2A3 Monoblocks

What people are working on at the moment
User avatar
Mike H
Amstrad Tower of Power
Posts: 20157
Joined: Sat Oct 04, 2008 5:38 pm
Location: The Fens
Contact:

#331

Post by Mike H »

Image Is that a 'keeper' now then?
 
"No matter how fast light travels it finds that the darkness has always got there first, and is waiting for it."
Cressy Snr
Amstrad Tower of Power
Posts: 10552
Joined: Wed May 30, 2007 12:25 am
Location: South Yorks.

#332

Post by Cressy Snr »

Mike H wrote:Image Is that a 'keeper' now then?
Aye.........it's a definite 'keeper'

Could do with a preamp or active buffer stage now to eliminate those dual volume controls.

Now that I know a bit about regulators I might be able to better my previous efforts with preamps.

I think that might be the best use of those wooden chassis our Ant built.
Cressy Snr
Amstrad Tower of Power
Posts: 10552
Joined: Wed May 30, 2007 12:25 am
Location: South Yorks.

#333

Post by Cressy Snr »

I had a little go with fixed bias on the mono blocks today.

I made a couple of full-wave rectified adjustable bias supplies, that took their juice from the 75V bias windings on the Danbury mains transformers.

OK, so the grid resistors could be no larger than 50K, which with 220nF coupling caps gave a crappy 15Hz -3dB point.

Another problem was the fact that dropping the 6B4G to 250V with -43V bias
compromised the driver stage swing as the HT to the drivers was reduced too.

However, despite all this, which of course led to a lack of grunt from the driver, the sound was extremely nice;
the best treble I have heard ever from this setup. It was a bit lifeless because of the above compromises,
but fixed bias seems worth pursuing, if the driver stage is supplied with a more suitable HT voltage.

Of course I've put it all back to cathode bias until I figure out a way to give the driver a higher HT voltage without exposing it once more to the sound of the PSU caps.
Sgt. Baker started talkin’ with a Bullhorn in his hand.
Cressy Snr
Amstrad Tower of Power
Posts: 10552
Joined: Wed May 30, 2007 12:25 am
Location: South Yorks.

#334

Post by Cressy Snr »

I've had a lovely day playing around with the driver stage.
In the last post I was examining the idea of running fixed bias on the 6B4g output stage. The main motivation was to remove the last electrolytic cap in the signal path,
ie the 100uF bypass on the power valve cathode.

Problem was, if the output stage was put in fixed bias, then the driver stage HT would be too low.
It was easy enough to up the driver HT by connecting its anode to the point before the regulators. However, this would have defeated the object
of removing the bypass off the 6B4G, as the driver would then be completely exposed to the influence of the power supply electrolytics.
With the 12HG7 driver then, we had a no-win situation. It was either ditch the driver or abandon the fixed bias idea completely,
but to ditch the driver would put the synergy of the whole amp at risk and there was a very real chance that the whole magic I had been used to,
would evaporate, leaving me with just another nice 2A3 type amp.

I had to find something that would work at a lower voltage than the 12HG7 was happy at. I like the sound of pentode drivers, but of the stock I had, there was nothing remotely suitable.
It was then that the idea came, why not make a cascode instead? I'd have all the advantages of a pentode front end but would be able to use a low voltage twin triode instead of the 12HG7.

So it was that the driver stage became an E88CC cascode.
A few calcs on a Post-It note and the cascode front end was born.
A couple of hours later, I had both amps running cascodes as drivers.
What I was not prepared for was the sound that came out of them.
I had been expecting a different presentation, but the sound was identical
to that obtained with the 12HG7. There were no differences that I could hear, which was a relief I can tell you.
The main difference was the current, the cascode ran at; 10mA for the E88CC rather than the 50mA, of the pentode.
The cascode had tons of gain; too much really, so I decided to take the upper grid feed from the upper anode, rather than the B+, altering
the potential divider to suit.
This gave a little bit of plate to grid feedback, which was sufficient to drop the gain down to about the same as I had had with the pentode driver - neat.
Might have reduced distortion a bit too, you never know.

