Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
Jones's servo bias
148 views
Skip to first unread message
Ian Iveson
Nov 27, 2011, 9:55:42 PM11/27/11
to
Following the 10c thread, I checked my slightly musty copy of the
quintessentially English Morgan Jones, "Valve Amplifiers", Newnes
1999.
I was curious because I thought his circuit uses a zener to restrict
the upswings of the signal from the cathode-current sensing resistor.
Using a fixed reference isn't ideal if the bias current is adustable,
because then the zener clipping point won't necessarily be correct. As
it happens, he doesn't use a zener. This is what he does instead.
First, he uses a 1R current sense resistor at the cathode of the
valve, followed by an opamp set to a gain of 100. This minimises the
effect of this part of the circuit on the operation of the valve. Each
10mA of cathode current thus results in 1V output from the opamp.
He generates two reference voltages, one twice the other, from a
convenient +ve voltage using a pot and two fixed resistors. The upper
reference is used together with a transistor/diode clamp to limit the
upswings of the signal from the opamp. A low-pass filter then smooths
this to a steady DC before it is compared to the lower of the two
reference voltages, using a second opamp.
The ensuing difference signal is then used to alter the grid voltage
via a second transistor.
Because the signal to the filter is always limited to twice the set
bias level, its AC component remains more or less symmetrical when the
valve cuts off during AB operation, so it smooths to a steady DC
level. With a triode, the servo will tend to very slightly raise bias
current during extended periods of AB, which is better than reducing
it as others tend to do.
That's two opamps, two transistors, a diode, and a bunch of resistors
and caps to maintain the fixed bias of one output valve. Morgan
doesn't use the circuit in any of his amps AFAICS, so there's no
example of how the scheme is extended to the one or more other output
valves.
It wouldn't be necessary to replicate the whole circuit for the other
valves, because they need to be set at the same level, to balance the
transformer current. The set reference voltages could be made common,
and the rest of the circuit replicated. Alternatively the unclamped
signal could be smoothed and used as the reference for the other
servos, which then wouldn't need clamps.
It's possible that reading the above will not result in a working
diagram in your head. The actual circuit does work. Buy the book. It
is British and we are short of money.
Some thought, in the details, must be given to stability. Each servo
is a feedback circuit, and all are related in various ways via various
delays to themselves and to each other.
Ian
quintessentially English Morgan Jones, "Valve Amplifiers", Newnes
1999.
I was curious because I thought his circuit uses a zener to restrict
the upswings of the signal from the cathode-current sensing resistor.
Using a fixed reference isn't ideal if the bias current is adustable,
because then the zener clipping point won't necessarily be correct. As
it happens, he doesn't use a zener. This is what he does instead.
First, he uses a 1R current sense resistor at the cathode of the
valve, followed by an opamp set to a gain of 100. This minimises the
effect of this part of the circuit on the operation of the valve. Each
10mA of cathode current thus results in 1V output from the opamp.
He generates two reference voltages, one twice the other, from a
convenient +ve voltage using a pot and two fixed resistors. The upper
reference is used together with a transistor/diode clamp to limit the
upswings of the signal from the opamp. A low-pass filter then smooths
this to a steady DC before it is compared to the lower of the two
reference voltages, using a second opamp.
The ensuing difference signal is then used to alter the grid voltage
via a second transistor.
Because the signal to the filter is always limited to twice the set
bias level, its AC component remains more or less symmetrical when the
valve cuts off during AB operation, so it smooths to a steady DC
level. With a triode, the servo will tend to very slightly raise bias
current during extended periods of AB, which is better than reducing
it as others tend to do.
That's two opamps, two transistors, a diode, and a bunch of resistors
and caps to maintain the fixed bias of one output valve. Morgan
doesn't use the circuit in any of his amps AFAICS, so there's no
example of how the scheme is extended to the one or more other output
valves.
It wouldn't be necessary to replicate the whole circuit for the other
valves, because they need to be set at the same level, to balance the
transformer current. The set reference voltages could be made common,
and the rest of the circuit replicated. Alternatively the unclamped
signal could be smoothed and used as the reference for the other
servos, which then wouldn't need clamps.
It's possible that reading the above will not result in a working
diagram in your head. The actual circuit does work. Buy the book. It
is British and we are short of money.
Some thought, in the details, must be given to stability. Each servo
is a feedback circuit, and all are related in various ways via various
delays to themselves and to each other.
Ian
arthr...@webtv.net
Nov 28, 2011, 12:10:23 AM11/28/11
to
Patrick Turner
Nov 28, 2011, 7:32:02 AM11/28/11
to
bias to lessen the effects of rising Ek across cathode RC bias
networks, and all are no good if Ek is allowed to rise more than about
10% from the idle value.
Some published schemes have sought to use a differential amp which
accepts the two Vdc signals across the 1ohm current sensors after
filtering out the Vac. I've tried out a few schemes like this, and the
simplest is done using two darlington pairs of PNP bjts for an LTP set
up with emitter CCS so that rise in Ek does not cause any increase in
grid bias voltage, but any difference in the Ek of either PP tube is
corrected. The LTP has good CMRR, and DC gain can easily be up to
1,000, which too much, so 100 is probably OK. I found such a circuit
worked fabulously to keep the Idc of each tube equal even in extreme
class AB, say with a 1 ohm connected across the 8 ohm labelled outlet.
In other words, it worked with the worst of circumstances. But then
when global NFB was connected, I could not stop LF oscillations no
matter how much open loop gain tailoring and network tweaking I did.
