Recapping. Bias trim Cap?

[Michael Allen]

I'm recapping my Sovtek Mig 50, and I'm curious about the electro on the bias supply trimmer.

In the Mig 50 it's 1500uF/150V. On a 1987 Marshall (quite close circuit) it's an 8uF - 15k - 8uF network. That's quite a difference in capacitance.

Is there an advantage to either method? I've already modded this amp a ton so i'm not worried about protecting the originality of the circuit. It would save me some cash on a bigger electro, is why I'm curious.

Any thoughts?

[Michael Allen]

While I'm at it, what's the point of so much filtering? There is mondo filtering in this amp, is some of it superfluous? Especially the 2x220uF caps.

[tubeswell]

Well back in the "old days" - apparently ;-) - when they built Marshalls and Fenders, 8uF was a pretty common (readily available) value and probably if they'd have had a bigger pile of larger-value caps they'd've used that. If you put a bigger value one in the Marshall, you'll get more residual charge between the rectified pulses (which will increase the -ve DC voltage slightly). For best long term results, the replacement should be about twice the voltage rating of what the bias supply is.

[Michael Allen]

Ah, that sounds reasonable. Nice to know!

[stokes]

Excessive filtering in the bias supply could be a bad thing.The cap takes longer to charge,leaving the power tubes with no bias voltage while the cap charges.I am not familiar with the Mig 50,but I have to assume the bias supply gets voltage even in standby mode,to make up for this.As for the 2X220uf caps,they are probably wired "totem poled" and equal about 110uf for the main filter,fairly common in the high voltage supply.More filtering here would give you a stiffer supply,equating to a tighter response,more perceived headroom,particularly in the bass response.Using a 20 to 40 uf here would give more sag,more of a Fender Tweed type sound.

[Enzo]

They charge up pretty fast. There is no current draw to speak of, so even though it is half wave, there is still not much ripple. I wouldn't try to overthink this filter. If the 8-15k-8 filter smoothes sufficiently, there won;t be any bias supply hum. If a single 1500 cap smoothes sufficiently, there won;t be any bias supply hum.

The engineers on one team didn;t likely bother to correlate their circuit with MArshall or anyone else. They needed a clean -60v supply (or whatever voltage) so they made one. More than one way to get a power supply filtered.

[Michael Allen]

took a look at the bias cap on my PCB and it's actually only 100uF. I guess that's enough eh?

Also, is there an advantage to wiring the capacitors in the filter supply in series, each cap paralleled with a resistor? Could I replace each set of 2Xcaps and 2Xresistors with one cap half the value and one resistor twice the value?

thanks for the responses!

[defaced]

100uf should be enough for your bias supply.

The reason you see series caps with parallel resistors is to filter high voltage supplies. For example, I don't know off hand of a 600v, 100uf filter cap for my HT, but I can take two 200uf, 300v caps, series them, and get what is effectively a 100uf, 600v filter.

In a series cap setup, the resistors are there to balance the voltage across each cap. If you have a single cap, you don't need the resistor, but you probably should have a resistor as a safety measure to bleed the high voltage off of the filter caps when the amp is off. I use 220k, 2w for this application, but there's nothing special about that value, just something I lifted from another circuit.

[Michael Allen]

ah got it. thanks again

[jrfrond]

In general, 100uF is plenty, and way more than most amps already have. As a general rule here in our shop, bias caps are 100uF/100V. That's what we use as replacements. The diode(s) always get swapped at the same time. It's cheap insurance.

[Enzo]

Parts are cheaper than labor, I like to tell folks.

I seem to recall a few schematics, Fender I think, where it calls for like 56 volts on the bias supply, and the caps are specs of 50v. Go figure.

I'm with John, 100u/100v makes a fine general all around bias filter.

[jrfrond]

Parts are cheaper than labor, I like to tell folks.

Which brings up a good point. If you even THINK that a part is bad, replace it, or at least sub it out with something close (for troubleshooting purposes) to make sure. Many a tech, including myself, has chased their tail trying to "do the right thing" and troubleshooting like a tech "should" (I am being facetious here folks!), rather than just tack-soldering or clip-leading a new part in and testing, which is a very valid form of troubleshooting. This is especially useful for mystery cases like OT's, where you are measuring resistance, injecting signal, etc., when you could just sub something in real quick with some clip leads and test the amp that way. No need to bolt in or anything, and no need for any fancy or correct transformers.

If you are a DIY'er, you DO have time, but in shops like mine and Enzo's et al, time can be BIG money!

[Dave Curtis, dB AudioTech]

Parts are cheaper than labor, I like to tell folks.


I'm with John, 100u/100v makes a fine general all around bias filter.

Ditto. Depending on the circuit.

I put a pair of 120u/100v in an Ampeg V4; took way too long to charge. Back to 10uf.