Thurlby Pl320 display repair

[WaveyDipole]

One of my Thurlby PL320's has developed a problem with its display. Individual segments of the digits have failed rendering a couple of digits rather difficult to read. I am contemplating whether and how to go about repairing this. I am sure I read about someone replacing the individual LEDs with SMD LEDs once. I was wondering however if I could replace the whole 7 segment module. I have a couple of questions though:

1. How do I know whether it is common annode or common cathode? From the circuit I can see that is uses a pair of AC573 ICs and a ULN2803 IC ADD3701 ICs to drive the display. Datasheets seem to indicate a positive output and hence common cathode but I am not sure. Can anyone confirm this please?

https://www.digchip.com/datasheets/parts/datasheet/477/AC573-pdf.php
https://www.sparkfun.com/datasheets/IC/uln2803a.pdf

2. The displays in the PL320 are connected with ribbon cable along the bottom, so that all pins are along one endge of the display. The closest I have to these are these:

https://www.ebay.co.uk/itm/7-segment-LED-Display-Single-Double-Digit-0-3-0-56-Common-Anode-or-Cathode/231631926172?hash=item35ee56b39c:m:mib257QA-kWsvZlww3HtVSw&var=530860231120

The other option that ocurred to me was to hook up a pair of these if I can figure out which point in the analogue circuit they would need to be attached to:

https://www.ebay.co.uk/itm/Mini-Digital-LED-Display-4-Bits-0-100V-Voltmeter-Panel-Volt-Voltage-Meter-T-K3B4/254327837713?epid=11027093952&hash=item3b371ed411:m:mCAQJ8YJc9vaeZm-kGGWooQ

Anyone replaced the 7 digit displays in these? If so, how did you do it and what parts did you use? How did you determine the order of the pins on the 7segment module?

Finally, I did wonder about forgetting the repair and buying either of one of these instead:

https://www.ebay.co.uk/itm/YH-3005D-II-2-CHANNEL-OUTPUT-ADJUSTABLE-LABORATORY-PRECISION-DC-POWER-SUPPLY/332597704112?hash=item4d705e35b0:g:1hAAAOSwNaRas0cI

https://www.ebay.co.uk/itm/DC-Regulated-Bench-Power-Supply-Variable-Output-0-60V-0-3A-0-6A-0-30V-5V-Fixed/333396141528?hash=item4d9ff565d8:g:s-4AAOSw7Z5dy~pq

The Yihua has the advantage of 3 voltage settings on the third supply. Any thoughts? Cheap and nasty or Ok? (I can't really afford or justify a Rigol!)

The Two Thurlby supplies I have both work, but one unit has some issues including a leaky DC smoothing cap and doggy fine voltage selection pot and the display problem which seems rather trickiest to fix. The other unit seems to be working fine so I would probably keep it or sell it.

EDIT: Changed circuit diagram to the correct (older) version.

[bd139]

Are this a brown or black/white or grey unit?  If it’s not the grey one I’d get rid of it and buy a newer one as the older ones are a complete bastard to work on. Not sure what type of display it is. I’d pull one and check it with a DMM diode tester.

Don’t buy the Yihua. It’s a pile of crap. There are plenty of the grey PL series units out there and they are pretty bomb proof and well engineered power supplies that will last years. The earlier ones, not so much!

[Andy Watson]

The displays are common cathode.  I replaced them with these: Kingbright CC56-12CGKWA
https://uk.rs-online.com/web/p/led-displays/8721900
(Other colours and suppliers are availabe.) The problem is that the original displays are relatively slim in comparison to the replacements.
If you only require one display I'll check my bits-box - although - I might have chucked-it already

[bd139]

Talking of bits box, looks like RS are out of stock. Is it these? https://www.bitsbox.co.uk/index.php?main_page=product_info&cPath=172_181&products_id=1386

[WaveyDipole]

Unfortunately (it would seem) both of mine are the the brown units.

Bd139, thanks for the comment on the Yihua. I thought there must be a reason for the comparatively low price. I notice that most modern supplies don’t have the sense connectors either, not that I have used them much. I am going to change the cap tomorrow. I am hoping that was the reason for the voltage sagging hen picking up again at random intervals. If it then tests OK I will change the pot and see what I can do with the display.

