kevin gilmore

ULTRA REFERENCE electrostatic headphone amp KGSSHV Carbon for Stax SR-007 SR-009 | eBay they are 1500uf caps as i remember.

the guy in vietnam is using even bigger caps. so no reason why it should not work.

I think that caps that big can be very dangerous. So if you are going to use them, find a way to make sure you can never come in contact with either of the terminals.

musican pegasus is probably in the same range

depends on which version of the pcb.

paradoxper settled on 50k feedback resistors and 4.2k ZF resistors. to increase the gain. both have to match so there is no volume difference between modes

SS mode its 25k (or 50k) /5k as long as there is enough gain in the front end. in ZF mode front end is a voltage to current converter. So 10k in parallel with whatever you pick for the ZF resistor. Much easier to do in spice.

i have the older model, never had any trouble with it

the harder the wood the easier it is to machine it. african blackwood has less issues than ebony.

the lathe is very decent, and capable of cutting screw threads if you ever need that. even though its early 1960's vintage. the other stuff is not capable of anything of high accuracy. at least not without digital readouts. And its heavier than it looks. just the lathe top is probably about 700 lbs

the 47uf caps have to be populated. the 600pf caps are optional, paradoxper decided not to populate them

this is what you want i think. kgdynalobal9 - CADCAM.ZIP

i went back and looked at everything. there is no servo in the amp part itself because its in the front end. the latest file which should be posted somewhere above is cfp3rssfixedsssw also attached here. cfp3rssfixedsssw - CADCAM.ZIP

which ever has the later date. definitely has the latest servo

so the idea is a 10 hour runtime. so a 3 cell lipo battery with 2.8 amp hour batteries is the target. if it has to be a bit bigger then it has to be a bit bigger. the amplifiers are 200 microamps output stage bias with 1.5ma current limit. Could certainly bump up the current limit a bit. the benefit of kerrys power supplies is that they are extremely small, max component height is .3 inch. if we change to a transformer based supply then the transformer is .8 inch. if we use the hvm supply for the bias it is .5 inch high. Smaller is better! For a given chassis size more room for the battery is a good thing. The final product will likely be a single circuit board with the power suppies on one side and the amplifiers on the other. The components with the largest height will be the stax jack and the volume pot.

the 300 volt supplies are about 110ma each which is acceptable. the bias supply is 200ma. This needs to get fixed. there is a resonant hvm module that does the bias for 25ma, but its a $135 part. if i could get someone to make us a custom switching transformer, we could do better. but no one will even return my calls or emails.

In progress Amp design by me, power supply designs and circuit board layout by kerry, support by birgir +/-300 volt power supplies. complementary push pull all bipolar design with cmos style multiplying current mirror.

pretty sure you have the pinout of the 2sa1699 wrong. emitter connects to source of the jfet. At least it does for srm212 and srm252

suitable replacements seem to be 2sa1552 / 2sc4027 toshiba tta004b / ttc004b also look compatible

amir is just another idiot with an audio precision. And is concerned about .000000001 thd. And is otherwise clueless. look at his review of glite.

not ideal, maybe find 20k?

+/-12v should be fine for a balnced setup. 50 amp power supplies would be perfect.

you need AC 120v not 12v and that is for low bias only

this is not a balanced input amplifier

wire the headphone grounds directly to the amp boards, not the power supply