kevin gilmore

balance pot has to be set in the middle or bad things may happen. normally the pots are set to middle out of the box. its usually pretty hard to do diode checks in circuit and get reliable info if all the leds are lit it is working and maybe the one pot is shorted

http://gilmore.chem.northwestern.edu/ubaltobal.jpg rest of the parts on monday

you need a better solder sucker. but what you can do instead is remove the 4 x 2.2k resistors set the pot for maximum resistance to the xlr jacks, and wire a new one that way

the pcb is wrong. but you need to check the pot to be sure the pins on the pot should be ground,out,in

knobs need to be 2 inches or more thick. reason is I have to have something to grab onto reliably to put the insert in. once the insert is in, I can bring it down to about .8 inch deep. machining into the insides of an expensive 2 inch collet chuck might get someone really pissed at me if I make a mistake. I have enough now for about 20 ebony knobs. but none are absolutely black like the few I had made previously. will be getting enough wood to make doug a pair of panels anyway. need to find someone that has been sitting on a piece of pure black ebony for 30 or 40 years.

this is how it should be wired, but the pot is actually not wired like that if the footprint for the alps is correct in their documentation http://gilmore.chem.northwestern.edu/chinesekgss.pdf

the pot is definitely wired wrong. what he probably did is turn the pot into a variable resistance for current output dacs. This would then require all the extra voltage gain. the whole input section needs to be re-wired. will draw up what I think it should be when I get home. the 4 resistors to ground by the xlr input jacks. 2 of them are 49.9k ?? . R40,R47 What are the other 2.R37,R42 the 4 resistors from the pot to the input fets are 2.2k?R9,R10,R11,R12 the 4 resistors on the bottom of the board, what is the value? the schematic I did showed only one channel. If you replace the 154k resistors with 100k, you have to replace a total of 8 resistors, 4 for each channel.

what is the finished size of the side panels if they were flat? going shopping for more ebony. (both kinds)

just got in by accident a piece of ebony 15 inches x 2 inches x .5 inches. probably not long enough for 2 end pieces

sure looks to me that the pot is wired wrong. input goes to the wiper, output goes to the amp... ??

rca and + input of xlr wired in parallel. the switch switches the - input of the xlr to ground for rca. guess they could not buy more stax jacks

needs an ebony knob.

real numbers. (single ended input) measured single stator to ground kgsshv with 450v power supplies and 2 x 100k feedback resistors with 572 millivolts rms input produces 800 volts peak to peak output (not clipping) Chinese kgss with 350v power supplies and 2 x 154k feedback resistors with 320 millivolts rms input produces 700 volts peak to peak output (just barely clipping) so with a 1.5v rms input the pot if it was linear would have to be at 20% or less to avoid clipping. and the matching on the pot is going to be lousy at 20%. if you change the 154k resistors out to 100k resistors, the input at clipping would be at 500 millivolts rms. Which was the original design spec. nothing is wrong with the pot. (other than its a cheap piece of shit and used poorly) you probably need a 10db input pad. and you need to cut the gain by changing the feedback resistors back to the original design spec.

its not the pot, but when you get it and take it apart, take a picture of the bottom of the board.

the gain is way high. I don't know why. The 100k feedback resistors are actually 154k. So its about 3.7db high in gain (unbalanced), 7.4db high (balanced) There are resistors across the pot on the bottom of the board, I never looked into the way they were wired. NO reason why those resistors would be changed, the output bias is normal, so they could be changed back to what they are supposed to be, then readjust the balance. http://gilmore.chem.northwestern.edu/chinesekgss.pdf birgir thinks its a linear pot that was converted to log via the resisistors soldered to it on the bottom.

I don't think it will. might be interesting to find out, but I have enough cheap meters, don't need another one. if you have 50 feet or more of cable you want to test you can drive it with a square wave 20khz or more, put a scope at the other end, put a pot at that end, and adjust for best looking wave. Then measure the value of the pot

sr-003 are the in ears with the standard pro-bias plug

only the input transistors are jfets. Everything else is bipolar unless you use the ixys parts which are depletion mode mosfets.

I'm sure the power supply is good, but you had better double check all the rewired transistors. Something very wrong, and I know that board works

very nice, like the color

the only amplifiers I know of that are absolutely flat THD vs frequency are CFA's. No way to build an electrostatic amplifier that way. my equipment is much older than justins gear, and the T2 when measured the same way that stax measures it comes out better than .01% No way am I shipping my T2 to Justin to have him measure it with his pile of late model Tektronix stuff. people over there lacking a clue do not understand that an electrostatic amp with a minimum gain of 54 db (eddie current) and a typical gain of 56 or 60db is going to have to perform differently than an amp with a gain of 6db.

high compared to what? stax amps are rated at .01% at 1khz. eddie current electra is between 1 and 2 % same thing with the wes if you really want .000000001% thd, you have to buy a LL

that is what the megatron is. All tubes. does keep the room warm in winter.

actually we should rename this thread KGITSOCH because this input section is virtually the kx-r with a few mods that make it even lower distortion, lower noise, and faster. No feedback of any kind. Should be doing a test run of boards very soon http://gilmore.chem.northwestern.edu/ubaltobal2.jpg

the heatsinks on the regulator transistors limit the energy to about 2 watts each. So what they do is limit the pre-reg voltage, on my line voltage, its about 380v not a lot of headroom. this is one of the many differences between this thing built well, but on a limited parts budget, and the massive overkill we typically do.