Dual power supply overshoot on one section only issue.

[Chris Wilson]

Thurlby PL320QMD dual section power supply 30V @ 2A with Revision 4 boards in it service manual is at  http://www.chriswilson.tv/PL_Series.pdf

The schematic for my exact PSU is on page 49 as seen with Adobe Acrobat reader.

The RH section, if I use it in isolated mode, shows a nasty overshoot nine out of ten times. This is when the supply is powered up on the master mains switch, and voltage set to a fixed value (5V in this case), and I then turn on the output switch for the RH section. I see an overshoot that slowly ramps down. This has cost me two DDS modules and 2 ATmega328PU chips before I stopped doubting my circuit wiring skills and began to look elsewhere.  Doing the same with the LH section I never see a spike, it goes to exactly where it's set and stays there. There are two "driver boards" with rectifiers, Zeners, blah blah, one for each side, and two identical control / display boards. Swapping driver boards does not cause the overshoot to change sides, swapping control boards sees the overshoot follow the RH control board to the LH side. I swapped out most, but not all, of the electrolytics on that board, none tested bad but I renewed them as a matter of course. I saw C11, which seems to be a tantalum at 1uF, so I swapped that for a 1uF electrolytic, as I don't have any tants in stock, but no difference.. Anyone care to advise where I might test in a more structured manner, given it is definitely a control board issue? Each Driver Board creates the supply voltages independently for each control board, so I do not believe it to be a board power supply issue to the control boards, as swapping the driver boards side to side made no difference. I am told normally these supplies have no real overshoot on either section. Thanks.

I edited the subject and text having been told what I am seeing is an overshoot, and not a spike as I originally described it.


http://www.chriswilson.tv/LH_Supply.jpg

http://www.chriswilson.tv/RH_Supply.jpg

[HalfSpace]

Hi Chris,

The circuit on page 43 seems like the adjustable output supply schematic for both channels. The circuit on page 49 is for a fixed 5V@7A supply.

Did you have a load on the supply when you did your scope measurement?

Check R60 (next to the load switch on page 43) if it is open circuit, +S (remote sensing) will be floating for a short period of time when the load switch is being switched causing the output to be unpredictable.

HalfSpace 

[Chris Wilson]

Hi, thanks for the idea. I had no load on the supply when getting the scope grabs. Not on either side, so was comparing apples to apples. I have checked R60 and R61 and both are good, unfortunately. I am wondering about R55, and will check that, and I am also wondering whether I ought to also attempt a calibration on the troublesome side of the supply?

[Chris Wilson]

Looking around this area has perhaps found two useful clues.


 On the faulty output side if I remove the link between + Out and + Sense, and either leaving in *OR* removing the - Out and the - Sense there's no overshoot at all. Replace the + sense to + Out link and immediately the overshoot returns.

On the faulty side, set to 5V with mains and independent power switches both on, measuring between linked - Out and - Sense and *UNLINKED* + Sense I see 5V
On the good side I see a couple of mV when testing in exactly the same way. So on the bad side the + Sense terminal is at 5V with that output on, and with no link to the + Out terminal


R60 and R61 appear good. In circuit testing of R55 shows just under 1 megohm. Hopefully this info will help get me closer, with your help? Thanks.

[Chris Wilson]

It's the ****** rocker switches!! Thought I was losing the plot, the other (good) side started doing the same, and that switch then felt "funny". I removed the switch on the original "bad" side and only one side of it was switching reliably. The other was sometimes staying closed, sometimes open. That would explain the odd time it would turn on without an overshoot. I took it apart and the hard copper rocking contact on one side was bent and displaced. I straightened it up, put a bit grease on the rocking / rubbing bits and it's OK. Going to look at the switch on the LH side next, that too seems to playing up now, and as I said, actually feels defective. I have been fiddling with it a lot of times. As you say, they are both clipped in the front panel, AND soldered via short round pins to a small PCB directly behind them, ideally I need like for like replacements, heaven knows if they are still available, or an equivalent. I think the contacts on them have been doing their own thing. I believe both boards have been OK all along, but intermittent mechanical / electric issues with the switches have caught me out. They always say suspect relays, switches and the like first, should have isolated them to start with I suppose I suspect unless I renew them it may dislodge the rocking contacts again, they may have gone soft with constant use. Thanks for all the advice gentlemen!

[HalfSpace]

Hi Chris,

Maybe replace the output load switching with some DPDT relays with 5A or 10A contacts and repair the front panel load switches so they can switch the relays. This will remove any high current surges at switch on from the load switches and extend their life.

HalfSpace 

[Chris Wilson]

 I have found two replacement switches via a friend, that fit perfectly. I had to file down the pins as they were eyeletted for feed through wire, not plain pins, but once I di that they dropped right into the PCB. I calibrated both sections today, and found the manual had an error. I had to use VR5 for setting the output voltage, not VR9 as stated in the manual. The machine seems to be working fine now, many thanks for your help and have a great Christmas, and a healthy and prosperous New Year.