Input Choke Too Small

[dre7]

Hi:

First a big thanks for this forum and all the information I've been able to glean here. I've been trying to electrocute myself for a few years now and in the process have built a few nice amps and repaired some others. The inevitable result of all this tinkering is a garage full of parts. So my idea was to build a "junkbox amp" out of them. PP 6B4G/ transformer splitter/6J5 driver/input transformer. Just got through breadboarding and the PS threw me a loop. Everything went fine cranking up the variac until I got past 100vac or so then the voltage starting runaway. Before runaway, B+ was hovering around 340v, a bit higher than I wanted, but I believe still within the capabilities of the Sovtek monoplates I'm working with. It's choke input PS and I think what's going on is the choke is too small (4 Hy) and "disappearing" past critical voltage at which point my PS becomes cap input.

I know the best solution would be a larger choke, but there are cost and chassis space issues as well as a garage to clean! So in that spirit, is there anything to do?

My thought was a bleeder resistor at the end of the voltage rail, but how do I figure the amount of current to pull to keep the choke from saturating?

Thanks for any help. This is my first go at trying to design something.

Will also try to post schematic.

[Eli Duttman]

A reasonable approximation for the critical current (in mA.) is given by V/L. It works out that a 4 KOhm bleeder resistor is what's needed with a 4 H. 1st inductor.

If you use filter caps. that would survive under cap. I/P conditions, you can increase the value of the bleeder resistor and allow signal circuitry draw to complete entry into the choke I/P regime. Slow starting rectification, like that provided by 5AR4s and damper diodes, is your friend. The signal tubes will be hot and ready to conduct by the time B+ rise occurs. That allows the WVDC of the filter caps. to be held down. Subtract the idle current of the signal tubes from the critical current. Use Ohm's Law to compute the value of the bleeder resistor needed. Round down to the nearest "standard" value.