I am trying to build a simple regulator to power the voltage and driver stages of a tube amplifier that I am building, but it doesn't seem to be working. I want 300 volts out. The front end of the amplifier draws about 50 ma. On the attached schematic, the voltages that I am getting (with all tubes in place)are shown in boxes with arrows pointing to the DMV attachment points. I have tried R values (the resistor between B+ and the zener string) from 18.2k ohm down to 500 ohm. I have swapped in a new IRF 840. I have swapped in new zeners. I have checked for shorts to the heatsink. I am still getting about 320 volts at the top of the zener string with all values of R. I am getting about 330 volts +/- 3 volts at the output. I have built this style of regulator before, with good results. (meaning that I get the voltage out that I expected) What am I doing wrong? What should I try next?
It should do. 5W Zeners - is a bit "over". Use 1-2 W Zeners instead of 5W. Consider Zeners current from 1 to 2 mA.
Use protective 15V G-S Zener too.
Have you measured a voltage across each Zener in that 300V line?
Use protective 15V G-S Zener too.
Have you measured a voltage across each Zener in that 300V line?
Last edited:
Try subbing a load resistor for the tube circuit, 300v/50mA = 6k, 20W.
Make sure the 5W Zeners run at over 5mA, and preferably more like 20mA.
Make sure the 5W Zeners run at over 5mA, and preferably more like 20mA.
You have no decoupling on the gate to ground, I think it might be a relaxation oscillator at the moment due to the large gate-source capacitance.
Be aware that this is not a regulator in the real sense, it doesn't regulate, it is a source follower. As such, the G to S drop depends on a particular device. Although in this case there seems to be a problem. I would first look at the pinout connection of the FET, if pins have been swapped you can get this.
The zener string must be OK because you measure the 320V at the gate, so no use to replace those.
Jan
Last edited:
- I'm not very sure because 1N5378 is 100 V rated, so it has to be close to 300 V but not 320 V.
Dry solder in the zener string, blown MOSFET? First thing to do is to protect the GS of the MOSFET with a 12 or 15V zener: in HV circuits, the slightest mishap results in molten silicon
But that is not the problem! The problem is that the output deviates too much from the gate voltages, so that has nothing to do with the zeners. Zeners are not very accurate so 320V from 3 zeners of 100V is not very strange (less than 10%). Anyway, it is not the problem, that should be clear.
Jan
Zener diodes have a tolerance rating spec for the rated voltage. The 1n5378B is available in ratings of 2, 5, and 10 percent. You can very easily get 330 volts when expecting 300. But that does not appear to be the issue.
You may want to try making a regulator circuit from a LR8n 3 pin regulator.
You may want to try making a regulator circuit from a LR8n 3 pin regulator.
You should get about 310v out with 320 on the gate.
How is the +/- 3% determined, load variation ?
And you do have a resistive load on the output when you measure it ? at least some bleeder R ?
If not, this could be it ...
How is the +/- 3% determined, load variation ?
And you do have a resistive load on the output when you measure it ? at least some bleeder R ?
If not, this could be it ...
Thanks to all who responded. I have tried some of the suggestions and have made some progress:
The zener string does have a 33 mfd capacitor across it. I forgot to draw it in the schematic.
I tried replacing the zener string with a resistor to make a voltage divider. It had no effect on the output. This suggested to me that the MOSFET was damaged.
I rebuilt the circuit with new resistors and a fresh IRF 840. I also changed the 18200 ohm resistor feeding the zener string with a 2200 ohm resistor, to draw more current through the zeners. I also added a 1N4007 diode across this resistor, from the drain to the gate.
This seems to have worked. I am now getting 304 volts out with approximately 340 volts at the input.
NOW, please help me diagnose a second regulator that is malfunctioning. I am using a second regulator in this amp for the output stage. It draws about 150 ma. and is supposed to output 450 volts, but was actually at 495 volts. Figuring that I had made the same mistakes in this regulator as the original, I rebuilt it the same way with the same parts values, except for a higher voltage zener string. It now outputs 460 volts, which is close enough, but the zener string is getting so hot that it is smoking. Is this because the resistor feeding the zener string is too small? I was careful mounting the parts to avoid shorts.
It feels like I am getting close to getting this sorted out. Any more suggestions would be appreciated.
The zener string does have a 33 mfd capacitor across it. I forgot to draw it in the schematic.
I tried replacing the zener string with a resistor to make a voltage divider. It had no effect on the output. This suggested to me that the MOSFET was damaged.
I rebuilt the circuit with new resistors and a fresh IRF 840. I also changed the 18200 ohm resistor feeding the zener string with a 2200 ohm resistor, to draw more current through the zeners. I also added a 1N4007 diode across this resistor, from the drain to the gate.
This seems to have worked. I am now getting 304 volts out with approximately 340 volts at the input.
NOW, please help me diagnose a second regulator that is malfunctioning. I am using a second regulator in this amp for the output stage. It draws about 150 ma. and is supposed to output 450 volts, but was actually at 495 volts. Figuring that I had made the same mistakes in this regulator as the original, I rebuilt it the same way with the same parts values, except for a higher voltage zener string. It now outputs 460 volts, which is close enough, but the zener string is getting so hot that it is smoking. Is this because the resistor feeding the zener string is too small? I was careful mounting the parts to avoid shorts.
It feels like I am getting close to getting this sorted out. Any more suggestions would be appreciated.
Definitely. Set the Zener current to between 10% and 40% of their rated current.
The lower their current, the more you have to worry about dropout.
Last edited:
Powerful Zeners are needed for bipolar transistors because of there low hfe comparing with Mosfets.
So, if you use Mosfets Zeners can be low powerful ones with the smallest minimum possible current.
So, if you use Mosfets Zeners can be low powerful ones with the smallest minimum possible current.
To calculate power dissipated in each zener you calculate P(zener) = V(zener) * I(zener).
To get I(zener) in case of mosfet you calculate I(zener) = (V(in) - V(zener string)) / R .
To get I(zener) in case of mosfet you calculate I(zener) = (V(in) - V(zener string)) / R .
Last edited:
Please think again. The zeners are not the problem. The problem is that the MOSFET doesn't work as intended.
Focus on the problem, don't go around haphazardly doing useless things.
Replace the MOSFET. Check the MOSFET pinout. Check that the voltage at the top of the zener string is actually the same as on the MOSFET gate pin.
Check that the input voltage is actually on the drain pin. Make sure you have a load of at least 10mA.
Make sure that what you measure at the output is actually on the source pin.
This should be fixed in 20 minutes!
Jan
Focus on the problem, don't go around haphazardly doing useless things.
Replace the MOSFET. Check the MOSFET pinout. Check that the voltage at the top of the zener string is actually the same as on the MOSFET gate pin.
Check that the input voltage is actually on the drain pin. Make sure you have a load of at least 10mA.
Make sure that what you measure at the output is actually on the source pin.
This should be fixed in 20 minutes!
Jan
Thanks, Jan. As usual, you are correct. I rebuilt the circuit using new resistors and a new IRF840. I also adjusted the pi filter feeding the regulator to a slightly higher voltage at the input of the regulator. It now outputs 299 volts. Close enough for government work. 😉 The amp that I am building uses 2 suggestions that you have made in the past. I installed the "diode trick": (2) UF4007 diodes feeding (2) 6CJ3 rectifier tubes for the B+. And combination bias for the output tubes: 100 ohm resistors from cathode to ground in combination with a regulated C- supply on the grids of the output tubes. The amp is working now and sounding good. I am going to let it play on the bench for several hours before I bring it to the living room. Again, thanks for your help.