2A3 or EL34 SE
ee
Is there really a big difference with the direct heated filiaments?
I like to try the 300B later but I hear the cheap tubes last only 3 months.
What do you nice folks on RAT think?
EE
Thomas Schick
If you take a 2A3 amp for a set of stereo speakers alone, these speakers
have to have Borg efficiency. 94dB+ .
An EL34 single should drive more common speakers as well.89dB+
As with everything, the cheapest thing available might have higher cost than
a quality product bought in first place for more money.
happy amp building
thomas
"ee" <ee1000@*not*hotmail.com> schrieb im Newsbeitrag
news:%H578.32861$_F4.6...@news0.telusplanet.net...
Preamp
<ma...@thomas-schick.com> wrote:
>
>If you take a 2A3 amp for a set of stereo speakers alone, these speakers
>have to have Borg efficiency. 94dB+ .
I'd like to know how you arrive at that figure. I use mine with 90 db
speakers and still get told to turn it down.
>An EL34 single should drive more common speakers as well.89dB+
In triode mode it doesn't do much better than a 2A3.
Preamp
>I am debating on making a 2A3 SE amp or EL34 SE in triode mode.
>Is there really a big difference with the direct heated filiaments?
Having built both (also 6V6 SE amps) I'd have to say the directly
heated tube sounds better, but that may be because it is a true triode
as much as anything else.
Russ Sadd
> I am debating on making a 2A3 SE amp or EL34 SE in triode mode.
> Is there really a big difference with the direct heated filiaments?
Hi,
I've been thinking about these things myself recently. Please note
that this is a work-in-progress and I'm missing some of the key
figures. If anyone's able to contribute more valve data here, I'd be
most grateful!
Triode-strapped output valves
=============================
It is a useful trick to triode-strap a pentode or a tetrode valve,
certainly as far as the hobbyist is concerned. For starters, 'true'
triode power valves are currently ridiculously expensive, with Russian
2A3s costing a good £15-£20 each. Bearing in mind that great-quality
Russian EL34s only cost between £5-£10, there is clearly something
wrong here. Or rather, the 'triode mystique' we seem to have at the
moment is pushing prices up for the simple reason that people expect
to pay more.
But price isn't the only consideration for turning, say, an EL34 into
a useful power triode. You also get the benefits of a standard octal
base, standard 6.3V indirect heaters, greater hum resistance, good
heat dissipation, and higher gain (mu) than our 2A3 might have. I've
listed a comparison of popular power triodes and triode-strapped
tetrodes and pentodes below.
Type Rating Va Ia Ra Mu RL Power
300B 40W 300V 60mA 700R 3.9 3K 6.0W
PX25 400V 63mA 1,000R 7.5 5K5 6.0W
PX4 300V 50mA 830R 6.0 3K5 4.5W
2A3 250V 60mA 800R 5.5 2K5 3.5W
6AS7G 2x13W 135V 125mA 280R 2.0 2K5 3.5Wx2
EL34 25W 340V 70mA 5.2 2K5 7.0W
KT66 25W 360V 63mA 1,450R 5.5 4K5 5.8W
[Rating = anode dissipation, Va = anode voltage, Ia = anode current,
Ra = internal impedance of valve, Mu = approx. voltage gain, RL =
transformer load, Power = single-ended power output]
As you can see from the table, the true triodes are more efficient
than triode-strapped pentodes and beam tetrodes. Typically,
triode-strapped pentodes have a higher anode impedance and require
higher voltages than standard power triodes. Yet the KT66, for
example, isn't too far off from, say, the PX25, nor does it appear to
be too much of a sacrifice compared to the 300B. Nor does the EL34,
for that matter.
Thus we can conclude that EL34s and KT66s / 6L6GCs are an acceptable
substitute for power triodes, as while their triode characteristics
aren't bad, they're simply so much cheaper and - in all likelihood -
more robust. I should probably add here that KT66 replicas are a
rather expensive alternative to DHTs like the Russian 6B4G, as they
cost the same at £25 each. But we might substitute Russian 5881s or
6L6GCs instead.
It should be noted that I quoted the 6AS7G double power triode in this
table. A single half is quite an acceptable alternative to a 2A3, with
the annoyance of its lower gain being offset by its much higher
damping factor (from its anode impedance. A single 6AS7G valve, with
both its internal triodes strapped together, is therefore an
intriguing alternative to a 300B - drive requirements withstanding.
The two triodes in a single 6AS7G can be used in push-pull for some
10W output power, or strapped in parallel for 7W single-ended. 'Hi-Fi
World' magazine once sold a kit using a single 6AS7G/6080 as a stereo
single-ended output stage. Russian 6AS7Gs are not just cheap, they're
rugged - and they look damn good in an amplifier!