Anyway here is about version 1000 of the circuit for the amplifiers :wink:

Image

As can be seen, each half of the cascode runs at around 100V, which is a nice region for the E88CC to work in. There are no bypass caps anywhere
at the front of the amp, so nothing detrimental has been introduced. The HT to the cascode could be reduced without any problems,
meaning that fixed bias could be employed at the output end of the amps, and the driver could still operate from the regulators, without being exposed to the PSU electrolytics.

The E88CC cascode drivers seem to have a particular affinity with the 6B4Gs. They compliment each other and together, make a bloody superb sound, though I say so meself.

Sod the high output Z, It works beautifully , I like, so there :)
Sgt. Baker started talkin’ with a Bullhorn in his hand.
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8867
Joined: Mon May 21, 2007 9:42 pm

#335

Post by Paul Barker »

well done for thinking outside the box.
User avatar
ed
retired
Posts: 5384
Joined: Thu Jun 21, 2007 4:01 pm
Location: yorkshire
Contact:

#336

Post by ed »

ooh a cold shiver, a cold breeze or something...

Steve, have you measured the voltage swing at the couple for any given input....I'm itching to know..

how do you determine the load....I've got some reading to do
There's nowhere you can be that isn't where you're meant to be
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8867
Joined: Mon May 21, 2007 9:42 pm

#337

Post by Paul Barker »

I am not sure about Steve's feedback way or what effect it has on the operation of the cascode.

But the operation of a cascode is that the bottom triode is a voltage to current converter so the gm of the valve determines the current flow. It sets the current which the upper triode operates at. the gain of the lower triode is very low, whereas the upper triode has it's normal gain so the two multiplied together give much greater gain. An e88cc would go over 100 an ecc82 would be around 70.

You can tune the sound by operating the upper valve with a lowish load resistor as though it were a pentode. Then you get more of the pentode sound. If you operate it as though it were a triode you get a more triode sound.

the bottom valve is inverting and the top valve is non inverting, so the phase is as though it were a triode or pentode stage.

I would suggest it would take a fair amount of experimenting to tune in the right anode load for the sound you like, it might actually pay to put a trim pot there and adjust on test.

the bias point for the upper triodes grid, which is the screen grid of the pentode should really be dealt with as if it were a screen grid with a voltage devider from B+ to ground, seting it up so that it provides the correct bias voltage of the upper triode.

It is the sort of circuit you have to adjust on test.
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8867
Joined: Mon May 21, 2007 9:42 pm

#338

Post by Paul Barker »

If steve's problem is gain then I would consider using the 5687 instead of the E88CC and also reduce the value of the load resistor to suit the gain required. this would also give the more pentode like sound.

The main reason for the circuit is that it supposed to be less noisy than a true pentode.

Morgan Jones used the e88cc as an example and I think he actually said the valve was purpose designed for that purpose. I imagine that in many uses the extra gain with low noise was the desired effect. Steve is unique in not requiring the gain. but as said that could be addressed by a different valve choice like 5687 and reduced anode load, aslong as you don't go round the anode bend of the pentode which cause horrible clipping. hoever if your voltage swing never takes it round the bend you would not suffer from the clipping noises anyway.
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8867
Joined: Mon May 21, 2007 9:42 pm

#339

Post by Paul Barker »

I would just add that if I were using this to drive a power valve I would intercept it with my new favourite that I can't imagine not using the 6em7 power section as a cathode follower, most likely direct coupled to the output valve.

It is now a standard in everything I do.
Cressy Snr
Amstrad Tower of Power
Posts: 10552
Joined: Wed May 30, 2007 12:25 am
Location: South Yorks.

