Large Farnell PSU trouble shooting question

[Chris Wilson]

0 to 60V at 0 to 50 Amps elderly linear PSU. Was working fine, tonight it emitted the magic smoke, and a LOT nof it. Now the volt meter flickers to 20v and down again. I have the schematics and manual for it, but I do not fully understand the transformer primary side SCR control. For testing purposes should I see the full uncontrolled secondary voltage at the smoothing caps after the FWB? Or should this dual SCR stuff limit the voltage unless the voltage control is wound right up? From where the smoke billowed out of the side fan for the paralleled pass transistor tunnel heat sink I suspect the trouble may lie in that region, but until I understand how the SCR stuff works I don't want to presume anything. I will try and attach the block diagram and the manual, but the block diagram may be too big to attach, so I will link to it as well. Thanks for any replies.

Block diagram:  http://www.gatesgarth.com/H6001.jpg

Manual with circuit explanations: http://www.gatesgarth.com/farnell_h60-50_0-60v_0-50a_power_supply_text_sm.pdf

[Andy Watson]

You seem to have the antipodean version of the service manual   Also, it's a bit short on actual schematic diagrams - which will make fault finding especially difficult.

The block diagram indicates that the SCRs control the primary of the main power transformer so I would not expect to see the full output voltage on the secondary - unless there's a problem with the SCR control.  Can you identify more exactly where the smoke came from?

[stj]

if this is the re-badged wayne kerr one, i know they have problems with RIFA film caps going bang.

[Chris Wilson]

Hi Andy, thanks (once again... ) for wading in for me. I have put the whole series of schematics at http://www.gatesgarth.com/supply.zip

I just took the top cover off after it died and couldn't see anything obvious. I need my beauty sleep to get my strength up to turn it over, it's extremely heavy, and look under the bottom cover, but I suspect it might be from the pass transistor area, and they are in a massive tunnel heat sink. Fom the way the V meter flicks up to about 20V for a fraction of a second and down to zero again I have a feeling there may be a short and it's going into its one Amp max protection mode. but just guessing at the moment.....Thanks again.

[Armadillo]

Check F1 fuse, and if it didn't blew, we have narrowed to the main capacitors and the diode bridge after transformer MT1. It seems to me that there is a cruised controlled current shorts, so likely capacitors breakdown are the suspect. Firstly Check all the capacitors for shorts or low ohms and their ESRs. And while you are at it, check all capacitors visually for bulge or anomaly. Consider replace all electrolytic caps.

Then measure the output resistance, hold the probes until the resistance peak. What is the value.?

[coromonadalix]

they seem like the sorensen / xantrex power supplies, schematics are available on the web ...   oups  just saw your zip files  loll

You could take an other power supply to feed yours past the rectifier bridge,  in parallel with c33, uncoupling the scr control board, but maintaining the secondary power supply section for the auxiliary supplies cct board

That way you'll know if your linear section is working,  the main input scr section is phase controlled ...

The more the scr are on, the full output become available  ...  it's a way to downsize and reduce heat and dissipation while not in full power,  an hybrid between full linear modes and smps modes.

You may have damage on the scr control board,  or scr2 and scr3 are bad or leaking,  or r99 and c25 have some damage

check for leaked caps , bursted ones,  for start

[CJay]

0 to 60V at 0 to 50 Amps elderly linear PSU. Was working fine, tonight it emitted the magic smoke, and a LOT nof it.

'A lot of it' usually means a capacitor has shuffled off its mortal inductor, on the face of it that may not seem hugely helpful I realise but it may help you narrow things down.

The SCR circuit seems to be only allowing half the mains waveform to the transfomer when the supply is set to certain parts of the output range, I'm not entirely sure though I have a sneaky suspicion you may be able to bypass it entirely and use the supply with slightly degraded performance for short periods of time (overheating risk for the series regulator).

Visual inspection is the first order, the mains filter (probably OK), then I'd go for the big electrolytics and bridge rectifier diodes, C36 on the output would be a suspect too.

After that, verify the +/-15V from the auxiliary supply board (it'll probably be OK)

[coromonadalix]

the two scrs are making some kind of triac,  maintaining the full ac phase.

Humm  maybe  not totally phase controlled, somekind of pwm controling

The supplies i have seen, are sniffing the ac phase, and adjust accordingly ...   doesn't seem the case here ?


Check the auxiliary supply unit, 

ground are #15, #14, #8
vcc on  point #16 should be around 50volts ??  , and after on point #2 around 30vdc

BUT there is a symetrical supply of +15vdc  and -15 vdc  as shown and written, but the common point ground is #5


I think they use floating grounds in the psu, be careful  you may have some falses values  because of that, it depends where you probe  on the boards.

The scr control board is floating on the - o/p output connector ... so does the two other boards, because of the second transformer.

Pin#5 on the control amplifier board is ground for that board, #8 is +15vdc, #1 is -15vdc


Your huge serie pass transistor  E, B C  should be tested disconnected from the circuit,  check your d23 and d24 output diode for short, conductivity or leakage ...


