kevin gilmore

I just for fun finished testing the new balanced dynahi into 8 ohms, and it actually works and works well. Unfortunately the power supply limits to about 1.5 amps, which ends up about 18 watts into the load, of which about 3 watts is pure class A. Everything gets very hot doing this. So if you put monster heatsinks on everything and beef up the power supply you can probably get more. The other thing is going to also run at +/-30 volts, but will put out 20 watts pure class A into 8 ohms. The power supply for this is going to be substantial.

sigma22's will work fine. Probably you want one power supply per channel. I never tried to drive speakers with the balanced board, that would be the equivalent of a 4 ohm load per side. Should work, might get very hot. Probably something like 18 watts. Birgir and I are working on a pure class A speaker amp (or lcd2/3,hf6) thing that is 20 watts into 8 ohms. Its a krell ksa50 clone down graded a bit but way more bias. All the new modern onsemi transistors.

1) i can do a board for 7189a's. But its going to be bigger because its now 8 tubes instead of 6. Also don't know what a 7189 is going to do at 800 volts. And its not a triode. 3) The bias pots take care of any differences. Self bias for these tubes rarely work, same as 6ca7 which mikhail tried to use with self bias. 4) bias pots are between -430 and -400 volts.

And you have the nice benefit that the supply won't blow up if one of the amp boards goes south. (or north if you live on the other side of the world) And its also current limited, so if one of the amp boards decides to short, no flying parts. I might be able to squeeze another .25 inch on the width if its that close.

AH yes, the age old question. About 12 volts. So an unreg of 412 volts. So if you pick the transformer correctly you want about 440 volts with the load. The 450v caps are rated 450v, 500v surge. If your AC line is 10% high, then you are pushing it. and if your AC line is 10% low, then you drop out of regulation. Better is to go with the 500 volt caps which are 550 volt surge. For a while the 500 volt caps were not available, but it seems people are making them again. Otherwise you have to go with the kgsshv supply which puts 2 x 450v caps in series, for lots of headroom. This also increases the physical size a fair amount. For the exstata you really want +/-300, so you set the zener strings for 300v and then you have plenty of headroom with the 450v caps.

updated power supply board with a pair of resistors for high voltage divider. impedance of the divider needs to be 1M or more. HE60 bias is 600V, so that should run fine on 580. Same as ESP950

here are the latest board layouts, if someone sees an error let me know http://gilmore.chem.northwestern.edu/llv2.jpg http://gilmore.chem.northwestern.edu/batepower.jpg http://gilmore.chem.northwestern.edu/bate.jpg

OK, now you are getting picky. Seriously, i want this one perfect before we make boards, and the more eyes (i.e group think) the better. Birgir is the "thought leader" on the project. I did make one mistake on the preamp board... But it works anyway And it works mighty well too. At least prototype runs are cheap these days. schematic http://gilmore.chem....u/batepower.pdf (yep more cut and paste) There are some 680uf/500V caps in stock at mouser, a little extra headroom always a good idea. not cheap though. http://www.mouser.com/ProductDetail/Kemet/ALC10A681DL500/?qs=sGAEpiMZZMukHu%252bjC5l7YSuaTY1g5Qh1YkGGiaJm46w%3d

I buy what i can find They did have a bottle of 13 year old, i saw no reason to buy both. Evidently they are getting a number of the buffalo trace things in early in the week.

It should certainly fit in neilvg's unit Its about time he got that thing fixed, he paid enough for it. All the ESX things have completely unregulated power supplies and no bias supply for the output tubes plus the electrostatic bias is the wrong voltage. You could always mount the power supply on the top of the box... The heatsinks are grounded. The square pads were that way on the kgsshv power supply, and this is just a cut and paste from it. Only took me about 1.5 hours to do this one. I can make the other ones square... wait a minute or two. I'm doubling up on the unreg and reg load resistors to cut the power and decrease the voltage across them to keep them within spec. The pappy van-winkle and the cuban are certainly helping. minute is over

Just scored a bottle of pappy van-winkle 23 year old family reserve. Yum.

power supply done, suitable for BATE and LL 7.8 x 6.12 inches, probably as small as i can get it. http://gilmore.chem....u/batepower.jpg This power supply also works if you are retrofitting an ESX, or want to build a point to point ESX. Finally finished the preamp board, tested all the audio stuff, and it measures below the noise level of my equipment. Will listen to it later, but that one seems now done too.

