kevin gilmore

its also a jfet. in any case i will try and buy a few and test them. i assume people are using them as current sources, would like to know how they actually perform.

true, but for $10 total more in parts it can be MUCH better.

re: 717, open it up and listen to it, when the sound goes away, look at the led's and see if all of them are lit. If none of them are lit, then you are having a thermal cutout inside the transformer... the srm1mk2 was made between 1985 and 1990. As such its more than 20 years old. More than enough time for the electrolytics to dry up. So people who still have them and don't like the sound should probably change out all the power caps. At this point the 717's are now 10 years old and probably need new caps too.

well there are definitely different kinds of distortion, mainly 2nd harmonic, or mainly 3rd harmonic. and sound signatures are different too. push pull sounds a particular way, whether solid state or tubes. same thing with the various kind of single ended things. There is also soft clipping vs hard clipping. all the stax amps are really singled ended with constant current sources for the solid state amps, and resistors for the tube amps. (the t2 is something else entirely) the exstata output stage is a constant current source. you can add a current source to the exstata gain stage and make it bunches better. The vhv amp i posted a bit ago is the only example of a solid state push pull amp that i know of that actually works. It depends on stacking transistors to meet the necessary voltage requirements, and its performance suffers as a result.

i'm thinking seriously of doing a version of the board with a simplified current source that performs just as good (1000 volt ixys part, thanks marc) and with standup heatsinks. Will make it lots easier and cheaper for people to build it. Power supply will be the same thing. Hopefully will get started on it friday. Will cut the power supply voltages to +/-450 just to make sure the current source has a bit of headroom. if anyone knows of a depletion mode fet that does more than 1000 volts and will run at 1ma, let me know.

There is more than one frank... I have a MIG2 dynamic amp (similar to BH) with 8 x 6c33's 6 filament transformers alone. Consumes about double the power of the T2. Even though i published schematics, no one but me ever built one. I have an all DHT direct coupled electrostatic amp. Its so expensive i can't even afford it. (plus i never finished doing the high frequency isolated filament transformers) $8k for the tubes alone.

i measure with pabbi1 listening levels... http://gilmore.chem.northwestern.edu/srm1mk2.pdf even so we are pretty close. -60db second harmonic is .1% -85db third harmonic is better than .01% i was quoting DC power supply voltages, not AC voltage swing. Ray samuels is the only moron i know that pushes 317HV's to those kind of voltages. Way in excess of their 150 volt ratings. (i think there is a 225 volt version now) And besides which there is no high voltage version of the negative regulator. Now what you could do is run the transformer on 120 vac, and set the jumpers for 100 vac, then use simple zener/fet regulators... That would work great. And the transformer is going to get a bunch warmer.

the SRM monitor is a different schematic... pretty sure... stock exstata +/-300 volts ... 1.0% THD stock srm1-mk2 +/-350 volts... .1% THD stock srm-313 +/-350 volts... .01% THD I would add an auxilliary power supply that is fully regulated. for the srm1-mk2 it would be +350, -350, -370.

now i remember... the srm-1 mk2 is the one before the srm-717... i think they are close, but i would pick the stax unit, then mod the power supply.

don't have the srm-1 schematics at hand at the moment, so i'll speak on the t1,t1s,006t,007t all of which are the same identical design, and are grid drive with the tube as the output stage. If you sub out the tubes with 6s4's things get much better. And if you add a current source instead of the plate resistors, things get better yet. exstata tube hybrid uses 6s4 in grounded grid as gain stage, and solid state output section which is the same as the srm212,313,717. And i assume the srm323. 727 is definitely different. which is better in stock (unmodified form) , srm313 vs exstata solid state... in my opinion, the srm313 is better. clear as mud.

well that depends on which datasheets you look at. most list the 6cg7 at 450 volts, some list 500 volts. on the srm600 the power supply voltages have been lowered to +/-300 volts to accomidate the ecc99 6s4's are rated for vertical flyback use in crt's, and easily run at 1kv.

Of course it does. Lets see, running a 450 volt MAX tube on 700 volts. Does no one ever bother to read the datasheets?

here is another one pointed to me by minivan before at TEN percent distortion. (yep completely open loop, no feedback) http://gilmore.chem.northwestern.edu/vhvtest1.pdf (well some local feedback around the output stage) and after at .01% THD. http://gilmore.chem.northwestern.edu/vhvtest2.pdf with added low impedance input and nelson pass super symmetry who is going to be the first to build it. 6 power supplies. Think T2 power supply only bigger... !!! +/-660 (650) +/-240 (250) +/-87 (90)

reasonable signal levels. 100 volts peak to peak. amp capable of about 600 volts peak to peak. (stator to ground) anything sounds better than a transformer box. unless you like that kind of sound.

