Hello (and the insides of a TTI / Thurlby Thandar PL330 power supply)

[chrissmeuk]

Hi,

New here! I'm an ex-EE in the UK who now, with much complaining, writes boring business software for a living as all the work went to China and into computers (the irony!). I miss the work and am repopulating my lab at the moment from a hobby perspective hence this post.

Thought I'd start with a partial teardown of a TTI PL330 power supply. I didn't want to take it to bits completely.
Snagged this and a huge pile of bits this morning from eBay for a mere £21 (Around 35 USD). The item was listed as untested and buyer collects. The seller appeared to be clearing out her deceased father's bits and had no idea what they were. RIP fellow engineer - your tools and spirit will live on! Collected it all this afternoon.

Cracked it open. Nice and easy. Was designed to be opened. Three screws on either side and the two handle screws on the top and you're in. Contained a lot of dead things (mainly spiders and their victims). Quick clean out with the vacuum cleaner and a large paint brush and it's nice and clean. Greeted by a large transformer. On the left is the array of power transistors and massive heatsink (convection cooled) and the right is the front panel control board.



Turning it around presents us with the control board. This contains a couple of regulators for the front panel and control functions on a heatsink. Smoothing caps as well. None of them bulging or burst - good news . Nice and tidy design IMHO.



Moving to the rear: a large array of power transistors (TO3 package). Surprised these are just convection cooled!



All looks ok. Checked for burned or bulging caps, quick continuity test on the fuses, check for signs of melted cables. What the hell - plugged it in and turned it on with a paint brush handle. No magic smoke released!

Chucked a dummy load on it (a whole pile of 220R resistors in parallel) and it holds up nicely. Current limiter works, voltage is stable, controls all work. Meters agree with my multimeter on the dot.



And there it is reassembled with current limiting at 8mA and voltage set at 5v for an LED. Voltage is held by the diode junction at 1.85v and the thing sucks 8mA and is being limited (hence the dots on the bottom of the right display). Awesome.

Included in the £21 lot was a pile of unused stripboard (RS branded stuff), about 2kg of hookup wire, some turned pin wirewrap boards, a few components (mainly mainstream logic ICs and op amps - all socketed on cleanly soldered boards), a velleman thermostat which had clearly been modified and more heat shrink than you've ever seen in your life.

All parts will be dedicated to the memory of the original owner who's name I assume was written on an antistatic band provided (and could solder pretty damn well - not a dry joint in sight).

[IanB]

TTi make nice instruments. That was a bargain at £21. I paid full price for my power supply.

Convective cooling is a big plus for me. Fan noise quickly gets intrusive in a quiet home environment.

[free_electron]

these are very good power supplies. i had at least 50 of them in my old lab at work (when i was still in belgium) in 14 years not a single failure. we got the 3 channel ones with gpib programmability PL330Q

[chrissmeuk]

Yep appears to be a pretty good piece of kit so far.

Can't find a service manual for it online though which is the only worry. The HP 6236B it sits next to has a wonderful manual and bog standard bits in it as well as very nice diagnostic charts.

IanB: agree with the cooling perspective now. I noticed the cheap Chinese units all seem to have a PC fan in them. I like to have a bit of music on when I'm working so that would annoy me

[mtdoc]

I'm fixing a TTi TSX1820  power supply which I picked up cheap on ebay as a broken, for parts unit.

I was able to get a service manual with schematics by contacting the US distributor. They were real helpful and contacted the home office in GB to get a PDF copy which they forwarded to me.

 I'd give TTi a call.

[chrissmeuk]

Excellent - will try that. Thanks for the suggestion.

[Mark Hennessy]

+1

I rang TTi only last week, and they were excellent - no problem to email me the manual I was after. Highly recommended!

Last time I made contact was about 10 years back, and they weren't so forthcoming then. But I was really pleased to see how they've improved their service since then, and it's nice that they are still doing well in a difficult market. They distribute a lot of brands in the UK, which presumably helps to support their own T&M products... Definitely a company that I would be happy to support

Mark

[chrissmeuk]

Well within 3 hours of firing off the support request, they just emailed me the service manual for it in a PDF file!

Wonderful service. Wish every company was like this.

[jpb]

Nice to see the inside.

I have the PL320QMD - it cost me £160 as new (old) stock - rather more than £21 but it is a double supply with tracking etc.

I've been very pleased with it but have never looked inside. The meters seem pretty accurate within about +-1 count.

One thing that I've just discovered is that the Voltage needs some current draw to set properly.