Driver triodes from small pentodes
----------------------------------
Of course, you can do exactly the same kind of thing with small-signal
pentodes, such as the 6AU6. These 'radio types' are often very
cost-effective, but when triode-strapped, do offer some intriguing
possibilities. Note in the case of the 6AU6/EF94 (in the table below)
that the anode impedance of 7K7 is comparable to that of a 6J5 or 6SN7
running at the same voltage, the increased anode current over the said
6J5, and even the higher mu of 33 as opposed to 20. Quite a nice
driver valve in place of a 6J5 in a standard ciruit, eh?
Type Rating Va Ia Ra Mu
6C4 250V 10.5mA 7K7 20
6SJ7 2.5W 250V 9.2mA 7K6 20
6AU6 3.5W 250V 12.2mA 7K5 33
EF86 1.0W 250V 3.6mA 26
I hope this has proved of interest.
Best regards,
Russ Sadd
Patrick Turner
Russ Sadd wrote:
And KT88 and 6550 can give 8 watts, from 30 watts of input.
>
>
> [Rating = anode dissipation, Va = anode voltage, Ia = anode current,
> Ra = internal impedance of valve, Mu = approx. voltage gain, RL =
> transformer load, Power = single-ended power output]
>
> As you can see from the table, the true triodes are more efficient
> than triode-strapped pentodes and beam tetrodes.
Eh, 300B, 6W from 40W, = 15% efficiency,2A3, 3.5W from 15W, = 23%,
but EL34, 7W from 25, = 28%,
and KT88 I have found to be up to 30%, with B+ = 500 v.
I have always found real triodes more inefficient.
> Typically,
> triode-strapped pentodes have a higher anode impedance and require
> higher voltages than standard power triodes. Yet the KT66, for
> example, isn't too far off from, say, the PX25, nor does it appear to
> be too much of a sacrifice compared to the 300B. Nor does the EL34,
> for that matter.
Yes,
>
>
> Thus we can conclude that EL34s and KT66s / 6L6GCs are an acceptable
> substitute for power triodes, as while their triode characteristics
> aren't bad, they're simply so much cheaper and - in all likelihood -
> more robust. I should probably add here that KT66 replicas are a
> rather expensive alternative to DHTs like the Russian 6B4G, as they
> cost the same at £25 each. But we might substitute Russian 5881s or
> 6L6GCs instead.
Yes,
>
>
> It should be noted that I quoted the 6AS7G double power triode in this
> table. A single half is quite an acceptable alternative to a 2A3, with
> the annoyance of its lower gain being offset by its much higher
> damping factor (from its anode impedance. A single 6AS7G valve, with
> both its internal triodes strapped together, is therefore an
> intriguing alternative to a 300B - drive requirements withstanding.
> The two triodes in a single 6AS7G can be used in push-pull for some
> 10W output power, or strapped in parallel for 7W single-ended. 'Hi-Fi
> World' magazine once sold a kit using a single 6AS7G/6080 as a stereo
> single-ended output stage. Russian 6AS7Gs are not just cheap, they're
> rugged - and they look damn good in an amplifier!
Yes,
>
>
> Driver triodes from small pentodes
> ----------------------------------
> Of course, you can do exactly the same kind of thing with small-signal
> pentodes, such as the 6AU6. These 'radio types' are often very
> cost-effective, but when triode-strapped, do offer some intriguing
> possibilities. Note in the case of the 6AU6/EF94 (in the table below)
> that the anode impedance of 7K7 is comparable to that of a 6J5 or 6SN7
> running at the same voltage, the increased anode current over the said
> 6J5, and even the higher mu of 33 as opposed to 20. Quite a nice
> driver valve in place of a 6J5 in a standard ciruit, eh?
>
> Type Rating Va Ia Ra Mu
> 6C4 250V 10.5mA 7K7 20
> 6SJ7 2.5W 250V 9.2mA 7K6 20
> 6AU6 3.5W 250V 12.2mA 7K5 33
> EF86 1.0W 250V 3.6mA 26
Some frame grid pentodes are quite exceptional.The king of them is the E280F, a nine pin tube with
a rated Gm maximum of 23 ma per volt....
But they are not common.
So 6EH7 and 6EU7, 6BX6 are also
gutsy pentodes, and they make gutsy triodes to.
>
>
> I hope this has proved of interest.
>
> Best regards,
>
> Russ Sadd
>
> www.griffon.dircon.co.uk
It ain't what you got that matters, its how youuse it.
Patrick Turner.
