I began my experience building single ended vacuum tube stereo amps, based on a number of kits, including EL34, KT88, and 300b. I have had great success with each of them resulting in amplifiers that sound fantastic. After about 3 years of bulding several of these amps, I became interested in OTL headphone amps. I have successfully built several of them based on the Morgan Jones design: https://headwizememorial.wordpress.com/2018/03/20/the-morgan-jones-mini-tube-headphone-amplifier/
Specifically I am building the "optimized amp with feedback" (which is the third schematic. For the B+ power supply I use a power supply with a 240vac secondary and full wave rectifier followed by a capacior, resistor, capacitor, resistor and anoter capacitor. The caps are 100uf/450v, the resistors are 1.8k/10W. My B+ voltage matches the spec, around 220v, althought it goes up to around 350v as the tubes are warming up. The filaments are 6.3vac with two 100 ohm resistors in series across the AC out and the center of the resistor is grounded (this is done to eliminate hum). The amp, in my opinion sounds great. I built 6 of them with no issue. For some reason, the seventh one I built has a spark in one one of the output tubes, lasting a fraction of a second and occurs about 3 seconds after turning the amp on. I have tried switching tubes, but the spark occurs no matter what tube I use. Also for some reason, the spark is only on one of the channels - the other output tube never sparks. I have been beating my head against the wall (figuratively of course) to figure out this problem. I measured the voltage of the top plate, top cathode/bottom plate, and bottom cathode (once the amp is warmed up) on both a working version and the defective version and the voltages are all very close to spec. I took the power transformer from the working one and appied it to the defective one and it still had the same problem. I took the power transformer from the defective one and applied it to the working one and it worked fine, so the power supply doesn't appear to be the problem. Once I get past the spart on the defective one, it works fine, but clearly this problem must be fixed as sparking is unacceptable. I am using a circuit board for this amp, that matches the design in the article. https://www.aliexpress.us/item/2251...2e9Bjft&gatewayAdapt=glo2usa&_randl_shipto=US
This vendor also sells a very nice looking case which I use, but I supply my own parts. I decided to build another board and completed it and replaced the defective board with the new one, but the new one also has the same problem. The only difference between the working one and the two defective ones is that I am using an audyn 0.47uf coupling cap in the working one, one of the defective ones uses a solen 0.47uf coupling cap and the other one uses a mundorf 0.47uf cap. I can't imaging that that could be causing the problem.
Any thoughts on how I can diagnose and/or fix this issue would be greatly appreciated. I have been working on this for several weeks and am running out of ideas.
Specifically I am building the "optimized amp with feedback" (which is the third schematic. For the B+ power supply I use a power supply with a 240vac secondary and full wave rectifier followed by a capacior, resistor, capacitor, resistor and anoter capacitor. The caps are 100uf/450v, the resistors are 1.8k/10W. My B+ voltage matches the spec, around 220v, althought it goes up to around 350v as the tubes are warming up. The filaments are 6.3vac with two 100 ohm resistors in series across the AC out and the center of the resistor is grounded (this is done to eliminate hum). The amp, in my opinion sounds great. I built 6 of them with no issue. For some reason, the seventh one I built has a spark in one one of the output tubes, lasting a fraction of a second and occurs about 3 seconds after turning the amp on. I have tried switching tubes, but the spark occurs no matter what tube I use. Also for some reason, the spark is only on one of the channels - the other output tube never sparks. I have been beating my head against the wall (figuratively of course) to figure out this problem. I measured the voltage of the top plate, top cathode/bottom plate, and bottom cathode (once the amp is warmed up) on both a working version and the defective version and the voltages are all very close to spec. I took the power transformer from the working one and appied it to the defective one and it still had the same problem. I took the power transformer from the defective one and applied it to the working one and it worked fine, so the power supply doesn't appear to be the problem. Once I get past the spart on the defective one, it works fine, but clearly this problem must be fixed as sparking is unacceptable. I am using a circuit board for this amp, that matches the design in the article. https://www.aliexpress.us/item/2251...2e9Bjft&gatewayAdapt=glo2usa&_randl_shipto=US
This vendor also sells a very nice looking case which I use, but I supply my own parts. I decided to build another board and completed it and replaced the defective board with the new one, but the new one also has the same problem. The only difference between the working one and the two defective ones is that I am using an audyn 0.47uf coupling cap in the working one, one of the defective ones uses a solen 0.47uf coupling cap and the other one uses a mundorf 0.47uf cap. I can't imaging that that could be causing the problem.