So I went right off the idea of having anything to do with grid bias
swinging to force the OPT stage to have balanced Idc on both halves of
the OPT. Electronics World even published a circuit using opamps to
perform exactly what I have described and the author said that with a
toroidal OPT, the iron permeability of the core could easily be
40,000, so it is critical to keep Idc balance in the OPT core or else
the core saturates all too easily even with say only 10mA dc
imbalance. If the OPT design has been done right, there will always be
enough primary inductance at any Va-a level above 5Vrms, and Fsat at
full power should be under 20Hz. Usually when these conditions are
met, the iron µ need rarely ever be more than 2,500, which is ideally
obtained with the use of GOSS C-cores with a small air gap, or E&I
laminations assembled with partial air gapping using bundles of Es and
Is facing same way, so that in a 50mm stack there may be 17 reverse
direction bundles interleaved. So the high µ of the GOSS is reduced,
and the saturation due to Idc imbalance becomes negligible. This being
the case, the bias servo isn't needed, and all because of issues not
raised in trendy magazines and books, and all because of good basic
knowledge within my father's generation. The brightest of them knew
all about air gapping PP OPTs, and using enough turns and large enough
cores to avoid the problems of balanced operation with transformers,
all of which tend to be unbalanced at some occasion as fixed bias
drifts away from its ideal setting as tube age.
But my father's generation never sought to have the convenience of
cathode bias while also triying to have Ek stay constant to allow AB
operation without mis-biasing.
I found my Dynamic Bias Stabilizer seemed to work well in the presence
of the CFB of the OP stage plus the global NFB.
Ian likes to pour buckets of poop on all my ideas, but we all must
remember he rarely ever solders up a circuit and spends days
analyising it, then publishing his consise findings. He has Flipper's
disease in a big way, and relies on simulation for his self validation
of righteousness without doubt. Well, I await the result of Ian's
soldering and measuring efforts. Nothing I've seen in Morgan Jones
makes me wanna copy anything Morgan suggests with regard to fiddling
with grid bias.
A number of amp makers have used toroidal OPTs over the last 15 years,
and those I have rewired because of smoke and unreliability all had
fixed bias and suffered because of attrocious methods used to adjust
bias and monitor bias as it is adjusted, thus leading owners to be
their own worst enemy. But not of their own intentions. Owners
much prefer cathode bias, just plug in OP tubes and let the RC do the
business. In a pair of Quad-II-Forty I have just completed re-wiring,
samples of Red Lion KT88 have given 4 different Ek voltages between
46Vdc and 37Vdc across the 630 ohm I have put in to replace the Rk =
390r originals which allowed the KT88 to overheat all to often and all
too easily. The range of Ek differences gives a maximum 15mA Idc
difference. Cores in Chinese made OPTs designed by Andy Grove are
probably E&I GOSS, and the Lp seems to be rather low, even though Fsat
at full PO is at 25Hz, 33Watts. So iron µ is maybe low, and whatever
it is I can't change that, but despite the maximum Idc imbalance the
core sat character as F is reduced is not alarming as it would be in
an ungapped strip-wound toroidal core. The answer to such uneven
matching amoung a given batch of supposedly NOS Red Lion KT88 is to
have some additional bias balancing adjust method as seen in another
pair of Quad-II I rewired where I used EHKT90 instead of KT66, see
http://www.turneraudio.com.au/quad2powerampmods.htm and proceed to
about 1/2 way down the page to the 2005 schematic.
There are more ways than one of doing something.
Anyway, although the Ek rises alarmingly in Quad-II amps when a sine
wave is used up to clipping level when say a 4 ohm load is connected
while the OPT is strapped for "8 ohms", most listeners only use a watt
of average power, so the load mismatch is well tolerated and Ek stays
fairly constant, and with only mild variations in the Idc in each 1/2
of the OPT.
I have a pair of Quad-II amps awaiting total re-wire, and this time I
will remove the OPTs from their cans and re-arrange the OPT sec
windings to give waste free connection to 4 ohms so that winding
losses will be optimal as they are when the OPT is strapped for 16
ohms. Kieth Snook's website explains what is inside a Quad-II OPT can.
Peter Walker could so easily have provided two extra board terminals
on his OPTs but he must have thought such a good thing would have
confused most ppl in 1955, most of whom didn't have any 4 ohm
speakers. So, Quad-II as they are in original condtion do not work
well with 4 ohms. Even with 8 ohms they work in AB, and the better
load for a decent 20 watts is 9ka-a, not bloomin 3k8. Quad-II work
best with OPTs set for 8 ohms, and speaker RL = 16 ohms.
I also have a pair of better OPTs wound especially for my pair of
Quads so the original Quad-II with correctly arranged OPT secs can be
used in another amp to replaced even worse OPTs.
Ive tried my DBS scheme 3 years ago in yet another pair of Quads with
KT88 outputs. Worked just fine, and with sine waves at clipping PO is
34Watts, low THD, and excellent sound. Ian can hate my scheme as much
as he likes, but my customers wouldn't agree with him.
Morgan Jones's book had several editions. The early editions were
riddled with frightful mistakes on nearly every page, most of which
the average groovy non technically trained audiophile never ever would
see. Later editions are better, but my copy sure has some real
clangers. So, with Morgan, you need be aware of his mistakes, the sort
of things that never would have been seen in RDH4, because there was
so much better proof reading and self criticism prior to publication.
Patrick Turner.
Ian Iveson
Nov 28, 2011, 10:53:45 AM11/28/11
to
>
> - Show quoted text -
Second. Circuit is on p294.
Ian
GRe
Nov 28, 2011, 4:31:18 PM11/28/11
to
"Patrick Turner" <in...@turneraudio.com.au> wrote in message
news:a25664c6-ab6d-4318...@u10g2000prl.googlegroups.com...
[...]
>The answer to such uneven
>matching amoung a given batch of supposedly NOS Red Lion KT88 is to
>have some additional bias balancing adjust method as seen in another
>pair of Quad-II I rewired where I used EHKT90 instead of KT66, see
>http://www.turneraudio.com.au/quad2powerampmods.htm and proceed to
>about 1/2 way down the page to the 2005 schematic.
