kevin gilmore

if the 8n80c shorts either drain or source to gate, then the zeners are going to pop. the 24v zener is strictly to protect the 8n80c on power turnoff, does not do anything for power on. suggest you purchase known real parts from mouser 8n80c needs thermal grease or thermal washer, but that does not have anything to do with the zeners blowing up.

i will look at the board layout later, but i thought it said 36 vac center tapped which is the same as 18-0-18 from now on, i will make that clear. the fat hvsic supplies are probably mislabeled too.

there was some discussion over there and calculations of the current needed to drive the headphones at full power at 20khz. That number is something like 18ma or so, the T2 is 16ma, a lot of the other amps are much less. at 20ma, the carbon has both a faster rise time and wider frequency response than anything else out there. So for a while it seems bright, then it seems more natural. Definitely a difference.

my guess is that there is a short on the 10m90 that i cannot see, or that part is dead resulting in way to much current to the zener string

its soldered, its just the back set of holes

so the default would be 25ma on the output stage and 15ma on the driver stage so 90 watts for the output stage and 54 watts for the driver stage (both at 900/450v) plus what ever differential on the power supply so close to 200W is probably right

you really want most of the power in the resistor and not the pot. that said, i don't think andy/steve have the servo turned on, but the thing is already so stable... i still have 2sa1968's and 2sc4686's available, i'm keeping the rest of the stuff. also have fqpf8n80c the circlotron is a must build, but way more complicated.

andy/steve have a the first working carbon from the group buy boards. (fat power supply boards) (400v) to hit 20ma, going to need to change the tail resistor on the offset pot probably to 100 ohms.

schematic updated with parts for latest board

correct, but I posted over there so people could understand the complete nature of the beast.

lousy excuse.

well, what are you waiting for.

absolutely fully direct coupled all dht amp makes the circlotron look tiny.

someone with a screen room and a 8.5 digit dvm needs to test

only if you stick them in backwards. (or if they are labeled backwards which I have seen) updated schematic... you gotta be kidding. probably this weekend. and then I am going to have to show how you calculate the trimmer resistors for other than 20v

unlikely it needs heatsinks, or just really small ones.

nope

that is exactly what I had in mind. except for the more basic units, no GR supply, and the rk50 replaced with a rk25 and its associated board. plenty of room for the transformer, xlr connectors and iec connector. the boards are 50th short on the width, so all the heatsink holes can be lined up end to end on .5 inch spacing. transistors on the bottom for the GR supplies, no problem, hanging in the air with the smallest of heatsinks also works. board layout for the rk50 board in the usual place.

you do have version 1.15 or later of the power supply board? the first version of the board had a 1khz noise problem that required adding a cap and resistor (x4) and I remember that knob

I have posted the goldenreference6 board, has options for center tap transformer, and dual windings for next time

the HV boards, the 3n254 is .15 spacing, and the board is .15 spacing any of the tiny 1 amp ones will work. on the LV board the spacing is .20, and its definitely a different bridge, the same one used in the ksa5 power supplies rated for 10A I think

this is a reasonably cheap version, probably a very good idea if you are doing T2's http://www.micromark.com/microlux-tapping-fixture,8363.html

the name of the game is to keep the tap completely perpendicular to the surface i got a super old tapping jig that does just that, so simple to use, and the 10 inch or so handle makes it easy to figure out how much torque to apply

i hate microsoft remotely administered computer requires an ok to change the kernel... (brilliant) fixed now. does not seem to happen very often