Thomas Schick
"Preamp" <pre...@antispam.dingoblue.net.au> schrieb im Newsbeitrag
news:3c60e729...@news.dingoblue.net.au...
> On Tue, 5 Feb 2002 18:53:44 +0100, "Thomas Schick"
> <ma...@thomas-schick.com> wrote:
>
> >
> >If you take a 2A3 amp for a set of stereo speakers alone, these speakers
> >have to have Borg efficiency. 94dB+ .
>
> I'd like to know how you arrive at that figure. I use mine with 90 db
> speakers and still get told to turn it down.
90dB is still efficient. However, the less effecient the speaker there are
bigger compression effects with flea powered amps. I love my britsh two way
monitors but with a 2A3 amp it is very limited. They hold something back
Breathtaking it gets with 104dB horns, there you have the ease and magic of
sound we're all after. Played at the same level, subjective of course :-)
>
>
> >An EL34 single should drive more common speakers as well.89dB+
>
> In triode mode it doesn't do much better than a 2A3.
Single ended EL 34 designs deliver 8-10W a channel. 2A3 single designes
around 3-4 aren't they?
Thomas
ee
I'm glad you suggest the 6AS7G but what to drive it?
I like to use 6922, 12AU7 or 12SN7.
BTW, if used in PP which phase splitter has better sonics, cathodyne or
long tailed pair?
I am liking the EL34 more and one can parallel EL34 for more power and low
cost. Or substitute with 6550 for non-horn speakers.
EE
Preamp
<ma...@thomas-schick.com> wrote:
>>
>> >An EL34 single should drive more common speakers as well.89dB+
>>
>> In triode mode it doesn't do much better than a 2A3.
>
>Single ended EL 34 designs deliver 8-10W a channel. 2A3 single designes
>around 3-4 aren't they?
>
EL34's only give 8 to 10 watts in pentode mode. A bit less for UL,
and only 3 to 4 watts in triode mode (just like the 2A3)
Preamp
>EL34 25W 340V 70mA 5.2 2K5 7.0W
That is more than the theoritcal limit of 1/4 plate dissipation. In
practice, I've found after many trial operating points and load lines,
that to keep distortion down to reasonable limits ( a couple of
percent) thaen 3 to 4 watts is all you can really use.
Preamp
>I am liking the EL34 more and one can parallel EL34 for more power and low
>cost. Or substitute with 6550 for non-horn speakers.
>
The problem with paralleling EL34's is that the bias current get's way
high and you need damn huge hunks of iron with big airgaps to cope.
Also, their heater current draw is a little thirsty too.
Choky
--
Choky
Prodanovic Aleksandar
choky*remove*@eunet.yu
YU
remove *remove* to reply!!
===================
"ee" <ee1000@*not*hotmail.com> wrote in message
news:wAf88.47159$_F4.9...@news0.telusplanet.net...
Patrick Turner
Preamp wrote:
You can get more than 4 watts from EL34 in SET.You do need to use a high B+,
about 500 volts,
and a highish load value, and then you can get
a wide voltage swing, as the plate resistance line
at Eg = 0 volts is in about the same place as when
a lower B+ is used, where there is not so wide volt swing.
With KT88, 6550, even more triode power is available,
and these rival the use of 300B, and
are easier to drive.
About 12 watts of SEUL power can be had from KT88/6550....
Patrick Turner
Russ Sadd
> I'm glad you suggest the 6AS7G but what to drive it?
> I like to use 6922, 12AU7 or 12SN7.
Hi,
When I built a push-pull 6AS7G amplifier, I deliberately wanted to use
a single 250V HT line. Therefore I used a transformer-coupled 6J5 with
a step-up transformer (1:2+2) to get the voltage swing. Other designs,
such as Svetlana's, use a second, higher HT line and conventional
driver stages. It should still be possible with 6922s or 6SN7s and a
'normal' HT line, and I'll hunt out some schematics to send you.
But I can only vouch for mine out of personal experience. (And if you
can wait 24 hours, there'll be a write-up of this design on my
website.)
> BTW, if used in PP which phase splitter has better sonics, cathodyne or
> long tailed pair?
Theoretically, they should be identical. In practice, I've only ever
had problems with the cathodyne, and find the long-tailed pair more
robust. Plus, of course, you can very easily tweak the long-tailed
pair for better performance (such as using a pentode on the cathode to
provide an 'infinite impedance' as Lowther used to do, or to have the
driver stage mounted on the anodes of the long-tailed pair in
cascode). Personally, though, I can't be bothered to fuss around with
these things, and go straight for a centre-tapped driver transformer!
> I am liking the EL34 more and one can parallel EL34 for more power and low
> cost. Or substitute with 6550 for non-horn speakers.
