Bias supply and grid current

[boywonder]

I have a power transformer with an extra 550V CT winding that I would like to use as a bias supply for Poindexter's EL34 Music machine. I'm assuming that I can use one half of this winding with a couple of resistors to derive the appropriate (-50V) bias voltage. The 550V CT winding is rated for 40ma.

How much current should I design this tap to source? If I use a relatively low R voltage divider to allow for some current, the resistors burn a lot of power. What's a typical design grid current for fixed bias trioded EL34's?

[aardvarkash10]

its a bias voltage, not bias current. Set to run real low current across the divider - 1 or 2ma will be fine. There is effectively NO current from the bias supply to the tube grid. A 550V, 40ma trannie is overkill!!! Hope you are doing something else with it!

[boywonder]

Originally posted by aardvarkash10
A 550V, 40ma trannie is overkill!!! Hope you are doing something else with it!

I'm using it for the power tranny, as it has a 720 CT @ 100ma also. I'm building monoblocks, and I have 2 of these. So no current capability is OK? none, zero, nada? Just supply the volts?

[oldmanStrat]

current always matters.

if you want -50 and you have 275 to work with you are going to be dropping a lot of voltage on the converter - and the amount of current had better be static (not variable) or a voltage divider won't give you what you want.

If you use a DC to DC be prepared for a massive heat sink.

standard disclaimers apply

V=IR, its the law.

[aardvarkash10]

whilst strictly correct, the old man is a little off subject imho. Your grid bias is effectively a voltage only issue - within the bounds of the normal grid leak resistor limits etc. I'd aim for a total resistance in the divider network of around 1/2meg (giving around 1ma current through the network and 0.5w dissipation), with the grid leak resistors at around 200k. Never worked with EL34s though, so I'm happy to be contradicted - its all a learning experience!

At worst, you will find the resistance is too high and you have to change four resistors - so what?

[SpreadSpectrum]

If you want to design it right, look at your max grid leak spec and consider what will happen if you have a tube that is pulling a little grid current. Your resistor divider has an output impedance that that appears in series with the grid leak resistor. That output impedance could be significant if you are dropping lots of volts in the divider and using large resistors.

Or you could drive the EL34s with MOSFETs and apply the bias voltage to the gates of the FETs through 10M or so. With FETs it really is just a voltage issue, tubes only nearly so.

[boywonder]

Thanks for the input, time to do a little more breadboarding.

Originally posted by SpreadSpectrum
If you want to design it right, look at your max grid leak spec and consider what will happen if you have a tube that is pulling a little grid current. Your resistor divider has an output impedance that that appears in series with the grid leak resistor. That output impedance could be significant if you are dropping lots of volts in the divider and using large resistors.

Got it, I'll check the max grid leak spec for fixed bias triode connection.

Originally posted by SpreadSpectrum

Or you could drive the EL34s with MOSFETs and apply the bias voltage to the gates of the FETs through 10M or so. With FETs it really is just a voltage issue, tubes only nearly so.

Sounds great, but still over my head. Without a schematic, I'm still at ohm's law and voltage dividers.

[aardvarkash10]

check this out http://www.geofex.com/Article_Folder...osfetfolly.htm

Treat it like a cathode follower triode. Does need a separate bi-polar power supply, but you have that to hand anyway.

[boywonder]

I just read the article and bookmarked it, very cool stuff, thanks. About 80% sunk in. I think I understand the IRF820 version of the baby huey a little better now also.

[boywonder]

I am finally building up my mono 6GK5/EL34 triode Music Machines, and I am still struggling with the bias supply design. I would like to use the extra 240V/40ma winding on my PT mentioned above with a voltage divider to get the -50V neg rail for the fixed bias, but I am concerned about the output impedance, ie it will be very high compared to the separate transformers that Poindexter has spec'd.

Poinz is specifying 44V transformers that are rated for around 32ma for the mono application, and 70ma or so for the stereo application. To keep the heat reasonable for the voltage divider R's, the extra winding I have will be capable of providing a couple of ma at best.

So, my basic question is: How many milliamps will the bias supply be expected to supply for a pair of triode connected EL34's? I am sure there is a reason that Poinz is specifying 32ma/70ma bias transformers, I just don't know what the reason is. The discussion above (which makes perfect sense to me) says that pretty much no current is needed (ie AB1). So is the bias voltage really just a voltage, with no need to provide any current?

My fear is that if the tubes need ma that the supply cannot provide, they will run away....