I also have a look out for a grey/white unit. There are a couple of double units on eBay at the moment but one is £100 and the other £150, although the latter has GPIB, so not particularly cheap.

EDIT: Seems I misread the shipping charge. There's a shipping cost and an import tax cost which when added up is a bit more than £50. More like £80-90, so total cost is nearer £200 which changes things somewhat.

[bd139]

Ahh. That would explain it. They’re ok when they work but have a number of flaws. The sagging probably is the capacitor.

The sense connectors are rather useful I find. My PL154 is used as a battery charger so it allows me to set the voltage at the battery and not the power supply terminals which can be quite different if you’re charging at 3-4A due to the I^2R loss.

£100 is about right for a double unit. Occasionally they pop up for less but it’s a rarity. Worth looking out for a full complement of shorting bars across the sense terminals - they seem to be mostly missing.

[WaveyDipole]

Ahh. That would explain it. They’re ok when they work but have a number of flaws. The sagging probably is the capacitor.

Except for the smoothing caps and one other cap next to the terminals which I missed, the unit was re-capped using Panasonic caps and calibrated when I first got it maybe a couple of years ago. I have now replaced both smoothing caps and replaced the cap next to the terminals which I had missed during my re-capping exercise a couple of years ago. I am not sure that this has fixed the random sag/overshoot problem as the voltage is still varying slightly (about 0.2v at 5.00v).

I then turned my attention to the 'fine' pot. This was removed from the chassis and tested on a DMM. There seemed to be one or possibly two drop-out points where the DMM momentarily displayed 'OL' and then back to a resistance reading. I opened it up and had a look inside. The track actually looked OK so it was cleaned with some contact cleaner, re-assembled and tested again, but the problem persisted, so it was replaced. This fixed the fine adjustment which now worked properly and without random voltage jumps. After running for 5 minutes or so, the PSU seemed stable, so I upped the current using a 10Ω 10W resistor which allowed me to test up to 1A at 10V and it appeared to remain stable. I turned it back down to 5v @ 0.5A approx and left it on soak test for a bit and after 20mins or so, the voltage had drifted from 5.00v to 5.09v. The current is also varying between about 430mA and topping out at 500mA. Some variation might perhaps be expected as the resistor warms up, but I expected it to remain stable thereafter.

The sense connectors are rather useful I find. My PL154 is used as a battery charger so it allows me to set the voltage at the battery and not the power supply terminals which can be quite different if you’re charging at 3-4A due to the I^2R loss.

That's a good usage case in point to ensure correct voltage is maintained at the battery terminals. As you say, at low voltages and high currents, the loss in the leads can make a significant difference.

£100 is about right for a double unit. Occasionally they pop up for less but it’s a rarity. Worth looking out for a full complement of shorting bars across the sense terminals - they seem to be mostly missing.

Thanks for that indication. I regularly see single units at similar prices and above. Making sure that all the shorting bars are present is also a very good point. I was fortunate that all were present on both of my units when I purchased them, but definitely something to keep in mind so thanks for the heads up on that. I found a couple of newer double units with LCDs rather then LEDs from the states for around that price. With the postage it would be around £150, perhaps almost worth considering despite the high shipping cost?

The pads on the display circuit board measure 1.32mm x 2.30mm, but the opening in the filter is rather smaller at 1.15mm x 1.94mm, so there is little leeway. If SMD replacements were used, they would need to be tiny and quite precisely placed. I have done some SMD work, but I'm not sure if I could manage on such a tiny scale!

The PSU is definitely more stable than it was, but I am still concerned about the variance so before I consider the display I will investigate that further and perhaps compare with the other unit.

EDIT: I did a comparison and the other unit stayed rock steady at 5.00v/10.00v. No movement whatsoever, so something is still not quite right with this one.

[WaveyDipole]

I cleaned up the PCB - there were some flux residues and a bit of fluff stuck to one corner. I also cleaned the sense shunts. They were tight but did have some corrosion on them. I also checked the calibration - at least for the voltage side. I couldn't do the current side because of the faulty display. It was pretty close and just needed very minor adjustment which I was able to carry out with no problem. I don't know whether it was removing the flux and  dirt, wiggling the pots or cleaning the sense shunts that did it, but I suspect it was the latter. Now after another 30 minute soak test it seems to be stable. All that remains is to sort the display.