6W as a Class A triode is damn good, so you don't need horns with an
EL34. I'm currently working on a 'single-ended acoustical' EL34 to get
10W out of a single valve, and that's perfectly okay for modern
loudspeakers. Triode-strapped 6550s aren't as good as EL34s, I should
add: the KT66 and EL34 are very unusual in that they were *designed*
to be triode-strapped, and thus are particularly linear. Or sweet, if
you prefer to think with your ears rather than with the figures.
Anthony Cirella
ee <ee1000@*not*hotmail.com> wrote:
Haven't seen this mentioned....
Use a EL34, and you can easily play with
the octal based 6.3v version of the 2A3...
the 6B4G, no socket change required.
6B4G is a bit hummier than the 2A3,
but keep the filament voltage to
6.2-6.3vac, and you'll minimize
the hum. NOS 6B4G are far cheaper than
NOS 2A3.
My experience seems to indicate
the new EH 300B is ok lifewise.
A.Cirella, Handmade
A.Cirella
Handmade Electronics
Parts, troubleshooting guide.
www.hndme.com
Preamp
<in...@turneraudio.com.au> wrote:
>
>
>Preamp wrote:
>
>> On Wed, 6 Feb 2002 19:22:22 +0100, "Thomas Schick"
>> <ma...@thomas-schick.com> wrote:
>>
>> >>
>> >> >An EL34 single should drive more common speakers as well.89dB+
>> >>
>> >> In triode mode it doesn't do much better than a 2A3.
>> >
>> >Single ended EL 34 designs deliver 8-10W a channel. 2A3 single designes
>> >around 3-4 aren't they?
>> >
>>
>> EL34's only give 8 to 10 watts in pentode mode. A bit less for UL,
>> and only 3 to 4 watts in triode mode (just like the 2A3)
>
>You can get more than 4 watts from EL34 in SET.You do need to use a high B+,
>about 500 volts,
>and a highish load value, and then you can get
>a wide voltage swing, as the plate resistance line
>at Eg = 0 volts is in about the same place as when
>a lower B+ is used, where there is not so wide volt swing.
>
What distortion figures do you predict. I found that it started to
look ugly up there :-)
Patrick Turner
Preamp wrote:
> On Thu, 07 Feb 2002 19:45:44 +1100, Patrick Turner
> <in...@turneraudio.com.au> wrote:
>
> >
> >
> >Preamp wrote:
> >
> >> On Wed, 6 Feb 2002 19:22:22 +0100, "Thomas Schick"
> >> <ma...@thomas-schick.com> wrote:
> >>
> >> >>
> >> >> >An EL34 single should drive more common speakers as well.89dB+
> >> >>
> >> >> In triode mode it doesn't do much better than a 2A3.
> >> >
> >> >Single ended EL 34 designs deliver 8-10W a channel. 2A3 single designes
> >> >around 3-4 aren't they?
> >> >
> >>
> >> EL34's only give 8 to 10 watts in pentode mode. A bit less for UL,
> >> and only 3 to 4 watts in triode mode (just like the 2A3)
> >
> >You can get more than 4 watts from EL34 in SET.You do need to use a high B+,
> >about 500 volts,
> >and a highish load value, and then you can get
> >a wide voltage swing, as the plate resistance line
> >at Eg = 0 volts is in about the same place as when
> >a lower B+ is used, where there is not so wide volt swing.
> >
>
> What distortion figures do you predict. I found that it started to
> look ugly up there :-)
You should obtain better overall figures.Say you set up so the power out is 7
watts in triode,
and 5% to 10% is the 2H thd.
Then at 4 watts, the amp will be lot better than the amp
which is struggling to make 4 watts, and often with the same %
of thd, at the lower B+, and lower value load.
Remember, that when you make the load line as flat as possible,
when you draw it up on the plate curves, the 2H distorion
gets less, and is minimum when the load line is
a horizontal line.
But the load line cannot be horizontal in a power amp, because with a horizontal
load line, there is no current change, and power is
only produced when there is a current change,
and PO = I x V, across the load.
Getting triodes to sing for their supper is always a compromise
and balancing act.
Patrick Turner.
ee
> Theoretically, they should be identical. In practice, I've only ever
> had problems with the cathodyne, and find the long-tailed pair more
> robust.
> 6W as a Class A triode is damn good, so you don't need horns with an
> EL34. I'm currently working on a 'single-ended acoustical' EL34 to get
> 10W out of a single valve, and that's perfectly okay for modern
> loudspeakers. Triode-strapped 6550s aren't as good as EL34s, I should
> add: the KT66 and EL34 are very unusual in that they were *designed*
> to be triode-strapped, and thus are particularly linear. Or sweet, if
> you prefer to think with your ears rather than with the figures.
