kevin gilmore

Pictures later, the bicote really came out nice, so did the cocobolo. Not really liking the redheart, and the olivewood is so so.

http://gilmore.chem.northwestern.edu/woodies.jpg the bocote looks a bit dark, who knows...

The challenge will be to make the thing as easy as a kgsshv to put together and get to work. Still the voltages are a bit high so i'm going to add foolproof testpoints for adjustments.

bocote, olivewood, cocobolo and redheart on the way. still can't find a piece of snakewood big enough.

I decided to finish off the one with the crack in it, and found out something very interesting. At the same speeds and feedrates that i use for the LV and Rosewood, and with the same carbide tools, the ebony gets stinkin hot. WAY WAY hot. Even with 5 thousandth passes.... Might need some kind of cooling... AH well, on to other hardwoods.

Only the front stator amplifier is shown. The rear stator amplifier is identical to the front and connects to C7. (pdf updated for clarity) R12,13,14 do make sense, its a differential amplifier. But the bias is a bit off and it can clip. It is correct as shown, at least for this version of the amplifier. There is no negative rail. There should be to bias the output stage. More on the darkstar... Input chip is great, i used it on my new preamp. But its bipolar and has significant offset current. I added a servo to drive the vocm pin to make sure the common mode voltage is zero. If ray had a regular pot in there instead of the dact, you would clearly hear woosh as you turn the knob, problems he has had on other stuff in the past. If you run the darkstar unbalanced, output DC may damage your headphones. Nice job ray.

Spritzer as usual is right on the money. From extremely high resolution pictures that were provided to us the following was generated, and has been verified by an owner of the amplifier. This is an early version before the addition of the balanced input. There is one error in the schematic, and the inductor has been measured at 25 millihenrys by an inductance meter of unknown accuracy. http://gilmore.chem....wooaudiowes.pdf And that is the issue with this amp. Replace the pair of inductors with a larger tapped inductor and then the thing turns into push pull and the peak to peak voltages actually are what they should have been in the first place. The power supply is a different issue entirely. And for those that have not seen it... http://gilmore.chem.northwestern.edu/darkstar.pdf The .5 ohm current limit resistor limits the thing to a maximum of 1 ampere... High resolution pictures supplied by our friend skylab...

water and hardige lathes is a really bad combination. i've seen the result and don't want to be blamed for this. Now lots and lots of oil, that would be possible. If i do it completely NC, then that lathe is stainless and designed to take that off white stinking mess that is mostly water. (the rest of it is bacteria)

The high voltage power supply is the kgsshv supply with all the pnp's replaced with 2sa1968's and the 2sc3840 replaced with ksc5026, and the pass fets replaced with the 900 volt versions. (and changes to the zener string and feedback resistors) Eventually due to the march to surface mount, i will be making DPAK to TO220 converter boards so that we can use the 2sa1413, or the ROHM parts in place of the 2sa1486. This will apply to the kgsshv too. The low voltage (well +/-85 is low voltage) is a dynahi supply with upgraded parts. Due to all this stuff, and the heatsinks required its going to be the size of the T2. The only person that really needs this is pabbi1. Practically, voltage swings above the bias voltage may do damage to the heaphones, so really +/-600 volt supplies are sufficient. Yes its a stacked transistor design. No other way to do it with solid state.

So birgir asked for this one, but truth be told, i have been working on it for quite a while, and i have a working prototype. Specs: DC to 200khz +0, -.2 db Slew Rate >350 volts per microsecond Thd < .01% Voltage Peak to Peak, Stator to Stator 2950 http://gilmore.chem....n.edu/togte.pdf This is not for the faint of heart. The amp actually can go higher in voltage by a bit, but anything over +/-750 for the power supply gets to be Very difficult to regulate well. The +/-250 is actually resistor dividers off the +/-750 Its the only push pull fully dc coupled electrostatic amp that i know of Since this schematic i have replaced the 2s3840 with ksc5026 which is a higher voltage transistor and a better match for the 2sa1968.

