What size capacitor I use for a 2A3 or 300B cathode bypass?

[sgerus]

300B Cathode Bypass Capacitor Value / Options?

What size capacitor I use for a 2A3 or 300B cathode bypass?

Simple question, Right?

Here is what I’ve seen:
1. 100uf seems to be the most popular
2. Seen values as low as 20-30uf for a 2A3
3. Some use 3 capacitors, one each on pin 1 and 4, and 1 on the hum pot. I’ve seen 3-50uf and 3-10uf… Sometimes with the caps on pin 1 and 4 bypassed with lower values.
4. Then some people bypass the main capacitor with 2 lower values:
100uf, then 1uf, then .001uf
5. And then there’s the Ultapath, which adds a capacitor from the B+ to the + side of the cathode bypass cap.
(This is what I used in my current amp)

So the choices are all over the map, 20-100uf with 1 to 5 capacitors… What do you folks prefer?

[arnoldc]

1. Because there's a Black Gate value that's 100uF
2. Because they're so big, like an Obligatto
3 & 4. Was discussed by Morgan Jones on the 3rd Edition
5. Because it's a Western Electric invention?

I prefer a "compromise" - Black Gate on the cathode

[Miles Prower]

Originally posted by sgerus
300B Cathode Bypass Capacitor Value / Options?

So the choices are all over the map, 20-100uf with 1 to 5 capacitors… What do you folks prefer?

Fixed bias. That way, I don't have to worry about it.

[dsavitsk]

I agree wth fixed bias in theory, though my 300b amps are ultrapath (50uF). But, to determine the standard bypass cap, use

(Rp + rp) || Rk

and plug this into

C = 1/(2 * pi * F * R)

where the above is R, and F is your 3dB roll off frequency.

[Miles Prower]

Originally posted by dsavitsk
I agree wth fixed bias in theory, though my 300b amps are ultrapath (50uF). But, to determine the standard bypass cap, use

(Rp + rp) || Rk

and plug this into

C = 1/(2 * pi * F * R)

where the above is R, and F is your 3dB roll off frequency.

Shouldn't that be:

Rk || rk where:

rk= (Rp + rp) / (1 + u)

[dsavitsk]

D'oh! My first day of law school, one of my professors was referring to some law and pulled out a copy of the statute to look at it to make sure he had it correct, even though he had written it in the first place. The lesson, of course, was not to assume you remember something but to look it up to be sure you are correct.

[sgerus]

Miles, Could you explain why you don’t have to worry about it with fixed bias?

Let’s look at the formula.
Given:
RL=3000 (OPT)
Ra=700 (JJ300B)
U=3.85 (JJ300B)
Cathode bypass Res=1K

Rk=(3000+700)/(3.85+1)=762R

r’k=1K//762R=432R

At 20K hz 1/(6.28*20*432)=18uf
At 10K hz 1/(6.28*10*432)=37uf
At 5K hz 1/(6.28*5*432)=74uf
At 4K hz 1/(6.28*4*432)=92uf
At 1K hz 1/(6.28*1*432)=368uf

Based on this, 18uf is good for 20hz, which is also the rating of most OPT’s

Using 100uf is good for <4hz, which seems like overkill?

My current amp is set up with a 100uF oil for the cathode bypass and a 20uf for the ultrapath….. Think I’ll do some testing over the weekend!

[kevinkr]

Originally posted by sgerus
Miles, Could you explain why you don’t have to worry about it with fixed bias?

Let’s look at the formula.
Given:
RL=3000 (OPT)
Ra=700 (JJ300B)
U=3.85 (JJ300B)
Cathode bypass Res=1K

Rk=(3000+700)/(3.85+1)=762R

r’k=1K//762R=432R

At 20K hz 1/(6.28*20*432)=18uf
At 10K hz 1/(6.28*10*432)=37uf
At 5K hz 1/(6.28*5*432)=74uf
At 4K hz 1/(6.28*4*432)=92uf
At 1K hz 1/(6.28*1*432)=368uf

Based on this, 18uf is good for 20hz, which is also the rating of most OPT’s

Using 100uf is good for <4hz, which seems like overkill?

My current amp is set up with a 100uF oil for the cathode bypass and a 20uf for the ultrapath….. Think I’ll do some testing over the weekend!

Because with fixed bias the cathode (or filament in the case of a dht) is grounded, there is no cathode resistor and hence no need to bypass it with a cap. Negative bias voltage is applied to the grid.

Incidentally when I go for cathode bias I usually use 100uF with a 20uF cap for hum cancelling alla Western Electric, but like Miles I much prefer fixed bias for this application.

By the way 100uF is not overkill, it really depends a lot on the primary inductance of the transformer, the L, R, C present in the output circuit conspire to create a series resonant circuit which at the secondary actually appears to look like a low to moderate Q parallel resonant circuit which at resonance can produce a significant boost in low frequency response - depending on where the boost occurs it can result in flabby, overblown bass, or worse can pump a significant amount of subsonic energy into your speakers if present at the input of the amplifier. (record warps, etc.)

The above is just another reason why I prefer fixed bias. (Transparency, ease of tweaking op point for "voicing" the amplifier, etc.)

By the way you meant Hz not kHz right?

[sgerus]

Here’s a follow up on some listening tests with different value cathode bypass caps for the 2A3:

I tested 3 values, from one extreme to another:
1. 200uf Mallory electrolytic bypassed with a with a 2.2uf Solen
2. 100uf Oil
3. 18uf E-Speaker Audio Cap

My test was rather crude…. Turn off the amp, jumper in the new cap, turn the amp back on and listen, then repeat with other caps.

My preference for right now after only 2 hours of listening is that the high quality 18uf cap sounds best. Of all the pentode driven (WE91 type) amp schematics I found, 16uf was the lowest value shown, with the average being 63uf.

In interesting side note: An “unofficial” schematic of the Wavelength Cardinal shows a 39uf Solen used for the 300B bypass.

Now I need to wire in a switch for the bypass cap’s and so some “Blind” listening tests.

I wonder if my crude test proved that to some point, quality is more important than a higher value uf?

[kevinkr]

Yes cap quality is critical here which is why I use fixed bias whenever feasible, and you may also be getting some "pleasant sounding" low bass boost with the 18uF that would be subsonic with larger values of bypass cap. Can you measure the frequency response of your amplifier to see?