Operating in class A or class AB? Any distortion measurements? The usual problem in class AB would be crossover distortion, and with no feedback this could be noticeable. I assume the bipolars are running off of reduced voltage rails, like maybe +/- 48V (and with no output xfmr).
There actually will be some local feedback to the tube if the screen grid is at a fixed voltage with respect to ground, since the output V will be effectively subtracting from that.
Is this using a Vbe multiplier to bias the bipolars for temperature compensation? You might want to consider NJL3281D and NJL1302D thermal-trak bipolar outputs for better control of the biasing versus temperature. (comp. diodes in the same package)
TubeCad had a hybrid ampl. using these (see bottom of page):
http://www.tubecad.com/2006/10/blog0084.htm
That design used several bipolar current gain pre-driver stages to drive a bank of outputs, but with a big EL86 you should be able to dispense with the pre-drivers. So you would only need one pair of thermal-traks and their two comp. diodes in series for biasing.
Don
There actually will be some local feedback to the tube if the screen grid is at a fixed voltage with respect to ground, since the output V will be effectively subtracting from that.
Is this using a Vbe multiplier to bias the bipolars for temperature compensation? You might want to consider NJL3281D and NJL1302D thermal-trak bipolar outputs for better control of the biasing versus temperature. (comp. diodes in the same package)
TubeCad had a hybrid ampl. using these (see bottom of page):
http://www.tubecad.com/2006/10/blog0084.htm
That design used several bipolar current gain pre-driver stages to drive a bank of outputs, but with a big EL86 you should be able to dispense with the pre-drivers. So you would only need one pair of thermal-traks and their two comp. diodes in series for biasing.
Don
Hi, operating in class AB(bias 80ma). Bipolars at +/-22v.
The bias circuit is simple, to 3k2 resistors , 2 transistors 2sa1265( (I have lots. Only Base and E in diode, termal coupled, bypassed with 100uf cap) and finaly 0r33 in E of each power tr.
EL86 (triode configuration) in a cathode follower configuration, drives directly these pushpull configuration via a 100uf cap in series with a 470r resistor .
npr
The bias circuit is simple, to 3k2 resistors , 2 transistors 2sa1265( (I have lots. Only Base and E in diode, termal coupled, bypassed with 100uf cap) and finaly 0r33 in E of each power tr.
EL86 (triode configuration) in a cathode follower configuration, drives directly these pushpull configuration via a 100uf cap in series with a 470r resistor .
npr
npr said:
I think you have current feedback through the finite beta of the BJTs.
Maybe you have an impedance converter. Maybe this helps with
crossover distortion also as the current feedback is coupled into your
40mA biased CF stage. Very interesting. Might be a nice 25 watt amp.
I never trusted beta... not constant or predictable enough. Unlike Mu...
But I never just tried a beta-dependent linear circuit. Wavebourne had
one also some months back, but in a BJT-tube darlington.
Drive the EL86 with the correct voltage swing and the correct fase (I use in classA a 6bm8\ecl82, triode section at 0,75ma, pentode at 10ma(CCS)).
I do not know the values of distortion. But I believe that for the gotten values of power(<25W), the bipolar ones are linear without feedback.
In summary, the circuit have 2 stages of voltage amplification and 2 of current.
I tested it, and please me the results.
npr
I do not know the values of distortion. But I believe that for the gotten values of power(<25W), the bipolar ones are linear without feedback.
In summary, the circuit have 2 stages of voltage amplification and 2 of current.
I tested it, and please me the results.
npr
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.