[Mark Hennessy]

Are this a brown or black/white or grey unit?  If it’s not the grey one I’d get rid of it and buy a newer one as the older ones are a complete bastard to work on.

No, you can't tell by the colour of the case.

As the extract of the schematic shows, this unit uses a PIC instead of a pair of ADD3701 DVM chip. This means it comes with the later innards.

The change was prompted by the ADD3701 going EOL, and it happened before the cases changed colour.

The way to tell at a glance is to look at the damping switch. On the later models, it's a small square push-button. The earlier "bird's nest" models used a larger rocker switch.

While not pretty, the older ones are actually pretty easy to work on. I have many of them, and have restored them all with no great difficulty. The only remaining issue is that they can drift a bit, which I suspect is the big (and expensive!) WW pot - it doesn't bother me enough to make me want to investigate further. We're talking about a few 10s of mV, so hardly an issue for most applications. Generally though, they are extremely reliable.

The newer ones are actually more problematic in my experience - just as well they're a bit easier to get apart. PCB contamination seems to be a common problem.

Pictures of my brown PL320 with the later innards: https://www.vintage-radio.net/forum/showthread.php?t=50989 - when I posted that, it was the first time I'd met a brown one with the PIC

[m3vuv]

dont know if it helps but with iffy/suspect pots,ive couloured the tracks in with a soft pencil,if the track is iffy it normaly sorts it,only a temp fix tho but can help in troubleshooting.

[WaveyDipole]

No, you can't tell by the colour of the case.

As the extract of the schematic shows, this unit uses a PIC instead of a pair of ADD3701 DVM chip. This means it comes with the later innards.

The change was prompted by the ADD3701 going EOL, and it happened before the cases changed colour.

The way to tell at a glance is to look at the damping switch. On the later models, it's a small square push-button. The earlier "bird's nest" models used a larger rocker switch.

That is a  much neater an more modular layout, and I see that the display as well as the pots are mounted differently. Mine is an older version with point to point wiring and the ADD3701 rather than a PIC. The two examples I have here do have different damping switches, but both are rocker types. I have added a couple of pictures which might help to identify the specific board.

dont know if it helps but with iffy/suspect pots,ive couloured the tracks in with a soft pencil,if the track is iffy it normaly sorts it,only a temp fix tho but can help in troubleshooting.

I had heard about that trick, but thank you for the reminder. I am reminded of an exercise in a vintage electronics learning kit where "resistors" could be drawn on a piece paper with a pencil. The resistance would depend on the length (and possibly thickness) of the line. In the exercise, the "resistor" would be hooked up to an oscillator to produce noises with varying frequencies. You kept the wire on one end and dragged the other one across the line of graphite! I can see how a bit of graphite on the track of a pot would work, at least of a period of time.

[BravoV]

My TTi PL330QMD has similar problem, one of the segment in the 7 segments display is died too. 

The board view

[Mark Hennessy]

No, you can't tell by the colour of the case.

As the extract of the schematic shows, this unit uses a PIC instead of a pair of ADD3701 DVM chip. This means it comes with the later innards.

The change was prompted by the ADD3701 going EOL, and it happened before the cases changed colour.

The way to tell at a glance is to look at the damping switch. On the later models, it's a small square push-button. The earlier "bird's nest" models used a larger rocker switch.

That is a  much neater an more modular layout, and I see that the display as well as the pots are mounted differently. Mine is an older version with point to point wiring and the ADD3701 rather than a PIC. The two examples I have here do have different damping switches, but both are rocker types. I have added a couple of pictures whuch might help to identify the specific board.

Good to clarify that, as your original post pointed very definitely to the Mk2, given the schematic extract from the later models. As I say, the colour of the case is meaningless with these - the damping switch (and position of the earth binding post) is the way to tell them apart from the outside.