I'm actually planning some PP also and maybe 4x EL34 work better than 2x
6550 and need 70W for ribbon full rangers.
The 6C33C looks more impressive than a 300B!
Must be a pig with regards to iron and runs hot.
Any caveats using this beast?
EE
Thermionic
<in...@turneraudio.com.au> wrote:
>
>
>>
>> What distortion figures do you predict. I found that it started to
>> look ugly up there :-)
>
>You should obtain better overall figures.Say you set up so the power out is 7
>watts in triode,
>and 5% to 10% is the 2H thd.
Yeah, that's what I mean... ugly!
>Then at 4 watts, the amp will be lot better than the amp
>which is struggling to make 4 watts, and often with the same %
>of thd, at the lower B+, and lower value load.
Maybe, but then it isn't a 7 watt amp, just a better 4 watt one.
>Remember, that when you make the load line as flat as possible,
>when you draw it up on the plate curves, the 2H distorion
>gets less, and is minimum when the load line is
>a horizontal line.
>But the load line cannot be horizontal in a power amp, because with a horizontal
>load line, there is no current change, and power is
>only produced when there is a current change,
>and PO = I x V, across the load.
>
>Getting triodes to sing for their supper is always a compromise
>and balancing act.
Yep, you can spend hours trying load lines and op points on a
spreadsheet trying to get that little bit better out of it.
Patrick Turner
Thermionic wrote:
> On Fri, 08 Feb 2002 00:55:13 +1100, Patrick Turner
> <in...@turneraudio.com.au> wrote:
>
> >
> >
> >>
> >> What distortion figures do you predict. I found that it started to
> >> look ugly up there :-)
> >
> >You should obtain better overall figures.Say you set up so the power out is 7
> >watts in triode,
> >and 5% to 10% is the 2H thd.
>
> Yeah, that's what I mean... ugly!
>
> >Then at 4 watts, the amp will be lot better than the amp
> >which is struggling to make 4 watts, and often with the same %
> >of thd, at the lower B+, and lower value load.
>
> Maybe, but then it isn't a 7 watt amp, just a better 4 watt one.
Say you have the tube set up with 500 volts, and you get 7 watts with7% with the
load.
At 3.5 watts, distortion is a lot better.
But average listening might be at 1watt, and then there is this
high ceiling for transient peaks, drum beats and so on,
and so that ceiling surely is worth having.
Alternatively, you could increase the load value, and get 4 watts,
at 3% and have not such a high ceiling, so average levels are
going to have to be kept to half a watt.
It depends on how loud you want to go.
Patrick Turner.
Thermionic
<in...@turneraudio.com.au> wrote:
>
>
>Thermionic wrote:
>
>> On Fri, 08 Feb 2002 00:55:13 +1100, Patrick Turner
>> <in...@turneraudio.com.au> wrote:
>>
>> >
>> >
>> >>
>> >> What distortion figures do you predict. I found that it started to
>> >> look ugly up there :-)
>> >
>> >You should obtain better overall figures.Say you set up so the power out is 7
>> >watts in triode,
>> >and 5% to 10% is the 2H thd.
>>
>> Yeah, that's what I mean... ugly!
>>
>> >Then at 4 watts, the amp will be lot better than the amp
>> >which is struggling to make 4 watts, and often with the same %
>> >of thd, at the lower B+, and lower value load.
>>
>> Maybe, but then it isn't a 7 watt amp, just a better 4 watt one.
>
>Say you have the tube set up with 500 volts, and you get 7 watts with7% with the
>load.
>At 3.5 watts, distortion is a lot better.
>But average listening might be at 1watt, and then there is this
>high ceiling for transient peaks, drum beats and so on,
>and so that ceiling surely is worth having.
Maybe, I'm not convinced a high ceiling is at all required. I have 2
watt 6aq5 'SET' that is super resposive to transients. Doesn't seem
to know it's ceiling is low...
Patrick Turner
Thermionic wrote:
> > >Then at 4 watts, the amp will be lot better than the amp
> >> >which is struggling to make 4 watts, and often with the same %
> >> >of thd, at the lower B+, and lower value load.
> >>
> >> Maybe, but then it isn't a 7 watt amp, just a better 4 watt one.
> >
> >Say you have the tube set up with 500 volts, and you get 7 watts with7% with the
> >load.
> >At 3.5 watts, distortion is a lot better.
> >But average listening might be at 1watt, and then there is this
> >high ceiling for transient peaks, drum beats and so on,
> >and so that ceiling surely is worth having.
