realistic case temperature of 100C, which is actually a design
> value for class A amplifiers, the mentioned BJT can dissipate 140W,
> and that MOSFET – only 40W.
I would *never* build any SS amp where you cannot place a finger on the device and hold it there,
so 45C is about it for me.
This means that a 40 watt class A amp has to have at least four TO3 or large flat pack devices,
preferably 6 devices, and mounted on a ginormous heatsink.
The heatsink should be a flat plate of 6mm Al, 300mm long x 150mm wide,
with 30 fins of 40 mm long, spaced 10 mm apart, mounted with fins vertically,
and exposed to good air flow, perhaps a fan, and with the devices
placed at equal distances along the heatsink, and with an air flow on the flat inside of
the heatsink.
Bigger is even better.
>
> Now, which amplifier can we build with a single pair of respective
> devices? I mention, of course, a class A design with wide range of
> load impedances and a possibility to avoid any kind of protection
> circuits. The output power of such amplifier should continuously
> increase with reduction of load impedance, and the SOA of the output
> devices will permit the continuous short circuit at the output with
> full uncompressed audio signal applied. Very much like a good tube
> amp.
> The answer is simple: 20W at 8Ohm with a pair of MJ15024/25 and just
> 7W with a pair of 2SK135/SJ50.
I disagree.
20 watts of Pd is quite OK with a TO3 mosfet, so two will give you
40 watts of Pd, so if max efficiency is 45% for the design for a given load value,
you get 18 watts of class A power.
The Nelson Pass Zen amp uses two mosfets, on eas the gain device,
and the other as a CCS.
There is 17volts across the gain device with 3 amps of idle current, for a total
of 51 watts of Pd.
the power output is around 17 watts from just one mosfet.
I personally disagree with Nelson that such dissipation figures are OK,
and I think 3 gain devices and 3 CCS pass elements should be used.
I also am not in total agreement with Phil Allison who seems to over emphasisze the reliability
of mosfets. I have seen them go phut pretty easy, but
everyone, even PA, knows how capricious and unreliable BJT output devices can be.
Anyone who does not use an OPT with the SS devices takes the risk
of burning woofer voice coils if they do not use protection circuits of some sort.
I like active protection for all direct coupled SS amps, and if there is more than 1 volt DV on the output
for longer than 3 seconds, the amp is automatically turned off at the mains immediately.
And since class A circuits run quite warm, the stress is much hgher than a class AB low bias amp.
There is repeated stress because the amp is asked to go from room temp to hot
everytime its used, which causes fatigue.
>
>
> Yes, even MJ15025 is roughly a half of Her Majesty EL509!
What on earth is so majestic about the EL509?
The old ones I have seen in EAR509 amps give the tubes a reputation for being
rather bad tempered commoners, and certainly not members of royalty.
I don't like using BJT output devices in any amps.
Its with reluctance I have used driver stages using signal BJTs.
My 50 watt class A PP mosfet amps with OPT coupling will soon have their bjt
input stages replaced by 2SK369 j-fets, for much lower noise,
and perhaps better sonic performance.
>
>
> Denis Afanassyev
Patrick Turner.
Chris Morriss
<denis_af...@mail.ru> writes
But even the best bi-polars suffer from secondary breakdown, which means
that under real-world conditions a 250w bi-polar has the same effective
dissipation as a 150W MOSFET.
How on earth do you get a class-A amp to increase power when the load
impedance decreases? The only way to do that is to over-bias it for the
normal load.
Class A has to keep both output devices in conduction so that the sum of
the current in the top device of the totem-pole, and the current in the
bottom device add up to the quiescent current.
If it increases power as the load drops it's VERY likely to be cutting
off one of the devices and entering class-B.
--
Chris Morriss
Patrick Turner
Chris Morriss wrote:
The class A PP amp reverts to class AB when the load value reduces, and hence you
get more po
without touching the bias.
A pure class A amp has to have its load value defined, it might be 8 ohms.
The idle current will permit that conduction is both devices still occurs in
each device right up to clipping.
Consider the PP amp with +/- 20 volt rails with 2.5A of idle current flow.
The input power does not vary at 100 watts.
The maximum class A po is 50 watts in theory, and about 45 watts in practice,
but let us keep the theoretical op in mind to gain the concept of what occurs.
The peak voltage load swing is then 20V, or 14.1 vrms, and the
current swing is 50/14.1 = 3.5arms, or 5 peak A, ie, twice the idle current
can flow in the load at the full voltage swing to either rail.
The load must be 14.1 / 3.5 = 4 ohms, and po = 50 watts.
With 2 ohms connected to the same amp with only 2.5 amps idle current,
the limits of class A are reached when the peak current in the load reaches 5A,
so peak V for class A is 10V, so there is 7vrms in the load, so
class A po = ( 7 x 7 ) / 2 = 25 watts.
But there is still an available voltage swing right up to the 20V rails,
so if we had the full swing of 14.1vrms, as in the case of the 4 ohm load,
we'd get 100 watts of output, which is class AB po.
The power input will have increased a lot as it does with all class AB amps.
With an 8 ohm load, the max voltage load swing is still 14.1vrms, so the po
is 25 watts but the current swing is only 2.5 peak amps max, so the amp
stays in pure class A and cannot operate in AB.
So all loads above the minimun rated load load for maximum pure class A
are subject to class A po, and all loads less than the rated load may produce
class AB.
SE amps are not able to produce the same load swing as RL reduces,
and their po declines either side of the load value
which theoretically produces the max po of 50% of the input power,
which will stay constant for all loads up to clipping.
>
>
> Class A has to keep both output devices in conduction so that the sum of
> the current in the top device of the totem-pole, and the current in the
> bottom device add up to the quiescent current.
>
> If it increases power as the load drops it's VERY likely to be cutting
> off one of the devices and entering class-B.
> --
> Chris Morriss
In my 50 watt mosfet class A amps, the class A po max is actually about 45 watts,
but definately all pure class A into 8 ohms, about 35 watts,
but into 3 ohms, I get 72 watts max of class AB.
The amp has a rail of about +34V feeding a CT with 2.8 amps.
The OPT I use has winding losses, and the rail sags a bit during tests on 3 ohms,
and the imperfections of the real world amp limit what is in available in theory.
Although I am using tube amp topology for my amp, and two NPN mosfets for
each 1/2 of the PP circuit, their operation is as desscribed above in principle.
Its easier to visualise and work out the op conditions for a tpotem pole NPN PNP
pair of devices.
Patrick Turner.
Denis
The second breakdown limit of the SOA does not derate with case
temperature at the same rate as heat dissipation. Also, the industry
standard TO-3 BJTs intended for linear applications are free from
second breakdown up to very high collector voltage even at Tc=25C.
For instance, the SOA of MJ15024/25 at this temperature is limited by
heat dissipation up to Vce=70V. At higher case temperatures this point
moves further towards the higher voltages together with power
dissipation derating. The thermal timeconstant of such transistors is
also considerably large. Therefore, as I already wrote, one such
device can be compared with no less than 3 mentioned MOSFETs. If this
would not be true, we could see the world of heavy duty professional
amplifiers totally dominated by Hitachi MOSFETs, and they surely would
not disappear from the production.
Actuallly we can see 5 or 6 pairs of the abovemetioned BJTs
operating at Vcc = +-90V in an amplifier capable of safely driving a 2
Ohm load. Such amplifier has typical short-circuit current of 30-40A
and each output transistor deals with the pulses of heat dissipation
of about 700W. The heatsink temperature limit determined by the
thermal protection circuit is about 120-130C. When such amplifier
drives a real world professional woofer, such pulses of heat
dissipation occur at virtually and signal cycle, and the typical
duration of such transients is up to 10-20ms. Such major manufacturers
of professional audio equipment as Electro-Voice and Dynacord cover
their products by 3-year parts and labor warranty. They definitely
know what they do.
The mean heat dissipation per transistor in the abovementioned
amplifiers is 60-80W. Of course, these amps run in very low bias class
AB, the idle current is typically 15-25mA per transistor and the SOA
protection circuits are often of trigger type and have a timeconstant
of 100-130ms.
If we want to make a class A amplifier without any protection
circuitry, we should allow much reduced output power for each
transistor . The heatsink temperature also will be much lower, say,
under 70C.
In such conditions a pair of the mentioned BJTs can provide 20-25 W of
output power at 8Ohm load with SIXFOLD reserve of short circuit
current still within the long-term sustaining capability. If we have
appropriate heatsinks and can afford the amount of energy drawn from
the mains, we can bias such amplifier to such idle current that the
full output power will be available in class A down to 2 Ohms! Who
not, if a typical single ended triode amp with such tube as 845 draws
the same power and has no current reserve.
Such heavy class A bias of a SS amp has many advantages. First, it
provides very stable operating temperature for the transistors. If the
heatsinks have enough area, the amplifier will never suffer from
overheat and, of course, no thermal protection is needed. Actually the
heat dissipation in class A reduces with increase in output level.