Any thoughts on how I can diagnose and/or fix this issue would be greatly appreciated. I have been working on this for several weeks and am running out of ideas.
I have two 100 ohm resistors connected in series. One end connects to one of the filament voltages and the other end connects to the other filament voltage (AC). The middle point where the two resistors are connected together, is connected to the ground. I did this to eliminate a hum in the circuit.
I measured the B+ voltage at the third capacitor. At the point I turn the amp on the DC voltage shoots up to about 350V and then starts dropping when as the filaments heat up. The spark occurs around the time the DC voltage is at it's max, which is about 3 seconds after turning it on. I was wondering if I could put a resistor between the B+ and ground to limit it's voltage to around 250V?
I measured the B+ voltage at the third capacitor. At the point I turn the amp on the DC voltage shoots up to about 350V and then starts dropping when as the filaments heat up. The spark occurs around the time the DC voltage is at it's max, which is about 3 seconds after turning it on. I was wondering if I could put a resistor between the B+ and ground to limit it's voltage to around 250V?
I think the issue is that the grid of the 'top' output tube is pulled to the full B+ voltage during startup. At the same time the cathode of that tube is being held near 0 V until the output capacitor has charged. So you get 350 V from grid to cathode.
One way to address this could be to add a diode from grid to cathode on that tube. Like this:
Tom
One way to address this could be to add a diode from grid to cathode on that tube. Like this:
Tom
BTW: If you have 350 V on B+ during startup, you'll want C4 to be a 400 V type, minimum. Preferably 450 V. You really don't want that capacitor to arc over and fry your headphones.
Tom
Tom
Tomchr writes "BTW: If you have 350 V on B+ during startup, you'll want C4 to be a 400 V type, minimum. Preferably 450 V. You really don't want that capacitor to arc over and fry your headphones."
Thanks for this note. I agree and did use 450V caps in this place
Thanks for this note. I agree and did use 450V caps in this place
Tomch also writes: "I think the issue is that the grid of the 'top' output tube is pulled to the full B+ voltage during startup. At the same time the cathode of that tube is being held near 0 V until the output capacitor has charged. So you get 350 V from grid to cathode."
Thanks for this explanation, but I have a follow-up question. So when the amp is warmed up, I measured the voltages and the plate on V1 and the cathode/plate in the middle of V2 are both around 110 volts, as in the spec. But at startup, the top half of the V2 plate is around 350V, which I would assume would make the plate on v1 around 240V so adding this diode would make the V2 Plate/Cathode connection around 240V. Is that the intended result of the diode?
Also this is what is puzzeling me. Why did the first 6 of these work perfectly fine without the diode, it is only versions 7 and 8 that have this spark?
Thanks for this explanation, but I have a follow-up question. So when the amp is warmed up, I measured the voltages and the plate on V1 and the cathode/plate in the middle of V2 are both around 110 volts, as in the spec. But at startup, the top half of the V2 plate is around 350V, which I would assume would make the plate on v1 around 240V so adding this diode would make the V2 Plate/Cathode connection around 240V. Is that the intended result of the diode?
Also this is what is puzzeling me. Why did the first 6 of these work perfectly fine without the diode, it is only versions 7 and 8 that have this spark?
When the tubes are conducting the cathode will always be at a lower potential than the grid, so the diode will always be reverse biased. Aside from a bit of depletion capacitance (maybe in the low pF range) it'll have no impact on the circuit.