Measured power on the amp channel is 15 watts, so the 2 amp channels are 30 watts. i'm going to do a smaller simple regulation power supply board to go with it, basically an updated version of the original BH board with ixys as the current sources for the zeners. No one is really seriously going to build a 4 x pa89 thing. I've used them a lot in the past, and any mistake, and they blow up real easy. While an ultra simple design this would be a bad idea, and perform like ray's darkstar. No one really wants that.

This i believe. I however take at least 10 minutes to drink one. I must be seriously out of practice.

dry ice is -80c. liquid nitrogen is -200C (or so) liquid helium is -270C and is way the hell to expensive to try this with. (but i might anyway) I have plenty of dry ice, will try that next. I wonder what that would do to the taste of the drink if directly dumped into the booze. Likely turn it into a fizzie. Although i have a really nice new 25L dewar that fits in my car for carrying around N2, The dry ice might be a better idea. The ceramic balls i have are for high vacuum/low temperature work, and i tried it in liquid N2, and they did not break. The guys at work had the idea to take copper balls and coat them with teflon to make them innert. This is going to take a while to try, and my guess is that the teflon is going to crack due to differences in coefficients of expansion.

The plate resistors will be on the top side of the board with the tubes. All the rest of the stuff is going to be on the bottom of the board. Would mount the same way the T2 amp board does.

Another electrostatic release by the stax mafia 2 just this week http://gilmore.chem.northwestern.edu/bate.zip this needs checking, for errors, but all the sizes of the components have been checked with mouser.

That is why you plate the copper with gold. Or use surgical stainless balls. Or ceramic. The idea is to cool them down a bunch (i.e. liquid nitrogen) to keep jim's sazerac's cold for the length of time it takes to drink the whole thing. I've tried soapstone in liquid nitrogen, and it just cracks and turns to powder.

So a couple of you have purchased whiskey rocks, which got me to thinking about better things to use with more thermal mass. I came up with the idea of 1 gram pure .9999 silver bars. Or gold plated pure copper balls. What you think?

layout updated to support 12ax7,6dj8,6n1p ... http://gilmore.chem....rn.edu/bate.jpg yeah i know working on a low-res screen is a bitch. will update with higher resolution later. Could add current sources everywhere. Including the output tubes. That is going to increase the size and require heatsinks. As far as the resistor to the negative rail, Lots of people do that, but i see no reason to, and besides unless you are going to use a delay relay to turn on the high voltage, you would end up with 800 volts across a cold 12ax7. Which can't be a good thing. This is great sounding, about as simple as possible, and something anyone could build and get to work. Birgir can voice the thing by picking the capacitors, or anyone else can add comments. The only ones that need to be >800v are the ones between the 2nd and 3rd stages. Picture will be updated soon.

Try this one on for size... specifically 7.6 x 6.25 inches http://gilmore.chem....rn.edu/bate.jpg 4 x 12ax7 2 x 6BL7 same teflon sockets as the T2.

no one seems in the mood to do the pa89 group buy. I already have mine. So what do people think of this http://gilmore.chem.northwestern.edu/bate.pdf

anyone going to be upset if i insist on the teflon sockets?

how about 12ax7,12ax7,6cm7. won't be much bigger than that. The power supply however is going to be at least that size or bigger.

Yeah, it was compact. And yes it blew up. Evidently quite often. i'm thinking 3 x 6bl7 per channel, +/-400 power supplies, gilmore triode amp style with cathode bias and DC coupled output. Local feedback on the input, and local feedback on the output. unbalanced input only. (i'm thinking about this last bit) 10 amps of filament current.