What's the reason you chose 560K for R11 while R10= 470K?
>matching amoung a given batch of supposedly NOS Red Lion KT88 is to
>have some additional bias balancing adjust method as seen in another
>pair of Quad-II I rewired where I used EHKT90 instead of KT66, see
>http://www.turneraudio.com.au/quad2powerampmods.htm and proceed to
>about 1/2 way down the page to the 2005 schematic.
Rgds,
Gio
>Patrick Turner.
John L Stewart
Nov 29, 2011, 12:22:01 PM11/29/11
to
'Ian Iveson[_2_ Wrote:
> ;943887']Following the 10c thread, I checked my slightly musty copy of
one of his books on guitar amps. Output is 4X 5881 running into a
Plitron OPT (toroid). The cathode current sampling resistor is 10R
driving point B on the schema. Each 5881 has an indepedant bias set cct.
Point A drives the OP tube grid bias.
Full loop NFB from the OPT secodary to the Grid 2 of the input dual
triode as is customary on these kind of amp.
Cheers, John
+-------------------------------------------------------------------+
|Filename: Kevin OConnor Auto Bias Control.jpg |
|Download: http://www.audiobanter.com/attachment.php?attachmentid=250|
+-------------------------------------------------------------------+
--
John L Stewart
Patrick Turner
Dec 1, 2011, 3:03:24 AM12/1/11
to
On Nov 29, 8:31 am, "GRe" <mynamespacedbyd...@xs4all.nl> wrote:
> "Patrick Turner" <i...@turneraudio.com.au> wrote in message
with common Rk resistance, the RLa seen by each gain tube of the LTP
should be slightly dissimilar to compensate for the unbalancing effect
of the common Rk value. If Rk was a CCS, then the total cap coupled
RLa on each side of the LTP needs to be the same for perfect balance,
even if the tube types used are quite different, eg, see the Radford
schematics. But where you have two pentodes as LTP gain tubes, gain is
close to gm x RL, and quite high, and the current change in one equals
the current cjange in the other plus the change of current in the
common Rk. You should be able to do the analysis of all working
voltages and currents in my circuit as easily as in any other.
In fact, the use of 560k and 470k is a bit high for Rg bias values,
because reverse grid current in ageing tubes can cause Eg1 to become
several volts more positive than the actual bias voltage at the end of
Rg away from the grid. In my example, if that happens, it unbalances
the Idc of the OP tubes thus causing the balance monitoring LEDs to
change from normal OK with both equally bright to having one brighter
than the other. An owner will turn the adjust pot with a thumbnail or
a dinner knife until Idc balance is restored, but if the problem of
positve grids gets so bad the balance pot cannot be adjusted to get
equal LED brightness, then he knows he has a tube problem, and needs
new tubes. For this reason, the high Rg values are not so bad, bearing
in mind the amp has cathode bias with individual R&C networks in each
OP tube cathode circuit.
In more recent rewiring of Quad-II-Forty amps completed last week, I
used 6SH7 in triode for V1, then used a 6SN7 for LTP with a CCS with
MJE340. RLa to B+ was 50k per 1/2 6SN7, and Rg = 120k taken to to the
OPT CFB tranny winding to partially bootstrap the Rg to have it appear
to to the 6SN7 as about 300k. The dc balance will remain fair but not
excellent with individual R&C bias networks, and with Rg at 120k, the
reverse positive grid current in ageing tubes won't cause too much
trouble. In two samples of the Red Lion KT88 in one Quad-II-Forty
monobloc, Ek = +38Vdc in one tube and +44Vdc in the other. Rk = 630
ohms, so Il for the two OP tubes is 60.3mAdc and 69.8mAdc, a
difference of 9.5mAdc, a considerable difference. But effects of the
Idc imbalance in Chinese made OPTs is minimal because although the OPT
cores seem large enough, the number of turns is dissapointing. But the
paucity of primary turns means the Np x Idc product in the equation
governing the Bdc within the OPT core remains fairly low, thus
allowing a high ammount of additive Bac before saturation occurs at
full power at about 25Hz, good enough. But Quad-II amps from the 1950s
have more Np turns and smaller core, and the use of a balancing pot
for Idc in each OPT tube isn't a bad idea.
My customer who had me re-wire his Quad-II-Forty amps has given me a
dozen Red Lion KT88 to test, all with unknown condition. I should be
able to find close pairs among them all, and then if he had an uncle,
his name would be Bob.
Generally, whatever I do in amps is done to help the tubes give their
best in terms of distortion, bandwidth, output resistance and
stability. I don't include ALL techniques in every pair of amps; not
all amps of mine have the same range of techniques. They don't need
to.
Providing the loading of tubes in tube amplifiers is technically
optimised, and the other 4 items done properly, with optimal DC
biasing, music will flow well without it becoming insiduously grubby
as the tube age, which is so common in Quad-II amps where their Uncle
Scrooge owners cannot bring themselves to replace 20 year old KT66.
Patrick Turner.
> "Patrick Turner" <i...@turneraudio.com.au> wrote in message
>
> news:a25664c6-ab6d-4318...@u10g2000prl.googlegroups.com...
>
> [...]
>
> >The answer to such uneven
> >matching amoung a given batch of supposedly NOS Red Lion KT88 is to
> >have some additional bias balancing adjust method as seen in another
> >pair of Quad-II I rewired where I used EHKT90 instead of KT66, see
> >http://www.turneraudio.com.au/quad2powerampmods.htmand proceed to
> news:a25664c6-ab6d-4318...@u10g2000prl.googlegroups.com...
>
> [...]
>
> >The answer to such uneven
> >matching amoung a given batch of supposedly NOS Red Lion KT88 is to
> >have some additional bias balancing adjust method as seen in another
> >pair of Quad-II I rewired where I used EHKT90 instead of KT66, see
> >about 1/2 way down the page to the 2005 schematic.
