Cathode bypass cap

[guavatone]

It’s been discussed ad nuaseum at diyaudio where there’s other very smart EE types there.

https://www.diyaudio.com/forums/tubes-valves/216543-eliminating-cathode-bypass-caps-set-amplifier.html

Hot topic of tubes built in feedback

https://www.diyaudio.com/forums/tubes-valves/178688-triode-plate-resistance-internal-negative-feedback.html

The problems with generalizations is that we are usually talking about at least 2 stages of amplification. And if you’re using feedback around those 2 stages one needs to consider the poles and zeros or at least know about bode plots/theory. 

Not to mention, impedance and that changing 1/2 of a whole circuit will affect the other 1/2.  I think CJ mentioned that tube preamps were designed to have less tubes bc of some VAT tax on each tube, so bypassing gave more gain for the bean counters.

The REDD.47 ef86 is partially bypassed and I read  about .5%  THD difference from the fully bypassed input tube. So yes  unbypassed achieves less THD.

Rod Elliot has some insight on partially bypassing too:
http://sound.whsites.net/valves/bias-gain.html#s2

 

[emrr]

  I think CJ mentioned that tube preamps were designed to have less tubes bc of some VAT tax on each tube, so bypassing gave more gain for the bean counters.

Nonsense.  Almost none of the RCA broadcast/film stuff has bypasses, and most things before a certain era don't either.  They were generally left out because electrolytics were the weakest link in an industry that needed to avoid weak links.  They only came into wider usage when advances proved more reliable, and lots of amps still didn't use them for the same reasons. 

Bypassing gave more gain for use of loop NFB without increasing the number of tubes to get it, combined with a move to higher mu triode tubes, and pentodes. 

 

[ruffrecords]

It’s been discussed ad nuaseum at diyaudio where there’s other very smart EE types there.

https://www.diyaudio.com/forums/tubes-valves/216543-eliminating-cathode-bypass-caps-set-amplifier.html

Hot topic of tubes built in feedback

https://www.diyaudio.com/forums/tubes-valves/178688-triode-plate-resistance-internal-negative-feedback.html

I notice that old chestnut is usually accompanied by the why does NFB produce higher order distortion (it doesn't)


[/quote]

 

[abbey road d enfer]

Yes, but... remember, most NFB theory makes simplifications based on the assumption of a relatively high open loop gain (A). In many tube circuits, especially individual stages, this assumption is not true.

I don't agree. NFB theory takes into account the actual value of the OLG. Consequences of lower OLG are very well documented.

 

[ruffrecords]

I don't agree. NFB theory takes into account the actual value of the OLG. Consequences of lower OLG are very well documented.

That is why I said 'most'

Cheers

Ian

 

[abbey road d enfer]

It’s been discussed ad nuaseum at diyaudio where there’s other very smart EE types there.

Is that a "very smart EE" that wrote "I think yo might lose some low frequency too, if you remove the cathode bias cap"?
And "With a 22 uF bypass, you are rolling off at about 100 Hz, and 20 Hz is gone"?
That is plain ridiculous. Removing the cath bypass cap actually sets the roll-off at DC. It reduces gain at all frequencies, so you may lose some 20Hz, but you lose the same at 1k or 20k, for a very linear response.

And rolling off at 100Hz means that 20Hz is somewhat attenuated, but not "gone".
See attached graph, that shows the effect of :

• No bypass cap (green)
• 22uF bypass cap (cyan)
• 1000uF bypass cap (red)

One can clearly see that said 22uF cap results in only 0.7dB of attenuation at 20 Hz

Hot topic of tubes built in feedback

That is a much more serious thread; however the claimed difference between triodes, pentodes, and even transistors is fallacious. The difference is just the order of magnitude of the effect, not its nature.

The problems with generalizations is that we are usually talking about at least 2 stages of amplification.

With two stages, its most of the times voltage-to-voltage NFB, and the usual saying "higher input Z, lower output Z" is true. In the case of unbypassed catode, it's current-to-voltage NFB, resulting in higher input AND output Z.

I think CJ mentioned that tube preamps were designed to have less tubes bc of some VAT tax on each tube, so bypassing gave more gain for the bean counters.

One shouldn't forget that tubes were always expensive (and still are); that's the reason for a number of radio tubes combining oscillator, mixer and one amplification stage, or the ECLL800 that made a complete push-pull amplifier with one single bottle.

The REDD.47 ef86 is partially bypassed and I read  about .5%  THD difference from the fully bypassed input tube. So yes  unbypassed achieves less THD.

Full cathode-bypassing in a REDD47 results in NO local NFB and NO global NFB; no doubt THD is higher then.

Rod Elliot has some insight on partially bypassing too:

The main (only?) reason for partial cath bypass is allowing extended NFB-controlled gain setting.