Cathode bypass cap

Okay. My tube mic preamp sounds better without cathode bypass caps. Far less distortion (measurable and audible). More open.

Convince me why I shouldn't take em out.

(Aside from the 9 dB loss.)

9dB loss in gain implies 9dB of NFB which will reduce distortion by 9dB and flatten the response a bit which explains why it sounds better. Connect the bypass cap via a switch and label it 'Pzazz'

Cheers

Ian

If there is audible distortion then it's possibly a gain staging issue where a stage is being overdriven. 

ruffrecords said:

label it 'Pzazz'

Click to expand...

It's been a long time since I had this coffee-through-nose-onto-keyboard laugh..

Thanks!

Jakob E.

sometimes whatever sounds different sounds better, like when i switch from led zeplin to the Allman Brothers,

ruffrecords said:

Connect the bypass cap via a switch and label it 'Pzazz'

Click to expand...

i vote for this.

Hmmm...y’all got me thinking.

My “ribbon” switch decouples the input transformer secondary resistor. I could also use that switch to engage the cathode bypass caps and give a 9 dB gain boost too.

...pondering.

5v333 said:

i vote for this.

Click to expand...

And a photo of Liberace next to the switch.

BluegrassDan said:

Far less distortion (measurable and audible).

Click to expand...

Apples to apples, gain matched?

EmRR said:

Apples to apples, gain matched?

Click to expand...

Yes. Gain matched. Input signals adjusted.

BluegrassDan said:

Yes. Gain matched. Input signals adjusted.

Click to expand...

Not sure what you mean by 'input signals adjusted'.  Up?  Down?  Why?  There are several ways you could look at that. 

Still 470mfd?  Try a film bypass?    I recall in one vintage circuit that used bypasses in one version, not in another revision, I had better results without them until I got up to something obscene like 1000mfd, so I left them out.  Plenty of things leave them out.  One argument for their presence is the ability to increase loop NFB, as trade-off to local.

Can you put some data to what 'far less' means? 

BluegrassDan said:

And a photo of Liberace next to the switch.

Click to expand...

have you thought about a frame to go with it?

5v333 said:

have you thought about a frame to go with it?

Click to expand...

Zircon encrusted, like my tweezers.

BluegrassDan said:

Convince me why I shouldn't take em out.

(Aside from the 9 dB loss.)

Click to expand...

I don't think there is a right answer for this, only trade-offs. In your case 9dB vs. minor change in THD. Also note the FFT distortion shape (2nd, 3rd orders and multiples) if you have a chance, it is a bit different with or without the cap.

Another example, I have a tube preamp with very high B+ (320VDC) but only relatively low gain low mu tubes in gain stages. Due to the nearly invisible 2nd order dominant distortion this set up produces, it sounds "better" or "smoother" with cathode bypass caps. And I have a switch for it. The more linear no-cathode-bypass sound has slightly less THD but spread across the nastier orders above 2nd. On paper it's more linear, but somehow "boring".

We're in the 0.1 - 0.5% THD area here with somewhat predictable distortion order spread ie. only sound matters and your choice is as good as any. If this was solid state all this would be extremely crappy and unacceptable of course.

Well here's where I'm at this afternoon. Most of the time I've found myself padding the input for less harmonic distortion.

Eliminating the cathode bypass capacitors from both 6072 gain stages: The second gain stage made little difference, but eliminating the first gain stage cap resulted in -6.3 dB in THD, more linearity, and better transients, with the downside of losing 6.3 dB of gain.

The "Ribbon" switch now serves three functions:

1. Disable 48v phantom (which it does currently)
2. Bypass the input transformer secondary resistor (which it does currently, and provides 1 dB of gain)
3. Insert the cathode bypass cap to bring back the 6.3 dB of gain, plus the 1dB, for a total of 7.3dB.

This means the normal settings will provide 58.5 dB MAX gain. Selecting the "Ribbon" switch allows 65.8 dB MAX gain.

Sounds like a good compromise in the feature set

ruffrecords said:

9dB loss in gain implies 9dB of NFB which will reduce distortion by 9dB and flatten the response a bit which explains why it sounds better.

Click to expand...

That is not entirely true. Un-bypassed cathode indeed introduces NFB; however, since it is current-sensing, it actually increases the output impedance, which in turn reduces the HF response.
The resulting smoothing effect may be preferred to a more linear response...

http://www.tubebooks.org/books/rdh4.pdf



http://www.tubebooks.org/Books/Crowhurst%20-%20Audio%20Handbook%20No2%20-%20Feedback%201952.pdf

Reading is good .

guavatone said:

http://www.tubebooks.org/books/rdh4.pdf



http://www.tubebooks.org/Books/Crowhurst%20-%20Audio%20Handbook%20No2%20-%20Feedback%201952.pdf

Reading is good .

Click to expand...

KILLER!

BluegrassDan said:

KILLER!

Click to expand...

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.

Cheers

Ian