I want to run my design's past the forum as this will be my first tube related build.
First the power supply.
The high voltage section:
Low voltage section:
The full power supply PCB:
____________________________________
Now the preamp design:
First the input is buffed to maintain a constant high input impedance and remain unaffected by the feedback divider that can be changed via potentiometer RV2A.
Simulated THD is nice and low and no sign of obvious problems.
However, I am also considering another design that also looks quite good:
Distortion is slightly higher, but now is almost entirely 2nd harmonic. High impedance can also be had (scale R9) and feedback amount can be changed via R14.
What do you think about this circuit? It seems like it might be worth a try. Do I need and additional protection for J1?
Maybe the first amp circuit if better?
Also what is the best way to switch the bootstrap capacitor in and out as this drastically changes distortion and would be nice to compare while listening.
Any feedback is much appreciated.
I can attach ASC files if anyone is interested.
First the power supply.
The high voltage section:
Low voltage section:
The full power supply PCB:
____________________________________
Now the preamp design:
First the input is buffed to maintain a constant high input impedance and remain unaffected by the feedback divider that can be changed via potentiometer RV2A.
Simulated THD is nice and low and no sign of obvious problems.
However, I am also considering another design that also looks quite good:
Distortion is slightly higher, but now is almost entirely 2nd harmonic. High impedance can also be had (scale R9) and feedback amount can be changed via R14.
What do you think about this circuit? It seems like it might be worth a try. Do I need and additional protection for J1?
Maybe the first amp circuit if better?
Also what is the best way to switch the bootstrap capacitor in and out as this drastically changes distortion and would be nice to compare while listening.
Any feedback is much appreciated.
I can attach ASC files if anyone is interested.
Last edited:
Slight improvements:
DC feedback has now been moved to the higher current output line, saving parts count and R16 dissipation.
I think this design will also be compatible with different tubes of the same pin-out as the bias can be set if R8 and R5 is a trimmer pot. Tube rolling fun?
There's also the fact that this is a non-inverting circuit, providing the correct phase (though its easy enough to swap speaker terminals) whereas the first circuit is inverting phase.
DC feedback has now been moved to the higher current output line, saving parts count and R16 dissipation.
I think this design will also be compatible with different tubes of the same pin-out as the bias can be set if R8 and R5 is a trimmer pot. Tube rolling fun?
There's also the fact that this is a non-inverting circuit, providing the correct phase (though its easy enough to swap speaker terminals) whereas the first circuit is inverting phase.
Haven't a lot to contribute as solid state design isn't my forte. The only thing I noted was the HV PSU has a lot of capacitance so make sure your power transformer can supply the current, you've no bleeder R on C1 & C2 & no fuse on the AC side, small issue's. My old eyes couldn't see the value of the caps too well, but it looks like you have two 100u caps in parallel after the IRF830 which is a bit overkill.
Apart from that, it looks a very good first design with valves, though there's more three legged fuses than valves, just my prejudice, Andy.
Apart from that, it looks a very good first design with valves, though there's more three legged fuses than valves, just my prejudice, Andy.
Nice PCB. But does the project really qualify as a vacuum tube design?
Just one tube, the rest is SS.
Just one tube, the rest is SS.
Power supply:
It is from bridge rectifier:
C > 100R > C > C > IRF830 > C.
C = 100uF
Yes, probably a little over done but simulation shows very little peak current change from adding more capacitance after 100R and seems to depend on the first capacitor after rectification. The only concern with it would be inrush current?
Anyway, it does seem I am overly concerned with ripple and seems plenty low enough when I remove the second 100uF. I shall play it safe and remove it from the board and but a bleeder resistor in its place.
________________
Preamp circuit:
I want to use 2 tubes, one per channel (Yes my supply can provide the needed filament current). As these tubes are dual triode, perhaps I could use them in parallel for the above circuits? What would be the benefits?
I have simulated with a SRPP configuration but was not as pleased with the results as with the bootstrap.
It is from bridge rectifier:
C > 100R > C > C > IRF830 > C.
C = 100uF
Yes, probably a little over done but simulation shows very little peak current change from adding more capacitance after 100R and seems to depend on the first capacitor after rectification. The only concern with it would be inrush current?
Anyway, it does seem I am overly concerned with ripple and seems plenty low enough when I remove the second 100uF. I shall play it safe and remove it from the board and but a bleeder resistor in its place.
________________
Preamp circuit:
I want to use 2 tubes, one per channel (Yes my supply can provide the needed filament current). As these tubes are dual triode, perhaps I could use them in parallel for the above circuits? What would be the benefits?
I have simulated with a SRPP configuration but was not as pleased with the results as with the bootstrap.
Last edited:
Ahh on the preamp section! I thought you were looking at the power supply section. D1 is 12V (should limit current to around 40mA with a VGS of 4V) and D2 and 3 are maybe 16V as the preamp can be seen more as VAS as it is also deigned to drive a unity gain output buffer before the speakers.
That would not work as the feedback would be in the wrong phase, that would be positive feedback.
This also simulates well but it is pushing 4.8mA though tube NH2 which seems high, may decrease the life of the tube?
Also may need to test the JFET J1 for IDSS.