I've been connecting two 6 1/2 DMMs in series to check their relative impedances - at 30V they both read 15V +/- a small difference. But if I reduce the output to 18V then the actual total voltage is only around 15V (the supply meters read 18V and 0amps). So I guess the Voltage control circuitry needs some (very small) current flow to work properly.

[Mark Hennessy]

One thing that I've just discovered is that the Voltage needs some current draw to set properly.

Not wishing to thread-jack here, but that doesn't sound right to me...

I have several of the older (brown-cased) PL320QMD, and none of them behave that way.

Yours has the newer control circuit. I do have a single PL320 that has that version (a PIC and a ICL7109 DAC to replace the ADD3701 meter driver chips), and it behaves just fine.

There is one "funny" with the older PL320QMD when in Tracking mode: if the slave voltage control is set to 0V, the slave voltage meter over-reads. The output from the slave is OK - it's just a metering problem. Turning the slave voltage control up cures it - on mine here, the voltage control needs to be set to about 7 volts before selecting Tracking, and then the slave meter is good over the full range.

I've no idea if the newer version has this bug - given that the control circuitry has changed quite a bit, I'd be a little surprised if it persisted - but I wonder if a similar, perhaps unrelated,  "bug" is present with yours? Are you using a single half in Isolated mode? Do both halves exhibit this bug?

My newer one had terrible troubles thanks to a leaking capacitor. I spent an age washing the PCB with IPA, but some residue remains - the symptom being that the voltmeter over-reads for the first 10-15 minutes of operation. It would probably be OK in a lab environment, but my workshop is in my cellar. It's fairly dry - I use a dehumidifier - but it's no-doubt more humid (~65% RH) than an air-conditioned lab would be. At least there are no static problems

Nice power supplies, though. The older ones are a pain to work on, and must have been incredibly labour-intensive to assemble; the newer ones are much better in that regard. With the older ones, I have re-capped them, and checked the torque of the screws holding the output devices to their heat sink - well-worth doing.

All the best,

Mark

[jpb]

One thing that I've just discovered is that the Voltage needs some current draw to set properly.

Not wishing to thread-jack here, but that doesn't sound right to me...

I have several of the older (brown-cased) PL320QMD, and none of them behave that way.

Yours has the newer control circuit. I do have a single PL320 that has that version (a PIC and a ICL7109 DAC to replace the ADD3701 meter driver chips), and it behaves just fine.

There is one "funny" with the older PL320QMD when in Tracking mode: if the slave voltage control is set to 0V, the slave voltage meter over-reads. The output from the slave is OK - it's just a metering problem. Turning the slave voltage control up cures it - on mine here, the voltage control needs to be set to about 7 volts before selecting Tracking, and then the slave meter is good over the full range.

I've no idea if the newer version has this bug - given that the control circuitry has changed quite a bit, I'd be a little surprised if it persisted - but I wonder if a similar, perhaps unrelated,  "bug" is present with yours? Are you using a single half in Isolated mode? Do both halves exhibit this bug?

My newer one had terrible troubles thanks to a leaking capacitor. I spent an age washing the PCB with IPA, but some residue remains - the symptom being that the voltmeter over-reads for the first 10-15 minutes of operation. It would probably be OK in a lab environment, but my workshop is in my cellar. It's fairly dry - I use a dehumidifier - but it's no-doubt more humid (~65% RH) than an air-conditioned lab would be. At least there are no static problems

Nice power supplies, though. The older ones are a pain to work on, and must have been incredibly labour-intensive to assemble; the newer ones are much better in that regard. With the older ones, I have re-capped them, and checked the torque of the screws holding the output devices to their heat sink - well-worth doing.

All the best,

Mark

I've not really had time to investigate it properly. I was actually trying to compare the impedance of the two meters on different ranges. At 30V (15V each) everything made sense - see attached photo.

When I went down to 18V so each meter would be in their 10V ranges (assuming an equal 9V each) I found that there was a disparity between the two meters and they added to a total of 12V while the meter on the supply stated 18V.
see attached two photos.

The supply has always worked ok on normal loads ie if connected to the two meters in parallel for instance. It may be something odd about the meters - it maybe that their impedances are so high that they are acting as capacitors - they read quite high voltages when just left open circuit.

When I get a bit of free time I'll try and get to the bottom of this. I plan to experiment with my 5 1/2 digit meter in 10Mohm mode in series with each meter in turn. I will also connect it in parallel with the two meters across the supply and check that it agrees ok then.