Second, even small emitted resistors are enough to completely
eliminate the possibility of thermal runaway. Since the voltage drop
on these resistors at idle current is as high as 200-250mV no
thermally compensated bias source is needed. These small resistors
also completely determine the dynamic transconductance. The distortion
becomes negligible without any global NFB. The transition to class AB
also is smooth, without sudden increase in high-order harmonics and
intermodulation products. As the SOA in this case is enough for
sustained operation at shorted output, the amplifier can be reliably
protected just by the fuses in the Vc rails.
Actually I’ve built the 18W/8Ohm and 40W/8Ohm class A BJT amps biased
to 4Ohm full power class A and 40W/8Ohm transformer-coupled
monoblocks, which deliver the full power (130W) in class A down to
2Ohm. The distortion spectrum of such amplifier is very soft,
absolutely like that of a good tube amp and is almost load
independent. At low levels (below few watts) the distortion products
are undistinguishable from noise.
Denis N. Afanassyev
Phil Allison
"Denis" <denis_af...@mail.ru
>
> The second breakdown limit of the SOA does not derate with case
> temperature at the same rate as heat dissipation. Also, the industry
> standard TO-3 BJTs intended for linear applications are free from
> second breakdown up to very high collector voltage even at Tc=25C.
> For instance, the SOA of MJ15024/25 at this temperature is limited by
> heat dissipation up to Vce=70V.
** 70 volts is not high at all.
> At higher case temperatures this point moves further towards the higher
voltages together with power
> dissipation derating.
** The SOA does not improve with higher case temps - you lunatic.
> The thermal timeconstant of such transistors is
> also considerably large.
** It is even longer with Hitachi mosfets - since the chip is bigger than
in a BJT.
> Therefore, as I already wrote, one such device can be compared with no
less than 3 mentioned MOSFETs.
** Shame that pile of misquotes and irrelevances does not tally with real
world amplifiers.
> If this would not be true, we could see the world of heavy duty
professional
> amplifiers totally dominated by Hitachi MOSFETs, and they surely would
> not disappear from the production.
** The tens of thousands of very reliable, high powered amps that were made
with Hitachi TO3 mosfets proves that Dennis is wrong. The number of mosfets
used was always the same or less than with a comparable BJT amplifier.
> Actuallly we can see 5 or 6 pairs of the abovemetioned BJTs
> operating at Vcc = +-90V in an amplifier capable of safely driving a 2
> Ohm load.
** But only a resistive one.
> Such amplifier has typical short-circuit current of 30-40A
** Dennis is just making numbers up - like he did with the phony
dissipation figures.
> and each output transistor deals with the pulses of heat dissipation
> of about 700W.
* At this power level a BJT chip will explode.
> The heatsink temperature limit determined by the
> thermal protection circuit is about 120-130C.
** Heatsinks in all BJT amps are limited to 75 - 80 degrees.C.
> When such amplifier
> drives a real world professional woofer, such pulses of heat
> dissipation occur at virtually and signal cycle, and the typical
> duration of such transients is up to 10-20ms.
** More fairy stories.
> Such major manufacturers of professional audio equipment as Electro-Voice
and Dynacord cover
> their products by 3-year parts and labor warranty.
** The first is a speaker maker and the latter has made many Hitachi TO3
mosfet amplifiers.
> They definitely know what they do.
>
** Yep - they used Hitachi TO3 mosfets.
.............. Phil
Patrick Turner
Denis wrote:
I guess one reason why I'll probably never use BJT outputs is because
I get so many amps brought to me which suffered from
a short circuit of the output transistors in one channel
due to shorted speaker leads, which went unnoticed by the owner
since he didn't have the loud enough to notice soon enough that one channel sounded crook,
or not working.
Or the owner connected extra speakers in parallel.
Not enough current to trigger the protection circuits, yet "POOF",
so much for "reliable" BJTs.
I am sick of the sight of bloomin BJT outputs.
I much prefer mosfets, and the Hitachi types I have used coped well
to make 120 watts into 8 ohms from 50v rails, with just one pair.
I don't see why one BJT could so easily always perform the job of 3 similar sized mosfets.
Maybe it depends what one is comparing.
There is no need for "current reserve" in any amp;
it simply has to be able to produce the voltage at the load used,
and its not at all necessary for good sonics for the amp to be able to
maintain say 4+ times the max current into the rated value of RL for the amp.
If the 8 ohm speaker has a dip in its impedance to 2 ohms, then it isn't really
an 8 ohms speaker, is it?
If the load Z does drop, its best it does so at some HF, like some ESL do.
This way the small voltages at HF do not cause huge currents to flow.
The huge amount of NFB will keep the voltage level linear with input signal,
and the designer has made his speakers to react so that an equal
level of accoustic energy is released when a flat source voltage is applied,
and the increasing power into the speakers where the Z drops is not his concern.
The owner's concern is to have enough power.
I have not noticed that any huge "current reserve" makes the music sound any better.
Of course with a tube amp one can parallel the two sec windings for 2 ohms on the OPT
instead of seriesing them for 8 ohms, thus allowing twice the current but only half the voltage
to achieve the same po into 2 ohms as 8 ohms without any increase in THD.
The SS amp has have its rails reduced, and perhaps device numbers increased to
achieve a similar thing, or be designed to cope with 2 ohms anyway.
Patrick Turner.
Phil Allison
"Patrick Turner" <in...@turneraudio.com.au>
> I am sick of the sight of bloomin BJT outputs.
>
> I much prefer mosfets, and the Hitachi types I have used coped well
> to make 120 watts into 8 ohms from 50v rails, with just one pair.
** There are any number of commercial amplifiers that do exactly that with
a single pair of Hitachi 2SK135 and 2SJ50s.
EG, the Electrovoice EV100M "Entertainer" 8 ch powered mixing desk - the
most popular such unit ever made - uses just such an output stage for each
of its two channels. It delivers 120 watts sine wave into 8 ohms and 140
watts into 4 ohms. In over 20 years of servicing the things I have yet to
see one dead Hitachi TO3 mosfet.
> I don't see why one BJT could so easily always perform the job of 3
similar sized mosfets.
> Maybe it depends what one is comparing.
** Denis is not interested in objective facts - he is peddling some
personal barrow for his own obscure reasons.
.......... Phil
Patrick Turner
Phil Allison wrote:
> "Patrick Turner" <in...@turneraudio.com.au>
>
> > I am sick of the sight of bloomin BJT outputs.
> >
> > I much prefer mosfets, and the Hitachi types I have used coped well
> > to make 120 watts into 8 ohms from 50v rails, with just one pair.
>
> ** There are any number of commercial amplifiers that do exactly that with
> a single pair of Hitachi 2SK135 and 2SJ50s.
>
> EG, the Electrovoice EV100M "Entertainer" 8 ch powered mixing desk - the
> most popular such unit ever made - uses just such an output stage for each
> of its two channels. It delivers 120 watts sine wave into 8 ohms and 140
> watts into 4 ohms. In over 20 years of servicing the things I have yet to
> see one dead Hitachi TO3 mosfet.
Yes Phil, I realise the gist of all this, and a colleague steered me towards
mosfets
when I began my interest in building solid state amps.
I built the Dave Tilbrook ETI design, with BF469 and BF470
in cascaded differential driver stages.
I found that they were fragile unless very carefully put together.
One mistake, and its sudden death for several devices, and I managed to blow
a couple of mosfets and driver transistors. But 0.001% thd was easy.
But its just as easy to blow BJT outputs and drivers.
So I scrapped the tilbrook design,
and I developed my own simpler circuit, and
I made a 300 watt /channel amp which I still have,
which uses six flat pack mosfets.
I don't use it much.
I have had to service one guitar amp which had a blown mosfet,
but which also had numerous other faults and sealed modules for the eq
section, so the client wrote it off, and I think he'd had trouble with it
before.
Another guy had me do some rewiring of his kit amp,
since he hadn't paid enough attention to earth paths, and the amp oscillated
badly
until a couple of the mosfets developed faults and it failed.
But he hasn't been back for 5 years, so it must still be OK.
But in 10 years, no other mosfet amps ahve arrived here with blown outputs.
>
> > I don't see why one BJT could so easily always perform the job of 3
> similar sized mosfets.
> > Maybe it depends what one is comparing.
>
> ** Denis is not interested in objective facts - he is peddling some
> personal barrow for his own obscure reasons.
>
> .......... Phil
I have known Denis for some years. He is not easily intimidated,
and patience is required to just get him to consider an alternative view.
I can't think when he was anything else except polite to
all of us here. But he, like everyone on the web, should be taken with a grain
of salt.
I am sure if there are any errors in what he is saying, then those who can
understand them will realise who is right
or wrong. Those who cannot understand about SOAs et all won't have any use for
any erroneous information.
What is being discussed here will go right over most folks heads.
This is a tube use newsgroup, and from the discussions
there are not many here who have designed and built diy SS amps.
Maybe there are more closit SS amp builders here than I think,
but so what, it matters not.