#340

Post by Cressy Snr »

Paul Barker wrote:
Morgan Jones used the e88cc as an example and I think he actually said the valve was purpose designed for that purpose.
Yep that's why I used the E88CC as it had all the necessary theory and
worked examples in MJ's book. Of course I did not run my cascode at the 1-and-a-bit milliamps that Jones did with his example.
I ran mine at 10mA which enabled me to understand better what I was doing, by scaling all the values by 10, except the
potential divider, which was altered when I incorporated the smidge of feedback.

I've tried the 6N1P and the 6N6P in the circuit with good results, but prefer the E88CC at the moment.
After a few hours listening, I have noticed a difference between the sound of the cascode and the sound of the 12HG7.
The cascode in its current configuration seems to give a slightly "easier" sound, maybe because at the moment it is pushed
more toward the triode end of its operation than the pentode. Whatever the cause I really do like it.
Sgt. Baker started talkin’ with a Bullhorn in his hand.
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8867
Joined: Mon May 21, 2007 9:42 pm

#341

Post by Paul Barker »

Well this is what it's all about finding the best sound with lots of trial and error.

I am sure that little ultralinear modification of yours is a big part of the triode nature you have discovered. I don't remember seeing that in Morgan. Where did you get that idea from? Your own? Very clever.
Cressy Snr
Amstrad Tower of Power
Posts: 10552
Joined: Wed May 30, 2007 12:25 am
Location: South Yorks.

#342

Post by Cressy Snr »

Paul Barker wrote:Well this is what it's all about finding the best sound with lots of trial and error.

I am sure that little ultralinear modification of yours is a big part of the triode nature you have discovered. I don't remember seeing that in Morgan. Where did you get that idea from? Your own? Very clever.
Ultralinear eh? Well I suppose it is.

I did not think of it as ultralinear. I was thinking of it more as Schade. What led me to the feedback option, to pull down the gain a tad,
was using the knowledge I gained from last year, when I was messing with single-ended pentode power valves, with plate-to-grid-feedback, a la Schade and latterly, Alexandar Kitic's SEP designs.

My thinking was, that if plate-to-grid feedback could be done to an output stage then why not an input stage? The fact that it worked first time and did what I wanted it to do was quite a nice bonus. :)

I suppose as is my wont, I've got it right but for all the wrong reasons :roll:
Sgt. Baker started talkin’ with a Bullhorn in his hand.
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8867
Joined: Mon May 21, 2007 9:42 pm

#343

Post by Paul Barker »

Just done a search on it and Broskie also cals it Ultralinear.

Image

What is most interesting is his explanation of the mechanism. It explains I think why you find it better. That is because you are using it as a driver and not just a voltage amplifier. (though I would still buffer it)

"The Ultra-Linear Cascode
This circuit approximates the ultra-linear pentode circuit by using two triodes. Like the pentode version it boasts both lower output impedance and distortion than otherwise, but differs in that the top triode's grid has a very high input impedance that is easy to drive with just a simple voltage divider.
A further advantage this arrangement holds over the stock Cascode is an increased maximum output swing, as the top triode's plate swings down, so too its cathode moves down. This increase in output headroom would not be needed in a preamp, but in the first stage of an OTL power amplifier it would prove significantly better than a stock Cascode circuit." Broskie, J (2000) Tube Cad journal Vol 2 No1 Pg4

What I find really interesting is the statement he made in the preceding UL preamp circuit where he uses an actual pentode, the top UL part of the voltage devider is a cathode follower it is biased through a potentiometer. The benefit is best said by him:

" The second tweak is to replace the voltage divider resistors with a single potentiometer. The potentiometer allows for some amazing adjustment to the pentode's mode of operation: at one extreme we have pure pentode; at the other, pure triode. In between these two extremes lies ultra-linear operation.
What is the best potentiometer position? Best for what? Best for rock and roll or best for opera? Best for what amplifier? What speaker? What day? What wine? What mood? This control gives the user a choice over the sound quality that is not frequency dependant. " As above for reference earlier in page.

