It's not clear where U1 will get its grid leak from in real life, but presumably you have something mind. Can't see any goofs.
Yep I would have a ac coupled input with a 100k to ground for U1. Its possible U2 may oscillate due to phase shifts t HF around tone control. You may need a small cap from plate to R7/R13 don't know. Hum may get in via R12. DC heater for U1. If switched on hot U3 can generate quite a voltage on outputs. Can damage SS amps. You can check this by making B+ a pulse. Some add back to back zener C2/R24 to ground or slow raise on B+ or try R12 on other side or R28 or mute relay.
Last edited:
Maybe 12at7 for U1 and U2. p331 of designing hi-fi tube pre-amps. Suggests you make the forward and feedback paths to the tone control the same input impedance by padding the lower of the two or make plate output impedance similar. Just suggestions no more - what you have works.
Last edited:
The front end is an anode follower, grid return thru the signal source.
The center part is a Baxandall tone circuit consisting of Pots R15, R16, R20, R21,
Final is just a CF. Very common circuitry in 50s & 60s kit amps, Eico & HK. 😀
The center part is a Baxandall tone circuit consisting of Pots R15, R16, R20, R21,
Final is just a CF. Very common circuitry in 50s & 60s kit amps, Eico & HK. 😀
You know I'm so used to AC coupling my followers, that I saw the 330R as the cathode resistor... My mistake! Of course it would be the biasing if it was cap coupled 🙂
One thing I do notice is I think there will be more gain than you need in a line stage using 12AX7, even with the insertion loss of the tone control which eats the gain of the 6DJ8. You might consider making the 12AX7 a 6DJ8 instead.
One thing I do notice is I think there will be more gain than you need in a line stage using 12AX7, even with the insertion loss of the tone control which eats the gain of the 6DJ8. You might consider making the 12AX7 a 6DJ8 instead.
I would change R15- R16 and R20-R21 into 250k lineair potmeters so you have a tone control; now they are just tone filters and there is nothing to adjust.
Great stuff. Thanks. Yes, when changing U1 to AC coupled FB rather than direct connection to the load I forgot to add the grid leak (my recollection is that DC coupled FB caused some problematic inter-stage interactions). Good catch.
R27 is a stopper.
Baudouin0 a couple of questions. Your idea of a cap on the AX7 plate would be to reduce high frequency response correct? I like the idea of clamping the output but I am not sure what you are suggesting with your last statement "or try R12 on other side or R28".
Yes R15:R16 and R20:R21 are linear pots.
Regarding the high'ish gain it is intentional as my main power amp has pretty low sensitivity.
I have to revisit R11. I don't remember if it was part of the original Bax tone control circuit I looked at or whether it is something I added. Looking at it right now it is not needed for ground reference so perhaps it was for impedance or divider purposes but at the moment it does seem like it might be superfluous.
R27 is a stopper.
Baudouin0 a couple of questions. Your idea of a cap on the AX7 plate would be to reduce high frequency response correct? I like the idea of clamping the output but I am not sure what you are suggesting with your last statement "or try R12 on other side or R28".
Yes R15:R16 and R20:R21 are linear pots.
Regarding the high'ish gain it is intentional as my main power amp has pretty low sensitivity.
I have to revisit R11. I don't remember if it was part of the original Bax tone control circuit I looked at or whether it is something I added. Looking at it right now it is not needed for ground reference so perhaps it was for impedance or divider purposes but at the moment it does seem like it might be superfluous.
You might model the circuit using 6N3 and 6N2 instead since they are less expensive 🙂 OTOH you're in the USA so you get free shipping from most tube shops like tube depot anyway. Still 6N3 (or 5670) I like a lot, and 6N2 looks more linear.
First image: 12AX7WB Sovtek. Second image: 6N2P Third: 6N3P.
Ignore the grid current as the test was done using a 47k shunt from grid to cathode because his test rig would oscillate...
From: http://www.klausmobile.narod.ru/testerfiles/index.htm
First image: 12AX7WB Sovtek. Second image: 6N2P Third: 6N3P.
Ignore the grid current as the test was done using a 47k shunt from grid to cathode because his test rig would oscillate...
From: http://www.klausmobile.narod.ru/testerfiles/index.htm
Last edited:
OK, I went back and added .47uF input cap and then DC coupled the feedback to the right hand side of C3 as well as deleting R11. Simulations still seems to function properly and the distortion simulation actually improved (probably because of reduced load on U1). Assuming that noise is acceptable it looks like a good improvement.
I played with a cap from the plate of the 12AX7 to the left side of the grid stopper and a 4.7pF cap rolls the output off over 33KHz so that is an option if there are any stability problems. I imagine that the Cmiller of the AX7 is probably enough to keep thing stable though.
I played with a cap from the plate of the 12AX7 to the left side of the grid stopper and a 4.7pF cap rolls the output off over 33KHz so that is an option if there are any stability problems. I imagine that the Cmiller of the AX7 is probably enough to keep thing stable though.
Lots of good advise here - the try idea was to slow up the b+ onto the plate so the time constant was longer than the output time constant to reduce the size of the thump. Probably won't help. All OK if you switch on from cold but if hot you may get quite a thump. Designing Hifi preamps by Merlin Bleneowe is a good read. Try and eliminate hum. You can simulate some 100Hz with the B+ on LTspice. Heaters are more of a problem and I would suggest DC in the first stage. Ground loops can also be an issue and you may wish to share the Left and right in the same valve to make the ground point easier rather than the mono-block approach. You can also generate a semi floating input by lifting the ground end of C5/R2 connecting this to say 15R to ground. The input screen then goes to that junction. This works well if you go with the mono-block method.