Bias Capacitor Question

[plumcrazyfx]

So I have a general question about the bias cap.  Later Fender likes one cap with anywhere from 25uf to about 100uf whereas the Tweeds and Marshall's tend to have 2x 8uf or 10uf.  Traynor liked one 8uf and one 64uf.


I guess I'm wondering what is optimal.  I've heard making them too large might lead to longer charging and more time without bias voltage on the tubes at turn on.  But is there a problem with only using one vs two (was Fender doing their standard saving money by eliminating one)?  Also what function do the serve- is there noise/ripple that can affect the circuit with too little capacitance?  Also, is it partially dependent on whether your bias voltage is coming from a dedicated bias tap, how does it work, they are working with + to ground correct?  Just something I've been wondering and I'd rather go better rather than rote copying in my non-clone builds.

[mresistor]

This will help with some of your questions.   https://el34world.com/charts/Biascircuits.htm

[tubeswell]

Anywhere between 8uF and 100uF. There isn’t much current drawn through the bias circuit, so the size of the cap isn’t critical.

[HotBluePlates]

... the bias cap. ...  Also what function do the serve- is there noise/ripple that can affect the circuit with too little capacitance?  ...

You're creating DC Volts from AC Volts; if the bias cap fails there will be massive 60Hz ripple injected right at the output tube grids.  Big Hum!

... Later Fender ... anywhere from 25uf to about 100uf whereas the Tweeds ... tend to have 2x 8uf or 10uf. ...

1950s amps tended not to have anything over 16µF anywhere.  So you're mostly seeing a capacitor industry develop larger-values in affordable production.

... There isn’t much current drawn through the bias circuit, so the size of the cap isn’t critical.

... I've heard making them too large might lead to longer charging and more time without bias voltage on the tubes at turn on. ...

Charging Time is about R x C.

Yes, bigger caps are slower to charge.  But Resistance limits charging current, and slows charging of any capacitor.  So there's 2 pieces to the puzzle.

You worry about "slowly building bias" but your tubes take a number of seconds to warm up & even pass current.  I'm gonna guess if you monitor negative volts at the bias supply, their value will steady before the tubes get hot enough to begin working.

As for current draw, think about overdriving the output tubes, which will result in output tube grid current & bias-shift.  Bigger bias caps tolerate brief overload with probably less bias-shift, and possibly help keep the shift from going as-far with severe overdrive.

... I'd rather go better rather than rote copying in my non-clone builds.

Then you will have to experiment A LOT to re-learn what the other designers knew, instead of just copying their designs.  You will be:

    Making Negative DC Volts for an available AC Voltage.
    Developing a sufficient voltage for the output tube being used.
    Enabling a range of adjustment (perhaps)
    Considering the supply's performance at start-up.
    Considering the supply's performance when the output tubes are overdriven (how overdriven? up to you to figure out)

Perhaps it's wise to copy, and see how the result changes.

[PRR]

....I guess I'm wondering what is optimal.  ....

So am I.

What is optimal:
* Least buzz?
* Buzz just-inaudible in a crowd?
* Lowest cost?
* Highest reliability?
* Least patent infringement?
* Best record sales?
* Award-winning innovation? (Who gives that award?)

[plumcrazyfx]

Least buzz and most reliability would be my targets.  The others, not so much.  Not interested in $ as long as we're not talking about the ideal adding $30 to the cost.  I was just wondering if there was one that was better/more modern.

[HotBluePlates]

Least buzz and most reliability would be my targets. ...

Then lower B+ volts and go for cathode bias.

As long as the cathode resistor is not too-small, you have most-reliability and least-buzz possible.  The tubes cannot conduct current without setting up their required bias voltage.

[plumcrazyfx]

I should have clarified.  The most reliable and least buzz fixed-bias setup.  Sometimes 25 watts from a pair of 6L6s isn't what I'm going for.

[SILVERGUN]

Grab a handful of these and you won't have to think about it again

[pdf64]

Might 100uFs result in a longish charge up time to reach full bias voltage in a typical Marshall 50 watter arrangement?
Eg https://el34world.com/charts/Schematics/files/Marshall/Marshall_1964_jmp_lead_and_bass.gif

[acheld]

By a split second.   BUT, the heaters will be way slower.  Even without the heater delay, should not be an issue.  My (non-expert) opinion only.

[shooter]

heaters are running with the ON switch; Bias won't get started til the STBY switch and everyone's insured by the 1/2A fuse 

[sluckey]

heaters are running with the ON switch; Bias won't get started til the STBY switch

Bad idea!

[mresistor]

OH  I agree bad idea     Marshall   

[pdf64]

Sorry that was a poor choice for an example schematic, as it included Marshall’s 2nd worst standby.

Try this one instead https://el34world.com/charts/Schematics/files/Marshall/Marshall_jcm800_4010_power.gif

I was looking at the 220k dropper feeding the bias reservoir, and wondering how things would work out if the 10uFs were swapped to 100uF caps.