During startup, the plate of the first tube will be pulled to the full B+ by R2. C4 will be fully discharged, thus, effectively act as a short circuit. This connects the cathode of the top output tube to ground via the headphones (or the 22 kΩ on the output). So without the diode you now have B+ from the grid to the cathode. The tube is rated for 200 V grid-to-cathode and you're applying 350 V. If this doesn't cause the sparking another possibility could be that the tube does something funky when it just begins to conduct and 'sees' significant positive grid-to-cathode voltage.
The diode limits the top tube to a positive Vgk of about 0.7 V (the forward drop of the diode). While the tube warms up C4 is now charged through R2, R7, and the diode.
This circuit will still produce a significant start-up burp. I would disconnect the headphones before turning the amp on or off. If this was a commercial product I'd add a muting relay on the output.
The reason it worked before but doesn't now could be tube-to-tube variation. It could also be that the B+ is higher in the new build for whatever reason. I think a relevant question to ask is: Why does the B+ droop by 130 V when the tubes start to conduct? Do you not have any reservoir capacitance in the power supply? The circuit only draws 13 mA per channel.
Tom
During startup, the plate of the first tube will be pulled to the full B+ by R2. C4 will be fully discharged, thus, effectively act as a short circuit. This connects the cathode of the top output tube to ground via the headphones (or the 22 kΩ on the output). So without the diode you now have B+ from the grid to the cathode. The tube is rated for 200 V grid-to-cathode and you're applying 350 V. If this doesn't cause the sparking another possibility could be that the tube does something funky when it just begins to conduct and 'sees' significant positive grid-to-cathode voltage.
The diode limits the top tube to a positive Vgk of about 0.7 V (the forward drop of the diode). While the tube warms up C4 is now charged through R2, R7, and the diode.
This circuit will still produce a significant start-up burp. I would disconnect the headphones before turning the amp on or off. If this was a commercial product I'd add a muting relay on the output.
The reason it worked before but doesn't now could be tube-to-tube variation. It could also be that the B+ is higher in the new build for whatever reason. I think a relevant question to ask is: Why does the B+ droop by 130 V when the tubes start to conduct? Do you not have any reservoir capacitance in the power supply? The circuit only draws 13 mA per channel.
Tom
Tom, again, thanks for your input. I plan to work on this amp this weekend and will try the diode solution and let you know how that works out. There is another solution I would like to ask you about. Since the amp works fine once the B+ stabilies, could I put a resistor or some other component between B+ (right before R4) and ground that would be sized to limit the B+voltage to never go past say 250V during startup? I had looked into a delay circuit, but unfortunately I am using a commercial case built for this amp and it is too small to include a delay circuit. Also, the power supply in the defective unit is OK because I wired it to power up the working amp and the working amp continued to work ok.
Another thought I had, I am using 100uf filtering caps in the voltage dropping circuit. If I used a smaller cap in the last cap or the others, would that prefent the B+ voltage from shooting up so high at startup?
Another thought I had, I am using 100uf filtering caps in the voltage dropping circuit. If I used a smaller cap in the last cap or the others, would that prefent the B+ voltage from shooting up so high at startup?
Last edited:
It looks like the problem is solved. I did two things. First, I was checking voltages and found a serious problem. For the many years I have lived in the same place my voltage was 125V instead of the standard 120V. So when I put together projects I took that into consideration. Specifically, for these tube headphone amps I had a small voltage dropping resistor in series with the filaments, to drop the voltage to 6.3v and that worked fine. For some reason, recently my voltage changed to 120v. Maybe someone in the neighborhood complained and the electric company fixed it. At any rate with the resistor in place now my filament voltage was 5.7v instead of 6.3v. So I took out the resistor and noticed that the filaments headed up much more quickly and the sparking problem did not occur. The filament voltage without the resistor is now 6.3v.
To be on the safe side I also implemented Tom's suggestion regardig putting in a diode to stabilize the voltage. The afternoon I will do some more diagnostics by measuring voltages, but it looks like the problem is solved. Thanks, Tom and lcscar for your suggestion.
To be on the safe side I also implemented Tom's suggestion regardig putting in a diode to stabilize the voltage. The afternoon I will do some more diagnostics by measuring voltages, but it looks like the problem is solved. Thanks, Tom and lcscar for your suggestion.