TubelabSE power supply trouble

[wicked1]

I'm hoping it is because it was damaged the 1st time my rectifier died, last month.. that resistor glowed! I tested it, and it was still alright, so left the resistor w/ burn marks on it in the circuit (won't do that again). So, my hope is that it got damaged that time, and finally gave out.
I tested the 2 caps down from that resistor, and all seem to be ok. I'm just testing w/ my DMM, but they aren't shorted and they charge.

It was easier to troubleshoot before I fused it.. I could trace the path of destruction . I'm not sure what I can look for other than a short circuit, since I can't power it up. Nothing is connected to ground, that shouldn't be. I guess I'll start checking for shorts between parts of the circuit that shouldn't be connected.

[Ty_Bower]

I don't see why a failed 5AR4 rectifier should take the bias supply circuit with it. I also don't fully appreciate what the MOSFET is doing on the grid of the output tube, but it seems to be sitting on the negative rail.

I prefer to power up my fixed bias amps with the B+ rectifier out of the circuit. I'll check the bias supply first and make sure all voltages along it are within spec. Do the startup directions for the Tubelab SE suggest this is possible? If it's safe to do, it might be a worthwhile test after you replace R5.

Generally, bias supplies source very little current.

[Evenharmonics]

wicked1, can you list the voltages at the “Checkout” points shown in Tubelab website? B+, B-, input tube plate and output tube plate.

After changing parts I always check those voltages in sequence shown in the website prior to putting the amp back in the system. They've been consistent so far but never know...

[tubelab.com]

I also don't fully appreciate what the MOSFET is doing on the grid of the output tube ..... Generally, bias supplies source very little current.

The mosfet is a source follower which provides a very low impedance drive signal to the output tube's grid. It eliminates the overload and recovery issue (blocking distortion) found in typical capacitor coupled amplifiers. In this case the bias supply feeds those mosfets a steady diet of 4 to 8 mA each. The complete explanation can be found here:

http://www.tubelab.com/powerdrive.htm

In order to reduce the possibility of "phase intermodulation distortion" the mosfets were selected to have minimal internal capacitance change with varying voltage and operated at a fairly high dissipation level.

The original design used 2SK2700 mosfets which were operated at about 6 mA. They ran hot but were completely reliable. The 2SK2700 was banished by ROHS laws becaused it contained a few micrograms of lead. I suggested several alternatives to the 2SK2700 mosfets but time has revealed that some of these were not as reliable as the 2SK2700 (especailly IXYS) when running under the same conditions. It is advisable to reduce the current through the mosfets if using something other than the 2SK2700 mosfet. The details are outlined on the home page of my web site. I would use the Toshiba 2SK3563 for replacement or new amp builds.

R5 is a 270 ohm resistor that is between the diodes and the input filter cap in the negative supply. Its only purpose is to reduce the peak current through the diodes and capacitor to eliminate "solid state diode noise". The only thing that could cause it to glow red is a shorted C6 (or something touching the back of the PC board). If there was a direct short down stream R6 would limit the current to about 40 mA which is still in the safe range for R5. If one of the diodes was shorted C6 might have been damaged.

R6 is a 10K 5 watt resistor. It drops the -350 volts to - 200 or so. It does get quite warm in normal operation.

I'll check the bias supply first and make sure all voltages along it are within spec. Do the startup directions for the Tubelab SE suggest this is possible?

Yes. The first step in powering up a new (or recently repaired) board is to power it up with no tubes installed. I usually have a meter on the B+ (across R30) and the B- supplies (across R7).

The negative voltage at R7 should be at least 100 volts and can be as much as 250 depending on the power transformer. If the voltage drop across R6 is high (over 200 volts) a shorted mosfet is possible.

Next install the rectifier tube only. The B- should appear as soon as the amp is turned on and the B+ will come up as the rectifier tube warms up. Measure the voltage at the grid pin of each output tube. Turn the bias pot for that tube and observe the voltage. It should vary from a few volts negative to -60 to -100 volts or more. If the voltage is positive or does not adjust the associated mosfet is likely dead.

[wicked1]

Thanks guys,
Parts are on order, and I'll update this thread when I get them and start testing. I've got a much better understanding of the situation with your help, and with the couple of days I've been thinking about it.

[wicked1]

I replaced the bad resistor, and all is good. That's all it was this time is some damage I failed to clean up after the diode blew last time it failed.

I had a st-35 filling in while this amp was out of service. While the dynaco is a good amp, I am very happy to have this one working again! It's beautiful!!!!!

[Evenharmonics]

Glad to hear the good news.