>
> What's the reason you chose 560K for R11 while R10= 470K?
>
> Rgds,
> Gio
To achieve Vac drive balance to each OP tube where there is an LTP
>
> What's the reason you chose 560K for R11 while R10= 470K?
>
> Rgds,
> Gio
with common Rk resistance, the RLa seen by each gain tube of the LTP
should be slightly dissimilar to compensate for the unbalancing effect
of the common Rk value. If Rk was a CCS, then the total cap coupled
RLa on each side of the LTP needs to be the same for perfect balance,
even if the tube types used are quite different, eg, see the Radford
schematics. But where you have two pentodes as LTP gain tubes, gain is
close to gm x RL, and quite high, and the current change in one equals
the current cjange in the other plus the change of current in the
common Rk. You should be able to do the analysis of all working
voltages and currents in my circuit as easily as in any other.
In fact, the use of 560k and 470k is a bit high for Rg bias values,
because reverse grid current in ageing tubes can cause Eg1 to become
several volts more positive than the actual bias voltage at the end of
Rg away from the grid. In my example, if that happens, it unbalances
the Idc of the OP tubes thus causing the balance monitoring LEDs to
change from normal OK with both equally bright to having one brighter
than the other. An owner will turn the adjust pot with a thumbnail or
a dinner knife until Idc balance is restored, but if the problem of
positve grids gets so bad the balance pot cannot be adjusted to get
equal LED brightness, then he knows he has a tube problem, and needs
new tubes. For this reason, the high Rg values are not so bad, bearing
in mind the amp has cathode bias with individual R&C networks in each
OP tube cathode circuit.
In more recent rewiring of Quad-II-Forty amps completed last week, I
used 6SH7 in triode for V1, then used a 6SN7 for LTP with a CCS with
MJE340. RLa to B+ was 50k per 1/2 6SN7, and Rg = 120k taken to to the
OPT CFB tranny winding to partially bootstrap the Rg to have it appear
to to the 6SN7 as about 300k. The dc balance will remain fair but not
excellent with individual R&C bias networks, and with Rg at 120k, the
reverse positive grid current in ageing tubes won't cause too much
trouble. In two samples of the Red Lion KT88 in one Quad-II-Forty
monobloc, Ek = +38Vdc in one tube and +44Vdc in the other. Rk = 630
ohms, so Il for the two OP tubes is 60.3mAdc and 69.8mAdc, a
difference of 9.5mAdc, a considerable difference. But effects of the
Idc imbalance in Chinese made OPTs is minimal because although the OPT
cores seem large enough, the number of turns is dissapointing. But the
paucity of primary turns means the Np x Idc product in the equation
governing the Bdc within the OPT core remains fairly low, thus
allowing a high ammount of additive Bac before saturation occurs at
full power at about 25Hz, good enough. But Quad-II amps from the 1950s
have more Np turns and smaller core, and the use of a balancing pot
for Idc in each OPT tube isn't a bad idea.
My customer who had me re-wire his Quad-II-Forty amps has given me a
dozen Red Lion KT88 to test, all with unknown condition. I should be
able to find close pairs among them all, and then if he had an uncle,
his name would be Bob.
Generally, whatever I do in amps is done to help the tubes give their
best in terms of distortion, bandwidth, output resistance and
stability. I don't include ALL techniques in every pair of amps; not
all amps of mine have the same range of techniques. They don't need
to.
Providing the loading of tubes in tube amplifiers is technically
optimised, and the other 4 items done properly, with optimal DC
biasing, music will flow well without it becoming insiduously grubby
as the tube age, which is so common in Quad-II amps where their Uncle
Scrooge owners cannot bring themselves to replace 20 year old KT66.
Patrick Turner.
Ian Iveson
Dec 1, 2011, 11:43:21 AM12/1/11
to
On Nov 29, 5:22 pm, John L Stewart <John.L.Stewart.
> John L Stewart- Hide quoted text -
I see this thread is quite a close replica of one a long time ago.
I don't see how the Kevin O'Connor circuit works. It looks like the
intention is to force symmetry of the AC component of the current
sense signal by clipping both positive and negative-going excursions,
but hasn't the signal already been smoothed? If so, then the one diode
limits the DC drift but doesn't prevent it, and the other diode is
redundant.
I wish I was better at understanding transistor and opamp circuits.
Looking at the other end of the servo, I don't see how that open
collector works. Doesn't it need a pull-down resistor to the negative
supply? And there's a resistor with no value shown...where's Flipper?
He seems to know about solid state stuff.
I guess Morgan got carried away with the elegance of his idea.
Clipping both ways isn't so clever, but it's a better solution I
think. The more the AC is chopped, the less the drift caused by
distortion.
Toroid in a guitar amp?
Ian
>
> - Show quoted text -
Hi John, good to see you're still here.
> - Show quoted text -
I see this thread is quite a close replica of one a long time ago.
I don't see how the Kevin O'Connor circuit works. It looks like the
intention is to force symmetry of the AC component of the current
sense signal by clipping both positive and negative-going excursions,
but hasn't the signal already been smoothed? If so, then the one diode
limits the DC drift but doesn't prevent it, and the other diode is
redundant.
I wish I was better at understanding transistor and opamp circuits.
Looking at the other end of the servo, I don't see how that open
collector works. Doesn't it need a pull-down resistor to the negative
supply? And there's a resistor with no value shown...where's Flipper?
He seems to know about solid state stuff.
I guess Morgan got carried away with the elegance of his idea.
Clipping both ways isn't so clever, but it's a better solution I
think. The more the AC is chopped, the less the drift caused by
distortion.
Toroid in a guitar amp?
Ian
flipper
Dec 1, 2011, 10:08:19 PM12/1/11
to
presume they were intended to you'd need to isolate them from the
filter. Like, for example, stick another 100k between them and the
-input so they can clip and then you average the result. Not saying
that would be ideal, just an example.