I know it works ok with the two meters in parallel so if their impedance is around 1Gohms (as a wild guess) then the power supply is ok with about 2nA output but unhappy with 0.5 nA.

Perhaps the central electrode is collecting charge.

Edit : I should add that the meter impedance is presumably affected by range so when measuring 15V they are in the 100V range whilst when measuring 5 or 6V they are in the 10V range.

Edit 2 : I realise that it can't be anything to do with current as it works ok with two in parallel down to around 100mV  yet with two in series it produces odd results even at 18V. It is all rather strange but I don't want to rush to any conclusions until I've time to repeat the experiments properly - perhaps this weekend.

Edit 3 : I've just turned everything on again - i.e. no warm up and having not gone up to 30V first and the meters are reading around 19V in total but are going all over the place - this may be because they are on auto and are switching ranges (and thus impedances) constantly - I'll switch them to manual mode and see what happens.

Edit 4 : If I manually set the ranges and set filtering and slow measurements then the total adds to 18V as it should. So what was happening is that the range switching was altering the impedance and in effect acting as a switched resistor amplifier/oscillator! It shows you shouldn't do experiments in a hurry.

Edit 5 : Thinking about it - the long wires I am using provide some inductance, rapid switching of the impedance must be causing some LdI/dt type voltage drops (back EMF) to quite a large extent for the total voltage to drop by 6V on average.

See attached picture - the total voltage is now 18.02V which is pretty close to the 18.04V on the power supply.

[Mark Hennessy]

It seems rather strange

If it helps, my later PL320 feeding a single 2015 works exactly as I'd expect. Sadly, I don't yet have two 2015s to play with

Let me know what you discover - it is most intriguing...

[jpb]

It seems rather strange

If it helps, my later PL320 feeding a single 2015 works exactly as I'd expect. Sadly, I don't yet have two 2015s to play with

Let me know what you discover - it is most intriguing...

See my edits above. What was happening is that the meters were on auto mode so where range switching which changed their impedances. As they voltages of around 9V where close to the 10V range switching point they were switching between two ranges which then changed the voltage distribution between them. The effect was like some sort of switched resistor oscillator (pump?) Edit : probably just the long wires behaving as inductors though I'm surprised at the size of the effect with such small currents.

[macboy]

+1

I rang TTi only last week, and they were excellent - no problem to email me the manual I was after. Highly recommended!

Last time I made contact was about 10 years back, and they weren't so forthcoming then. But I was really pleased to see how they've improved their service since then, and it's nice that they are still doing well in a difficult market. They distribute a lot of brands in the UK, which presumably helps to support their own T&M products... Definitely a company that I would be happy to support

Mark

I had the same experience getting a service manual from TTi. I e-mailed both the Canadian and the home office, and both replied with the manual in under 24 hrs. I am impressed with the design and construction of the power supply (TSX1820) and I love the very detailed circuit descriptions in the service manual, similar to the way old HP manuals were written. I definitely do not hesitate to buy TTi.

[PTR_1275]

Hi Guys,

Anyone out there with a schematic for the PL330 that they don't mind sending through to me??

I have contacted TTI roughly 4 weeks ago (Mindful of the Christmas break etc...) But still haven't heard from them.

Some advice from people more experienced than myself would be greatly appreciated if you can give me somewhere to start

I have 2 x PL330's, one of them fried itself a few years ago and I didn't have the time to look at it back then so it went to the back of the cupboard. I pulled it out the other month to try and fix it and this is where I am at so far.

The current limiting function isn't working which caused the 1? 1% resistors leading into the power transistors to burn out. They have been replaced and the unit is giving voltage no problems, but it isn't reading the current (Hence it cannot current limit). I took the control board out of the working one and put it into this and it works a treat, so I know the problem is within the board. No visible damage, heating etc... Shunt resistor measures fine and is giving the same voltage drop as the good board (With the same load / voltage on it).

That is about where my component knowledge ends. Without a schematic it is very difficult to trace things back and see where they go. I traced some tracks back to the LM 324 op amp, I had a few spare in the parts boxes so replaced it and still no luck.

Any help would be greatly appreciated.

Cheers
-Chris

[gfiber]

I got lucky and purchased a TTi TSX-1820 on Ebay for $145.00 plus $20.00 shipping within the last month, no one bid on it.
It looks brand new just wrote and asked TTi for the service manual and schematic for it.
It puts out volts just fine, have yet to put a nice load on it, maybe this weekend.

Gary