So any talk about solid state is OT.
I don't mind, since there is so little on the tube design issues at present.
I suspect if mosfets hadn't been as good as they are, then they wouldn't
have ever become as popular the way they did against the established
habit of using BJTs.
What do Crown and Crest use?
Halcro use mosfets in their state of the art amps,
and get a claimed 0.0001% thd at 200 watts.
I don't know what the op conditions are though.
I have had to repair a few Sugden class A and class AB amps
which use BJTs.
But their faults were mainly with blown output capacitors
and dry joints on the rather poor pcbs.
Why they run the fins on their heatsinks horizontally to impede the flow
of air is beyond me. Serviceing them includes the BS of having gain access by
stripping the
darn thing apart, rather than being able to simply remove a bottom cover to
access tracks.
I have had to repair Soundcraftsmen pro amps and Denons,
they all too easily blow their BJTs, and the protect circuits left something to
be desired.
When I saw the high brow amps, it inspired me to
make my own, and mosfets seemed logical, and since I have no bean counters
working here,
I have had no problems with marginally sized heatsinks, or hard to service
boards.
Patrick Turner.
Sander deWaal
>If the 8 ohm speaker has a dip in its impedance to 2 ohms, then it isn't really
>an 8 ohms speaker, is it?
Umm.......I think you know that NO real world speaker is just 8 ohms
over the entire audio range, Patrick.
Dips to 2 ohms are not that uncommon, especially under dynamic
conditions.
>If the load Z does drop, its best it does so at some HF, like some ESL do.
>This way the small voltages at HF do not cause huge currents to flow.
An ideal amplifier should be able to reproduce the same voltage over
any load value, be it resistive, inductive, capacitive or a
combination.
There has to be a resistive part to dissipate power, however.
>I have not noticed that any huge "current reserve" makes the music sound any better.
I did, that's why I used 4 P and 4 N MOSFETs in each channel and a
huge power supply reserve, only to produce a mere 50 watts in 8 ohms.
Voltage stays nearly the same in 1 ohm, though.
Patrick Turner
Sander deWaal wrote:
> Patrick Turner <in...@turneraudio.com.au> said:
>
> >If the 8 ohm speaker has a dip in its impedance to 2 ohms, then it isn't really
> >an 8 ohms speaker, is it?
>
> Umm.......I think you know that NO real world speaker is just 8 ohms
> over the entire audio range, Patrick.
> Dips to 2 ohms are not that uncommon, especially under dynamic
> conditions.
I very rarely ever test an 8 ohm speaker and discover dips
to 2 ohms in the impedance. This *is* the dynamic impedance.
I assess the average impedance between 100Hz and 1 kHz, because that's where
most of the power goes.
>
>
> >If the load Z does drop, its best it does so at some HF, like some ESL do.
> >This way the small voltages at HF do not cause huge currents to flow.
>
> An ideal amplifier should be able to reproduce the same voltage over
> any load value, be it resistive, inductive, capacitive or a
> combination.
> There has to be a resistive part to dissipate power, however.
Most speakers are mainly resistive in the main power band.
Ppl say this isn't so, but models of the impedance suggest otherwise,
along with the measurements.
>
> >I have not noticed that any huge "current reserve" makes the music sound any better.
>
> I did, that's why I used 4 P and 4 N MOSFETs in each channel and a
> huge power supply reserve, only to produce a mere 50 watts in 8 ohms.
> Voltage stays nearly the same in 1 ohm, though.
Something is wrong with an 8 or 4 ohm speaker design with a dip to
2 ohms let alone 1 ohm.
I don't need to make sure my amps will drive such low loads as a paper clip,
but most can, because the impedance matching avialable on the OPT
allows the match to say 2 ohms, and a load mismatch is OK down to 1 ohm.
A PP tube amp is thus capable of serious amps.
If the instantaneous current from a pair of KT88 is 2 amps,
and the turns ratio is 15 to1 from 1/2 the primary to the sec in a class AB2 situation,
then you can get 30 amps of output current. Enough? Surely it is.
But why do such weakling SET amps still sound wonderful?
Its because they are operating within their capabilities.
There is simply no need to be able to produce 100 times the current actually
being used to drive a speaker at normal listening levels.
10 times, maybe, but 100 or 1,000 times makes SFA difference to the sound, imho.
Some ppl like the idea of having some "muscle" available in their
system, so they must have Krell or Mark Levinson or a Gryphon before they feel secure
about the sound being reproduced properly.
I know a guy who used two Yamaha 2200 amps to drive his JBL
monitor speakers which had a total of 4 x 15" woofers, and horn loaded tweeters in huge
cabs,
and in a room 12' square. Their sensitivity was 95 dB/w/m.
He used maybe 100 milliwatts of power.
He also went to the gym to preen his muscles, and I am sure he just liked muscly things.
I think a pair of 300 B amps would have sounded great!
I recently repaired a clients SE stereo amp which uses 6M5 outputs,
to make just 2 watts per channel. It sounds quite OK into my 88 dB/w/m speakers
at low levels, sitting close.
Admittedly, they ain't as good as my SE pentode amps good for 38 watts
at much much lower thd than the 6M5 at 2 watts,
but then to make 200 millwatts, the 6M5 is working as hard as my
more powerful amps are when making 3.8 watts, which is deafening to me.
The 6M5 amps will shine better if my client gets some sensitive speakers,
and if he let's me talk him into 6BQ5 tubes instead of 6M5s.
In these considerations, I never think about current ability.
Patrick Turner.
Denis
> Patrick Turner <in...@turneraudio.com.au> said:
>
> >If the 8 ohm speaker has a dip in its impedance to 2 ohms, then it isn't really
> >an 8 ohms speaker, is it?
>
No such thing as "8 Ohm speaker" does exist anyway.
> Umm.......I think you know that NO real world speaker is just 8 ohms
> over the entire audio range, Patrick.
> Dips to 2 ohms are not that uncommon, especially under dynamic
> conditions.
>
> >If the load Z does drop, its best it does so at some HF, like some ESL do.
> >This way the small voltages at HF do not cause huge currents to flow.
>
> An ideal amplifier should be able to reproduce the same voltage over
> any load value, be it resistive, inductive, capacitive or a
> combination.
> There has to be a resistive part to dissipate power, however.
>
Moreover, when the amplifier deals with a compex waveform of a music
signal, the current drawn or pumped into the output of the amplifier
as a responce to the voltage and its derivatives is described by a
solution of some very complex integro-differential equation, the very
definition of the complex impedance loses its applicability. Generally
speaking, the crest factor of the current is much greater than that of
the voltage. All energy released back from the voice coil into the
output of the amplifier is indeed dissipated in its output devices.
The current pulses come in any phase in respect to the voltage, so the
amplifier MUST tolerate 100% reactive load and have substantial
current headroom in ourder to develop its full dynamics without being
blown out. The most conservative current reserve as determined by the
measurements carried out on various loudspeakers is at least THREEFOLD
in respect to the value defined by the rated power at rated load. Note
that the modern Hi-End loudspeakers are often more demanding, and
their modulus of impedance also drops down to 2 OHms and sometimes
even lower.
The vacuum tube is No1 in the ability to withstand such abuse. The
plate has very large thermal timeconstant and can average the
dissipation over the time period measured in minutes. If a tube at the
same time can develop a substantial plate current, it is possible to
built an amazing amplifier.
For instance, EL509 has rated plate dissipation of just 35W, but it is
capable of 1.5A or even higher peak emission current in the power
stage of audio amplifier. During the IHF tone burst test (20ms on,
480ms off) it withstands peak dissipation of 800W for prolonged
period. A pair of these amazing tubes can give 40W of class A power
with a possibility to continuously increase at lower loads, and soft
transition of class AB too. The only SS devices able to compete with
such majesty are TO-3 linear BJTs manufactured by ON Semiconductors
(formerly the division of Motorola). The linearity issues are
concerned too.
Chris Morriss
The energy put back into the amplifier from the voice coil of the
speaker is trivial. Why? Because in an 8 Ohm loudspeaker, the dc
resistance of the voice coil is likely to be 6 Ohms. Even if the back
EMF was 30V (which it will never be this high), then the maximum current
will be 5A, well within the capacity of any reasonable amplifier to
absorb.
The reason for the very low impedance dips in POORLY designed multi-unit
loudspeakers is to do with the crossover network, not the drive units.
And EL509 are horribly non-linear valves, so they aren't really good
examples. Try a DA100 or an 845 if you want linearity.
--
Chris Morriss
Denis
> "Denis" <denis_af...@mail.ru
> >
> > The second breakdown limit of the SOA does not derate with case
> > temperature at the same rate as heat dissipation. Also, the industry
> > standard TO-3 BJTs intended for linear applications are free from
> > second breakdown up to very high collector voltage even at Tc=25C.
> > For instance, the SOA of MJ15024/25 at this temperature is limited by
> > heat dissipation up to Vce=70V.
>
>
> ** 70 volts is not high at all.
