Home Brew SVT Bias Issues

[jakejoseph]

I have built a homebrew SVT with exactly the same components and schematics as the commonly available 6550 schematics on the web.
I plugged it in to start to debug last night, and all the voltages on all components EXCEPT the Filament Transformer seemed fine.

715 B+
365 Screen
350 preamp supply
-47 Bias
5.0 (!) V Filament

I took the filament transformer out, and it read 5.3V No load not in circuit - so obviously something is going on there. I ordered a replacement and contacted the original supplier to find out what was up.

SO prior to discovering the problem with the filament transformer, I attempted to bias the tubes statically. I could measure the bias, and every thing was fine UNTIL I adjusted the bias controls - then everything went wacky - the tubes started to red plate, and go tremendously over rated current specs.

I shut it down, and tried again - everything was OK until I tried to adjust the bias - wacky response.

Could this be because of the low filament voltage? I am waiting to trouble shoot more,, because I know that 5.0 volts is too low. But I wanted to start looking elsewhere - any suggestions?

J

[g1]

Maybe the bias pot is bad? The filament transformer is proof enough that "new" does not always equal "good".
May be unlikely, but worth checking anyway.

[jakejoseph]

I don't think it's the bias pot - the bias seems pretty steady with no tubes plugged in. And the bias is fine until I adjust it. As soon as I adjust the bias pots, the tubes red plate. Doesn't matter what direction I do it in, either.

The bias voltage goes to 0, but only with tubes plugged in.

[Enzo]

With no power tubes, does the bias voltage adjust smoothly up and down?

[jakejoseph]

yep - it adjusts perfectly smoothly from about -31V to about-65V. Spec is -47V.

[Steve Conner]

Are you sure you have no parasitic oscillations? Check with a scope.

What happens if you adjust the bias to say, -50v with no tubes installed, then shut it down, put the tubes in, and power it up again? Stable or unstable?

If you answered yes to the above: Will it amplify a signal without going unstable?

Does it still misbehave with only one pair of power tubes? Two pairs?

Does disconnecting the negative feedback cure it?

Heater to cathode insulation OK on your cathode followers? The 12AU7s or whatever (I forget) that drive the power tube grids.

[jakejoseph]

Could be parasitic oscillation - didn't check yet because of the low filament voltage.

Could low filament voltage be causing some parasitic oscillation?

The Bias voltage drops to zero with power tubes installed - only after you adjust the bias. Would parasitic oscillation cause that?

Thanks for all of the help. Would a photo of lead dress help?

[Steve Conner]

Low filament voltage will reduce the gain and power output of all the tubes. It should make parasitics less likely, not more.

If the bias voltage drops to zero, it seems more as if you have a bad tube shorting it out, or a fault in the bias supply that's preventing it from delivering the required amount of current. Parasitics would probably make your meter go crazy rather than just read 0.

Do all 6 tubes redplate, or just three on one side? Do both sides lose bias voltage? Try to narrow it down to a single tube causing the problem, by pulling tubes until it goes away.

Also recheck your wiring: you do have those 47k resistors going from the 12BH7 cathodes to -150V or whatever that rail is? Is the -150V rail present and correct? Does it droop when the problem starts?

[jakejoseph]

Thanks -

I am waiting to get the new filament transformer in until I do further debugging -but your suggestions are the first I will try.

[kg]

jake, i hate to sound like your mother, but it doesn't sound like you have a ton of experience troubleshooting tube amps...

please be careful.

with these kinds of voltages your margin for error is razor thin. if not zero.

[jakejoseph]

Thanks for your concern, kg, but actually, I've been working on tube amps for about 10 years, having built about 6 home-brews, many from scratch designs rather than copies. I have debugged many amps.

My questions stem more from the unusual way that the SVT drives the 6550's - using the 12BH7a's as drivers, rather than a typical phase splitter. This combined with the curve ball associated with the low filament voltage may me question whether this could be the cause.

The grids on the 6550's on an SVT are driven by the cathodes of the 12BH7's and I wondered if the low voltage on the filaments could make the way that the cathodes drove the grids unstable.

Again, thanks for your concern.

[kg]

yup, quite familiar on the SVT design as that's what i modeled the BAGA after so many moons ago.

it sounds like you have a way of firing up everything but the main B+ and g2 supply.

i suggest you thoroughly suss out everything up to each and every one of the output stage grid pins WITH the 6550s in place, but WITHOUT the HV applied.

you should see smooth and linear bias voltage as you sweep the bias pots. bias voltage should have the same range of motion for both phases. double check the dissipation THROUGH the bias pots... most pots are 1/4w max.

if your bias voltage is fed off the wiper, consider adding a "failsafe" resistor from wiper to raw -ve bias supply in case the wiper lifts off the pad or the pad open circuits.

if you have an oscope and a sig gen make sure that the output g1 pins see a solid (ie non oscillating) signal waveform. use a 10x probe to avoid circuit loading.

make sure you have stoppers on g1 and g2 on each output tube.

consider that you're imposing a voltage from the 12bh7 cathode to heater circuit. i've seen a lot of H-K insulation breakdown occur across many different tube types and it can be hard to troubleshoot.

finally do not be surprised if those 6550s aren't matched after red plating them, even if only for a few seconds. cathodes get damaged very quickly at such high potentials and dissipations, due to outgassing and/or excessive current.

[jakejoseph]

kg

Fantastic - thanks a lot for the advice. As soon as I get the filament transformer in and installed, I will give it a shot, and I promise to be careful!

J