Also, I've just read that in the filaments 'warm up' period, it can draw 2/3x as much current!
This may be OK as the voltage will slowly ramp up, but I may need to increase capacitor C13 to provide a slow enough ramp to mitigate the warm up time.
Also may need to test the JFET J1 for IDSS.
Also, I've just read that in the filaments 'warm up' period, it can draw 2/3x as much current!
This may be OK as the voltage will slowly ramp up, but I may need to increase capacitor C13 to provide a slow enough ramp to mitigate the warm up time.
I missed a 10k resistor before R14. Its purpose would be to decrease AC feedback thus increasing the gain of the amp.
Want a top sounding tube preamp. Loose the hv reg,and put a high xl low ohm choke there.make sure its high quality,high b strip wound double c core best. You will have dynamics, fluidity,harmonic resolution,all the good stuff we audio nuts love. Probably get ride of the silicon in the amp. Keep it simple,use best quality parts.good luck.
So the HV PSU needs 240V AC for 230V DC? That is 340V DC for 230V DC. A 110V (!) voltage drop really is not OK.
Using 12V AC so 16V DC for 6.3V DC is also worth re-evaluating as 10W is burnt like it is now. Besides the tubes your PSU designs will burn energy for no good reason and they will become very hot and likely also unreliable. They will definitely need pretty large and expensive heatsinks. A nice solution is to use tubes with 12.6V filaments and have just 12V DC for tubes filaments and buffer.
- You can use non linear parts to limit output DC voltage at power on but a muting relay will have a benefit also at power off. Both at power on and off complete silence and no risk at all for connected equipment. This will add to ergonomics and will give a luxury touch.
- An input 100 kHz low pass filter is not luxury just as GND referencing of inputs. Today's situation is not like in the tube era.
Using 12V AC so 16V DC for 6.3V DC is also worth re-evaluating as 10W is burnt like it is now. Besides the tubes your PSU designs will burn energy for no good reason and they will become very hot and likely also unreliable. They will definitely need pretty large and expensive heatsinks. A nice solution is to use tubes with 12.6V filaments and have just 12V DC for tubes filaments and buffer.
- You can use non linear parts to limit output DC voltage at power on but a muting relay will have a benefit also at power off. Both at power on and off complete silence and no risk at all for connected equipment. This will add to ergonomics and will give a luxury touch.
- An input 100 kHz low pass filter is not luxury just as GND referencing of inputs. Today's situation is not like in the tube era.
Last edited:
I think you are assuming 240V RMS? where as I am meaning peak AC voltage. I get 240V DC after rectification and drop around 15V over R2/R3 voltage divider driving Q1, then the VGS.
I get 14VDC and the 6N2P tubes require 340mA 6.3V each so the dissipation is around 5W. Still not ideal but is manageable.
I have considered wiring in series but don't like the idea as one tube may hog more current than the other, I have considered a circuit to 'balance' the current:
A buck converter is another option, but I don't really like the idea of switching and potential EMF problems.
A relay delay circuit is on the schematic. R38 will be large enough for the required warm up delay.
The circuit for a basic input switch, low pass and volume control:
I'm leaning towards this circuit:
The zener is 24V.
Each tube is at 700uA with a 151V plate voltage, higher plate voltage seems to bring down distortion.
This is the FFT at 15V (30Vpp) output (0.056% THD):
Seems like a good result and am thinking of designing some test PCB's.
The zener is 24V.
Each tube is at 700uA with a 151V plate voltage, higher plate voltage seems to bring down distortion.
This is the FFT at 15V (30Vpp) output (0.056% THD):
Seems like a good result and am thinking of designing some test PCB's.
Normally RMS values are used. BTW the muting relays contacts should be over the 47 kOhm resistors to GND otherwise they will always be in the signal path.
Would a 1 ...10 µF cap behind the MOSFET not be a better choice? Large value caps do not bring lower ripple at this spot.
When you would use a 2 x 7V transformer you could use 2 LDO regulators for the 6.3V filaments. What you gain in efficiency you will not loose on heatsinks. AFAIK there is no gain in building a discrete regulator as most 3 pin regulators do a fine job feeding a filament.
Just suggestions.
Would a 1 ...10 µF cap behind the MOSFET not be a better choice? Large value caps do not bring lower ripple at this spot.
When you would use a 2 x 7V transformer you could use 2 LDO regulators for the 6.3V filaments. What you gain in efficiency you will not loose on heatsinks. AFAIK there is no gain in building a discrete regulator as most 3 pin regulators do a fine job feeding a filament.
Just suggestions.
Last edited:
Yes, I was thinking that.. This should be correct now:
As for providing filament current, I think using schottky diode bridge rectifier to 12V LDO then to the 'balance' circuit (repurposed virtual ground) would be a smart solution.
A quick search and I found the MIC29300-12WT LDO TO-220 5A seems to be a suitable choice.
(Datasheet: https://ww1.microchip.com/downloads...urrent-Low-Dropout-Regulators-DS20005685B.pdf)
If I understand this correctly, I would only need only 220mV higher than 12V at 750mA for it to regulate properly?
Would the valves operate optimally at 6V each?