A little cathode bypass

[Rhodesplyr]

I just finished refurbishing a ~1953 Multivox Premier 120 for a customer. This is the version in the overbite cabinet that uses 12AX7s, not 6SL7s.

I decided to do a little experimenting with it before sewing it up, so I tried clipping in a couple of output stage bypass cap values. The 300 Ohm shared cathode resistor for the two 6V6s is unbypassed with the amp in stock form.

What I found I liked was a bypass of just 1uF. It gives the amp a little chime in the highs without the excess bass of a larger cap. I suspect that the output of this amp is purposefully limited by negative feedback from the shared cathode resistor (as the output stage goes into AB1) in part to prevent speaker damage. The musical effect is compression at all frequencies.

I think I may have to let the amp's owner hear it to decide whether he likes the amp dead stock or with a bit of extra transient response in the highs.

It's amazing how much different cathode bypass values can change the character of an amplifier.

[Rhodesplyr]

Coming back to it with a fresh ear, I'm changing the value I like to just 0.39uF in parallel with the 300 Ohm shared cathode resistor. It lets the amp sing when pushed, but limits the bass, which sounds better and protects the speaker (original) from excessive excursion.

The thing I like about this is that the capacitor doesn't really have any effect until the amp goes into AB1 operation.

[EFK]

I agree with you on this 1000%! I usually settle on a value around .68uf to 1 uf on my own, most of which are based on old Supro circuits. That's when I use them - I used to just slap a big 30 or 50uf cap across anything and everything, after all that's what we're supposed to do right?, but I have really come around to appreciate the rounder, less harsh and less muddy sound of an unbypassed cathode R.

[Gtr_tech]

I have in a little Blackheart 15w head for some mods right now and they use *huge* bypass caps...100u on each (seperate) cathode resistor. This is excessive and causes the thing to be extremely bass heavy. I talked with the owner about lowering the value to tighten low end response and....he likes it that way. Go figger....

[pdf64]

In a cathode biased p-p amp, seperate cathode resistors is a very different situation to having a shared resistor, with regard to bypassing.
With a shared resistor, at lower signal levels, the tubes effectively bypass the resistor, due to the inverted signals at either grid appearing on the cathode and cancelling out. Only when the signal level goes high enough for one side to cut off (for some part of the wave) is that effect lost.
Whereas with seperate cathode resistors, there's no self cancelling whatever the signal level.
Bass boost? If there's no global negative feedback on the power amp then the amp's output impedance can be high, which results in the speakers movement at bass resonance being undamped, plus the amp can act as a current amp rather than a voltage amp, so putting more power into the much higher than nominal load.
But not bypassing the output tube cathodes would tend to increase the output impedance even higher, so I would have thought that would tend to defeat the low end attenuation of a smaller bypass cap. It's a tricky thing to measure on the bench though.
Pete.

[enorbet]

Hello
I'm considering making a thread regarding this subject of cathode bypass caps, but I'm going to start by replying here. In addition to choosing a value that allows for some Hi-Pass Effect, which is almost always just a few mfd's or less (even in most preamp circuits) there is a tone bonus that it seems few consider, and that is symmetry and the lack of it. When using capacitors above 1.0mfd it is common (nearly universal) that manufacturers, including amateur builders, opt for the easy to find electrolytics. It is my opinion this is a mistake if one's goal is a natural, "live" sounding amplifier.

The reason I consider this a mistake, beyond the empirical experience of listening to the difference, is that electrolytics do not have the same impedance in both directions. This introduces variables that are not unlike "dead" strings that have lost their ability to vibrate freely and evenly.

Solution - Instead of using asymmetrical electrolytics find non-polarized substitutes (not too bad to do in lower values) or choose two electrolytics of half the capacitance value (and I recommend at least double the voltage rating) and solder them in parallel plus to minus and minus to plus, to create a substitute that is for all intents and purposes, virtually identical to a non-polarized cap.

Note - While the effect of even one stage can be heard by a trained ear, ideally all bypass caps should be non-polarized, symmetrical units. If you can't hear and even feel the difference, then obviously opt out, but most players do notice the effect is much like putting on a new set of strings - very lively and responsive.

Enjoy

[pdf64]

'electrolytics do not have the same impedance in both directions'
I've seen that idea put forward previously, but not a verifiable source for it.
Do you have one, or have you made your own measurements somehow?
Pete.

[Chuck H]

I don't buy it. The difference can't be more than a couple of ohms in a circuit with an impedance that is typically between 820 and 10,000 ohms! I think the only way to hear a difference is by believing you can.

