801a maximised class A2 excursion.
- Paul Barker
- Social Sevices have been notified
- Posts: 8879
- Joined: Mon May 21, 2007 9:42 pm
#121 Re: 801a maximised class A2 excursion.
Like my dyslexic mind this topic/thread has weaves around until if I were to rethink the title it would change.
So sub title now is:
Paul had a near breakdown over using transistors after unexpectedly poor results compared with self professed accolades by the designer(s), so Paul has taken refuge back with thermionic valves. Anyway I now find out FETs might not be advantageous for line level as their capacitances are high vis a vis valves. So with local feedback not required to lift upper level bandwidth above 10khz.
So now the obstacle I’m allowing my unusual functioning brain to chew on is my decision for which cf driver.
The ideals are:
1/ current flow to cope with whatever grid current the worst case scenario of all output valves considered.
2/ High transconductance (to lower output impedance)
3/ Voltage swing to facilitate maximum power with headroom.
4/ the wattage to satisfy the voltage which provides for parameter 3/ and the current for parameter 1/.
If it were just for 801a 5687 easy, Stephie used it, so hard work done already.
But it isn’t! It would be subjected to 811-10 and 211 maybe GK71 (rather than gm70 because it more fits in with A2 operation; the method of the thread).
5687 lacks current and power to drive 50mA grid current.
6bx7 fails at the power hurdle. It may be 10 watt each side but you can only have 6 watts per side if paralleling, which quite honestly is a disappointment. I can’t get my headroom and at worst any powerful output valve would be crippled by the driver.
Which makes it a job for parallel 6AS7. Correct me if I’m wrong but no power limit on paralleling. So we can now satisfy condition 4/.
So sub title now is:
Paul had a near breakdown over using transistors after unexpectedly poor results compared with self professed accolades by the designer(s), so Paul has taken refuge back with thermionic valves. Anyway I now find out FETs might not be advantageous for line level as their capacitances are high vis a vis valves. So with local feedback not required to lift upper level bandwidth above 10khz.
So now the obstacle I’m allowing my unusual functioning brain to chew on is my decision for which cf driver.
The ideals are:
1/ current flow to cope with whatever grid current the worst case scenario of all output valves considered.
2/ High transconductance (to lower output impedance)
3/ Voltage swing to facilitate maximum power with headroom.
4/ the wattage to satisfy the voltage which provides for parameter 3/ and the current for parameter 1/.
If it were just for 801a 5687 easy, Stephie used it, so hard work done already.
But it isn’t! It would be subjected to 811-10 and 211 maybe GK71 (rather than gm70 because it more fits in with A2 operation; the method of the thread).
5687 lacks current and power to drive 50mA grid current.
6bx7 fails at the power hurdle. It may be 10 watt each side but you can only have 6 watts per side if paralleling, which quite honestly is a disappointment. I can’t get my headroom and at worst any powerful output valve would be crippled by the driver.
Which makes it a job for parallel 6AS7. Correct me if I’m wrong but no power limit on paralleling. So we can now satisfy condition 4/.
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
#122 Re: 801a maximised class A2 excursion.
Hi Paul,
Stating the obvious - the big problem with Class A2 is driving the grid through the large and rapid input impedance change with low distortion - the classic approach at audio frequencies is to engineer a low impedance driver and a dc connection or an IT based step down to avoid 'blocking' through a capacitor. But still the large and rapid load change on the driver changes its distortion level and profile...
AT RF they do things differently, for more than one reason, but it also beneficial at audio in that they drive the cathode not the grid (I know you know this but bear with me) and this provides two advantages - one it provides a constant impedance input so the driver generates the same level of distortion and with the same profile at all signal levels (assumes the driver is well engineered through the signal levels used). And it separates the dc bias applied to the grid from the driving circuit so its easy to experiment with the operating point. My A2 designs generally use cathode drive rather than grid drive and I found it easier to get it to sound consistent and good - incidentally the output valve generates higher gain and higher Gm from the output valve too - not a lot but anything helps
My N43 amplifier designs uses cathode drive - not because its needed as I'll keep this in the A1 regime but because I can well its a bit different and I like using signal chokes.
In terms of driver I might go for E55L in pentode mode as that makes a great CF with G2 cap coupled to the output and a lower output impedance if you used a step down IT as the CF load the output impedance could be seriously low...
Hope this provides further food for thought and intrigued to see how you solve this
Sorry if this has been covered before as I've haven't caught up with the whole thread yet...
ciao
James
Stating the obvious - the big problem with Class A2 is driving the grid through the large and rapid input impedance change with low distortion - the classic approach at audio frequencies is to engineer a low impedance driver and a dc connection or an IT based step down to avoid 'blocking' through a capacitor. But still the large and rapid load change on the driver changes its distortion level and profile...