The circuit's application is also limited because it depends on bias
being in the right range for the diodes although one might be able to
fix that with a gain buffer off the sense resistor. Well, you could
vary the sense resistor too but you usually want that as low as
possible and a buffer amp let's you do that and then gain to suit.
(That's 'modern times' solid state thinking, you see. What the heck,
throw in an opamp, they're cheap).
I presumed the open collector went into something else because it
doesn't do anything all by it's lonesome, unless he's depending on
grid leakage for some nominal current. Looks upside down to me anyway
because I'd end with a low impedance follower.
GRe
Dec 2, 2011, 7:13:58 AM12/2/11
to
"Ian Iveson" <iani...@gmail.com> wrote in message
news:6c1d3dcb-d4bf-48f5...@q30g2000yqj.googlegroups.com...
On Nov 29, 5:22 pm, John L Stewart <John.L.Stewart.
91ad...@audiobanter.com> wrote:
> 'Ian Iveson[_2_ Wrote:
[...]
91ad...@audiobanter.com> wrote:
> 'Ian Iveson[_2_ Wrote:
But you can make a few assumptions and see if that works out.
Assume Vref (see schematic in John's post) is just above 400mV(say
410mV), -Vb is 40V and Rx= Ry<< Rg.
Also assume the -Vg/G1 drive is as in the schematic below but other
configurations possible/imagineable.
Points "A" and "B" in the ASCII schematic go to the identical points in
John's schematic.
If (one of) the 5881 idle's at Ic= 40mA than V-Rc= 400mV, so Vout of OP-A1
is positive(assuming bipolar supply) resulting in T1(MPSA56) turned off and
T2(MPSA92) turned on, latter resulting in -Vg= ~1/2-Vb= 20V.
Now, if Ic rises above 41mA, resulting in V-Rc> 410mV, OP-A1/Vout will ramp
negative* until T1 starts to conduct thereby robbing base-current from T2
which again results in a -Vg value that is somewhat more negative than
before.
Et Voilà. My 2 euro-cents (that's even less than before).
+---> OPT
|
|
===== 5881
-----
-----
Vin >---+---- -- -- G1
| O
| |
+-+ V Ic
| |Rg |
+-+ +---> B
| |
+-------+ -Vg | ---
| | | ^
+-+ +-+ +-+ |
| |Ry | |Rx | | Rc 10R |V-Rc
+-+ +-+ +-+ |
| | | |
| | | |
| | --- GND ---
^ |
A -Vb
It would be nice to know a few more details of John's schematic.
- Vref
- -Vg/G1 drive circuit details
- -Vb
- The missing value resistor
- The supply voltages* of A1 (bipolar, bipolar a-symmetric, unipolar)
Do you have that information John?
Gio
*Ramp-negative to T1-conduct time, as well as ramp-positive to T1-off time,
will depend on the supply voltage of A1(besides V-Rc).
John L Stewart
Dec 5, 2011, 11:44:14 AM12/5/11
to
'Ian Iveson[_2_ Wrote:
> ;944335']On Nov 29, 5:22*pm, John L Stewart <John.L.Stewart.
> 91ad...@audiobanter.com> wrote:-
> > 'Ian Iveson[_2_ Wrote:
> >
> >
> >
> >
> >-
> > > ;943887']Following the 10c thread, I checked my slightly musty copy
> of
> > > the
> > > quintessentially English Morgan Jones, "Valve Amplifiers", Newnes
> > > 1999.-
> of
> > > the
> > > quintessentially English Morgan Jones, "Valve Amplifiers", Newnes
> >-
> > > I was curious because I thought his circuit uses a zener to
> restrict
> > > the upswings of the signal from the cathode-current sensing
> resistor.
> > > Using a fixed reference isn't ideal if the bias current is
> adustable,
> > > because then the zener clipping point won't necessarily be correct.
> As
> > > it happens, he doesn't use a zener. This is what he does instead.-
> restrict
> > > the upswings of the signal from the cathode-current sensing
> resistor.
> > > Using a fixed reference isn't ideal if the bias current is
> adustable,
> > > because then the zener clipping point won't necessarily be correct.
> As
> >-
> > > First, he uses a 1R current sense resistor at the cathode of the
> > > valve, followed by an opamp set to a gain of 100. This minimises
> the
> > > effect of this part of the circuit on the operation of the valve.
> Each
> > > 10mA of cathode current thus results in 1V output from the opamp.-
> > > valve, followed by an opamp set to a gain of 100. This minimises
> the
> > > effect of this part of the circuit on the operation of the valve.