Steve this sort of control is right up your street! In your existing circuit you could make your anode load variable but adjecent to a fixed resistor to prevent adjustment too far, and the same for the 100k + 100k voltage devider. Wiper to top triode (screen) grid cap from wiper to ground or bottom triode cathode.

Tune in your favourite sound on the day for the genre you are playing.

Here is Broskies pentode version with cathode follower supporting the grid control but guided by the pot into position. In your case as the top triode grid is an easy load the Cf is non essential. It gets it;s UL connection fia the CF's control grid, it is effectively the same as a wiper direct to the screen grid but gives better drive.

Image

Notice he buffers the output!
Andrew
Eternally single
Posts: 4206
Joined: Thu May 24, 2007 2:18 pm

#344

Post by Andrew »

Steve,

You night find this interesting info on your circuit.

http://www.tubecad.com/january2000/page5.html

Andrew
Analogue, the lost world that lies between 0 and 1.
Cressy Snr
Amstrad Tower of Power
Posts: 10552
Joined: Wed May 30, 2007 12:25 am
Location: South Yorks.

#345

Post by Cressy Snr »

Paul Barker wrote:Just done a search on it and Broskie also cals it Ultralinear.

Image

What is most interesting is his explanation of the mechanism. It explains I think why you find it better. That is because you are using it as a driver and not just a voltage amplifier. (though I would still buffer it)

"The Ultra-Linear Cascode
This circuit approximates the ultra-linear pentode circuit by using two triodes. Like the pentode version it boasts both lower output impedance and distortion than otherwise, but differs in that the top triode's grid has a very high input impedance that is easy to drive with just a simple voltage divider.
A further advantage this arrangement holds over the stock Cascode is an increased maximum output swing, as the top triode's plate swings down, so too its cathode moves down. This increase in output headroom would not be needed in a preamp, but in the first stage of an OTL power amplifier it would prove significantly better than a stock Cascode circuit." Broskie, J (2000) Tube Cad journal Vol 2 No1 Pg4

What I find really interesting is the statement he made in the preceding UL preamp circuit where he uses an actual pentode, the top UL part of the voltage devider is a cathode follower it is biased through a potentiometer. The benefit is best said by him:

" The second tweak is to replace the voltage divider resistors with a single potentiometer. The potentiometer allows for some amazing adjustment to the pentode's mode of operation: at one extreme we have pure pentode; at the other, pure triode. In between these two extremes lies ultra-linear operation.
What is the best potentiometer position? Best for what? Best for rock and roll or best for opera? Best for what amplifier? What speaker? What day? What wine? What mood? This control gives the user a choice over the sound quality that is not frequency dependant. " As above for reference earlier in page.

Steve this sort of control is right up your street! In your existing circuit you could make your anode load variable but adjecent to a fixed resistor to prevent adjustment too far, and the same for the 100k + 100k voltage devider. Wiper to top triode (screen) grid cap from wiper to ground or bottom triode cathode.

Tune in your favourite sound on the day for the genre you are playing.

Here is Broskies pentode version with cathode follower supporting the grid control but guided by the pot into position. In your case as the top triode grid is an easy load the Cf is non essential. It gets it;s UL connection fia the CF's control grid, it is effectively the same as a wiper direct to the screen grid but gives better drive.

Image
Hey that's very interesting Paul!

I thought it would have had a high output impedance, which is why I said "sod the high output Z" when actually it does not. It's easy-drive all the way through. Told you I tend to get things right but for the wrong reasons.
I just did the P-to-G Fb to correct one particular gain problem, hence my not realising the whole thing was now ultralinear in nature and that there were quite a lot of possibilities from this arrangement.
I'll have a good read of the Broskie. :)
Tune the amp to the sound you like?.....now that's what I call a tone control :lol:
Last edited by Cressy Snr on Sun Mar 04, 2012 9:42 am, edited 1 time in total.
Sgt. Baker started talkin’ with a Bullhorn in his hand.
Post Reply