Thurlby PL310 various issues (brown unit)

[MrBungle]

I've obtained another PL310. I've got a nice working perfectly later grey model unit already so decided to pair it up with another one when one came along.

Well I got a cheap one but it's not without problems. Issues:

1. In output off (current/voltage setting mode), it is outputting ~ 0.21 volts. Sense bars have been installed across the sense/output terminals for ref.

2. The current limit pot is wobbly (this is the one that scares me - hoping the board isn't knackered)

3. The current damping switch is broken and stuck on.

4. It's got a Supplied by RS sticker on it (this annoys me, I don't know why )

5. Needs a good clean up.

Ergo, does anyone have service information for the older PL310 units. I've got the service manual for the newer ones already. I haven't cracked it open yet.

[MrBungle]

Scrap this. I found it: https://www.arklite.net/thurlby%20pl%20xxx.pdf (don't think this violates any forum rules)

I'll keep the thread open to post any repairs if anyone else comes across this.

[mhennessy]

Ah - the early "birds nest" edition

There's a fair bit about them on the forum. I've repaired a few as well. I'm not immediately sure why there's a voltage when off, as the output is isolated properly by the DC switch - perhaps it's leakage in the switch itself? Presumably the voltage drops like a stone when loaded with something like a 1k resistor?

The current pot is a cheap 10k log pot - no worries there. Unlike the voltage control, which is a hideously expensive wirewound job (that one needs to be good quality, and can be responsible for voltage drifting). When you get inside, you'll see that unlike the later ones, the pots aren't mounted on the PCB.

Most of mine are RS branded. Back then, RS branded practically everything they sold...

Good luck

[MrBungle]

Excellent I will have a read around here on them.

Absolutely no idea what the output voltage issue is. Will disassemble it later and have a poke around.

Glad to hear it's a normal pot not a PCB mounted one - I hate them! Will extract, clean check it etc. The main voltage control pot is good.

Thanks for the info.

[MrBungle]

And typical. RS has no suitable replacement rocker switches until March the 30th!

Edit: appears that 13x16mm panel cut out rocker switches are rare as anything.

[Oldtestgear]

It might be helpful to some people on the forum to know that TTi have a very reasonable spare parts policy. The minimum order charge is £5 +VAT and postage is free. Compared to many of the far Eastern companies that are really easy to deal with and accept card payments too.

Phil

[MrBungle]

That's pretty good. I've back ordered the switch from RS in the end. It's here for reference:

http://uk.rs-online.com/web/p/products/7109378/

I went through the calibration steps and the issue still persists with the output having a measurable voltage. The schematic shows that this should not be the case so I'm going to look at the output switch.

There are two rather suspicious looking axial electrolytic capacitors that look after-market attached to the output terminals which may explain this...

[MrBungle]

Definitely suspicious. There are two. One from each output terminal to the DC out switch. If I discharge the caps, the voltage on the output disappears. Wonder if these are for stability. They're not in the service manual though. This is the SMM schematic:



This is reality:



Also a nasty surprise: the main filter cap has no bleeder resistor on it! Nasty. Glad I checked that. Adding one later.

Fixed the current limit pot. The thread end was mostly stripped but I took a spacer ring out and it is fine now. Sticks out about 1mm more than it should but I can live with that.

Edit: thinking of soldering a 10k resistor across the output. That will null the voltage. I will re-cal the current monitor after that as it'll add a 1.5/2mA offset to it.

[MrBungle]

Added a 22k across it. That gives well over 100v across it inactive before the 0.6W metal film resistor is out of spec.

Have re-calibrated it and all is good. Drops to 0.0v straight away and is rock solid stable when limiting or not.

[mhennessy]

That capacitor is C9 - why are you suspicious of it?

They are placed as physically near to the output terminals as possible to optimise the transient response.

Sounds like leakage in/around the switch to me... If you add that resistor, the current error will change with output voltage. 30k will limit the worst-case error to 1mA.

Incidentally, if you short that cap while the unit is on, you're shorting the regulated output. Obviously the supply will current-limit, but you won't be aware of that because the decimal points are illuminated when the DC output if off, so they won't be able to flash.

[MrBungle]

C9 is fine - this was an error on my part. I was seeing capacitance on the output when it was isolated as in output off. There are two capacitors after the switch that are not on the schematic. There is no leakage through the switch - I measured it on my Fluke 8024B.

The 22k gives approx 1.45mA error however there is near 100% linearity on the current sampling from my measurements so I have calibrated that out. The current limit output matches my measurements externally as well.

All in all it adds up. I'm not totally happy with the solution but it makes the power supply behave a little better.

Realistically speaking the moment a nice older HP unit makes an appearance these are being ejected

[mhennessy]

I can't see capacitors after the switch - sorry to ask what might be a daft question, but you're not mistaking the diodes for capacitors?

[McMurdo]

I'm sure we had these same models where I first worked. The user manual says something similar to 'in common with all good power supplies, the thurlby has a capacitor connected directly across the output terminals'.

I'm sure I did not imagine that.

[mhennessy]

No, you didn't imagine it:

4.6 Instantaneous Current Output
The current limit control can be set to limit the continuous output current to levels down to 1mA. However, in common with all precision bench power supplies, a capacitor is connected across the output (isolated by the output switch) to maintain stability and good transient response. This capacitor charges to the output voltage, and short circuiting of the output will produce a short current pulse as the capacitor discharges which is independent of the current limit setting.

The capacitor in question is C9, clearly visible in post #8. Note the bit in brackets, which confirms it's prior to the output switch, which is why I'm confused by the mention of "two capacitors after the switch".

The warning it gives is quite useful - it's saying that if you're tempted to use a PSU like this to test semiconductors with no series limiting resistor, you're likely to damage them (as I did with some LEDs many decades ago )

[MrBungle]

Check the photo attached above

[mhennessy]

I did (see post #12).

[MrBungle]

My mistake. Apologies!

I have traced the whole circuit again after throwing my notes and prior analysis in the bin and it is indeed as per the manual. C9 is actually two caps in parallel.

This does not explain however how that is producing a voltage across the output. However the addition of the resistor has fixed it so I'm happy (for now)

[dave cox]

Junction EMF created by a chemical reaction somewhere (if you were you measuring with a high impedance volt meter) ?

dc

[MrBungle]

Yeah 10M impedance Philips PM2425.

Test cases I performed before modification were:

1. Turn supply voltage to 0.00v. Turn output on. Turn output off. Voltage measures 0.0v
2. Turn supply voltage to 32.00v. Turn output on. Turn output off. Voltage measures 0.455v.
3. Turn supply voltage to 32.00v. Turn output on. Turn output off. Voltage measures 0.455v. Short output. Voltage measures 0.0v.
4. Turn supply voltage to 32.00v. Turn output on. Turn output off. Voltage measures 0.455v. Short capacitors. Voltage measures 0.0v.

So the condition is charge storage after the output of some description (confirmed by 1,2,3 above) which does involve the capacitors in the photo above (indicated by 3,4).

Even on a 10M impedance meter that should drop to 0 like my newer PL310, Heath and HP supplies do, ergo the fix involved ensuring that they were discharged properly with a medium value resistor. It was that or reverse engineer the REAL schematic of the output which I'm not 100% sure does match what is in the service manual at the moment.

[dave cox]

I asked because I got caught out and wasted and hour or two chasing stray voltages due to the fact that I forgot my voltmeter had a > 50G ohm input, being able the measure the voltage on a charged 100pF capacitor put me straight

(PM2425 actually 20M ohm on the low ranges!)

dc