kevin gilmore

actually i did one amp this way with multiple taps off the zeners... worked fine and all the parts were already there, just add one output resistor caps in parallel with zeners usually is a bad thing

actually WAY back then there were ceramic spacers for doing that. speaking of knobs, got in a piece of ebony, completely coated in wax. took the wax off and turned it round. completely black. then 3 days later 4 major cracks due to the wood being very wet. so pissed off. a piece of kiln dried ebony on the way. another vendor says there is no such thing as kiln dried ebony. precious metals are so much easier... want r-core transformers so bad...

wow, all the holes on the bottom in a straight line. good job. resistors really did not have to be that far off the board, but looks nice, all even. I'm sure birgir has a picture of a particular persons build (cough cough henyo cough cough) where the bottom of the thing was attacked with a hand drill and he should post it.

there is no hum whatsoever in either of mine, but I'm using power supplies that are probably 15 years old. most of it I would expect is radiation from the transformer. so with a really good torroid, you should be able to make it absolutely quiet.

you really want to use the dual linear supply, or something like amb's supply. Works much better at 16+ volts

absolutely no artic silver. The stuff is electrically conductive.

if that schematic is correct, the moron is reverse biasing the diaphragm. massive bad

the mpsa parts cannot handle the 1W for the output devices. everywhere else they are ok. same silicon, smaller package

that is slightly excessive, plus its not in stock. building up a parts order now. getting tired of $5 in parts and $7.50 in shipping

use plenty of heatsinking... 320ma on each it should be noted that the distortion figures above are unbalanced in, unbalanced out. balanced out reduces the second harmonic by about 6db resulting in a thd of less than .002% at these levels its silly and very expensive to go any further. singlepower squarewave into 50 ohms is .055% (no matter what Mikhail said, never designed to drive 50 ohms)

finally updating my test equipment... with a 30db attenuator and 50 ohm load... http://gilmore.chem.northwestern.edu/ssdynalobalthd.jpg surprisingly close to my 30+ year old antique

kgdynalobal7 is the all surface mount version. no one that I know of has made this one yet, so maybe check for errors. will look at the current mirror later, but the board would probably have to get bigger as its stuffed.

identical. pzta56 silicon exactly the same as mpsw56 only difference is package and heatsinking the silicon to the tab

for the surface mount version which can handle a bit more heat, 20ma per transistor 160ma on each rail per board

the current sources see the tails of the differential inputs. So really its very close to a dc level that does not change with input signal. maybe microvolts if that much, so the current sources really have very little difference. changing the current sources may change the current a bit which changes the output stage current, so you may be hearing that.

there will be a new version of the boards posted when i get home. difference is bigger mounting holes.

the pass transistors go thru the bottom, so you would have to remount them on the new chassis the transformer is 2.04 inches, so the 2.1 inches is right

make sure you use 500v caps, and it will work at 400v. obviously change the zener string and transformer. in other news, kgsshvcarbon boards by joamat posted. mirror images for the truly insane.

massively sweet. definitely needs a wood knob.

this should give you all a headache... certainly gave me one. http://gilmore.chem.northwestern.edu/kgsshvcarbonstereo.pdf

yep, grounded grid. There is much more gain in the output stage, so all the extra parts in the BH are not needed. So the amp is faster because of less parts in the signal path. 3 stages instead of 5. board is 6.3 x 2.75 so it fits on a 3 inch high heatsink. (do I want to do 1500 lbs of aluminum again... highly unlikely)

someone really needs to check this hard. http://gilmore.chem.northwestern.edu/kgsshvcarbonright1.pdf changed to multipage to make it easier.

of course, thats the point, i'm working on adding connector blocks now. joamat did 2 boards, left side and right side, similar to the t2, that way everything is symmetrical. but it also means that there are twice as many boards.

another reason why many of the boards are on sockets is rework. If everything was on one board, and there was a problem, fixes could be tough. this way, one mini board does not work and cannot be fixed, it gets tossed.

joamat is doing the layout on this. all of the stax brain trust use the same software which makes things easier.