Before or after spritzer modifies the hell out of it?? A 6s4a is a WAY better tube than any 6fq7/6cg7! not sure about the ecc99

I should also mention that this is a relative thing. many people would be happy with an exstata in stock form. spritzer and i are not in that group of people. Not after listening to the T2, BHSE, BH... whether it is fair to compare an amp that is $500 in parts to an amp that is $4000 in parts is not my problem

OK, lets see if i got this right, since my dsl is down, i think i brought all the right stuff to work. So as spritzer says, before CJ, i built by hand an exstata solid state version. Perfboard version. And it sounded lousy, and it tested lousy. ONE PERCENT thd at reasonable signal levels. Fast forward a bit, i got alex's boards, built them with exactly the same parts as the BOM and guess what, exactly the same performance. synthesized here. http://gilmore.chem.northwestern.edu/exstatas.pdf sure enough slightly more than ONE PERCENT THD (second harmonic, top trace in distortion graph) and third harmonic pretty miserable too. so i started investigating and came up with this. http://gilmore.chem.northwestern.edu/exstatas1.pdf way better, .01% THD. and the sound is night and day better too. but when you clip the thing, it got nasty. SO... http://gilmore.chem.northwestern.edu/exstatas2.pdf but really, the more i play with it, the more it looks like what it actually came from, which still performs almost 5db better. http://gilmore.chem.northwestern.edu/srm313orig.pdf of course with modern day parts you can bump up the srm313 to higher voltage... http://gilmore.chem.northwestern.edu/srm313.pdf and of course you could do the same with the exstata's, but why bother. fact is that the 2sj74 is a lot less trouble than the j271 and you don't have to buy 50 parts to get 4 that work. and guess what the same thing for the exstata tube version. http://gilmore.chem.northwestern.edu/exstatat.pdf and with fixes... http://gilmore.chem.northwestern.edu/exstatat1.pdf fact is that the kgsshv still performs better than all the above and at higher voltages. i'm working on it to make it even better now.

Don't make any mistakes with this amp... Doing so would cause the output transistors to fail, and any modern replacements are going to sound completely different. As i remember there is virtually no short circuit protection on this thing. Should be a great headphone amp, or if you have k-horns... which is what it was designed for.

Yep, that works too... where did you find those white caps ??

some schematics say 230 volts for low bias. Some say 260. the resistor is indeed 5 meg. except for one very old schematic that says its 2 meg. i would go with 5 meg and 250 volts.

Well the power supply was not really designed with low bias in mind. but a rearrangement of the zeners allows for the right low voltage, and extra wire, and probably a little circuit board with 2 resistors and 2 caps needs to be added to the power supply. This is corrected on the kgsshv board and power supply. The alternative is actually quite easy. The 250 volt power supply can be used as the bias and is already very low noise and accessable. As there are two different wires for the two different stax jacks, this is the no modification approach. So you can have one high bias, and one low bias jack at the same time.

You already have 2 stax jacks for that, and you know it. But i will send you one more to replace the defective one plus the 2 knobs and the dact bracket as soon as the stuff shows up. The solid single piece BH is indeed great. I hope to do even better with the solid single piece T2. I think we are both in a race to see who can spend more on massive solid chunks of aluminum.

Not exactly. In fact mikhails amplifier was a fully dc coupled amplifer with a HUGE amount of dc drift over time and temperature. And the drift was not necessarily positive or negative, and it was well over 1 volt or more. So mikhail had to put in an output cap, and since he did not know the direction of the drift, he had to put in 2 caps back to back in series. But then because of the caps he picked, and the time it took them to actually charge up, he biased up the center pin, so that it did not suck so bad (meaning no bass for the first hour or so) and then told people to keep it turned on all the time. Serious clusterfuck, every part of the amp. http://gilmore.chem.northwestern.edu/dinglepower2.jpg

well i have some of the connex teflon ones, but those are very hole size specific and don't match any standard pattern. Also have a couple of the old ceramic ones for the original kgbh, similar to the ones used in your amp...

a stax of staxjacks http://gilmore.chem.northwestern.edu/staxjacks.jpg i wanted to make 40, but friday got in the way... this is enough for a while, the 4 people getting chassis this month, let me know whether you want low bias or high bias jacks. If the material comes in today, i will start on more angle pieces. All of the parts for the 4 are done, so they will ship as soon as the stuff gets back from anodizing.