>
Enough to forget about the second breakdown in these transistors in
all carefully engineered applications. The thermal timeconstant helps
too.
> >
> > The thermal timeconstant of such transistors is
> > also considerably large.
>
>
> ** It is even longer with Hitachi mosfets - since the chip is bigger than
> in a BJT.
>
Doubtful. The increase in thermal timeconstant requires special
measures. The TO-3 BJTs intended for different purposes have this
timeconstant varying within an order of magnitude. Mentioned are the
LINEAR BJTs specially designed for audio amplfication, servo systems
and like purposes. They are considerably different from the general
application and switching transistors.
Now manufactured Hitachi lateral MOSFETS (plastic) have published SOA
for pulse conditions revealing the thermal timeconstant of about 50ms.
This looks good for audio applications, but I've seen them easely
gouing PFFF when such BJT as MJ15024 will sustain until you'll release
it from the torture.
>
>
> ** The tens of thousands of very reliable, high powered amps that were made
> with Hitachi TO3 mosfets proves that Dennis is wrong. The number of mosfets
> used was always the same or less than with a comparable BJT amplifier.
>
Can't agree. There was indeed a period when some manufacturers tried
to beleive in those japanese advertisements.
>
> > Actuallly we can see 5 or 6 pairs of the abovemetioned BJTs
> > operating at Vcc = +-90V in an amplifier capable of safely driving a 2
> > Ohm load.
>
> ** But only a resistive one.
>
Not only, actually such ampifier will tolerate a reactive load with
modolus of impedance of 2Ohm power factor of at least 0.5 at any
audio frequency. This is a professional requirement for driving a
loudspeaker with nominal impedance of 4 Ohm.
>
> > Such amplifier has typical short-circuit current of 30-40A
>
>
> ** Dennis is just making numbers up - like he did with the phony
> dissipation figures.
>
> > and each output transistor deals with the pulses of heat dissipation
> > of about 700W.
>
Look at service manual of, say, Dynacord S1200.
>
> * At this power level a BJT chip will explode.
>
Indeed does not.
>
> > The heatsink temperature limit determined by the
> > thermal protection circuit is about 120-130C.
>
>
> ** Heatsinks in all BJT amps are limited to 75 - 80 degrees.C.
>
Simply wrong. Imagine the heatsink area and/or cooling fan capacity of
a 1kW+ professional amp occupying only 2HU of rack space. Even the
plastic case ON Semiconductor BJTs operate at case tempearatures up to
110C. Again look at the user or service manuals. Hard job, indeeed.
>
>
> > Such major manufacturers of professional audio equipment as Electro-Voice
> and Dynacord cover
> > their products by 3-year parts and labor warranty.
>
>
> ** The first is a speaker maker and the latter has made many Hitachi TO3
> mosfet amplifiers.
>
The first is involved in audio electronics as long as loudspeakers.
Now Electro - Voice and Dynacord are the members of the same group of
companies and produce the wide range of professional amplifiers using
only industry grade BJTs at the output.
The same devices are used in pro amplifiers manufactured by Crown,
JBL, Crest, QSC, to name just major market players.
>
Sander deWaal
>> Dips to 2 ohms are not that uncommon, especially under dynamic
>> conditions.
>I very rarely ever test an 8 ohm speaker and discover dips
>to 2 ohms in the impedance. This *is* the dynamic impedance.
Do you test sweeping with sine bursts?
In that case, the phase difference as caused by the complex load will
reveal impedance dips along the frequency range.
Not all speakers will go down to 2 ohms, however.
It is possible though that an *impedance* below the DCR of the
individual units is measured, due to the phase shifts.
>Something is wrong with an 8 or 4 ohm speaker design with a dip to
>2 ohms let alone 1 ohm.
4 ohms designs are likely to fall below 2 ohms somewhere.
We have many German loudspeaker designs floating around here in
Europe, and they're mostly rated as 4 ohms.
There's a Swiss brand, Piega, with ribbon tweeters that go down to 1
ohm at some point.
They later added a transformer to correct for the load, as some
amplifiers don't like this kind of behaviour.
>If the instantaneous current from a pair of KT88 is 2 amps,
>and the turns ratio is 15 to1 from 1/2 the primary to the sec in a class AB2 situation,
>then you can get 30 amps of output current. Enough? Surely it is.
I don't think the DCR of the average tranny's sec. will allow for 30
amps to be delivered. I measured an Imax of 6...8 amps in 2 ohms on my
PP KT88 amplifiers. That was on the 4 ohms tap, no GNFB.
You must have extremely good transformers!
>But why do such weakling SET amps still sound wonderful?
>Its because they are operating within their capabilities.
That depends on the speaker, though.
Try driving a pair of Maggies, Quads or MLs with a SETA.......
Maybe if you're using 211s or 845s in PSE :-)
SETAs are not my thing, though. Neither are horns.
This opinion must make me an outcast in RATubes <g>
>In these considerations, I never think about current ability.
You're absolutely right that current reserve isn't the only
consideration, but to me it is an important one.
Must be my speakers.........
Sander deWaal
>The reason for the very low impedance dips in POORLY designed multi-unit
>loudspeakers is to do with the crossover network, not the drive units.
But in real-worlds scenarios, we have to work with the speaker as it
is, so we have to take into account the behaviour of the units WITH
the crossover network.
Patrick Turner
Denis wrote:
> Sander deWaal <nonon...@citroen.demon.ln> wrote in message news:<42s3c0turgbp78124...@4ax.com>...
> > Patrick Turner <in...@turneraudio.com.au> said:
> >
> > >If the 8 ohm speaker has a dip in its impedance to 2 ohms, then it isn't really
> > >an 8 ohms speaker, is it?
> >
>
> No such thing as "8 Ohm speaker" does exist anyway.
But 8 ohms speakers do exist. Its simply a conventional way of describing
the average impedance of a speaker, but quite wide Z exists.
>
> > Umm.......I think you know that NO real world speaker is just 8 ohms
> > over the entire audio range, Patrick.
> > Dips to 2 ohms are not that uncommon, especially under dynamic
> > conditions.
> >
> > >If the load Z does drop, its best it does so at some HF, like some ESL do.
> > >This way the small voltages at HF do not cause huge currents to flow.
> >
> > An ideal amplifier should be able to reproduce the same voltage over
> > any load value, be it resistive, inductive, capacitive or a
> > combination.
> > There has to be a resistive part to dissipate power, however.
> >
>
> Moreover, when the amplifier deals with a compex waveform of a music
> signal, the current drawn or pumped into the output of the amplifier
> as a responce to the voltage and its derivatives is described by a
> solution of some very complex integro-differential equation, the very
> definition of the complex impedance loses its applicability.
Huh?
> Generally
> speaking, the crest factor of the current is much greater than that of
> the voltage. All energy released back from the voice coil into the
> output of the amplifier is indeed dissipated in its output devices.
Ah, the old back emf idea again.....
>
> The current pulses come in any phase in respect to the voltage, so the
> amplifier MUST tolerate 100% reactive load and have substantial
> current headroom in ourder to develop its full dynamics without being
> blown out.
But a speaker would be silent if RL was purely reactive.
Silent speakers don't exist.
careful measurements of the currents into nearly all speakers don't show any
surprising current peaks.
> The most conservative current reserve as determined by the
> measurements carried out on various loudspeakers is at least THREEFOLD
> in respect to the value defined by the rated power at rated load. Note
> that the modern Hi-End loudspeakers are often more demanding, and
> their modulus of impedance also drops down to 2 OHms and sometimes
> even lower.
The speakers dip to two ohms in the main power range only if
they have been poorly designed.
Most amps can tolerate a mismatch at some frequency range.
What they won't tolerate is more device dissipation causing
over heating.
So one cannot use enormous power levels on 2 ohms speakers
when the amp is designed for 8 ohms.
>
>
> The vacuum tube is No1 in the ability to withstand such abuse.
Not necessarily. It depends on the design of the amp.
I will grant you that most tube amps give owners a little more time to realize
that the quiet speaker in the room is due to a short circuited speaker lead.
But often owners don't realize anything is wrong, and don't notice the tubes in one channel
are glowing red hot, and only notice when the amp blows a fuse, or a when cloud of smoke
makes all the heads turn at the party.
The transistor amp fails a lot faster with the same fault, and my shed
is weighed down with blown SS amps with fused transistors.
Not many mosfet amps ever come in for a repair.
> The
> plate has very large thermal timeconstant and can average the
> dissipation over the time period measured in minutes. If a tube at the
> same time can develop a substantial plate current, it is possible to
> built an amazing amplifier.
The maximum current from EL509 is limited by the reflected winding resistance
of the OPT.
If the load is a rated 2k a-a, then with 10% winding resistance of 200 ohms,
and with a supply of 300 volts, and with the tube momentarily
shorted, the max current into the primary is 300 / 200,
so 1.5 amps max.
But a tube which is turned on fully, even with class AB2 does have some
finite resistance, so its doubtful if 1.5 amps into the P winding is ever possible.