AT RF they do things differently, for more than one reason, but it also beneficial at audio in that they drive the cathode not the grid (I know you know this but bear with me) and this provides two advantages - one it provides a constant impedance input so the driver generates the same level of distortion and with the same profile at all signal levels (assumes the driver is well engineered through the signal levels used). And it separates the dc bias applied to the grid from the driving circuit so its easy to experiment with the operating point. My A2 designs generally use cathode drive rather than grid drive and I found it easier to get it to sound consistent and good - incidentally the output valve generates higher gain and higher Gm from the output valve too - not a lot but anything helps
My N43 amplifier designs uses cathode drive - not because its needed as I'll keep this in the A1 regime but because I can well its a bit different and I like using signal chokes.
In terms of driver I might go for E55L in pentode mode as that makes a great CF with G2 cap coupled to the output and a lower output impedance if you used a step down IT as the CF load the output impedance could be seriously low...
Hope this provides further food for thought and intrigued to see how you solve this
Sorry if this has been covered before as I've haven't caught up with the whole thread yet...
ciao
James
#123 Re: 801a maximised class A2 excursion.
For what its worth I use a el84 as a current production valve in my 211 to drive the grid.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
#124 Re: 801a maximised class A2 excursion.
Nick,
On your amps did you look at the how the distortion spectrum changes through the A1 to A2 transistion? I remember doing the two channel oscilloscope output minus input display to look at what happened and was quite surprised at how it changed through the transition region so that I couldn't 'get it right' for all signal levels... of course this was only done at a few different signal levels as the 'scope has to be re-balanced at each level...
Now-a-days its probably easier to use a PC based spectrum analyser - I always wanted to use the residue signal output from my HP 334 but haven't built an A2 amp since I got it...
James
On your amps did you look at the how the distortion spectrum changes through the A1 to A2 transistion? I remember doing the two channel oscilloscope output minus input display to look at what happened and was quite surprised at how it changed through the transition region so that I couldn't 'get it right' for all signal levels... of course this was only done at a few different signal levels as the 'scope has to be re-balanced at each level...
Now-a-days its probably easier to use a PC based spectrum analyser - I always wanted to use the residue signal output from my HP 334 but haven't built an A2 amp since I got it...
James
#125 Re: 801a maximised class A2 excursion.
Sorry if you've posted this before Nick, but do you run the el84s in pentode mode?
Also (and this is not an entirely serious question, but I found myself smiling when I read your post) when you say 'as a current production valve' do you mean that they are still being produced, or that you are using them to provide the current drive for the A2 grids?
Last edited by Max N on Thu May 28, 2020 12:27 pm, edited 1 time in total.
#126 Re: 801a maximised class A2 excursion.
I did at the time, I remember that the main thing was 3rd increased, that’s why I found a relatively high level of 2nd improved things. If its of interest I could do it again this weekend as I am using the 211 mp at the moment as someone wants to borrow it when the madness ends. What would you want to see, I can do distortion against level, and I also have a distortion magnifier that would show the distortion signal.
I think a feedback based driver line the ACF will do a "better" job, but I seem to remember that even if you use cathod edrive the change in Ra after +ve grid will mean a change in the spectra.
I found a E130L was another good choice to drive a grid (75-th). Probably drive a cathode just as well.
I think a feedback based driver line the ACF will do a "better" job, but I seem to remember that even if you use cathod edrive the change in Ra after +ve grid will mean a change in the spectra.
I found a E130L was another good choice to drive a grid (75-th). Probably drive a cathode just as well.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
#127 Re: 801a maximised class A2 excursion.
Yes, as a pentode cathode follower. Yes, current production as its being made. I decided to try and only use those sorts for ethical reasons.Max N wrote: ↑Thu May 28, 2020 12:26 pmSorry if you've posted this before Nick, but do you run the el84s in pentode mode?
Also (and this is not an entirely serious question, but I found myself smiling when I read your post) when you say 'as a current production valve' do you mean that they are still being produced, or that you are using them to provide the current drive for the A2 grids?
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
#128 Re: 801a maximised class A2 excursion.
Hi Nick,
The things to measure would be distortion spectra and Intermodulation but mainly the spectra as that would give absolute levels and relative levels - What the 'scope showed readily was that in A1 it had the mainly 2nd harmonic but through the transition 3rd became noticeable - also the absolute level of distortion increased i.e. 2nd harmonic increased as well - this corresponded to a thickening and increasing hardness of the sound i.e. a more SE pentode type sound rather than SE triode...
The intermodulation distortion is a new thought that this is likely to increase as the transfer curve increases in curvature and adds higher order curvature - not worked that through so its just speculation that might lead to a hypothesis...
James
The things to measure would be distortion spectra and Intermodulation but mainly the spectra as that would give absolute levels and relative levels - What the 'scope showed readily was that in A1 it had the mainly 2nd harmonic but through the transition 3rd became noticeable - also the absolute level of distortion increased i.e. 2nd harmonic increased as well - this corresponded to a thickening and increasing hardness of the sound i.e. a more SE pentode type sound rather than SE triode...