In the earlier units, there's no driver chips - the display is driven directly from the ADD3701, which includes LED drivers as part of its design (just needs cathode drivers, which is simply 4 transistors in the PL320 design). Just in case you haven't already seen it, the manual for the earlier models is here:

https://web.archive.org/web/20140408222743/http://k1.dyndns.org/Vintage/Sinclair/Other%20Inventions/Other%20Electronics%20Products/SINCLAIR%20THURLBY%20PL%20SERIES%20POWER%20SUPPLY.pdf

And if anyone needs the later one:

http://www.chriswilson.tv/PL_Series.pdf

And the ADD3701 datasheet:

http://pdf.datasheetcatalog.com/datasheets/560/502474_DS.pdf

No idea if it's possible to find the original display type new today - you might be lucky and find a NOS item, or perhaps someone has a spare from a donor item - they were widely used back in the day, so the odds are good. Thurlby used them in their logic analysers and some of their bench multimeters, for example. But modern 4-digit common-cathode displays are widely available - the challenge being getting the right physical dimensions to fit the opening in the front panel. Wiring them up is the easy part, though might not be pretty - but least that's in keeping with the rest of the instrument

Another thing to check with a "drifty" Mk1 - the presets can get iffy with age. Unfortunately, most of them have far too much range, so the setting of them is very difficult to get right. PR1 is worth checking if the output voltage is genuinely drifting (deciding if it's the PSU or the meter is obviously the first step). If you're feeling keen, it's worth changing resistors either side of the pre-sets to give the pre-sets much less range. The trouble with this design is the pre-set has to include enough range to cover the tolerances of the tracks in the voltage pots, which could easily be +/-20% (though the NOS Colvern WW pot I've got standing by actually measures within 2% of 50k, which is pretty respectable). Even so, there's still far too much range. The Mk2 removes the pots from the feedback network, replacing them with a fixed resistor - this means the pots can be cheap and nasty consumer-grade types, and the "set max" pre-set can have less range, leading to better "setability" and stability.

These earlier units might be a bit ugly internally, but they are generally very reliable once restored, and are well-worth the effort. I've worked on far worse.

Oh, you mentioned units with LCD displays - presumably these were a TS series? The ones I've seen use ICL7106-like display chips, that are 2000-count units. That means that they auto-range to 100mV resolution as they get close to 20V. I can't remember what the current meter does - it either has 1mA resolution, but can't read over 1.999A, or it has 10mA resolution (think it's the former, but would have to dig mine out of storage to check). Only the 4 amp TS units came with an auto-ranging current meter. The PL series maintains the 10mV/1mA resolution across the whole range, which was unheard of at that price point when they were introduced, and is still very respectable today. I think the TS series was a "cost-reduced" version of the PL models - I don't think the dual models came with the series/parallel/tracking modes that the PL QMD have (mine doesn't, at least). Nice enough, if the metering is good enough for your needs - the build is not all that dissimilar to the later PL models, with sensible modularity, etc.

[Renate]

I don't have one of these, but I did have a few thoughts.
My first would be just to use standard single digit displays (best choice of color) and knock off a PCB.
Since thickness is a problem, my more radical solution would be to make a project out of this and use 2.2" 128x32 OLED displays.
It would look great.
Ok, you would need to reverse engineer the PIC and write your own code to drive the OLEDs.

[WaveyDipole]

Good to clarify that, as your original post pointed very definitely to the Mk2, given the schematic extract from the later models.

Thanks Mark for spotting that. I did indeed post the later version schematic in error. I have now corrected that oversight... 

Another thing to check with a "drifty" Mk1 - the presets can get iffy with age. Unfortunately, most of them have far too much range, so the setting of them is very difficult to get right. PR1 is worth checking if the output voltage is genuinely drifting (deciding if it's the PSU or the meter is obviously the first step). If you're feeling keen, it's worth changing resistors either side of the pre-sets to give the pre-sets much less range.

There did seem to be quite a bit of range on one preset in particular which did seem a bit tricky to adjust but it was acheivable - just! It could have done with having better precision. Changing the resistors is and interesting idea and I might look at that at some point.

The Mk2 removes the pots from the feedback network, replacing them with a fixed resistor - this means the pots can be cheap and nasty consumer-grade types, and the "set max" pre-set can have less range, leading to better "setability" and stability.

That sounds like a mod project!

Oh, you mentioned units with LCD displays - presumably these were a TS series? The ones I've seen use ICL7106-like display chips, that are 2000-count units. That means that they auto-range to 100mV resolution as they get close to 20V. I can't remember what the current meter does - it either has 1mA resolution, but can't read over 1.999A, or it has 10mA resolution (think it's the former, but would have to dig mine out of storage to check). Only the 4 amp TS units came with an auto-ranging current meter. The PL series maintains the 10mV/1mA resolution across the whole range, which was unheard of at that price point when they were introduced, and is still very respectable today.