Its sure ain't desirable!!
Say you have an EL509 pair powering a 2k to 5 ohm OPT.
The TR is 20:1 but in the case of a class AB amp where the turn ratio from an "on" tube
is half the normal TR, you have 10:1, so if the tube is shorted, and 1.5 amp
is the max I, then all you get is 15 amps from the sec, and not a milliamp more.
One exicon mosfet can do a lot better.
>
>
> For instance, EL509 has rated plate dissipation of just 35W, but it is
> capable of 1.5A or even higher peak emission current in the power
> stage of audio amplifier. During the IHF tone burst test (20ms on,
> 480ms off) it withstands peak dissipation of 800W for prolonged
> period.
800 watts equates to 300 volts x 2.66 amps across the tube, and that I have not seen.
What prolonged period do you suggest a tube rated at 35 watts could withstand
800 watts of Pd?
I have never ever seen any example of any tube in any audio amp
using normal octal based tubes in any amp or using nine pin EL509
ever dissipate 800 watts.
> A pair of these amazing tubes can give 40W of class A power
> with a possibility to continuously increase at lower loads, and soft
> transition of class AB too.
That AB transition isn't too "soft". Its full of distortions.
The gain abruptly changes 6 dB lower when one tube sharply cuts off,
and the other takes over alone in the "B" part of the AB cycle.
Oddles of odd order harmonics are produced.
The transfer curve for pentode amps looks as straight as a dog's hind leg,
or a zebra stripe, and never straight.
In triode, they are better, because triodes have lower Ra, and the
cut off is more gradual, resulting in mainly 3H only being made
during the AB transition. The low Ra results in far less gain change due to the load changing
when the AB transition occurs.
The EAR509 amp is very hard sounding, and despite its total of 44 dB of NFB
in 3 loops, it measures no better than many other UL AB amps with only 16 dB of NFB.
But it operates in class B mainly, so 100 watts max is possible, but unreliable.
Sure when set up for 40 watts class A, a pair are a lot better,
but from what I have seen, I much prefer KT88/6550 any day.
I use CFB on the output of my amps using 6550/KT88 etc, and
like Quad II, they measure exceptionally well while in class A, and I rate my
amps to have all class A into 8 ohms, say 35 watts from a pair of KT88,
and then there is still 20 watts of class A into 4 ohms, but a 53 watt AB ability.
99% of listeners never use the AB part of the power output ability.
> The only SS devices able to compete with
> such majesty are TO-3 linear BJTs manufactured by ON Semiconductors
> (formerly the division of Motorola). The linearity issues are
> concerned too.
I disagree strongly.
You are entitled to think those BJTs are the best, but it won't make the world believe you.
Mosfets do it very well thankyou, especially the Exicon types,
rated for a max of 20 amps. And they are so much easier to set up because like tubes,
they are transconductance devices, which turn the input voltage into output current
which does not affect the signal input.
BJTs need a current input, so if hfe is 10, and output current is 4 amps,
then 0.4 amps input to the base is needed, and even a darlington pair
has has a low impedance input compared to a mosfet, and the
variations of output current means you get a continually
varying current drive requirement, which weighs down the collector output
of the voltage gain stage so lots of NFB is needed to
maintain linearity.
The mosfet has changing gain with load, but I think the voltage gain driver stage has an easier
job to drive the mosfet gates.
Halcro achieve 0.0001% thd, even at 20 kHz, at 200 watts into the rated load.
They use mosfets. I suggest you look at the website,
http://www.halcro.com/
They have secret circuitry, but it cannot be anything else except
NFB which gives the good THD figures.
Douglas Self is another expert on SS amplifiers.
Before trying to convince the world of you particular view of electronics,
have a really good long read at
http://www.dself.dsl.pipex.com/ampins/ampins.htm
The short term high current supply ability of tubes is commendable,
but under normal circumstances the ability for high current
is simply irrelevant.
I have heard a few SE 300B amps which provide better sounding music than
many high current BJT amps.
The BJT amps all have high levels of NFB, and very good measured
performance on linearity, like most mosfet amps, but the 300B
or the SE trioded KT88 without any loop FB can still compete on sound quality,
and there is simply no reliance on the short term high current ability.
I know a guy who connected 8 x 300B in parallel for 80 watts,
and it merely went louder, but had no trouble doing so,
and also it had adequate current ability.
8 times more than one 300B.
I doubt the owner of the 80 watter ever used more than 3 watts, so
each tube had only to make 3/8 of a watt, where the linearity
of the 8 watt capable class A triode is very good,
well within specs for what is the technical definition of hi-fi.
I heard an audiophile said, " Ah, Halcro, ...like 300B,
but just go louder..."
Patrick Turner
Patrick Turner
Sander deWaal wrote:
Many speakers use second order LC filters, suitably damped by the
R of the speaker, to give the wanted 12 dB/octave roll off.
Say the crossover filters at 250 Hz results in the
an impedance drop to about 5.6 ohms for an 8 ohm speaker at the Xover F, and
since there
are two such Xovers, one LPF for bass, and one HPF for mids,
you end up with two 5.6 ohm impedances at parallel at 250 Hz.
A similar thing occurs at the MF to HF crossover if LC filters are used.
Patrick Turner.
Phil Allison
"Denis" <
** Also know as the LIAR.
> "Phil Allison" <> > >
> > ** 70 volts is not high at all.
> >
>
> Enough to forget about the second breakdown in these transistors in
> all carefully engineered applications. The thermal timeconstant helps
> too.
** Lie #1.
> > >
> > > The thermal timeconstant of such transistors is
> > > also considerably large.
> >
> >
> > ** It is even longer with Hitachi mosfets - since the chip is bigger
than
> > in a BJT.
> >
>
> Doubtful.
** Lie #2. Go look at one.
> The increase in thermal timeconstant requires special
> measures.
** Lie #3. Bigger chip = bigger time constant.
>
> Now manufactured Hitachi lateral MOSFETS (plastic) have published SOA
> for pulse conditions revealing the thermal timeconstant of about 50ms.
> This looks good for audio applications, but I've seen them easely
> gouing PFFF when such BJT as MJ15024 will sustain until you'll release
> it from the torture.
** Lie #4.
> >
> > ** The tens of thousands of very reliable, high powered amps that were
made
> > with Hitachi TO3 mosfets proves that Dennis is wrong. The number of
mosfets
> > used was always the same or less than with a comparable BJT amplifier.
> >
>
> Can't agree.
** Lie #5. I have posted details of this - Denis has no answer.
> >
> > > Actuallly we can see 5 or 6 pairs of the abovemetioned BJTs
> > > operating at Vcc = +-90V in an amplifier capable of safely driving a 2
> > > Ohm load.
> >
> > ** But only a resistive one.
> >
>
> Not only, actually such ampifier will tolerate a reactive load with
> modolus of impedance of 2Ohm power factor of at least 0.5 at any
> audio frequency.
** Lie #6.
The vast majority of BJT amps VI limit and generate huge spikes when
loaded like this.
See my article in the topic: http://sound.westhost.com/vi.htm
>This is a professional requirement for driving a
> loudspeaker with nominal impedance of 4 Ohm.
>
** Lie #6. Few BJT pro amps can even drive 4 ohm modulus loads with 45
degree phase angles.
> >
> > > Such amplifier has typical short-circuit current of 30-40A
> >
> >
> > ** Dennis is just making numbers up - like he did with the phony
> > dissipation figures.
> >
> > > and each output transistor deals with the pulses of heat dissipation
> > > of about 700W.
> >
>
> Look at service manual of, say, Dynacord S1200.
>
** So some ***stinking Dynacord service manual** is Denis's amp design
bible !!!!
> > * At this power level a BJT chip will explode.
> >
>
> Indeed does not.
** Lie #7. It does.
>
> >
> > > The heatsink temperature limit determined by the
> > > thermal protection circuit is about 120-130C.
> >
> >
> > ** Heatsinks in all BJT amps are limited to 75 - 80 degrees.C.
> >
>
> Simply wrong.
** Lie #8 . That is the rating on the thermal cutouts that are used.
> Imagine the heatsink area and/or cooling fan capacity of
> a 1kW+ professional amp occupying only 2HU of rack space. Even the
> plastic case ON Semiconductor BJTs operate at case tempearatures up to
> 110C.
** Case temp is NOT the same a heatsink temp !!!! What a moron.
> Again look at the user or service manuals. Hard job, indeeed.
** I service pro amps for a living - you cretin.
> >
> > > Such major manufacturers of professional audio equipment as
Electro-Voice
> > and Dynacord cover their products by 3-year parts and labor warranty.
> >
> >
> > ** The first is a speaker maker and the latter has made many Hitachi
TO3
> > mosfet amplifiers.
> >
>
> The first is involved in audio electronics as long as loudspeakers.
** Totally irrelevant to the question.
> Now Electro - Voice and Dynacord are the members of the same group of
> companies and produce the wide range of professional amplifiers using
> only industry grade BJTs at the output.