The intermodulation distortion is a new thought that this is likely to increase as the transfer curve increases in curvature and adds higher order curvature - not worked that through so its just speculation that might lead to a hypothesis...
James
#129 Re: 801a maximised class A2 excursion.
Will see what I can do this weekend.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
- Paul Barker
- Social Sevices have been notified
- Posts: 8879
- Joined: Mon May 21, 2007 9:42 pm
#130 Re: 801a maximised class A2 excursion.
James: Don’t you use grounded grid on output valves for RF ? Surely my bandwidth not that demanding?
I’m juggling many things don’t really want to add my first grounded grid output stage into the mix. But if on the scope hf shows problems I’ll explore.
I’ve understood all your’ (James and Nick) points.
A2 is not under any doubt as to whether I am put off by the distortion characteristics which are unique to A2, it’s happening regardless of distortion theory. A2 always sounds better to me than the dull sound of A1 (comparatively).
I’ve always enjoyed a well made A1 with A2 amp. The less shut in sound reminds me of a good OTL. (Bad OTL,s not tolerated here)
I’m juggling many things don’t really want to add my first grounded grid output stage into the mix. But if on the scope hf shows problems I’ll explore.
I’ve understood all your’ (James and Nick) points.
A2 is not under any doubt as to whether I am put off by the distortion characteristics which are unique to A2, it’s happening regardless of distortion theory. A2 always sounds better to me than the dull sound of A1 (comparatively).
I’ve always enjoyed a well made A1 with A2 amp. The less shut in sound reminds me of a good OTL. (Bad OTL,s not tolerated here)
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
#131 Re: 801a maximised class A2 excursion.
Hi Paul,
Yes but RF use is mainly due to the grounded grid shielding the input signal from the output signal and hence removing the Millar effect and associated millar capacitance extending the HF response of the triode for the rf boys. The constant input impedance is also a benefit as they are often using class C and the driver is easier to design with a constant load as it has to deliver real power into the cathode
I've not noticed a big difference between Class A1 and class A2 sound when both are well engineered or at least I've attributed the sound differences to that between the valve types rather than the class its driven in... Need to do the comparison on the same valve driven in the two different modes I guess but its likely to be damn loud in the A2 mode if constantly there... need to look a=for a valve that only operates in A2 - can't think of one of the top of my head...
ciao
James
Yes but RF use is mainly due to the grounded grid shielding the input signal from the output signal and hence removing the Millar effect and associated millar capacitance extending the HF response of the triode for the rf boys. The constant input impedance is also a benefit as they are often using class C and the driver is easier to design with a constant load as it has to deliver real power into the cathode
I've not noticed a big difference between Class A1 and class A2 sound when both are well engineered or at least I've attributed the sound differences to that between the valve types rather than the class its driven in... Need to do the comparison on the same valve driven in the two different modes I guess but its likely to be damn loud in the A2 mode if constantly there... need to look a=for a valve that only operates in A2 - can't think of one of the top of my head...
ciao
James
#132 Re: 801a maximised class A2 excursion.
I have noticed a big difference in sound between amps with the ability to provide current into the grid and those that cant drive even the miller capacitance, but not specifically 1/2
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
- Paul Barker
- Social Sevices have been notified
- Posts: 8879
- Joined: Mon May 21, 2007 9:42 pm
#133 Re: 801a maximised class A2 excursion.
With my dismal results with FET s I decided to buy a transistor tester.
Pretty isn’t it?
SDidnt realise the kit didn’t include the electronics!
Now duly ordered, another £6.
At least the kit box is Perspex ....
Pretty isn’t it?
SDidnt realise the kit didn’t include the electronics!
Now duly ordered, another £6.
At least the kit box is Perspex ....
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
- IslandPink
- Amstrad Tower of Power
- Posts: 10041
- Joined: Tue May 29, 2007 7:01 pm
- Location: Denbigh, N.Wales
#134 Re: 801a maximised class A2 excursion.
It's got that retro home-made airing cupboard look about it. A lick of paint will smarten it up !
"Once you find out ... the Circumstances ; then you can go out"
#135 Re: 801a maximised class A2 excursion.
I bought one from a radio rally for a fiver.. seems to work but not tried it yet..
If it's as useful as my valve tester I'll be very pleased.
Nice little project building a tester Paul.
You will have to hear my take on the mofo I'm sure you would be impressed. Building my own board has helped my transistor understanding, which was at zilch when i started.
If it's as useful as my valve tester I'll be very pleased.
Nice little project building a tester Paul.
You will have to hear my take on the mofo I'm sure you would be impressed. Building my own board has helped my transistor understanding, which was at zilch when i started.
The tube manual is quite like a telephone book. The number of it perfect. It is useful to make it possible to speak with a girl. But we can't see her beautiful face from the telephone number