[Tom Phillips]

... choose two electrolytics of half the capacitance value (and I recommend at least double the voltage rating) and solder them in parallel plus to minus and minus to plus, to create a substitute that is for all intents and purposes, virtually identical to a non-polarized cap...

I don't agree with that. If there is a DC component across the resulting cap then one of the individual caps will be in a reverse voltage condition and will have a very short life. If there is pure AC across the resulting cap then each will be exposed to high ripple and will be in a reverse voltage condition 50% of the time. Again, this will equal short life. Making a "non-polarized" cap as described may even result in some interesting entertainment if you like hissing and/or exploding things.

[Chuck H]

Yup. Not sure where enorbet got info, but IIRC the way to make a non polar circuit with polar electros is to put two caps of double the uf's in series with like leads connected. Plus to plus or minus to minus. It shouldn't matter. Of course then you have additional ESR because of the series connection!!! If you can hear a couple of ohms difference in a 1000 ohm circuit I guess this idea is straight out the window!?! Then again, non polar electros typically have poor ESR compared to polar electros. So neener, neener.

AFAIK the polarity of the cap is WRT DC!!! AC obviously plays off the 0V reference. Any difference in the positive and negative swings of the waveform due to polar impedance differences has to be just tiny. Way too many SS devices using polar electros for coupling caps and sound just fine with respectible distortion figures (as measured on big fancy machines that enorbet doesn't own (or even know about apparently).

[enorbet]

Yup. Not sure where enorbet got info, but IIRC the way to make a non polar circuit with polar electros is to put two caps of double the uf's in series with like leads connected. Plus to plus or minus to minus. It shouldn't matter. Of course then you have additional ESR because of the series connection!!! If you can hear a couple of ohms difference in a 1000 ohm circuit I guess this idea is straight out the window!?! Then again, non polar electros typically have poor ESR compared to polar electros. So neener, neener.

AFAIK the polarity of the cap is WRT DC!!! AC obviously plays off the 0V reference. Any difference in the positive and negative swings of the waveform due to polar impedance differences has to be just tiny. Way too many SS devices using polar electros for coupling caps and sound just fine with respectible distortion figures (as measured on big fancy machines that enorbet doesn't own (or even know about apparently).

Greetings and thank you for reminding me of the proper (series) combination and correcting my brain freeze in that regard. It is possible to do this in parallel if one chooses (as stated) a much larger voltage rating but it is simply not as practical, even though we are talking about voltages that commonly vary between 1 and 40vdc. I haven't done this on a large scale in years. However the basic premise is still accurate and the bottom line is that it works, amps sound better as I stated and it has nothing to do with belief.

It might be useful to recall that few "believed" in TIM until it became measurable by electronic devices (I might add, devices many times less discriminatory than the human ear/brain connection) or for that matter notice that for over 70 years many still argue that there is no difference in sound reproduced/translated by solid state devices and vacuum tubes. On the latter if there is none, then why are you here?

It is my contention, backed with years of tweaking, that amplifiers with more than a single stage have additive and even multiplicative effects, so yes, a few ohms in a ~1000 ohm circuit, especially when there are several of them makes an audible (and in the case of instrument amplifiers which are not simply reproductive devices but musical instruments in their own right) "feelable" difference.

Also please consider that layout, connection methods vis a vis circuit board vs/ point-to-point, and many other so-called "non-measurable" mechanics have an effect on the end product. If you're satisfied with "sound just fine" then why even bother with tubes, Class A, direct-coupled cathode followers, etc. If you are concerned only with "respectable distortion figures" that "big fancy machines" (that yes, I do not own but AM quite aware of) then you have either forgotten or totally disregarded that even the so-called "cleanest" guitar players prefer amps of a particular sound color brought about by dynamic distortion.

On the tail end, consider that until Celestion developed the Laser Interferometer, nobody had in any way measured speaker distortion dynamically. One might ask oneself why there are so many models if some sort of generic 12 inch would suffice. Also until Thiele and Small delved deeper into the dynamic properties of impedance of speakers and air columns we were stuck with a formula for porting cabinets that did not give consistent results and speakers that required MUCH larger cabinets than we use today.

As far as I can tell the responses here serve only to show the age old gulf between scientists and artists, and if my memory serves, innovation is commonly created by trying ideas and concepts out, often by trial and error or even serendipity. It is worthy of note that written music, designed before recording devices, serves to allow someone to reproduce/translate the mechanics of music and thus reproduce it. It doesn't make one create like Mozart and most often even inhibits it.

Bottom Line - A/B it. Take two identical (as much as that is actually possible) amplifiers, preferably with at least four stages of amplification, and have many players try them. A trial group of 100 should show the trend, as long as the majority are not merely "weekend warriors". Or, you can continue, safe, in assumptions and "just fine".