>
> Maybe, I'm not convinced a high ceiling is at all required. I have 2
> watt 6aq5 'SET' that is super resposive to transients. Doesn't seem
> to know it's ceiling is low...
Well, if it sounds right, stay with it.
Patrick Turner.
Denis
>
> Triode-strapped output valves
> =============================
<snip>
>
> Type Rating Va Ia Ra Mu RL Power
> 300B 40W 300V 60mA 700R 3.9 3K 6.0W
> PX25 400V 63mA 1,000R 7.5 5K5 6.0W
> PX4 300V 50mA 830R 6.0 3K5 4.5W
> 2A3 250V 60mA 800R 5.5 2K5 3.5W
> 6AS7G 2x13W 135V 125mA 280R 2.0 2K5 3.5Wx2
> EL34 25W 340V 70mA 5.2 2K5 7.0W
> KT66 25W 360V 63mA 1,450R 5.5 4K5 5.8W
>
> [Rating = anode dissipation, Va = anode voltage, Ia = anode current,
> Ra = internal impedance of valve, Mu = approx. voltage gain, RL =
> transformer load, Power = single-ended power output]
>
> As you can see from the table, the true triodes are more efficient
> than triode-strapped pentodes and beam tetrodes. Typically,
> triode-strapped pentodes have a higher anode impedance and require
> higher voltages than standard power triodes. Yet the KT66, for
> example, isn't too far off from, say, the PX25, nor does it appear to
> be too much of a sacrifice compared to the 300B. Nor does the EL34,
> for that matter.
>
These triode-strapped pentodes are always more nonlinear than classic
triodes.
It is evident from their sets of output I-V curves. They are obviously
not parallel to each other and are unevenly displaced for equal grid
voltage increments.
Such triode as 300B is nonlinear too, but to lesser extent. At least
it's curves look like those of ideal triode for a first glance. The
2A3 is the same.
Nevertheless this is another circuit topology, which can give
excellent results with pentodes, it is cathode feedback circuit.
I think it is possible to construct a CFB amplifier, which will
develop 10W at 2.5% of THD(almost H2) on a single EL34. The distortion
at typical listening levels(below 1W) will be just a fraction of a
percent, and this will be almost pure second harmonic. The driver
requirements also will be moderate, below 35Vrms, so it seems possible
to construct an amplifier with just two stages, using something like
6SL7 as a driver.
Denis Afanassyev. www.cortmi.com.ua/omak
Patrick Turner
Denis wrote:
Not necessarily so
Take a look at a 12BH7 plate curves.
Then look at 12AT7,
and then look at 6SL7,
or 12AX7.
Despite the fact these are all classic triodes,
there is a great deal of variation in the shape of each plate
resistance line at each value of Eg bias.
The 12AX7 is far more linear than a 12BH7.
> Such triode as 300B is nonlinear too, but to lesser extent. At least
> it's curves look like those of ideal triode for a first glance. The
> 2A3 is the same.
And so does an 807 when strapped as a triode.
> Nevertheless this is another circuit topology, which can give
> excellent results with pentodes, it is cathode feedback circuit.
>
> I think it is possible to construct a CFB amplifier, which will
> develop 10W at 2.5% of THD(almost H2) on a single EL34. The distortion
> at typical listening levels(below 1W) will be just a fraction of a
> percent, and this will be almost pure second harmonic. The driver
> requirements also will be moderate, below 35Vrms, so it seems possible
> to construct an amplifier with just two stages, using something like
> 6SL7 as a driver.
I agree entirely, my first amps were SE EL84 in an old stereoradiogram in which I connected the output
secondary of the
OPT into the cathode circuit. Then since that worked
so easily, I was able to fit in another layer of wire around the OPT,
and added that into the feedback loop to make about
15 % of cathode feedback, and with a 12AX7, I had a reasonably
low Zout amp.
For the average DIY person, this is too hard, and finding
suitable OPTs is impossible, so the alternative is to use
a resistor network from the plate of the pentode,
to the cathode of the driver tube,
and the same results are gained, with the added benefit that the driver tube
does not have to produce a large drive voltage, hence produces low distortion,
and the driver is included in the feedback loop, so whatever distortion is produced
is by the driver is also reduced.
However, the output tube isn't getting the effective screen feedback when there is no
cathode feedback, and a pentoad distortion is worse in high order
products than a "de facto ultralinear" amp which gets its screen feedback
by means of the difference of signal voltage between cathode and screen,
due to the CFB.
In fact the original Hafler and Keroes article on UL in about 1953
talked around all this, with PP circuits.
Quad II has 10% of the total voltage fed back at the cathode,
so the power of the tetrode is maintained,
and the low Ro of the triode is gained.
