kevin gilmore

switcher is completely open loop. double the input voltage, and it will try and make +/-500 or so. And parts will start flying.

never had a schematic of the power supply in that unit but i'm pretty sure its the filter resistor and should be the same as r41 i think the srm212 is very similar, will look later trace out the power suppy from the transformer, should be a couple of diodes directly to a pair of caps, then resistors from those caps to the second set of caps

6V filaments are AC the 300V comes from the 450V thru a resistor. no need for extra power supply

plate voltage with respect to the cathode is about 340 volts

no current source in the input stage of the exstata just resistors

triode mode no autobias, if the tubes are reasonably well matched and the power supplys are correct the most offset I have seen is about 5 volts c1/c2 couple from the differential input stage to the second gain stage. c6 and c8 do the same thing to drive the outputs

1mm is .039, and the holes are already .040

6sn7gta or 6sn7gtb only, otherwise arcy sparcy how big you want the holes on the caps?

James lin came up with this, I turned it into a circuit board. cascaded current sources et all. http://gilmore.chem.northwestern.edu/srxschematic.jpg http://gilmore.chem.northwestern.edu/srx6.jpg http://gilmore.chem.northwestern.edu/boards/srx6.zip I did make a few changes including balanced input

lets not forget the cavalli .000000001 % thd at 20hz to 20khz at any level including clipping

Those were SIT's, and evidently had a short life because all the manufacturers went out of business quickly. Static induction transistors. These Silicon Carbide Mosfets (and one company makes jfets) act like standard mosfets except are very high voltage, very high current and can run up to silly high temperatures. Also negative temperature coefficient, so no thermal runaway. And lower capacitance than the crap cavalli uses, so much easier to drive at high frequencies.

would not do 1oz copper. too thin. 2 oz minimum, the original boards were 3oz

about 120ma, measure across the 10 ohm resistors

does not need heatsinks. very low power.

yep, that would work.

the one i had used the 2sc3381 parts same silicon, one was a 5 pin, the other a 7 pin

James Lin insists that srx can be made much better with 10m90s cascaded into dn2540. I have a board pending verification. I have not heard it, so I don't know. The difference between being better in theory and actually being beter... ?

mpsa06 is fine for those, very low power. Fairchild version should be available for a long time.

The problem with that one is that it is incompatible with the way I make my knobs. The same guy had the standard shaft model which is what I cleaned him out of. would love quad black beauty or quad noble pot. new goldpt 47 position quad pot is over $450 might as well buy a P&G

Low input impedance amp. 10k quad is best. There was a guy on ebay selling them, but I cleaned him out. And I need more. Group buy thru partsconnexion is 100 pieces minimum. Which is stupid.

board version with surface mount outputs is done, just need to verify. people with current boards should buy mpsw06/mpsw56 now

mpsa06/mpsa56 can be substituted everywhere except the outputs on the kgdynalobal (have to check power levels on dynahi)

you install either the LO resistor, or the DT resistor. one is for dual tracking, the other is for local feedback don't install the current limit transistors.

I would hurry. last order date dec 31 last delivery date in may but the following parts are identical silicon in surface mount pzta06,pzta56 so will have to make adapters because they require 6cm2 of heatsink copper

well I just bought a lifetime supply of the mpsw06 and mpsw56 just to make sure. And at 1000 pieces, dirt cheap. TT electronics is probably the best bet, have to get a quote from mouser so group buy kind of thing. going to search available Japanese companies now. looking for 80v npn and pnp, 1 watt