> Each
> >-
> > > He generates two reference voltages, one twice the other, from a
> > > convenient +ve voltage using a pot and two fixed resistors. The
> upper
> > > reference is used together with a transistor/diode clamp to limit
> the
> > > upswings of the signal from the opamp. A low-pass filter then
> smooths
> > > this to a steady DC before it is compared to the lower of the two
> > > reference voltages, using a second opamp.-
> > > convenient +ve voltage using a pot and two fixed resistors. The
> upper
> > > reference is used together with a transistor/diode clamp to limit
> the
> > > upswings of the signal from the opamp. A low-pass filter then
> smooths
> > > this to a steady DC before it is compared to the lower of the two
> >-
> > > The ensuing difference signal is then used to alter the grid
> voltage
> > > via a second transistor.-
> voltage
> >-
> > > Because the signal to the filter is always limited to twice the set
> > > bias level, its AC component remains more or less symmetrical when
> the
> > > valve cuts off during AB operation, so it smooths to a steady DC
> > > level. With a triode, the servo will tend to very slightly raise
> bias
> > > current during extended periods of AB, which is better than
> reducing
> > > it as others tend to do.-
> > > bias level, its AC component remains more or less symmetrical when
> the
> > > valve cuts off during AB operation, so it smooths to a steady DC
> > > level. With a triode, the servo will tend to very slightly raise
> bias
> > > current during extended periods of AB, which is better than
> reducing
> >-
> > > That's two opamps, two transistors, a diode, and a bunch of
> resistors
> > > and caps to maintain the fixed bias of one output valve. Morgan
> > > doesn't use the circuit in any of his amps AFAICS, so there's no
> > > example of how the scheme is extended to the one or more other
> output
> > > valves.-
> resistors
> > > and caps to maintain the fixed bias of one output valve. Morgan
> > > doesn't use the circuit in any of his amps AFAICS, so there's no
> > > example of how the scheme is extended to the one or more other
> output
> >-
> > > It wouldn't be necessary to replicate the whole circuit for the
> other
> > > valves, because they need to be set at the same level, to balance
> the
> > > transformer current. The set reference voltages could be made
> common,
> > > and the rest of the circuit replicated. Alternatively the unclamped
> > > signal could be smoothed and used as the reference for the other
> > > servos, which then wouldn't need clamps.-
> other
> > > valves, because they need to be set at the same level, to balance
> the
> > > transformer current. The set reference voltages could be made
> common,
> > > and the rest of the circuit replicated. Alternatively the unclamped
> > > signal could be smoothed and used as the reference for the other
> >-
> > > It's possible that reading the above will not result in a working
> > > diagram in your head. The actual circuit does work. Buy the book.
> It
> > > is British and we are short of money.-
> > > diagram in your head. The actual circuit does work. Buy the book.
> It
> >-
> > > Some thought, in the details, must be given to stability. Each
> servo
> > > is a feedback circuit, and all are related in various ways via
> various
> > > delays to themselves and to each other.-
> servo
> > > is a feedback circuit, and all are related in various ways via
> various
> >-
> > > Ian-
> >
> > Here is a successful auto bias cct used by Kevin O'Conner, published
> in
> > one of his books on guitar amps. Output is 4X 5881 running into a
> > Plitron OPT (toroid). The cathode current sampling resistor is 10R
> > driving point B on the schema. Each 5881 has an indepedant bias set
> cct.
> > Point A drives the OP tube grid bias.
> >
> > Full loop NFB from the OPT secodary to the Grid 2 of the input dual
> > triode as is customary on these kind of amp.
> >
> > Cheers, John
> >
> > +-------------------------------------------------------------------+
> > |Filename: Kevin OConnor Auto Bias Control.jpg * * * * * * * * * * *|
> > Here is a successful auto bias cct used by Kevin O'Conner, published
> in
> > one of his books on guitar amps. Output is 4X 5881 running into a
> > Plitron OPT (toroid). The cathode current sampling resistor is 10R
> > driving point B on the schema. Each 5881 has an indepedant bias set
> cct.
> > Point A drives the OP tube grid bias.
> >
> > Full loop NFB from the OPT secodary to the Grid 2 of the input dual
> > triode as is customary on these kind of amp.
> >
> > Cheers, John
> >
> > +-------------------------------------------------------------------+
> > |Download:http://www.audiobanter.com/attachment.php?attachmentid=250|
> > +-------------------------------------------------------------------+
> >
> > --
> > John L Stewart- Hide quoted text -
> >
> > - Show quoted text --
> > +-------------------------------------------------------------------+
> >
> > --
> > John L Stewart- Hide quoted text -
> >
>
> Hi John, good to see you're still here.
>
> I see this thread is quite a close replica of one a long time ago.
>
> I don't see how the Kevin O'Connor circuit works. It looks like the
> intention is to force symmetry of the AC component of the current
> sense signal by clipping both positive and negative-going excursions,
> but hasn't the signal already been smoothed? If so, then the one diode
> limits the DC drift but doesn't prevent it, and the other diode is
> redundant.
>
> I wish I was better at understanding transistor and opamp circuits.
> Looking at the other end of the servo, I don't see how that open
> collector works. Doesn't it need a pull-down resistor to the negative
> supply? And there's a resistor with no value shown...where's Flipper?
> He seems to know about solid state stuff.
>
> I guess Morgan got carried away with the elegance of his idea.
> Clipping both ways isn't so clever, but it's a better solution I
> think. The more the AC is chopped, the less the drift caused by
> distortion.
>
> Toroid in a guitar amp?
>
> Ian
Hi Ian & Others- Yeh, I'm back again after another 5 year stint in
> Hi John, good to see you're still here.
>
> I see this thread is quite a close replica of one a long time ago.
>
> I don't see how the Kevin O'Connor circuit works. It looks like the
> intention is to force symmetry of the AC component of the current
> sense signal by clipping both positive and negative-going excursions,
> but hasn't the signal already been smoothed? If so, then the one diode
> limits the DC drift but doesn't prevent it, and the other diode is
> redundant.
>
> I wish I was better at understanding transistor and opamp circuits.
> Looking at the other end of the servo, I don't see how that open
> collector works. Doesn't it need a pull-down resistor to the negative
> supply? And there's a resistor with no value shown...where's Flipper?
> He seems to know about solid state stuff.
>
> I guess Morgan got carried away with the elegance of his idea.
> Clipping both ways isn't so clever, but it's a better solution I
> think. The more the AC is chopped, the less the drift caused by
> distortion.
>
> Toroid in a guitar amp?
>
> Ian
hitech sales. But I guess for the last time at my age!
Kevin O'Connor is adventurous & perhaps does some things others may not
approve of. Thus a toroidal OPT in a guitar amp. No problem with servo
bias holding the lid down.
This one not as elegant as that found in the Brook, easily the best I've
seen in my very long career. It is unique & sure makes a lot of sense.
How else to get 30 Watts at low distortion out of a pair of 2A3's in
AB2?