Under large signal excursions, in class AB,
there are some horrid high odd order products still made,
in PP, but I have found a set of screen taps for UL operation
as well as CFB , will work with the best linerarity of all.
But you need a good OPT.
With SE, and all class A, such glitches in the transfer curve
from class AB operation are not present.
There's only one bother, and that is that some people say
it doesn't sound as good as a 2A3, and they also
say Hi-fi stopped advancing
when tetrodes and pentodes were invented.
Who am I to argue, as I don't like having mud thrown at me.
Patrick Turner http://www.turneraudio.com.au
S. Rahaman
2A3 over anything for the money.
Eike Lantzsch
What about General Electric 8417? I have a pair of those available.
Didn't decide what to do with them up to now.
Regards, Ike
--
. . Eike Lantzsch ZP6CGE
<___> CdC 1519 ASUNCION PY
____|\|\ ____ .ooo../ www.geocities.com/cgelantzsch/index.html
_________\ \_____/ Found a rattlesnake on my backdoor yesterday
Denis
> >
> > These triode-strapped pentodes are always more nonlinear than classic
> > triodes.
> > It is evident from their sets of output I-V curves. They are obviously
> > not parallel to each other and are unevenly displaced for equal grid
> > voltage increments.
>
> Not necessarily so
>
> Take a look at a 12BH7 plate curves.
> Then look at 12AT7,
> and then look at 6SL7,
> or 12AX7.
> Despite the fact these are all classic triodes,
> there is a great deal of variation in the shape of each plate
> resistance line at each value of Eg bias.
> The 12AX7 is far more linear than a 12BH7.
>
Yes, the majority of small double triodes are not the best ones
considering the linearity. They can operate with quite low distortions
in small-signal stages with carefully chosen resistive loads. But the
large-signal mu nonlinearity is important issue for such applications
as the output and driver stages. I prefer such triodes as 6SN7/6N8S,
6N23P-EV(6922) and 6N1P-EV for use in large-signal gain stages.
> > Such triode as 300B is nonlinear too, but to lesser extent. At least
> > it's curves look like those of ideal triode for a first glance. The
> > 2A3 is the same.
>
> And so does an 807 when strapped as a triode.
I haven't tried 807 in CFB circuit, but it seems that this tube also
will be better in CFB than triode strapping.
The most significant advantage of CFB is that all job is done within a
single gain stage. And there is a freedom in choosing the screen
voltage. The reduction of screen voltage makes fantastic changes in
THD spectrum of a pentode, and allows one to reduce the 3rd harmonic
together with other odd ones almost to nothing.
I can design and make the wideband CFB transformers for any special
application for competitive prices. Therefore all interisting DIY
enthusiasts can write to me.
> In fact the original Hafler and Keroes article on UL in about 1953
> talked around all this, with PP circuits.
> Quad II has 10% of the total voltage fed back at the cathode,
> so the power of the tetrode is maintained,
> and the low Ro of the triode is gained.
> Under large signal excursions, in class AB,
> there are some horrid high odd order products still made,
> in PP, but I have found a set of screen taps for UL operation
> as well as CFB , will work with the best linerarity of all.
> But you need a good OPT.
> With SE, and all class A, such glitches in the transfer curve
> from class AB operation are not present.
The outburst of odd harmonics is typical for class AB operation either
with pentodes or triodes. The nonperfect OPT has also some
contribution to it.
Again the CFB tecnology allows one to reduce the high-order harmonic
content for pentodes even in class AB.
>
> There's only one bother, and that is that some people say
> it doesn't sound as good as a 2A3, and they also
> say Hi-fi stopped advancing
> when tetrodes and pentodes were invented.
>
From other hand I know people who claim quite opposite.
Denis Afanassyev. www.cortmi.com.ua/omak
Patrick Turner
Denis wrote:
> Patrick Turner <in...@turneraudio.com.au> wrote in message news:<3C6515B3...@turneraudio.com.au>...
>
> > >
> > > These triode-strapped pentodes are always more nonlinear than classic
> > > triodes.
> > > It is evident from their sets of output I-V curves. They are obviously
> > > not parallel to each other and are unevenly displaced for equal grid
> > > voltage increments.
> >
> > Not necessarily so
> >
> > Take a look at a 12BH7 plate curves.
> > Then look at 12AT7,
> > and then look at 6SL7,
> > or 12AX7.
> > Despite the fact these are all classic triodes,
> > there is a great deal of variation in the shape of each plate
> > resistance line at each value of Eg bias.
> > The 12AX7 is far more linear than a 12BH7.
> >
>
> Yes, the majority of small double triodes are not the best ones
> considering the linearity. They can operate with quite low distortions
> in small-signal stages with carefully chosen resistive loads.
