Spud and a half

[mashaffer]

Playing around with current drive amp designs I got an inspiration for something rather different. One could rework it as a current amp but this one is a more conventional voltage source amp.

The 6GV8/ECL85 seems to be somewhat limited as and single bottle per channel amp in part due to the triode not being up to the task. However it has pretty good gm so i though why not use the triode section as the bottom half of a cascode and add a high mu triode for the top tube.

Presto: Spud and a half.

Any merit?

mike

 

[BinaryMike]

Couldn't you make up enough gain by bypassing U1 cathode instead of adding U5?

C5 is superfluous. No need to bypass an ideal voltage source.

 

[mashaffer]

Oops, forgot to put U1 bypass back in (left over from my current source experiments). Of course it provides some local feedback as is so might still be desirable. Will have to play with this more. My recollection is that running everything open loop proved unsatisfactory. Definitely need to play some more.

 

[mashaffer]

Bypassing the output stage bias resistor ups gain some making the circuit as drawn able to produce full output with 2Vrms input which it couldn't do before. Also distortion spectrum appears better. Without the cascode gain would be insufficient for applying FB as near as I can tell.

Oh, and C5 is just there as a reminder. It will of course be feed from rectified AC not a true DC voltage source.

 

[mashaffer]

Tweaked.

 

[Kay Pirinha]

C6 is waay too big. The RC high pass including R4 has a 1Hz roll-off. For stability reasons this has to be set to a bigger value than the R16/C9 low pass roll-off of 1.6 Hz. Get 0.1µF as C6, instead, maybe even less.

Best regards!

 

[mashaffer]

Made that change thank you. I do note that it increased the infrasonic bump by a couple dB. Of course decreasing C3 brings it back down.

 

[Ketje]

TweakTweak
Mona

 

[kevinkr]

I designed a distantly similar amp back in 2010.. Pentode vas, tetrode/triode mode output stage, two feedback loops; output tube plate to cathode of vas, and global from secondary of output transformer.

For some reason it works well with Altec coaxials. Still in use today. Thread title is a bit misleading.

http://www.diyaudio.com/forums/tubes-valves/173454-6v6-behavior-w-fixed-bias-se-amp.html

 

[Osvaldo de Banfield]

The screen grid unbypassed increases undesirabily the output impedance looking from the transformer, decreasing the damping of the mechanical resonance of the speaker. Dislike this feature.

 

[mashaffer]

TweakTweak
Mona

Thought of that but specifically wanted to keep OPT out of the loop as in this case it is likely to be a distinctly non-highend OPT.

 

[mashaffer]

The screen grid unbypassed increases undesirabily the output impedance looking from the transformer, decreasing the damping of the mechanical resonance of the speaker. Dislike this feature.

You are referring to a cap from screen to cathode? I did forget to consider bypass cap but did not know this effected Zout of the amp. The main reason for bypass is noise avoidance right?

I am not necessarily averse to higher Zout in some applications. in fact this idea came to me when working on a current source amp idea.

 

[Osvaldo de Banfield]

Yes. It is desirable to have the screen grid bypassed to cathode. The screen external resistance can be viewed as a cathode resistance µ(g2-g1) times less than the actual g2 resistor, but "amplified" by the [µ(g1-p) +1] factor, added to the anode resistance. Try this in the simulations below, with and without screen byapss capacitor, as bigger as you can.

 

[Osvaldo de Banfield]

Unfortunately I could't find a mathematical explanation for the fact I said below, I only find an expression for the gain. Look at this, from G. Valley, H. Wallman - "Vacuum tube amplifiers" - 1948.

 

[Osvaldo de Banfield]

Also for the gain, I found in Radiotron's Book, the following paragraph:

 

[mashaffer]

Thanks will take a look.