i'm on a roll... the kgsshv

[s_r]

I let it stay on for another hour or so and those measurements didn't budge.

[jwzhan]

don't worry about it then. 569 for bias is plenty accurate. It's much more accurate than what STAX amps have.

 

EDIT: 9V difference off 450V is only 2% difference and 11V difference off 580V is only 1.8%. Take into account that your meter most likely has a 1% tolerance and your resistors have 1% tolerance, you really shouldn't worry about these too much, unless explosions take place.

Edited May 16, 2013 by jwzhan

[palchiu]

Got the dissipante in from modushop, now I can start the hard part of the build.

 

P1070107.JPG

 It's looks beautiful! Nice "KGSSHV" logo.

[spritzer]

That small a difference is nothing to worry about.  The zeners are probably 10% or 5% at best. 

[nopants]

Just booted up the positive rail, 503V. This was after powering on the LV section and installing the bridge backwards (thought the + sign for the cap was for the bridge), luckily I caught a whiff of something acrid. However, this wasn't before the negative regulator died.

 

Edit: negative rail is -504V, bias before 5M resistor is 583V

 

Time to stress test

[TMoney]

Another HV (mine) goes down to a shorted SumR trafo. Grrrr.

[GeorgeP]

That sure sucks. Though I thought the other SumR had sustained some fairly significant abuse before succumbing (apart from the T2 ones that had issues). Assumuming you tested it for shorts already and have concluded so. Which SumR trafo were you using?

[shipsupt]

Bummer.  

 

Makes me think about using one of my Toroidy's instead of the SumR.

[nopants]

Anyone know how to determine if the electrolytics in the LV section are still viable even though my bridge was backwards for a bit? Can they still go bad even if there's no pop/leak? I've no CLR meter with which to check...

[eggil]

Another HV (mine) goes down to a shorted SumR trafo. Grrrr.

That really sucks!

Could you describe what happened? Yours had been running longer than my off board build.

[johnwmclean]

Not good news...

 

I know Richard has refined this design since a few earlier TX’s fucked up with thicker insulation between the secondaries.

Were you using thermistor/s?

[justin]

I know Richard has refined this design with thicker insulation between the secondaries since a few earlier TX’s fucked up .

 

just wondering if this is what you meant

 

on the original Aristaeus transformers, Plitron made them with no insulation between secondaries.  they all shorted after a few days.  their engineer said they should have been fine because the enamel on the wire is rated for 800V They all had to be sent back to be stripped down 

Edited May 17, 2013 by justin

[johnwmclean]

Indeed enamel rated at 800V, you'd have to have a good dose of blind faith...

 

I can’t say if the transformers from Sumr had a similar scenario or not, I know when I ordered my second TX he used “thicker” insulation and had a better understanding of the current demands of the circuit... memories vague on this, long time ago now.

[spritzer]

Ahhh magic coating supposed to save the day.  How ignorant people can be about high voltage behavior never ceases to amaze me... 

[s_r]

So after testing the boards I got one to balance & offset fine. However there's a problem with the other board. Only one LED lights up (D13). I checked beforehand for possible shorts and didn't see anything, also made sure to hook the signal inputs to ground. Any obvious spots I should check first?

[eggil]

What is the voltage on the + and - outputs?

[s_r]

I shut it off right away when I didn't see every LED light up, but remeasuring I got +457V on both +/- outputs of the problem board (while I get -a few volts on the good board since I set the balance/offset while it was warmed up).

[jwzhan]

use diode test to see if any of the transistors are blown and compare the diode test values to the one that's working. It's a pretty simple and straight forward way to find obviously blown transistors. also look for burn marks and chippings on resistors. It happens when resistors are overloaded.

[s_r]

Well, there are no burn marks or anything visual that would indicate a bad part. I'll try and see if any of the 4686/1968's are blown (sure hope not though).

 

Edit: It doesn't seem like the output transistors are the problem, their hfe all check out according to the spec sheet and unless I'm reading my multimeter wrong they seem to pass the diode test. Only D13 lighting up seems a little weird to me though, shouldn't D10 be lit as well?

Edited May 18, 2013 by s_r

[eggil]

Are the balance and offset pot in the middle?

[spritzer]

No amount of offset would cause +475V on both outputs so it isn't the pots.  I can't ever remember the numbers but isn't D13 next to the 2SA1968's?  That will light up is the CCS is working, regardless of what's going on elsewhere in the circuit. 

 

Check the 1486's too as they could be the culprits.  I did have one board that behaved in this fashion and I never managed to fix it, even after swapping out pretty much everything.  Gave up and built another one... 

[s_r]

D13 is right beside the 2SC1815 (Q20). I'll try swapping a few 4686's with spares, and D10 too since that looks suspicious to me (since it's right beside the +15 input).

 

Edit: Well it's not any blown LEDs, they all light up after applying voltage directly to them. Still puzzled as to what would cause only D13 to light up. Also, measuring the -HV input on the problem board I get about -63V (-459V on the working board).

Edited May 18, 2013 by s_r

[palchiu]

Question solved 

[spritzer]

What ever that fits really.  400V would probably fit but it is overkill. 

[palchiu]

Spritzer, thank you!