Blackstar HT-Delay repair

[Yazoo]

I am trying to repair this pedal. Something fairly major happened in the power section and toasted the board around the section which creates the high voltage for the 12AX7 plate voltage. I have repaired this section but the replacement IRF830 also gave out magic smoke. I used a bench power supply to test it and I could see the pedal was drawing up to 2 amps when it is only supposed to draw around 1 amp.

Unfortunately, there are no full schematics available, though I did find a schematic for a more generalised Blackstar high power section which broadly matches this pedal.


The other puzzler is there is an LM7915 voltage regulator (as well as an LM7815 and an LM7805) but the power brick is 22 volts DC and this is confirmed in the manual and on the pedal itself. I am trying to trace the power supply out but in the past, I have only ever seen bipolar voltages where the power supply was AC. When I test the board without the IRF830, Both the positive regulators work but the 7915 is doing nothing, no voltage on the input or output pin.

[merlinb]

I have repaired this section but the replacement IRF830 also gave out magic smoke.

Sounds like the UC3843 is dead and not producing the correct control voltage, so the MOSFET is stuck 'on'.

I have only ever seen bipolar voltages where the power supply was AC.

Maybe it's using an artificial ground? Or it has another switching inverter somwhere, or it is cap-coupling off the IRF830.

[Yazoo]

Thanks, I have got a replacement uc3843 chip on order. I think I'll try jumpering the IRF830 until I get this sorted out (I hope!).

[vigilante397]

Interested to see what you find out. I always found it interesting that Blackstar used 22V supplies for a 12AX7 pedal when a 12V would be more intuitive, but they sound good so it's hard to argue with their results.

[Yazoo]

I have traced out the negative power section and this is what it looks like. If you hadn't mention cap-coupling, I wouldn't have had a clue. I had never heard of this so I looked it up and it does make sense in this circuit.

[Yazoo]

I have replaced the uc3843 and I am pleased to say I now have -15v on the LM7915. I am going to socket the IRF830 this time. I don't think the solder pads could take another desolder, so I am just waiting for the socket to arrive.

[Yazoo]

I have spent the last week trying to debug the high voltage part of this without any real success. I have narrowed the problem down to the Comp and VFB pins on the uc3843. I tried to replace the components in the damaged section of the board but after checking and re-checking, I get the same results. The MOSFET gets hot and burns out. If I have understood the circuit correctly, the MOSFET should turn on and off at about 66Khz. I get the saw tooth pattern on pin 4 of the uc3843, but I don't see the square wave on pin 6, the output pin, so I think the MOSFET is permanently on and so burns out.

Today, I set the uc3843 up on a breadboard and used the open loop example from the uc3843 data sheet, using a trimpot to adjust voltage to the VFB pin and I was able to get a square wave but the MOSFET still got very hot, even with a heatsink attached.

I just don't have the expertise to debug this so I have ordered an NCH6100HV power module which runs off up to 24V DC and outputs up to 235V, about 10% lower the the stated plate voltage for the 12AX7. I am hoping to wire this in to the power board. Would this cause any problems do you think?

[iainpunk]

i have run 12ax7's on only 60v, Eddie van Halen used a variac to lower his amps voltage to way lower than normal, about 20% lower, so i wouldn't worry about it.

cheers

[vigilante397]

Agreed, 235V is what I typically run my tube pedals at, that's plenty. You won't hear a difference.

[Yazoo]

I have spent the last fortnight trying to debug this. There is something really peculiar going on. From what I have picked up, the VFB pin of the UC3843 should see 2.5 volts. There is a voltage divider from the 1N4937, 1 meg to the VFB pin and 8k2 from the VFB pin to ground. Doing the calculations, this means there should be 307 volts. The schematic I have states 270 volts. When I removed and checked the original 1 meg resistor, which looked a bit weedy, it read 1.7 meg. I replaced it with a 600mW resistor and rebuilt the circuit, but the MOSFETs still smoked or got very hot.

I measured the high voltage circuit several times, but measuring from the 1N4937 to the empty socket of the UC3843 never gave me 1 meg. I just could not track down the fault. It seems the HV circuit in these pedals does fail. I have come across 2 other examples on the web.

I wired in the NCH6100V but still had problems. Under load, the -15v dropped and I only got very distorted sound. By chance, I had an LT1054 chip, so I breadboarded a voltage inverter. Under load, I get 14.5v but it seems happy and now works.

The plan now is to extend the bottom of the box using a 3D printed "shim" and put both the NCH6109V and the LT054 in it.

[merlinb]

Which parts have you actually replaced, I've lost track. Presumably you replaced the main rectifier diode already?

[Yazoo]

I rebuilt the circuit in the damaged section, replacing the 1N4937, the 47N capacitor, the first 10uf capacitor and the dodgy 1M resistor. All the other resistors measured correctly. I went through 2 new IRF830s which both burned out. I also replaced a couple of capacitors but the circuit still did not work. As I said, the voltage divider to the VFB pin did not measure correctly so there must be a fault somewhere else I couldn't find. When I used my bench power supply, the circuit drew over 2 amps, the maximum for the bench power supply. The actual pedal is rated at 1.1 amps.

I am going to disable the high power circuit on the board and replace it with the NCH6100V and replace the -15v circuit with the LT1054. I would prefer to find the fault but I just don't have the skills to do this, and it does seem to be a problem with these pedals

[Yazoo]

I went ahead and 3D printed a 25mm extender for the enclosure for the NCH6100HV and voltage inverter and fitted it all back together. I removed the uc3843 and tapped the voltage for the replacement circuits from what had been the uc3843 power pins.

I had to slightly enlarge the base plate holes to get all the bolts to go back in. You have got to laugh though - I put everything back together but then the looper footswitch would not work. The footswitches are heavy duty but connect to pcb switches using springs and one of the springs had not aligned properly. So I took it to pieces again and sorted that out. Everything is now working. The tube voltage is 235v, not 270v but there is no appreciable difference.

Definitely worth doing and I now have a looper. Any ideas why the high voltage part would burn up though? The diode and the circuit around it had really toasted, in parts right through the pcb. I have seen the same fault in other Blackstar pedals when I did web searches.