Did you connect something  on it before it gone smoking ??  had bad experience with inductive loads before, my Kepco psu went berserk, modified to add supplemental protections ..

[CJay]

Looking at the block diagram there are definitely floating grounds, it looks like a hugely powerful upgrade of the L30 bench power supplies except I think it's got two transformers, one for the floating supplies to the control boards and then the main power transformer (makes sense as you'd not want to be running your control circuitry from a transformer you were playing tricks with on the primary)

[Chris Wilson]

Thanks for the replies, I am going to get it on its side on the bench and have a look under the bottom cover later. The sheer amount of smoke and the smell of it suggests a LOT of burnt insulation or a cap. I am hoping cap.... It does indeed have a separate dual secondary transformer for the control circuits. I was actually given it years ago with a shorted pass transistor, and I later found a control PCB was missing altogether. I bought a very cheap baby brother to it and hand drew the tracks on PCB and etched a new board and populated it, and it worked fine for ages. I'll see what I can find and report back, cheers.

[CJay]

Thanks for the replies, I am going to get it on its side on the bench and have a look under the bottom cover later. The sheer amount of smoke and the smell of it suggests a LOT of burnt insulation or a cap. I am hoping cap.... It does indeed have a separate dual secondary transformer for the control circuits. I was actually given it years ago with a shorted pass transistor, and I later found a control PCB was missing altogether. I bought a very cheap baby brother to it and hand drew the tracks on PCB and etched a new board and populated it, and it worked fine for ages. I'll see what I can find and report back, cheers.

Well, as long as it's not the main transformer that's burned pretty much anything else should be repairable, the small transformer should be easy to replace as well, one from the previously mentioned L30 bench supplies might well do the job.

[coromonadalix]

it can be converted to a simple linear mode psu, it will dissipate more in some case,  but it will work ...

[Chris Wilson]

A screw that holds the bottom cover on has penetrated part of the loom, it looks like these screws are a bit too long, and have probably been replaced at some time in its history. I have repaired the loom nice and neatly, and powered up. The 50 Amp fuse is open, two resistors burnt on the control boards as well. R5 is smoked, R7 looksto0 have got uncomfortably warm. I have a mark by R5, that's the fella' I changed before, years ago, when a pass transistor failed...The loom that was damaged was to the pass transistor block.... I need to pull the pass transistor block on its big tunnel heat sink out, which was not trivial. Still, if I managed it years ago I guess it will all come back to me I wonder if this loom was always damaged and has chosen to short to ground again....?

[Chris Wilson]

Here are two more photos

[Armadillo]

The DC bus  ought to be heat shrunk ..... man! talk about physical separation of hv/lv.... I won't be putting my hand in. LOL

[Chris Wilson]

I am curious as to why a shorted pass transistor (and at least one *IS* shorted), takes out R5 on this "Control Board" every time. I am not advanced enough to work out why R5 lights up, and R7 seems to have got pretty darned hot. Thanks. The schematic is here for that board and the other link is the block diagram

http://www.gatesgarth.com/H60001.jpg

http://www.gatesgarth.com/H60002.jpg

Thanks.

The last time it did this it did NOT have an inductive load on it, as far as I can recall.


Another oddity is when I turn the main front panel mains power switch off it often trips the house mains distribution panel breaker, which is 30 A , but not the RCD.

[Chris Wilson]

the two scrs are making some kind of triac,  maintaining the full ac phase.

Humm  maybe  not totally phase controlled, somekind of pwm controling

The supplies i have seen, are sniffing the ac phase, and adjust accordingly ...   doesn't seem the case here ?


Check the auxiliary supply unit, 

ground are #15, #14, #8
vcc on  point #16 should be around 50volts ??  , and after on point #2 around 30vdc

BUT there is a symetrical supply of +15vdc  and -15 vdc  as shown and written, but the common point ground is #5


I think they use floating grounds in the psu, be careful  you may have some falses values  because of that, it depends where you probe  on the boards.

The scr control board is floating on the - o/p output connector ... so does the two other boards, because of the second transformer.

Pin#5 on the control amplifier board is ground for that board, #8 is +15vdc, #1 is -15vdc


Your huge serie pass transistor  E, B C  should be tested disconnected from the circuit,  check your d23 and d24 output diode for short, conductivity or leakage ...


Did you connect something  on it before it gone smoking ??  had bad experience with inductive loads before, my Kepco psu went berserk, modified to add supplemental protections ..

I'll check on your suggestions, many thanks for taking the time to type all that up!

In fact there was a load on it,  LF amplifier in Class D push pull on 136kHz and it turns on and off with the load applied.  I guess it could be seen as inductive, which is a major PITA if that's the case, as i wanted this hefty variable supply working specifically for this usage......

[coromonadalix]

well thats a nasty loom burn  outch   you did have to rebuild this ....