>
** Denis has snipped my pointing out that EVEN his precious f.....g
Dynacrod have used Hitachi TO3 mosfets.
> The same devices are used in pro amplifiers manufactured by Crown,
> JBL, Crest, QSC, to name just major market players.
>
> >
** All irrelevant to the question of how well Hitachi TO3 mosfets perform.
............. Phil
Phil Allison
"Denis" <
>
> No such thing as "8 Ohm speaker" does exist anyway.
>
** How pedantic = lie #9.
>
> Moreover, when the amplifier deals with a compex waveform of a music
> signal, the current drawn or pumped into the output of the amplifier
> as a responce to the voltage and its derivatives is described by a
> solution of some very complex integro-differential equation, the very
> definition of the complex impedance loses its applicability.
** Pure gobbledegook.
>Generally speaking, the crest factor of the current is much greater than
that of
> the voltage.
** Wrong terminology.
> All energy released back from the voice coil into the
> output of the amplifier is indeed dissipated in its output devices.
** What energy ?? All loudspeakers are very inefficient generators.
> The current pulses come in any phase in respect to the voltage, so the
> amplifier MUST tolerate 100% reactive load and have substantial
> current headroom in ourder to develop its full dynamics without being
> blown out.
** More lies ........
> The most conservative current reserve as determined by the
> measurements carried out on various loudspeakers is at least THREEFOLD
> in respect to the value defined by the rated power at rated load.
** Denis reads crank hi-fi articles and * BELIEVES * them.
Note
> that the modern Hi-End loudspeakers are often more demanding, and
> their modulus of impedance also drops down to 2 OHms and sometimes
> even lower.
** The Quad ESL 57 drops to 1.5 ohms - but at 19 kHz where the power
demand is puny.
> The vacuum tube is No1 in the ability to withstand such abuse.
> The
> plate has very large thermal timeconstant and can average the
> dissipation over the time period measured in minutes. If a tube at the
> same time can develop a substantial plate current, it is possible to
> built an amazing amplifier.
** But not nearly as good in that respect as nearly any mosfet amp made.
( snip verbal diarrhoea - I have had enough )
............ Phil
Patrick Turner
>
> The vast majority of BJT amps VI limit and generate huge spikes when
> loaded like this.
>
> See my article in the topic: http://sound.westhost.com/vi.htm
Your observations and ideas conveyed via the Rod Elliot website are interesting,
( for a change ).
I have to say I have never seen such "reverse current phenomena", perhaps
because I haven't
risked testing some client's BJT amps with an inductor in series with a load R
of the rated value,
and I wasn't to know if the amp had VI limiting.
And I have only once built a kit BJT amp for a subwoofer project,
to feed a Peerless 12" XLS, only equipped with active filtering at the amp
input.
What is the behaviour of ME amps, in your experience?
I will try your test sometime when I get the opportunity.
I find it difficult to see how a positive *voltage* spike at the output
exteninding briefly
to the positive rail voltage can occur when the output is being pulled to a
negative rail by a BJT which is
being turned on, and the output resistance is supposedly low.
Are you saying the applied base voltage is out of phase sufficiently to
cause the phenomena?
Why does the NFB not correct the situation?
I guess when the device turning on and pulling the output voltage negative has
no gain due
to excess emitter current, then the NFB is rendered useless.
None on the mosfet amps I have built use the same sort of VI limiting, and don't
seem to need it.
I do have about 9 volt zener diodes to limit the gate to source voltage on my
300 watt amp
so that with a 3 ohm load or below, power in limited to less than 350 watts.
This is only 10.8 amps rms of output current, and when 1/2 of the 6 output
mosfets are
turned on, their current ability is a lot more than this.
Mosfets are transconductance devices like tubes, and the amplitude of the
voltage applied between
gate and source to generate drain to source current varies with the load.
This is another way of saying the voltage gain varies almost directly with load
value, like a pentode.
Reducing load value means a higher g-s voltage is required for a given gain.
So if a shorted output is the load, even if the g-s voltage is +12 volts, which
is the limit set by
the inbuilt zener diodes in a Hitachi mosfet, then a max current of Gm x 12v is
all that could flow.
But that is about 0.8 x 12 = 9.6 amps peak, which exceeds the rated max.
Having 8 volt zeners would current limit the mosfet to about 7 amps.
I haven't tested many mosfets to destruction to find out everything about them;
it can be expensive to blow them up to watch them die,
and its years since I played with any.
Patrick Turner.
Phil Allison
"Patrick Turner"
> The vast majority of BJT amps VI limit and generate huge spikes when
> > loaded like this.
> >
> > See my article in the topic: http://sound.westhost.com/vi.htm
>
> Your observations and ideas conveyed via the Rod Elliot website are
interesting,
> ( for a change ).
** The Turneroid Parrot only believes things when he sees them on a web
site - or in the RDH4.
Even **my** words take on a whole new significance when so presented.
>
> I have to say I have never seen such "reverse current phenomena", perhaps
> because I haven't risked testing some client's BJT amps with an inductor
in series with a load R
> of the rated value, and I wasn't to know if the amp had VI limiting.
** If it has no VI limiting - there are no spikes possible.
>
> What is the behaviour of ME amps, in your experience?
>
** ME amps are a closely guarded secret - they all return to their
maker for service.
Even TW only does monkey work on them
> I will try your test sometime when I get the opportunity.
** Got to get an amp with VI limiting first.
>
> I find it difficult to see how a positive *voltage* spike at the output
> exteninding briefly to the positive rail voltage can occur when the output
is being pulled to a
> negative rail by a BJT which is being turned on, and the output resistance
is supposedly low.
** That would be difficult to see if it were happening - but it is not.
> Are you saying the applied base voltage is out of phase sufficiently to
> cause the phenomena?
** The VI limiter staves the output stage of drive current at some
particular point in the cycle, then the NFB loop comes into play and causes
the voltage spike with help from the inductive load.
> Why does the NFB not correct the situation?
** The output stage is not responding to drive at that moment - it has
become a current source.
> I guess when the device turning on and pulling the output voltage negative
has
> no gain due to excess emitter current, then the NFB is rendered useless.
>
>
> None on the mosfet amps I have built use the same sort of VI limiting, and
don't
> seem to need it.
** Only a wanker adds a VI limiter to a Hitachi mosfet stage.
............. Phil
Denis
>
> The vast majority of BJT amps VI limit and generate huge spikes when
> loaded like this.
>
> See my article in the topic: http://sound.westhost.com/vi.htm
>
Everything said in this your article about VI limiters is true. But
what will you say about a professional amplifier WITHOUT such damn
thing? Dynacord and EV amps are particular examples. And they have
just peak limiting circuit, which permits about 7A per transistor. The
more sophisticated protection circuit mutes the signal and
disconnects the load when severe overload conditions exist for certain
time. Then after some seconds the protection is released and amplifier
tries to continue operating. The timeconstant of such circuit is
about 130ms. The mentioned amount of current reserve guarantees that
the current clipping is unlikely to occur BEFORE voltage clipping and
back-EMF pulses mentioned in your article never occur.
A professional amplifier of my design uses 8 pairs of MJ15024/25 BJTs
in bridge output stage and produces 1100 Watts at 4Ohm and about 1500
Watts at 2Ohm. This amplifier has simple diode network for peak
current limiting at about 50A and the trigger protection circuit which
totally shuts the amp down when pulse SOA is repeatedly approached for
certain time. The actual parameters of this protection setup were
verified during extensive tests with real loudspeakers and musical
signals. And I can say, all these reserves appeared necessary,
otherwise the protection was triggered from time to time. By the way,
the heatsink temperature threshold for activating the thermal
protection in this amplifier is 110C. The TO-3 transistors are mounted
to the heatsink without dielectric washers and the temperature
difference between the cases and the heatsinks is under 10C. For
several years we had but one such amp returned for service and only
because it was stuffed inside by dirt and dust like a vacuum cleaner
belly. The heatsinks become inefficient and the thermal protection
started triggering. No output transistors were damaged. Needless to
say, this particular amp was extensively abused in a disco every night
for 3 or 4 years.
Otherwise I saw several MOSFET pro amps including some with EXICON
output devices, and I saw how those MOSFETs easely go PFFFF. This, of
course might happen due to wrong design, or the original Hitachi
MOSFETs might be more healthy as they were advertised. Indeed, why
those TO-3 devices have disappeared from the Hitachi production range
if they are so good performers? It is known that successful products
have long life. A good example is 2N3055.
Of course, my class A BJT amps have large external heatsinks, which
never become hotter than 60-70C, and together with the huge SOA
reserve they can live without any protection circuit. Of course, there
are fuses at the proper places.
Phil Allison
"Denis"
> "Phil Allison"
> >
> > The vast majority of BJT amps VI limit and generate huge spikes when
> > loaded like this.
> >
> > See my article in the topic: http://sound.westhost.com/vi.htm
> >
>
> Everything said in this your article about VI limiters is true.