I got interested in servo bias some time ago while doing my thing with
6AS7/6080 PP circuits. Some may remember the bootstrapped driver schemes
with & without the UL OPT. But the heat from the cathode bias resistors
under the chassis is a serious problem. So lets try somekind of servo
bias. Perhaps this coming Winter.
I've been busy on the bicycle (like Patrick), more than 4000 Km since
May 1st. And a lot of physical labor here on the acres. I would rather
be outside than slaving over a hot soldering iron!
For anyone who would like a complete schema of the Kevin O'Connor guitar
amp referenced email me direct as follows-
jh dot stewart at sympatico dot ca
Gotta go now. Today its time to service the summer equipment & put it
away till Spring! And here is that easy to drive 6AS7/6080 amp for the
curious.
Cheers, John
+-------------------------------------------------------------------+
|Filename: 6BQ7 6080 Amp.jpg |
|Download: http://www.audiobanter.com/attachment.php?attachmentid=251|
GRe
Dec 6, 2011, 7:30:42 AM12/6/11
to
news:f00b9799-c00e-4f50...@q16g2000yqn.googlegroups.com...
Hi Patrick,
It took me some time to digest the comprehensive info, especially when I
started reading the QuadII mod's article on your site, hence the delayed
answer.
I don't think I digested it all, yet I learned something, thanks.
B.t.w., do yourself a favor, leave the narcissists doing what they do best;
gazing at their own belly button.
Gio.
It took me some time to digest the comprehensive info, especially when I
started reading the QuadII mod's article on your site, hence the delayed
answer.
I don't think I digested it all, yet I learned something, thanks.
B.t.w., do yourself a favor, leave the narcissists doing what they do best;
gazing at their own belly button.
Gio.
Ian Iveson
Dec 6, 2011, 9:30:32 AM12/6/11
to
On Dec 5, 4:44 pm, John L Stewart <John.L.Stewart.
saturating OPT is a desirable tone option.
> This one not as elegant as that found in the Brook, easily the best I've
> seen in my very long career. It is unique & sure makes a lot of sense.
> How else to get 30 Watts at low distortion out of a pair of 2A3's in
> AB2?
>
> I got interested in servo bias some time ago while doing my thing with
> 6AS7/6080 PP circuits. Some may remember the bootstrapped driver schemes
> with & without the UL OPT. But the heat from the cathode bias resistors
> under the chassis is a serious problem. So lets try somekind of servo
> bias. Perhaps this coming Winter.
>
> I've been busy on the bicycle (like Patrick), more than 4000 Km since
> May 1st. And a lot of physical labor here on the acres. I would rather
> be outside than slaving over a hot soldering iron!
I quite fancy a little rocket. Bradford is full of hills and dangerous
drivers with no insurance. Even walking is scary. I don't have room
for a helicopter.
> For anyone who would like a complete schema of the Kevin O'Connor guitar
> amp referenced email me direct as follows-
>
> jh dot stewart at sympatico dot ca
I replaced Jones's clamp with O'Connor's. Perfect.
> Gotta go now. Today its time to service the summer equipment & put it
> away till Spring! And here is that easy to drive 6AS7/6080 amp for the
> curious.
Summer equipment? In Europe, we have a fire in the sky controlled by
druids with henges. Every time the druids get locked up on drug
charges, the weather falls apart.
> +-------------------------------------------------------------------+
> |Filename: 6BQ7 6080 Amp.jpg |
> |Download:http://www.audiobanter.com/attachment.php?attachmentid=251|
> +-------------------------------------------------------------------+
Thanks. Interesting. I guess there's some maths that goes with the
weirdness? Like the load seen by, and output resistance of, the short-
tailed pair's anodes.
Ian
922c...@audiobanter.com> wrote:
>
> Hi Ian & Others- Yeh, I'm back again after another 5 year stint in
> hitech sales. But I guess for the last time at my age!
>
> Kevin O'Connor is adventurous & perhaps does some things others may not
> approve of. Thus a toroidal OPT in a guitar amp. No problem with servo
> bias holding the lid down.
That might be hotly disputed by those who feel that a softly-
>
> Hi Ian & Others- Yeh, I'm back again after another 5 year stint in
> hitech sales. But I guess for the last time at my age!
>
> Kevin O'Connor is adventurous & perhaps does some things others may not
> approve of. Thus a toroidal OPT in a guitar amp. No problem with servo
> bias holding the lid down.
saturating OPT is a desirable tone option.
> This one not as elegant as that found in the Brook, easily the best I've
> seen in my very long career. It is unique & sure makes a lot of sense.
> How else to get 30 Watts at low distortion out of a pair of 2A3's in
> AB2?
>
> I got interested in servo bias some time ago while doing my thing with
> 6AS7/6080 PP circuits. Some may remember the bootstrapped driver schemes
> with & without the UL OPT. But the heat from the cathode bias resistors
> under the chassis is a serious problem. So lets try somekind of servo
> bias. Perhaps this coming Winter.
>
> I've been busy on the bicycle (like Patrick), more than 4000 Km since
> May 1st. And a lot of physical labor here on the acres. I would rather
> be outside than slaving over a hot soldering iron!
drivers with no insurance. Even walking is scary. I don't have room
for a helicopter.
> For anyone who would like a complete schema of the Kevin O'Connor guitar
> amp referenced email me direct as follows-
>
> jh dot stewart at sympatico dot ca
> Gotta go now. Today its time to service the summer equipment & put it
> away till Spring! And here is that easy to drive 6AS7/6080 amp for the
> curious.
druids with henges. Every time the druids get locked up on drug
charges, the weather falls apart.
> +-------------------------------------------------------------------+
> |Filename: 6BQ7 6080 Amp.jpg |
> |Download:http://www.audiobanter.com/attachment.php?attachmentid=251|
> +-------------------------------------------------------------------+
weirdness? Like the load seen by, and output resistance of, the short-
tailed pair's anodes.
