Happy this worked out, well done Tinkerer.

Congrats!  The amp looks nice.

congratulations on a fun build.

The harder the work, the sweeter the sound.

Congrats

Congrats.

We're always glad to help, just not in the way you sometimes expect....

Well, after everything running fine for a couple weeks, something gave out in the power supply. No pop or spark or smoke but the +500V just dropped down to nothing.

Amp boards themselves are fine. On the PSU, the 15V+/- and the -500V are fine. The transformer is good. All the sand and the diodes were replaced on the +500. The traces look fine. No scorch marks under any caps or resistors I can see. I think it's become connected to the ground plane somehow past the R2 and R3 power resistors on the main +500 rail.

I made a crude measured voltage map of the working and nonworking sides to compare. Working one might be helpful in the future for anyone else troubleshooting a 500V PSU on a KGSSHV. I know that sort of thing was asked about earlier in the thread. Mostly posting see if I'm on the right track or missing something really obvious.

Sorry about the orientation. It was just easier for me to compare left to right.

...and your rectified voltage is?

Yeah, close to zero volts DC doesn't make any sense unless the diodes are toast.  If they were shorted to ground then it would all explode. 

First thing I figured when it went out was HV rectifier bridge, so I replaced that and the zener string at the same time just because I had the parts laying around. Didn't fix it. So then I replaced all the other transistors and diodes at the same time. Must've blown the new parts piecemeal then when I powered it up. Going to have to do the whole thing at the same time then now I suppose just to be sure. Well, I had to make a new mouser order for some other stuff anyway. Thanks. I'll give it another try in a few days.

Have you checked the AC voltage out of the transformer without it being connected to the power supply board?

Just now, wink said:

Have you checked the AC voltage out of the transformer without it being connected to the power supply board?

Yeah. Both the 450VAC windings are fine. Same result if you switch them. Negative rail on the PSU ramps up to spec and the positive sits within a volt or two of zero.

11 hours ago, Tinkerer said:

Yeah. Both the 450VAC windings are fine. Same result if you switch them. Negative rail on the PSU ramps up to spec and the positive sits within a volt or two of zero.

...the stth512 has 2V drop, and you have excatly 4V after rectifying ....and you ac is 450V .... I'd try if the connector terminal for the AC is ok?

Edited November 21, 201510 yr by sorenb

Welp, soren called it. Cold joint on the terminal block. Must have missed it when I reflowed the components originally. That was a lot of desoldering and hassle over nothing. Thanks for helping me work through things again guys.

This must be a record slow build, but I now have a working PSU (+448v -448v) and all is fine - Pro Bias also OK. I need normal bias as I currently only have Sr5. For normal bias, I tapped the circuit where the 500v option jumper connects (on the same side as the pro bias connection) and duplicated the circuit from the 10m90S to outpuit, with a 230v zener string. I presume if I was going to tap the circuit on the other side of the jumper (straight off the bridge without the voltage doubler) it would need another 15k? resistor in series Which side should I tap?

43 minutes ago, DHTGUY said:

This must be a record slow build, but I now have a working PSU (+448v -448v) and all is fine - Pro Bias also OK. I need normal bias as I currently only have Sr5. For normal bias, I tapped the circuit where the 500v option jumper connects (on the same side as the pro bias connection) and duplicated the circuit from the 10m90S to outpuit, with a 230v zener string. I presume if I was going to tap the circuit on the other side of the jumper (straight off the bridge without the voltage doubler) it would need another 15k? resistor in series Which side should I tap?

I'd rather change the values off the zener-string a bit an tap the NB from there. Then you only need to add three resistors and a couple of caps.
Like Kevin has done here:

As you may notice, there are two NB BIAS connections corresponding to the two pins on the jack.

/S

For NB I thought you could tap straight from B+ with a simple voltage divider.

Lots of ways to skin this cat.

I'd just tap into the B+ with a voltage divider and a ballast resistor. 

Like this?

and thanks for your help

Yep that will work, I put the cap after the divider but it doesn’t really matter. You could reduce the two 178k into a single 355k. But check your calculations here:

http://www.raltron.com/cust/tools/voltage_divider.asp  

 

 

Thanks John. What is the reason for the 10M90S (I should work it out from the data sheet I suppose), if a divider works OK? (the 2 x 178k are just what I have available). I suppose the calculated 225v above is near enough to 230 to be fine

You can put a tiny cap after the divider I like the safety the divider brings so as little storage after it as is possible. 

0.01uF 250v tiny enough? Should it be on the bias out side of the 4.7M ballast resistor (then the storage is important agreed) - or on the divider side? I'm interested to know the advantage of the 10m90s

Never put a cap after the ballast resistor.  That will damage the headphones over the long term.  Stax used to do it back in the 70's but have since learned their lesson. 

That cap would work before the ballast but I'd still leave it out in favor of a cap before the divider.  The 10M90 is worthwhile to do to fully regulate the bias. 

Hmm - perhaps I'll do another board with the 10m90s again (I managed to damage (fry) the one I made by not waiting long enogh for cleaning solvent to evaporate ). Thanks all - just a cap before the divider makes sense.