** " Thuuuumpppp !! " ......... sound of Phil falling on floor in
shock.
> But what will you say about a professional amplifier WITHOUT such damn
> thing?
** It is one of two things:
An Hitachi TO3 mosfet amp or an accident looking for somewhere to happen.
> Dynacord and EV amps are particular examples. And they have
> just peak limiting circuit, which permits about 7A per transistor.
** Such an amp is vulnerable to failure with to low Z reactive loads and
near shorts.
> A professional amplifier of my design uses 8 pairs of MJ15024/25 BJTs
> in bridge output stage and produces 1100 Watts at 4Ohm and about 1500
> Watts at 2Ohm.
** That is way fewer devices than is wise or normal commercial practice.
> This amplifier has simple diode network for peak
> current limiting at about 50A and the trigger protection circuit which
> totally shuts the amp down when pulse SOA is repeatedly approached for
> certain time. The actual parameters of this protection setup were
> verified during extensive tests with real loudspeakers and musical
> signals.
** So you say.
> And I can say, all these reserves appeared necessary,
> otherwise the protection was triggered from time to time. By the way,
> the heatsink temperature threshold for activating the thermal
> protection in this amplifier is 110C. The TO-3 transistors are mounted
> to the heatsink without dielectric washers and the temperature
> difference between the cases and the heatsinks is under 10C.
** So you repeat your errors with heatsink design.
> Indeed, why those TO-3 devices have disappeared from the Hitachi
production range
> if they are so good performers?
** Hitachi abandoned all T03 pack devices at about that same time.
Does this prove TO3 is no good ??
By your mad logic it does.
............ Phil
Patrick Turner
Phil Allison wrote:
> "Patrick Turner"
>
> > The vast majority of BJT amps VI limit and generate huge spikes when
> > > loaded like this.
> > >
> > > See my article in the topic: http://sound.westhost.com/vi.htm
> >
> > Your observations and ideas conveyed via the Rod Elliot website are
> interesting,
> > ( for a change ).
>
> ** The Turneroid Parrot only believes things when he sees them on a web
> site - or in the RDH4.
>
> Even **my** words take on a whole new significance when so presented.
Don't let your head swell up and explode like an over driven BJT when I actually
find something you have had posted on the web as being fit to read and non
offensive.
I have many books here, and your page on VI limiting is probably
something which is elsewhere addressed, so like me, regard yourself as just
another
voice in the cacophany of noise which is the flow of information.
I don't think Denis is a liar many times over like you accuse him of being,
anymore than I am really a certain personoid with feathers, or
any more than you ride around the land on a donkey charging at windmills.
A liar is someone who *knows* the truth about something, but deliberately
states a conflicting story, thus decieving the recipient with malice, or
spurious
or mischievous intent.
Perhaps Our dearest Denis has simply got his wires crossed, ie, he is genuinely
unaware of the reasons behind your claims to what are facts as you consider them
to be.
But I have digressed, now back to voltage spikes.....
>
> > I have to say I have never seen such "reverse current phenomena", perhaps
> > because I haven't risked testing some client's BJT amps with an inductor
> in series with a load R
> > of the rated value, and I wasn't to know if the amp had VI limiting.
>
> ** If it has no VI limiting - there are no spikes possible.
So the transistor emitter trying to go negative is switched off by the excess
back current, and the back current produces a voltage spike at the output?
> >
> > What is the behaviour of ME amps, in your experience?
> >
>
> ** ME amps are a closely guarded secret - they all return to their
> maker for service.
I have serviced two ME amps in the last 5 years.
Not everyone tries to return them to the maker for service.
They have a complex lot of stages, requiring that the output stage
has its own loop of FB and the driver stage does too,
and it seems it would be impossible to have just one loop of global FB
because the amp would never be HF stable.
TW loves to say the ME amps have no global FB.
He gets on the audiologically correct bandwagon, well, he thinks he's on it,
but the wagon is a very dilapidated old bomb....
Sure, ME don't have GFB, but they have a lot of stages, and two loops of FB
and the total amount of FB is quite high.
ME became liquidated last year, and went out of business, and
ppl wanting service would simply take them to where is convenient.
>
>
> Even TW only does monkey work on them
>
> > I will try your test sometime when I get the opportunity.
>
> ** Got to get an amp with VI limiting first.
>
> >
> > I find it difficult to see how a positive *voltage* spike at the output
> > exteninding briefly to the positive rail voltage can occur when the output
> is being pulled to a
> > negative rail by a BJT which is being turned on, and the output resistance
> is supposedly low.
>
> ** That would be difficult to see if it were happening - but it is not.
>
> > Are you saying the applied base voltage is out of phase sufficiently to
> > cause the phenomena?
>
> ** The VI limiter staves the output stage of drive current at some
> particular point in the cycle, then the NFB loop comes into play and causes
> the voltage spike with help from the inductive load.
So with no base current, there cannot be collector-emitter current in the normal
direction, so no
gain, but would not the voltage spike be fed back and cause a massive
correctional voltage?
But then since there is a voltage phase difference between the open loop signal
and
output signal, then a spike is possible.
I need to look at things awhile on a CRO to get what's happening.
>
>
> > Why does the NFB not correct the situation?
>
> ** The output stage is not responding to drive at that moment - it has
> become a current source.
>
> > I guess when the device turning on and pulling the output voltage negative
> has
> > no gain due to excess emitter current, then the NFB is rendered useless.
> >
> >
> > None on the mosfet amps I have built use the same sort of VI limiting, and
> don't
> > seem to need it.
>
> ** Only a wanker adds a VI limiter to a Hitachi mosfet stage.
I have a zener voltage limiter to make sure the gates don't cop too much Vg-s.
and to make sure the collector current from the voltage amp stage doesn't
blow the 12v zeners in the mosfets.
There are actually two zeners and two diodes, to clamp the voltage
swing in either direction.
I also have 6 amp diodes from the output to each rail to make sure that back
voltages
from the output never make a current try to flow backwards into the emitter
circuits
and to the rails.
BJTs sure don't like having current forced through them backwards.
In my 300 watt amp, I would not be happy with just the existing 12 volt internal
zener diode
vgs limiters, lest too much power be allowed to be made into low value loads,
with a correspondingly too high amount of dissipation in the devices.
I leant it to a guy for PA use and he couldn't kill it.
The large school hall was filled with excellent sound, we thought.
Just occasionally, it went up to clipping, indicated with an led,
but not heard.
Patrick Turner.
Phil Allison
"Patrick Turner" <in...@turneraudio.com.au
> A liar is someone who *knows* the truth about something, but deliberately
> states a conflicting story, thus decieving the recipient with malice, or
> spurious or mischievous intent.
** Not so. A lie is a falsehood, a liar is a purveyor of falsehoods -
all the worse in a public forum.
When someone shows a reckless disregard for facts and logical reasoning in
spite of compelling, contrary evidence being presented they thoroughly
deserve the epithet.
> Perhaps Our dearest Denis has simply got his wires crossed, ie, he is
genuinely
> unaware of the reasons behind your claims to what are facts as you
consider them
> to be.
** Our Denis is pushing some personal barrow - the content of that messy
barrow is dribbling out presently.
> > >
> > > What is the behaviour of ME amps, in your experience?
> > >
> >
> > ** ME amps are a closely guarded secret - they all return to their
> > maker for service.
>
>
> I have serviced two ME amps in the last 5 years.
** Wow - many as that !!
> Not everyone tries to return them to the maker for service.
** They may have discovered that their maker charges like the proverbial
wounded bull.
> TW loves to say the ME amps have no global FB.
** Bit like saying the earth is flat and we should all fall off if it were
not.
> He gets on the audiologically correct bandwagon, well, he thinks he's on
it,
> but the wagon is a very dilapidated old bomb....
** TW is a parrot - but of a different feather.
He is Peter Stein's well trained parrot.
> ME became liquidated last year, and went out of business, and
> ppl wanting service would simply take them to where is convenient.
** ME refuse to assist outsiders with service - even TW is an outsider.
Magician's must not reveal their tricks to the hoi polloi.
> > ** Only a wanker adds a VI limiter to a Hitachi mosfet stage.
>
> I have a zener voltage limiter to make sure the gates don't cop too much
Vg-s.
> and to make sure the collector current from the voltage amp stage doesn't
> blow the 12v zeners in the mosfets.
** The choice of zener voltage is important if you expect mere fuses to
save TO3 mosfets internal fuses from near shorts under full drive.
>
> I also have 6 amp diodes from the output to each rail to make sure that
back
> voltages from the output never make a current try to flow backwards into
the emitter
> circuits and to the rails.
** Mosfets have built in, structural reverse diodes - your additional
ones serve no purpose.
>
> BJTs sure don't like having current forced through them backwards.
>
** Must be a bunch of rabid homophobes.
> In my 300 watt amp, I would not be happy with just the existing 12 volt
internal
> zener diode vgs limiters, lest too much power be allowed to be made into
low value loads,
> with a correspondingly too high amount of dissipation in the devices.