Ian
Patrick Turner
Dec 8, 2011, 6:12:50 AM12/8/11
to
On Dec 6, 11:30 pm, "GRe" <mynamespacedbyd...@xs4all.nl> wrote:
> "Patrick Turner" <i...@turneraudio.com.au> wrote in message
>
> news:f00b9799-c00e-4f50...@q16g2000yqn.googlegroups.com...
> On Nov 29, 8:31 am, "GRe" <mynamespacedbyd...@xs4all.nl> wrote:
>
>
>
>
>
> >> "Patrick Turner" <i...@turneraudio.com.au> wrote in message
>
> >>news:a25664c6-ab6d-4318...@u10g2000prl.googlegroups.com...
>
> >> [...]
>
> >> >The answer to such uneven
> >> >matching amoung a given batch of supposedly NOS Red Lion KT88 is to
> >> >have some additional bias balancing adjust method as seen in another
> >> >pair of Quad-II I rewired where I used EHKT90 instead of KT66, see
> >> >http://www.turneraudio.com.au/quad2powerampmods.htmandproceed to
> "Patrick Turner" <i...@turneraudio.com.au> wrote in message
>
> news:f00b9799-c00e-4f50...@q16g2000yqn.googlegroups.com...
> On Nov 29, 8:31 am, "GRe" <mynamespacedbyd...@xs4all.nl> wrote:
>
>
>
>
>
> >> "Patrick Turner" <i...@turneraudio.com.au> wrote in message
>
> >>news:a25664c6-ab6d-4318...@u10g2000prl.googlegroups.com...
>
> >> [...]
>
> >> >The answer to such uneven
> >> >matching amoung a given batch of supposedly NOS Red Lion KT88 is to
> >> >have some additional bias balancing adjust method as seen in another
> >> >pair of Quad-II I rewired where I used EHKT90 instead of KT66, see
very well indeed. But if they just throw rocks at me, what's the
point?
I just done a pair of Quad-II-Forty amps with a basic 5050 schematic
instead of the original Quad horror.
6SH7 is a really good pentode when Ia becomes high enough, ie, much
more than in any Quad. But the 6SH7 becomes a really super dooper
triode with high gain and low Ra, and it works with 6SN7 LTP very
well.
During the same months I did tha last Quad mods, I also did a re-wire
of an Earl Weston amp which had about the worst damn schematic I have
ever seen in any amp, but after my singing lessons with a 5050 type
schematic, the owner said it sounded better than the Quads. Earl is a
bloke in Victoria making more amps than I ever will, but man, although
his work looks so pretty, there's a mess under the bonnet. My customer
got lotsa smoke instead of music, I could not just repair it; it
needed complete gutting and re-wiring. Earl's transformers were quite
OK and I could give the amp unconditional stability with just the
right phase tweaking networks. Tube line up ended up 1/2 6SN7 for
input SET and then 6SN7 for LTP, then chinese KT66 with 40% UL
operation, it gave about 30Watts, very nice.
When I get time, all the schematics of these mods to amps will be
posted at my site.
At the CERN accelerator, they are looking for a Higs Boson or somink
like that, and today I read a whole page in the newspaper about it. I
didn't understand a word. But, maybe it'd be nice to know what
particle causes Time, and Mass, as then we might manipulate it, and
fly to elswhere at a speed of C squared, or C to the power of
infinity, and we might amble up to God, and ask him how the universe
doth go, and he might tell us he's busy, because he has to look after
an infinite number of universes besides our little one we find
ourselves within. It would not seem unreasonable to think that if some
God does exist who we might comprehend, then He, She, or It might just
know NOW what's going on, and that means light must travel infinitely
faster than it does, and just why it seems like it is stuck at about
300,000km/sec and restricted to this speed is a deep question indeed.
I also read today about nuclear power and how Green Politics is often
so backward in thinking, like they never wanna give up old beliefs
based on BS rather than answers to questions based on facts observed.
As a result of much questioning, blokes have come up with IFRs,
Integral Fast Reactors which use up leftover waste from old fashioned
nuclear power plants. With IFRs, we could avoid greenhouse effects,
and get off fossil fuels, thus be able to use fossil deposits sensibly
without burning them for transport, hot water, cooking etc.
Of course, if we had cheap electricity and no greenhouse effect,
nothing would stop economic growth and population growth so we'd have
to farewell the environment anyway. Food production has to double in
35 years, well, its what ppl will want, and anyway, lots more
questions will need answering as we muddle along, fixin' one problem,
only to find we have another.
But I digress.
Patrick Turner.
rrus...@hotmail.com
Dec 15, 2011, 1:15:37 AM12/15/11
to
> Of course, if we had cheap electricity and no greenhouse effect,
> nothing would stop economic growth and population growth so we'd have
> to farewell the environment anyway. Food production has to double in
> 35 years, well, its what ppl will want, and anyway, lots more
> questions will need answering as we muddle along, fixin' one problem,
> only to find we have another.
Eugenics.
flipper
Dec 15, 2011, 6:41:45 PM12/15/11
to
Patrick Turner
Dec 18, 2011, 5:49:27 PM12/18/11
to
future, and I don't care about it; I won't be in it. If humans alter
their genes to cope with future circumstances ( if they can afford the
therapy under USA style medical system ) then that's their choice
then, not mine now.
Meanwhile, the radioactive shit from Fukashima power plants circles
the Earth in its oceans, and that shit will be joined in future by
lots more shit, because everyone knows shit happens.
It seems most unfortunate that the Japanese bore the brunt of nuclear
madness after copping a dose of genocidal atomic bombings in WW2.
( Boy, if thay don't start a flame war, nothing will.)
"Sir, would you like sunglasses to eat your tuna caught last Friday?
We do realise they glow in the dark..."
Patrick Turner.
0 new messages