With the 10k input resistors on the T2, the 22 ohms would have no effect. The 22 ohms should really be 50 ohms to be able to better drive high capacitance interconnect cables.

I'm efficient I had been thinking about how to do this from the very start, so once i got the cardas parts, it was actually pretty easy. The fact that cardas can't even supply a dimensional drawing of what is supposed to be their own part is very sad.

well this should confuse most people http://gilmore.chem....inputboard6.jpg Stuff it with one set of parts and its balanced, and you need 2 boards, stuff it with another set of parts and its unbalanced and you only need 1 board, to keep it the same size, for unbalanced you give up an input. The unbalanced input buffer is an lme49720 dual opamp, or your favorite in a gain of 1 setup. Needs a bit of checking now.

Fine, i'll make a single ended version. I may be able to get both on the same board if i make it a bit bigger. Then you stuff it one way for single ended and another for balanced. The input chip is not compatible with single ended. So there will have to be 2 different single ended front ends.

the first thing that happens is that all inputs get converted to balanced. So to use the thing as single ended you would need a bunch of dual rca to xlr adapters, and you have to jumper across the input amp, then it actually would work. The tape out would be unbuffered this way. I do have the cardas stacked rca jacks in house, so i could do a single ended board.

http://gilmore.chem.northwestern.edu/inputboard5.jpg the pass thru is much more easily done in software given the posibility of unbalanced input, balanced output. maybe a completely single ended version in the future.

I put in a bid on ebay, i'm sure i'll be overbid soon. I really hate ebay for this kind of stuff. I'll contact the other place on monday.

Finding the snakewood turning out to be very hard. I need at minimum a 2 inch cube. But 2 x 2 x 5 or better desired. If anyone finds anything interesting of an appropriate size, please email me the link to purchase. Air dried and no cracks... Also trying to find some merchaum and not having any luck there. I have found nice pieces of zebra wood and bocote...

lil-knight has the board files now, we are going to do a prototype board run to make sure everything is perfect. Same thing with the black hole.

cocobolo is also on the list. practically due to the setup time on the lathe, it makes much more sense to do 5 or so at a time, it goes so much quicker that way. Cut 5 on the band saw, live center and spin plate on the lathe, then step collet gripper for inside hole, then boring bar, then do up the aluminum pieces. Assemble, then micromachine. doing a marble one requires diamond tools, probably not ready to do that at the moment.

http://gilmore.chem.northwestern.edu/ebony9.jpg The "Knob of Voltron" is done and ready to ship. (the ebony one) I kind of like the rosewod one for now. Trying to find a decent size piece of snakewood.

We will soon find out how well the ebony knob tolerates the heat. Clearly the titanium knob and the aluminum knobs don't care. Machining corian turns out to be really miserable, dulls even carbide bits real quick. Which is one of the many reasons that wilson speakers cost so much. I did some corian stuff a while back, not going to do that again. I think i have a line on a piece of wooly mammoth ivory that is big enough to make a knob out of. I refuse to buy black market ivory for the obvious reasons. I think i can do marble... Lets find out. Voltron's knob should be done tomorrow when i have some more time. Then i have to find a suitable wood presentation box for it. Afterall you have to have a box to put your knob in

i'm beginning to hate ebony. http://gilmore.chem.....edu/ebony8.jpg tried a different direction for the 3rd ebony knob, broke after machining, and had 2 cracks by the time i got it home. The other one is perfect, so unless i screw it up, senior voltron will get his knob. I'm also really liking the rosewood knob.

There is a shortage of all sorts of transistors these days, and the floods in taiwan probably influence this very little, who knows. Its a massive pain in the ass to source parts these days.

i'm already using 8 bits for the relays, so the direct option is not going to happen via hardware. Via software its easy, just pick the input, set the attenuator for no resulting gain...