** Correct.
> I leant it to a guy for PA use and he couldn't kill it.
** PA folk have great difficulty killing mosfet amps - only the most
determined have ever succeeded.
............. Phil
Patrick Turner
Phil Allison wrote:
> "Patrick Turner" <in...@turneraudio.com.au
>
> > A liar is someone who *knows* the truth about something, but deliberately
> > states a conflicting story, thus decieving the recipient with malice, or
> > spurious or mischievous intent.
>
> ** Not so. A lie is a falsehood, a liar is a purveyor of falsehoods -
> all the worse in a public forum.
>
> When someone shows a reckless disregard for facts and logical reasoning in
> spite of compelling, contrary evidence being presented they thoroughly
> deserve the epithet.
But I don't think Denis perhaps knows anything else than what he knows.
He may decide like so many in the public forum to never question his own views,
lest they loose face.
But it was Denis who eventually persuaded me to consider triodes as devices
with inherent
feedback within them.
Before I knew him, I used to think of SET amps as being free of NFB.
Completely free of the trickery of corrective circuitry.
Denis didn't need to call me a liar to get me to see what I now think is the
truth with triodes,
ie, that they have NFB from the anode to the electron stream at all frequencies
due to the anode's electrostaic field.
The grid goes negative, and shuts down the current, the anode goes positive
because less current flows in the load, and the anode's voltage tries to
increase the current.
So the anode voltage and grid voltage do a tug of war to control current.
Then I read Terman's 1937 book which deals with Child's law, and he describes
what happens
in every triode a lot better than elsewhere in the universe, so I became a
believer
in the "self regulation" in a triode, as Terman describes it.
>
> > Perhaps Our dearest Denis has simply got his wires crossed, ie, he is
> genuinely
> > unaware of the reasons behind your claims to what are facts as you
> consider them
> > to be.
>
> ** Our Denis is pushing some personal barrow - the content of that messy
> barrow is dribbling out presently.
He's always focused on what he makes and sells, and to anyone who isn't a man
of the world,
that can be tiring. He'd be better off leaving any mention of his gear out of
it.
But what he makes *is* his frame work for discussion, just as your repair
experience is yours,
and my amp making is mine.
Both he and I must surely know that there is SFA profit extolling the virtues
of what we make,
because nobody from rec.audio.tubes buys our products.
But I get weekly emails from diy ppl wanting explanations of my info at my
site,
and lots of folks have tried the circuits.
I am sharing my knowledge, not hawking it.
>
> > > >
> > > > What is the behaviour of ME amps, in your experience?
> > > >
> > >
> > > ** ME amps are a closely guarded secret - they all return to their
> > > maker for service.
> >
> >
> > I have serviced two ME amps in the last 5 years.
>
> ** Wow - many as that !!
But there are thousands of ME amps around, and not all owners are kind to
them...
>
>
> > Not everyone tries to return them to the maker for service.
>
> ** They may have discovered that their maker charges like the proverbial
> wounded bull.
>
> > TW loves to say the ME amps have no global FB.
>
> ** Bit like saying the earth is flat and we should all fall off if it were
> not.
Well the last time I tried to proove the earth was flat, I walked across
Sydney,
and the hills were quite tiring, and when I walked across the floor of the
Harbour,
I stumbled on bottom of the harbour schemes, and came up a smellin....
The earth ain't flat, or very clean.....
>
> > He gets on the audiologically correct bandwagon, well, he thinks he's on
> it,
> > but the wagon is a very dilapidated old bomb....
>
> ** TW is a parrot - but of a different feather.
>
> He is Peter Stein's well trained parrot.
>
> > ME became liquidated last year, and went out of business, and
> > ppl wanting service would simply take them to where is convenient.
>
> ** ME refuse to assist outsiders with service - even TW is an outsider.
>
> Magician's must not reveal their tricks to the hoi polloi.
There are no incredible secrets within an ME amp. I happen
to think they are fairly well made.
But the clients I mainly get don't like what they do with music.
One tried to trade in a preamp for a tube one.
I joked that I would find the ME pre to be impossible to re-sell
after folks heard a good tube pre.
But another customer loved his ME, all that SS grunt was a real plus.
>
>
> > > ** Only a wanker adds a VI limiter to a Hitachi mosfet stage.
> >
> > I have a zener voltage limiter to make sure the gates don't cop too much
> Vg-s.
> > and to make sure the collector current from the voltage amp stage doesn't
> > blow the 12v zeners in the mosfets.
>
> ** The choice of zener voltage is important if you expect mere fuses to
> save TO3 mosfets internal fuses from near shorts under full drive.
I have 5 amp fuses on my rails, and 3 amps on the output.
That's OK for me.
I never listen loud.
I'd rarely use more than 2 watts.
So I use zener voltages lower than 12v.
>
> >
> > I also have 6 amp diodes from the output to each rail to make sure that
> back
> > voltages from the output never make a current try to flow backwards into
> the emitter
> > circuits and to the rails.
>
> ** Mosfets have built in, structural reverse diodes - your additional
> ones serve no purpose.
But what is their current rating?
Anyway, additional diodes can't do any harm.
>
>
> >
> > BJTs sure don't like having current forced through them backwards.
> >
>
> ** Must be a bunch of rabid homophobes.
*
>
>
> > In my 300 watt amp, I would not be happy with just the existing 12 volt
> internal
> > zener diode vgs limiters, lest too much power be allowed to be made into
> low value loads,
> > with a correspondingly too high amount of dissipation in the devices.
>
> ** Correct.
>
> > I leant it to a guy for PA use and he couldn't kill it.
>
> ** PA folk have great difficulty killing mosfet amps - only the most
> determined have ever succeeded.
The guy should have killed it.
He ran an old tubed Lennard amp with 8 x EL34 into two bins with
a total of 12 x 12" woofers and some weakling tweeters, which he didn't notice
were
blown for about 20 years when he brought the then failing amp to me to
rewire and re-design, etc.
The Lennard amp had 900v anode supply and 450v screen supply,
and during a brief test I measured 500 watts at clipping.
After two repairs including some output tube replacements,
I said before you waste my time again, I want to examine those speakers
to see if they are 8 ohms like they are supposed to be.
He reckoned they were 8 ohms allright, and they were too!
But he'd hooked all of them up in parallel, when he'd replaced a couple of
drivers
many years before. He'd gone all these years like that.
Must have blown many tubes, there were 4 different brands of EL34 within
and a couple of burned tube sockets.
When I re-wired the speakers and replaced the tweeters, used plenty wool
inside the boxes, lowered the B+ to 670v, added 12 dB NFB, the system
sounded extremely healthy for its 250 watts.
He ain't been back for 6 years now.
He also bought a Yamaha 2200 for a low price from Cash Converters.
The metering boards were full of dry joints, but when I got that going properly
it must also have proved reliable. It'd already seen 20 years of service in a
dyke nightclub down here,
with weekly gigs of all sorts.
Patrick Turner
Phil Allison
"Patrick Turner"
> He also bought a Yamaha 2200 for a low price from Cash Converters.
** If you want to see a powerful amp produce some *mighty* back EMF spikes
when driving an inductive load there are none better for that than the
Yamaha P2200 - also known as the " 22 - oh - boom". Try a 4 ohm resistor
plus 5 mH inductor as a load at 115 Hz for some entertainment.
The P 2200 has no thermal protection save for a red warning "temp" LED
next tot the power switch - about 45 seconds after the red LED lights the
output stage **totally expires** - thence the AC fuse and no more sound.
Very inscrutable.
> it must also have proved reliable. It'd already seen 20 years of service
in a
> dyke nightclub down here, with weekly gigs of all sorts.
** Should be donated to the Smithsonian.
......... Phil
Patrick Turner
Phil Allison wrote:
> "Patrick Turner"
>
> > He also bought a Yamaha 2200 for a low price from Cash Converters.
>
> ** If you want to see a powerful amp produce some *mighty* back EMF spikes
> when driving an inductive load there are none better for that than the
> Yamaha P2200 - also known as the " 22 - oh - boom". Try a 4 ohm resistor
> plus 5 mH inductor as a load at 115 Hz for some entertainment.
Its been soldiering on for a long time, as I said, 20 years in a dyke night
club.
Maybe its a lesbian amp.
It was not in my brief to modify the amp against spikes.
>
> The P 2200 has no thermal protection save for a red warning "temp" LED
> next tot the power switch - about 45 seconds after the red LED lights the
> output stage **totally expires** - thence the AC fuse and no more sound.
> Very inscrutable.
I didn't advise him to buy this thing at CC. He just saw it there for a low
price,
finding it to be irresistable.....
Maybe he never pushes it hard, and maybe the previous owners at the club never
did either,
since female watts are easier on the amp than male watts.
I used to specialise in repairs to PA and musician's amps at first,
but the clients were always broke, and there didn't seem to be a great deal of
profit
in dealing with that line of business.
Patrick Turner.