Measured DC voltages are needed to know if the EF184,s are working at the correct operating points. There should be approx. 10 -14 mA of current flowing for each valve and the cathode voltage should be approx. 1.8 - 2 volts. A check should be made to establish the anode voltage with and without the 220kohm Schade feedback resistor connected.
miniwatt,
Your CCS (constant current Sink) should work well, as long as the burden voltage across the CCS keeps the NPN out of saturation when the music signal is at its maximum.
I designed a CCS very similar to your CCS.
A resistor from B+, and a cap to ground, "powers" the LED to 1.7V.
There was a 100 Ohm base stopper from the LED to the NPN base.
The emitter had a resistor to ground. With about 1V across that resistor, the resistance dictated the current. 1V / I = R
Set R to the desired combined 2 cathode current (calculate and wire it in, no rheostat needed).
There was plenty of voltage across the NPN collector to emitter, to work well with a 12AU7, and with an ECC99 too.
One thing about a phase splitter with CCS to the cathode to cathode connection (no resistor between the cathodes), as long as the plate loads are very well matched, the plate signal amplitudes will be equal (as long as there is no grid current; if there is, turn the volume down, and if there still is grid current, replace the tube(s)).
That means there does not need to be any potentiometer or rheostat in the CCS and cathode circuit (be sure the next stages Rg resistors are very well matched too).
The other thing about that kind of phase splitter, is that with signal voltage swings, the cathodes only move 1/2 as many volts as the control grid (that means the CCS does not need to have as small of a burden voltage, becuse the cathodes do not swing very far).
The disadvantage is that the CCS phase splitter gain is only 1/2 of what a paraphase pnase splitter has.
Your CCS (constant current Sink) should work well, as long as the burden voltage across the CCS keeps the NPN out of saturation when the music signal is at its maximum.
I designed a CCS very similar to your CCS.
A resistor from B+, and a cap to ground, "powers" the LED to 1.7V.
There was a 100 Ohm base stopper from the LED to the NPN base.
The emitter had a resistor to ground. With about 1V across that resistor, the resistance dictated the current. 1V / I = R
Set R to the desired combined 2 cathode current (calculate and wire it in, no rheostat needed).
There was plenty of voltage across the NPN collector to emitter, to work well with a 12AU7, and with an ECC99 too.
One thing about a phase splitter with CCS to the cathode to cathode connection (no resistor between the cathodes), as long as the plate loads are very well matched, the plate signal amplitudes will be equal (as long as there is no grid current; if there is, turn the volume down, and if there still is grid current, replace the tube(s)).
That means there does not need to be any potentiometer or rheostat in the CCS and cathode circuit (be sure the next stages Rg resistors are very well matched too).
The other thing about that kind of phase splitter, is that with signal voltage swings, the cathodes only move 1/2 as many volts as the control grid (that means the CCS does not need to have as small of a burden voltage, becuse the cathodes do not swing very far).
The disadvantage is that the CCS phase splitter gain is only 1/2 of what a paraphase pnase splitter has.
My Y chromosome is actually Dutch, first in America 1634 in New Amsterdam as "Hoornebeeck" as expressed without Umlauts. Had a wonderful German teacher in high school, Frau Guenther, originally from Czechoslovakia. She told us harrowing stories that have shaped my life.
Yours is an important point about language in the current wide world, but also here on diyAudio. Americans, in general, can't speak or write in anything except English, so everybody has to speak English. It's depressing to Americans how much better the English of so many Europeans and other non-native speakers is than our English. Well, tough luck, we've got all the aircraft carriers. For a while.
Arf!
Chris
#41, my error ; that should be the 100kohm resistors from anode to anode, not the 220kohm grid resistors.
I also have severe doubts that there is enough voltage over the CCS's. Remember the EF184's very high transconductance (it's a frame grid pentode!) that usually results in rather low control grid quiescent voltages at the operating points. In your case, it works in one channel, because it's just enough, but not in the other one, due to tube tolerances maybe. Tie the CCS's to the negative bias voltage rail - and leave out those potentiometers between the CCS's and the cathodes, as yet said before. They only make things worse.
Best regards!
Best regards!
If the CCS goes to the negative rail, the LED current can come from ground. Agree about the balancing pot being unnecessary for AC balance, but can help with DC balance (get the same anode voltage on each EF184), the frame grid EF184s can have quite a lot of variation between them.
I want to thank everyone for the ideas and suggestions. I changed a few things and the amps work like a dream now.
I changed the CCS power supply from CRC to C and a 7805. Seems to help the CCS do it's job, and it got rid of residual hum..
I threw out all the trimpots for balance etc.
I changed the anode resistor on the bottom driver from 10K to 12K. That pretty much solved the balance problem.
With the 100K feedback R sound was great, but a little harsh and tiring. No bueno.
Changed that to 56K. Bingo! Definitely hit a sweet spot there. I may play around some more with that value, small changes make a lot of difference.
I don't know how to calculate the amount of feedback yet, but I do know it is A LOT. Some tube heads may not like that.
Thanks again for all the help, I learned a lot!
I changed the CCS power supply from CRC to C and a 7805. Seems to help the CCS do it's job, and it got rid of residual hum..
I threw out all the trimpots for balance etc.
I changed the anode resistor on the bottom driver from 10K to 12K. That pretty much solved the balance problem.
With the 100K feedback R sound was great, but a little harsh and tiring. No bueno.
Changed that to 56K. Bingo! Definitely hit a sweet spot there. I may play around some more with that value, small changes make a lot of difference.
I don't know how to calculate the amount of feedback yet, but I do know it is A LOT. Some tube heads may not like that.
Thanks again for all the help, I learned a lot!