My inductives problems on my Kepco where triggering the output short protections, meaning the supply was stuck in protection even the load output disconnected

Mine psu has a preregulator section who was getting hot as hell, while the regulator section was warm,

I have a special resistor who is build like an dryer home appliance  loll

In your special output transistors (10x in parallels), even with balancing resistors, if one of the loot is somewhat more frail/under rated than the others, it will sacrifice itself  loll

In an old psu desing, still running after 10 years,  i matched all the transistors gains (hfe) to have an almost perfect 150 amps

[coromonadalix]

In a big psu like this  they should have added some time delayed primary transformer section,  the capacitors must give an hell of kick when first charging on power up.

[Chris Wilson]

It does fire up with a manly thump

[Kevin.D]

I am curious as to why a shorted pass transistor (and at least one *IS* shorted), takes out R5 on this "Control Board" every time. I am not advanced enough to work out why R5 lights up, and R7 seems to have got pretty darned hot. Thanks. The schematic is here for that board and the other link is the block diagram

http://www.gatesgarth.com/H60001.jpg

http://www.gatesgarth.com/H60002.jpg

Thanks.

Hi Chris .
  I think you asked this before and I answered in this post https://www.eevblog.com/forum/beginners/would-a-shorted-pass-transistor-cause-this-failure/msg432022/#msg432022.  I also suggested later what to do to prevent R5 being taken out in future) .

Regards

[Chris Wilson]

Hell, so you did Kevin. I am definitely getting senile I thought someone a long time back had explained the reasons, but for some reason I thought it was on an antique radio / test equipment forum here in the UK. I searched there but (obviously) couldn't find it. Forgive me...... I am old! Thanks, given what has occurred again I will make the mods you suggest. No offense! One day you'll know what i mean i guess <LOL> Thank you.

[Chris Wilson]

I am afraid I am back again!

I intended to modify the board with the R5 resistor and C10 cap change, but discovered the schematic I have is different to the reality of the board... It doesn't actually have C10, it seems like I do not have the correct schematic for this supply. An earlier version of the supply used discrete components on this board instead of IC's, and that version doesn't have C10. So it looks as if this board is an interim between all discrete and IC op amps. Not being sure what to do I replaced all the pass transistors for new 2N3055 and checked their DC voltage gains. I had some variations, and the gain on all of them was much higher than the ones I swapped out. The supply ran fine for a few days, but just now, running at 50V and about 40 Amps it went bang again, taking out R5 again. I have to assume another pass transistor has gone. I did check the balance resistors were all OK and they seemed fine. Googling shows some people who restore 70' and 80's gear say that the dies for the 2N3055 were changed and the current stocks are inferior to the earlier devices. It is mooted in several places to swap them out for MJ15003 devices with higher voltage and current capability. Is that worthwhile and logical? The other thing I have at the back of my mind is that I am driving a powerful 2kW low frequency RF amp (actually two 1kW amps into a combiner), so there's lots of 136kHz RF about. I wonder if it's getting back into the supply? Is the supplies topology likely to make it sensitive to this?

I link to a screen shot of the troublesome board from the earlier version of this supply, that does not use C10 as my own does not, but it also uses discrete components instead of IC's...

If patience with me has not run out I would be hugely grateful for more ideas, over the weekend I traveled a couple of hundred miles and bought another one of these beasts, I haven't had the covers off yet to see if it is exactly the same as this one, or a slightly different revision, but will want to fettle the second one up with new electrolytics and probably pass transistors.

The earlier variation of the schematics is at http://www.gatesgarth.com/H60-50_Diagrams(2).pdf

Thanks!

[oldway]

It is difficult to help you without the right schematics.

1) the phase control of the primary is regulating the voltage drop on the pass transistors (probably 5 to 8V...see technical documentation for the right value)...this must be checked when you repair this power supply before using it with high current.

2) There is no true soft start but there is a delay before firing the scr's.(not the best option, I did not like this circuit)...on the newer version, this is made by C12 (4.7µf 25V) of the scr contol board ...replace it !

3) You can't use this power supply to feed a high power load  (2KW)  which has high 136Khz ripple. You must use high value, low ESR external capacitors to reduce ripple current in the power supply.

Without external capacitor, the high frequency ripple will kill your pass transistors each time.

[Chris Wilson]

Thanks, all noted, I have just pulled the front panel on the "new" supply and it seems this control board DOES have C10 (10nF)added, across D1. It also has one less wire link on the PCB. The one that blows links the anode of Z2 to board connector finger 12, which goes among other places to the anode of Z3. The "new" supply does not have this link. Otherwise, at a fairly close look, then PCB and components are the same, and it has IC  Op Amps like my board, not discrete components.Apart from D1 and D2 being a 1N4003 on the new board, and the blown board has / had small glass diodes that look like 1N4148's


Each amplifier has its own low ESR 4700uF across the supply with a paralleled 0.1uF ceramic as well.

What's the best way to check the pass transistors are being kept in a sensible voltage drop range by the SCR's please?

Thanks again.