kevin gilmore

you cannot correctly setup this amp by measuring these voltages. you need a signal generator and a scope. leave the phase splitter adjustment alone and set the output plate voltages to about 320v. but this is only approximate. what you want for the output tubes is maximum peak to peak voltage with symmetrical clipping. what you want for the phase splitter is balanced peak to peak voltages between the 2 output tubes.

here is another koss circuit. this one seems built better.

you mean like this there was another koss circuit on ebay from a guy in ukrane, i did not save that picture. it was even worse. the soviet era toroid transformer is a real gem.

here you go pa443D

jumper connector pins 7 and 8 of the phono board connector. but really it should work fine with the board removed.

a little high voltage opamp porn

the d10 absolutely does use the pa443. same style of bent heatsink to get the heat to the case. birgir can post pictures of the d10 if he wants, and there are other pictures of the d10 with the heatsink removed. you might want to look at the datasheet, a single resistor programs the output stage bias current

pa79 mouser price is $56 each. but limited to +/-175 (its a dual opamp) i can't imagine there are any fake parts in that package so likely parts are disappearing out the wrong end of the factory. similar to the shure earbuds. mouser price on pa97 is $150 each. there are better apex parts to use in that voltage range for about the same price. you should be able to do a pair of carbon boards for the same $600, probably less. that $250 price for the portable includes the headphones. what it does not include is a lithium ion battery pack.

back board in the power supply next to the high voltage opamps has what looks to be a pair of 3 terminal regulators. one low voltgage power supply needed to drive all the input relays.

that does look like the right thing. 12au7 for gain driving 300b cathode follower. 70ma definitely pushing the tube pretty hard. not balanced.

i've been searching for over 24 hours now and cannot find a single example of a 300b used as a cathode follower. All 300b preamps seem to use an output transformer. And this completely makes sense as the voltage grid to cathode for a 300b is 75 volts. used as a cathode follower its performance has to be poor. and the opamp used for the headphones, from the apex website, this device runs class C. no output stage bias. furthermore the maximum voltage swing bandwidth is 3khz. and the slew rate is 8 volts/us. this has got to be a miserable sounding thing. there seems to be another amp just like this (for over $30k) that likely has some significant amount of solid state on the internal heatsink. 300b are great tubes when used as outputs in a push pull circuit. thats about it.

to do 10 amps you are going to need to boost the drive current. probably lowering the value of r1 to 120 ohms. same with r4.

golden reference was designed for only 5 amps.

measuring stuff like this correctly while avoiding ground loops and transformer radiation is very hard.

yep

here you go cfp3largeT.ZIP

m2's owned by someone else. too big for my room.

the 45k miles is a joke. i had 4 sets of the as3,3+ over the years and never got more than 30k. completely even wear. different car of course, but would expect the suv to be worse. those high mileage tires are not going to ride as nice.

its still 2 boxes. but sure if you want to build it that way.

nope, going to need two boxes. no easy way to stack the front end board onto the heatsinks. the 2 boards will take up much of the bottom. zero feedback generally ends up with a lower damping factor. so to boost the damping factor need more output transistors in parallel. and the sanken parts are actually 2 die in each package. and are to3p. i could go to 3 or 4 pairs of output transistors for even lower output impedance. the idea is to drive jbl m2's with the amplification they deserve. not a pair of crown industrial amps. and things like the susvara.

its already a split board. the power resistors make it impossible to shrink. joamat may come up with a solution but i doubt it. but you can download some free layout software and try to come up with something smaller.

won't fit in a 2u. if you do all the transistors on 1 edge its too long for the case. unless you want to do fully custom

soo to get it into a dissipante 3u chassis

actually 12.7 x 3.85 at the moment. size may change a bit

the board will be about 15 x 4.