Or even better, when they are in balanced circuits,with a CCS common cathode load, or when they
have CCS plate loads, ie, with a virtually horizontal load line.
> But the
> large-signal mu nonlinearity is important issue for such applications
> as the output and driver stages. I prefer such triodes as 6SN7/6N8S,
> 6N23P-EV(6922) and 6N1P-EV for use in large-signal gain stages.
So do I.
> > > Such triode as 300B is nonlinear too, but to lesser extent. At least
> > > it's curves look like those of ideal triode for a first glance. The
> > > 2A3 is the same.
> >
> > And so does an 807 when strapped as a triode.
And so do most other pentode and tetrode output tubes,
> I haven't tried 807 in CFB circuit, but it seems that this tube also
> will be better in CFB than triode strapping.
It will be almost identical to KT66, 6L6, etc.CFB and UL applications can be used with all power
pentodes and beam tetrodes.
I don't see much evidence than very many folks
can tell the difference with their ears, in realistic
tests.
One thing is for sure, if you look at the curves for 300B, 2A3,
and most pentode and tetrodes when in triode,
I really don't see how anyone could tell which tube curves
were which, going on the shape and spacing
of lines alone, with all the other numbers missing off the graphs.
And the THD graphs for each and every harmonic
in a similarly loaded condition don't reveal too much either,
between "real" triodes, and "fake" ones.
Then there are the various manufacturer variations in THD.....
Not necessarily so I am afraid. And biasing has a huge effect,ie, class of operation, on the 3H and other
odd orders.
I have found keeping the screen supplied at the same DC potential
rather than have a lower screen voltage keeps THD to the minimum.
But again, there is no general rule, as some tubes MUST
have a lower screen supply when used as pentodes, or tetrodes,
like the plain old 6L6, when a high plate voltage supply, of say
600 volts is used.
CFB usage to its best advantage depends on optimised OPT geometry
so that the the cathode windings are tightly coupled to the plate
windings and the secondaries as well, and placement
of the CFB windings relative to both other windings
is the result of careful choice, and nobody knows
much about how to do it.
I would only suggest that people try to use local CFB
if they know the quality of the proposed OPTs is blameless.
So one has to know all the details of the wind up,
and how all the windings are arranged, layer by layer,
with all the materials and distances and turn nubers given.
Rarely do you see this info, because hardly anyone
understands any of it, or they would misinterpret what was said.
I don't mind telling the world my OPT secrets,
and I hope somebody winds some samples,
perhaps they can sell them,
but it is unlikely, as these items require a very dedicated
approach, which was routine in 1955, but all too rare these days.
>
>
> I can design and make the wideband CFB transformers for any special
> application for competitive prices. Therefore all interisting DIY
> enthusiasts can write to me.
>
>
> > In fact the original Hafler and Keroes article on UL in about 1953
> > talked around all this, with PP circuits.
> > Quad II has 10% of the total voltage fed back at the cathode,
> > so the power of the tetrode is maintained,
> > and the low Ro of the triode is gained.
> > Under large signal excursions, in class AB,
> > there are some horrid high odd order products still made,
> > in PP, but I have found a set of screen taps for UL operation
> > as well as CFB , will work with the best linerarity of all.
> > But you need a good OPT.
> > With SE, and all class A, such glitches in the transfer curve
> > from class AB operation are not present.
>
> The outburst of odd harmonics is typical for class AB operation either
> with pentodes or triodes. The nonperfect OPT has also some
> contribution to it.
> Again the CFB tecnology allows one to reduce the high-order harmonic
> content for pentodes even in class AB.
But only by the amount of gain reduction.If you have 5% of THD and the CFB reduces the gain by 6dB,
then that 5% gets reduced to about 2.6%, and the mix
of high order products INCREASES.
The Zout will be reduced far more because Zout change
is related to Mu times beta, not gain times beta,
and if the mu of the pentoad is high, then the
platr resistance, or Zo, will be reduced a lot.
Triode output tubes don't benefit much from CFB,
their Zout is already low, and 10% of CFB might only reduce
Zo from 2 kOhms to 1.4 kOhms, in a typical amp.
So triodes don't need CFB, they are already
nicely linear, but suffer power limitations.
But they make simple nice amps.
> >
> > There's only one bother, and that is that some people say
> > it doesn't sound as good as a 2A3, and they also
> > say Hi-fi stopped advancing
> > when tetrodes and pentodes were invented.
> >
>
> From other hand I know people who claim quite opposite.
>
> Denis Afanassyev. www.cortmi.com.ua/omak
It all depends on how you use what you have got.
Patrick Turner.