Man, I have been using those terminal blocks and have never had any problems.I would cut down on the amount of solder if I were you.

http://www.farnell.com/datasheets/1998954.pdf

Rated 600V (IEC) 300V (UL)

I'm going to order some today, have a good inspection and see what I think before deciding to install.

I'm watching with interest and this got me thinking I've made PCB's with mains on but just used the spacing on the connectors reasoning they must be far enough apart. The proper answer is this https://www.google.co.nz/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&ved=0ahUKEwiPs6i8jOXOAhXGj5QKHXGmD4kQFggcMAA&url=http%3A%2F%2Fsisko.colorado.edu%2FCRIA%2FFILES%2FREFS%2FElectronics%2FIPC_2221A.pdf&usg=AFQjCNEa4EGy9i_1N9DFkSvA-G7WRvYoug&sig2=1bL_pBagjRyixszFPnuroA&cad=rja

Care of google :-) This is the IEC document that expressly stipulates trace spacing vs Voltage. I've been rebuilding my old Quad ESL's and the EHT  was potted in beeswax 60 years ago. Some of the newer ones had 'conformal' coating instead which is a special varnish. The IEC did experiments however and they found breakdown was purely a function of spacing. The coating must help with humidity though?  I'm a newbie and I've joined the buy group to build a KGSSHV Carbon.  I'm getting excited about getting started so sorry to butt in unannounced :-)

I tested +400V GRHV today and found it was not working. It outputted 403V immediately when powered up, and quickly dropped to about 200V after 5 seconds. It would continuously dropping and won't stop. Where should I begin to check with? -400V with same batch of parts work perfectly(-403V)

Edited August 29, 20169 yr by joehpj

10 minutes ago, joehpj said:

I tested +400V GRHV today and found it was not working. It outputs 403V immediately when powered up, and quickly dropped to about 200V after 5 seconds. It will continuously drop and won't stop. Where should I begin to check with? -400V with same batch of parts work perfectly(-403V)

Remove the load from the PSU, and start from there ... sounds like the current limiter kicks in

No, it's unloaded.

19 minutes ago, joehpj said:

No, it's unloaded.

you might poke around the amplifier then ... check the ksc5026 ...too much solder can hide a bad joint

Do you have 10V on each side?  (base of both ksc5026) ....check the two 390k resistors and the .047uF cap

Coming back to insanity's problem with the terminals/tracks.  It might be worth masking the area and spraying with conformal coating to further increase the insulation strength locally.

http://uk.farnell.com/electrolube/hpa200h/coating-conformal-aerosol-200ml/dp/298281 

It can be found cheaper elsewhere but if Farnell are delivering to you maybe not.

PS alternatively you can spray a little into a jar and paint it on with an artist's brush

Edited August 29, 20169 yr by headinclouds
add PS

A stupid thing... or not.

If the problem is the floor connector, Why you don´t use a thermal silicone (same that we used between headsink and trasistors) but between the connector and the pcb? You only need to make the holes on the silicone.

 

Tried to turn up +400V several times, it always gave +403V in the first and quickly dropped to 300V. I hooked the fluke on LT1021 and found LT1021's voltage would slowly increase from 4V initially to about 4.5V when the output voltage dropped to 300V , which was obviously wrong. Where should I continue to check with? I am sure it was marked LT1021-10 on the chip.

Did diode test on ksc5026, which seemed be fine. 6V on both sides and it was not shorted.

Could it be the LT1021 that is the problem? It's supposed to be 10 volts as long as the regulated voltage is some 20V or higher.

6 hours ago, joehpj said:

I am sure it was marked LT1021-10 on the chip

meaning?

On ‎8‎/‎29‎/‎2016 at 5:57 AM, joehpj said:

I tested +400V GRHV today and found it was not working. It outputted 403V immediately when powered up, and quickly dropped to about 200V after 5 seconds. It would continuously dropping and won't stop. Where should I begin to check with? -400V with same batch of parts work perfectly(-403V)

didn't you build one of these a while back? Just do a diode test between the two PSUs and you should find the problem fairly quickly.

Being really dumb, and ready to put on my hard hat. Do we need the screw connectors? Can't we simply solder the wires directly to the PCB? If you need to remove a board just clip the wires or de-solder. Only asking.

8 hours ago, sorenb said:

meaning?

It's 10V not sth like 5v version. Switched Lt1021, nothing changed. LT1021, 10M90s, DN2540 already switched. Still the same. Voltage dropping speed became a little bit slower, but it's still not working. 

8 hours ago, GeorgeP said:

didn't you build one of these a while back? Just do a diode test between the two PSUs and you should find the problem fairly quickly.

-400v was exactly the same circuit and was working perfectly. It's a little bit hard to measure because all the fets are directly mounted.

After turned off, I couldn't feel any part is hotter than room temperature. I think it's because there's no current at all. (current limiter?) 

1 minute ago, joehpj said:

It's 10V not sth like 5v version. Switched Lt1021, nothing changed. LT1021, 10M90s, DN2540 already switched. Still the same. Voltage dropping speed became a little bit slower, but it's still not working. 

Can you please try to clean the joints in the psu-amp with some wick and re-solder those not using too much solder?

Thanks. Do you mean there might be short between fets? I think I will replace  STN9360, KSC5026, BC557 one by one. Unregulated voltage was fine(450V). 

16 minutes ago, joehpj said:

Thanks. Do you mean there might be short between fets? I think I will replace  STN9360, KSC5026, BC557 one by one. Unregulated voltage was fine(450V). 

No. If there are any shorts I expect you to have found those already? I suspect bad joints rather ...too much solder can hide a bad joint ...I would clean up the the joints and re-solder those before beginning to replace parts ... 

yes, you might want to check the continuity of the pins on the chip carrier for the lt1021 to see if the joints are good.

I am in the process of repairing by carbon. I now have the problem, that the offset seems to be stuck at -60V and I cannot adjust it in either direction. 

I have checked the current sources and they work fine. Where would you start looking? 

Of course the servo jumper is not connected. 

Edited September 1, 20169 yr by insanity

4 minutes ago, insanity said:

I am in the process of repairing by carbon. I now have the problem, that the offset seems to be stuck at -60V and I cannot adjust it in either direction. 

I have checked the current sources and they work fine. Where would you start looking? 

check you CCS'es if they are putting out what ever you aims for 18-20mA, if you still have the problem you probably need to adjust the tail resistor a bit

Looks like the voltage on the psu +400 drops to 60 when I put on the amp board. I will have to test if the same happens with the other amp board. 

Will do it tomorrow. Too tired to work on HV. 

Edited September 1, 20169 yr by insanity

33 minutes ago, insanity said:

Looks like the voltage on the psu +400 drops to 60 when I put on the amp board. I will have to test if the same happens with the other amp board. 

Will do it tomorrow. Too tired to work on HV. 

ohm test the amp boards  HV+ - GND, and HV- - GND , around 250kohm-ish as far as I remember ...probably a short somewhere

@joehpj: did you try measure the voltage drop on the current limiting resistor on the B+ side through the power-up and voltage sagging lapse? 

Just a thought.

Changed several sands, and resulted in BC557s fried. Suspecting .047uf cap leakage because the voltage of 20K kept raising (~15V) and opposite side LT1021 voltage dropping. Will change all the sands and caps. I think this is the fastest way now for me could thought of. Luckily I still have same spare parts.

How do you guys think?