All the Bias options

[jkeny]

As a noob, I have been looking into cathode bias options and what initially started as just a few to choose from quickly multiplied into many.

So far, I have changed the Rogers Cadet output cathode biasing from single R (130 ohm bypassed with 40uF) shared between both output tubes to individual R (130ohm bypassed with 47uF cerafine). This made a noticable difference.

I rebuilt Cadet's power stage into Baby Huey stage and left this cathode bypassing in place - excellent sound from this new circuit - compared to the Cadet, it's more accurate, not as wooly as the Cadet.

I changed the biasing to battery (9v alkaline - I know it should be NiMh or NiCd) but didn't notice any diff in sound.

I'm wondering about other biasing schemes - I've seen mention of

CCS,
LEDs,
SY's Red Light District bank of LEDs
Low R from cathode to ground supplemented by additional current
80% bypassing where a low R & larger R are used with the Large R bypassed with cap

Are there others I should look at/try?

John

[jkeny]

Oh forgot to mention - the low R cathode bias won the European Triode Festival Listening test beating Battery Bias, LED bias, etc.

http://db.audioasylum.com/cgi/m.mpl?...er&r=&session=

and here:http://www.tnt-audio.com/shows/etf04_e.html

Is this worth a try next - has anybody got first hand experience?

John

[ray_moth]

130ohm bypassed with 47uF

That would have a turnover frequency of 26Hz - much too high, especially for an output stage. 470uF would be more like it. Also, if two tubes needed 130 ohms between them, then separate resistors would need to be 260 ohms each.

I think the reason there are so many types of bias circuit around is because no one is best, they all have their strong and weak points. For me, it's fixed bias for the OP stage and cathode bias everywhere else - but that's just me

[rdf]

Originally posted by jkeny
Oh forgot to mention - the low R cathode bias won the European Triode Festival Listening test beating Battery Bias, LED bias, etc.

An interesting result. I'll throw out a conjecture, it's been a consistent experience on the bench that the higher distortion components respond much more strongly to local degeneration than the predominant 2nd. A little impedance in the cathode circuit can knock 15 dB off, for example, the 4th or 5th with little to no visible impact on the 2nd. The ETF result correlates with my experience using bad - as in relatively high impedance - LEDs for cathode biasing, usually greens. Bypassing a green with a large cap often results in an increase in higher harmonic distortion components.

In your case though, which is an output, I agree 100% with everything Ray wrote.

[SY]

I'll caveat Ray's comment with the sad trade-off: when we make that cathode cap bigger to reduce LF phase shift, we degrade recovery time after overload.

[sorenj07]

Originally posted by SY
I'll caveat Ray's comment with the sad trade-off: when we make that cathode cap bigger to reduce LF phase shift, we degrade recovery time after overload.

could this effect be reduced by using lower-ESR caps such as metalized polyester?

[SY]

No, it has nothing to do with the ESR, it is completely determined by the cap value. The recovery time is set by the time constant between the bypass cap and the parallel equivalent of the cathode resistor and the cathode impedance. At cutoff, the latter is quite high, so the cap can only discharge through the cathode resistor. Correct bias isn't restored for several time constants. So, you're stuck: if you increase the cap size for bass, you increase the recovery time.

[jkeny]

Yes Ray,
Im sorry my figures were wrong in the posting - original shared R was 130ohm which I changed to individual 270ohm & indeed it was bypassed with 470uF starget cap.

I understand that there is no best bias scheme but there are theories behind each scheme:

For instance SY's LED matrix is premised on the idea that fast recovery time is at the heart of an excellent amp.

And rdf's comment on low R bias:

I'll throw out a conjecture, it's been a consistent experience on the bench that the higher distortion components respond much more strongly to local degeneration than the predominant 2nd. A little impedance in the cathode circuit can knock 15 dB off, for example, the 4th or 5th with little to no visible impact on the 2nd.

Does anybody know the theories behind the other bias schemes?

John

[jlsem]

Are there others I should look at/try?

http://home.att.net/~chimeraone/axiom300bschematic.html

http://home.att.net/~chimeraone/axiom300btestdata.html

[Sheldon]

Originally posted by jkeny
CCS,
LEDs,
SY's Red Light District bank of LEDs
Low R from cathode to ground supplemented by additional current
80% bypassing where a low R & larger R are used with the Large R bypassed with cap

Originally posted by jkeny
understand that there is no best bias scheme but there are theories behind each scheme:

For instance SY's LED matrix is premised on the idea that fast recovery time is at the heart of an excellent amp.

And rdf's comment on low R bias:

I don't have much more experience here than you do John, but I'll hazard a guess.

The CCS is essentially the opposite of an LED - very high vs. very low dynamic impedance. The CCS can be used to set the current operating point, as was done in a recent thread on 12B4 based preamps. It substitutes for a cathode resistor but with a more stable operating point. It would be bypassed and so would have the same considerations regarding blocking, capacitor selection for sound characteristics, etc.. - Not talking here about application beneath a long tailed pair.

The low R with extra current feed is an interesting one. I would guess that the downside would be it's sensitivity to the supply quality (or maybe upside, if you can work it out so that it cancels noise on the plate). A harder downside to get around, is that for an output stage, you could be burning off a lot of power in a low value cathode R. I would think it could be readily configured to avoid blocking problems by feeding it with a good voltage source, but again, that source is square in your audio path. It doesn't require a negative supply, but would it really be easier in most cases than fixed bias?

The 80% solution (could be any percentage) allows some degeneration at the cathode. rdf's comments apply here, so you might get a better distortion profile.

I'd be interested in other's comments/experiences.

Sheldon