I have the following phase splitter designed for a 12AU7.
What components would I need to change for it to use a 12AX7 so it can drive a quad of 6V6 in push pull parallel unless with feedback I can get a 1V input sensitivity for 25 watts out with the 12AU7?
The 6V6 has a max grid resistance of 100k when operated fixed bias so the load on each output of the phase inverter will be 50k unless it's ok to use a single 100k grid resistor for two 6V6 in parallel.
What components would I need to change for it to use a 12AX7 so it can drive a quad of 6V6 in push pull parallel unless with feedback I can get a 1V input sensitivity for 25 watts out with the 12AU7?
The 6V6 has a max grid resistance of 100k when operated fixed bias so the load on each output of the phase inverter will be 50k unless it's ok to use a single 100k grid resistor for two 6V6 in parallel.
Something like this?
6V6 is pretty forgiving, I wouldn't expect problems with 100k for two tubes in parallel.
I use this one
I prefer the 6H30 for V2, very linear.
The entire circut is linear because the tubes; also the Zout of the splitter is readonable low.
I drive also a double p-p of 300B or KT150
I prefer the 6H30 for V2, very linear.
The entire circut is linear because the tubes; also the Zout of the splitter is readonable low.
I drive also a double p-p of 300B or KT150
I currently use a long tailed pair inverter.
What needs to be changed so each output can drive a 100k load?
Currently it takes 1V for full 25 watt amp output.
What needs to be changed so each output can drive a 100k load?
Currently it takes 1V for full 25 watt amp output.
Change the tube
With one with less Rp and good gain
But you need to re-design the circuit
One interesting is 12BZ7 the Rp is half of AX7 with same gain
Another can be 5965 with lower gain
Another can be 5965 with lower gain
I might just have a 12BZ7.
Would that be applicable for running the 6V6 tubes fixed bias?
So for a 100k impedance I'd need 200k resistors.
I'd then be able to use the current coupling cap value of 1uF. Would just need to buy two more.
Would that be applicable for running the 6V6 tubes fixed bias?
So for a 100k impedance I'd need 200k resistors.
I'd then be able to use the current coupling cap value of 1uF. Would just need to buy two more.
Got the 12BZ7 installed. Just dropped right in.
What resistors need to change for the long tailed pair? Would I use 200k for the 300k and 100k for the 200k then adjust the gain to keep the 1V input sensitivity?
I'm surprised the 12BZ7 isn't more popular for amp designs given it could eliminate one stage in some amps given it can drive a lower impedance with the gain of the 12AX7.
What resistors need to change for the long tailed pair? Would I use 200k for the 300k and 100k for the 200k then adjust the gain to keep the 1V input sensitivity?
I'm surprised the 12BZ7 isn't more popular for amp designs given it could eliminate one stage in some amps given it can drive a lower impedance with the gain of the 12AX7.
Last edited:
The one downside to using 12BZ7 instead of 12AX7 is that 12BZ7 has significantly worse linearity. 12AX7 is actually one of the cleanest triodes there is, as long as it's lightly loaded. 12BZ7 has roughly half the rp of 12AX7, which should be 25k to 30k ohms or thereabouts.
If you really need each triode to drive a 100k load, then the rp of the triode should be no higher than 33k ohms. That would follow the rule that the load impedance should be no less than 3X the source impedance.
12AT7 is the tube people usually use for driving a 100k load. 12AT7 rp is roughly 15k to 20k ohms. 12AT7 gain is almost as high as 12AX7 or 12BZ7. You would usually use about 47k plate load resistors on the 12AT7. The original silver-face Fender tube amps used that (12AT7 LTP with 47k ohm plate load resistors).
If you really need each triode to drive a 100k load, then the rp of the triode should be no higher than 33k ohms. That would follow the rule that the load impedance should be no less than 3X the source impedance.
12AT7 is the tube people usually use for driving a 100k load. 12AT7 rp is roughly 15k to 20k ohms. 12AT7 gain is almost as high as 12AX7 or 12BZ7. You would usually use about 47k plate load resistors on the 12AT7. The original silver-face Fender tube amps used that (12AT7 LTP with 47k ohm plate load resistors).
If you use a 12BZ7, you could decrease the plate load resistors to something like 56k or 47k, and still get your approximately 35X gain. You wouldn't need to change any other parts values.
12BZ7 could be biased to about 1.5mA plate current per triode (3mA for the pair). Zout would be about 30k to 35k ohms (per triode). THD is pretty high, but maybe this is for a guitar amp? If this is for a hi-fi amp, you could use a constant current source in the 'tail' instead of the 10k ohm resistor. That would reduce THD by a LOT in this particular circuit.
If you use a 12AT7, you could use those same plate resistor values (probably 47k) with plate current of about 2mA per triode. Zout would be about 20k to 25k ohms. Gain would be about 25X. THD is about 30% lower.
EDIT TO ADD.... Here's what I was thinking of:
12BZ7 could be biased to about 1.5mA plate current per triode (3mA for the pair). Zout would be about 30k to 35k ohms (per triode). THD is pretty high, but maybe this is for a guitar amp? If this is for a hi-fi amp, you could use a constant current source in the 'tail' instead of the 10k ohm resistor. That would reduce THD by a LOT in this particular circuit.
If you use a 12AT7, you could use those same plate resistor values (probably 47k) with plate current of about 2mA per triode. Zout would be about 20k to 25k ohms. Gain would be about 25X. THD is about 30% lower.
EDIT TO ADD.... Here's what I was thinking of:
Last edited:
I'll test it with a 12AT7 with the stock resistor values first to see what gain I get.
That way I'll know if a 12AT7 will work with the current feedback arrangement.
I'll then swap to 50k resistors as I can make them from two 100k resistors in parallel as I have those already and no 47k.
I also have some 120k resistors so I could parallel one of those with the 100k resistor as well.
That way I'll know if a 12AT7 will work with the current feedback arrangement.
I'll then swap to 50k resistors as I can make them from two 100k resistors in parallel as I have those already and no 47k.
I also have some 120k resistors so I could parallel one of those with the 100k resistor as well.
So I tried the 12AT7 and it doesn't have enough gain as with it I need the 10k gain pot set to 0 ohms in order to get 1V input sensitivity.
The 12AX7 does work in the new circuit still. So does the 12BZ7, although now it has slightly more gain.
So for the CCS in the inverter would a standard adjustable regulator work?
Here's the circuit with the changed values. I didn't have four 200k resistors nor two more 1uF caps so I just used a single 100k per pair of tubes.
Would the CCS increase the gain to where I can use a 12AT7?
The 12AX7 does work in the new circuit still. So does the 12BZ7, although now it has slightly more gain.
So for the CCS in the inverter would a standard adjustable regulator work?
Here's the circuit with the changed values. I didn't have four 200k resistors nor two more 1uF caps so I just used a single 100k per pair of tubes.
Would the CCS increase the gain to where I can use a 12AT7?