Yes, the one I looked at is indeed a TS range, Thandar branded rather then Ti, and it is a dual 2A rather than 4A unit. It does need a bit of work as well, but nothing major and at least one of these sense jumpers is missing. Now that I have got this one stable, if I can get the display sorted, then I don't need to buy another. I was thinking worst case that maybe one of the ICs (which might now be unobtanium) may have started to break down.

I don't have one of these, but I did have a few thoughts.
My first would be just to use standard single digit displays (best choice of color) and knock off a PCB.
Since thickness is a problem, my more radical solution would be to make a project out of this and use 2.2" 128x32 OLED displays.

True, I did discover that it was possible to get white, green, yellow or blue 7-segment displays in single digit packages, but apparently only red in the 4 digit versions, although I did find a green one on Bitsbox. A number of cheap blue ones seemed to be clock modules and there was one from Adafruit (complete with a driver IC) but it was £13 plus postage and I would have needed two of them. The only thing with single units is that there will be an awful lot of pins to parallel and that is bound to look a bit messy. TI did find the ones in the link below, but the driver chip and pins would need to be removed and soldering the connections might be a bit tricky, but they are cheap:

https://www.ebay.co.uk/itm/TM1637-4-Bits-Digital-LED-Anzeige-Clock-Tube-Display-0-56-For-Arduino-Blue/292793177429

BTW, why do they call them "tube" or sometimes even "nixie" displays? 

[macboy]

If you are serious about replacing the (newer version, PIC microcontroller driven) 7-segment displays with something more modern, I'll point out that I have reverse engineered the firmware. Actually, I reverse engineered the firmware for the TSX1820/TSX3510, but their display board is essentially identical. The firmware is only slightly different, but not so much that it would be difficult to apply my work to. You just need to be a PIC assembly guru    The firmware stores the displayed digits in BCD format in dedicated registers. It is quite easy to implement a bit-banged UART to blast this data to another microcontroller for display on a fancy display (no memory/MIPS available on the original micro to do anything fancy).  I used a TL866 to read the program from the original micro (they didn't set the code protect bit), and used a PIC16F57 as a drop-in replacement device with FLASH memory.

[WaveyDipole]

If you are serious about replacing the (newer version, PIC microcontroller driven) 7-segment displays with something more modern, I'll point out that I have reverse engineered the firmware. Actually, I reverse engineered the firmware for the TSX1820/TSX3510, but their display board is essentially identical. The firmware is only slightly different, but not so much that it would be difficult to apply my work to. You just need to be a PIC assembly guru    The firmware stores the displayed digits in BCD format in dedicated registers. It is quite easy to implement a bit-banged UART to blast this data to another microcontroller for display on a fancy display (no memory/MIPS available on the original micro to do anything fancy).  I used a TL866 to read the program from the original micro (they didn't set the code protect bit), and used a PIC16F57 as a drop-in replacement device with FLASH memory.

I do want to replace the display, but unfortunately, my unit is the older version using the ADD3701 which drives the 7-segment display directly, rather than a PIC. However, thanks for the insight. That would have made for an interesting way forward. Those displays I linked above do accept BCD input and would probably have been easy to implement, but, as you point out, more sophisticated displays (OLED etc) might have been possible.

As it is, I did wonder about using an Arduino (Nano or Micro) to read the 12 signals and output to a BCD display. A Mini Mega board might possibly drive both. The easiest option seemed to be to just remove the driver chip and access the LED display directly. but, unfortunately, according to the datasheet the TM1637 chip drives a common anode display, but the one in the PL320 is apparently common cathode. A pair of 4069's might be one way to deal with that.

With regards to driving display that takes a BCD input, there are plenty of BCD to 7-segment drivers, but, perhaps understandably, there doesn't seem to be anything to convert the other way around.

[WaveyDipole]

Anyone know much about Digimess power supplies?
Seen a good deal on a PM3006-3 0-30V 6A dual PSU.

[bd139]

Run a mile. They are all rebranded Mastech. Also they have a store on eBay where they sell the broken ones. Lots of them.

[WaveyDipole]

Thanks. I did wonder how a "new" HY3003-3 could be less than half price. This one is apparently brand new, never been out of the box, although the seller offers no warranty. The PM3006-3 is described as "used" so perhaps a bit more credible. As the adage goes:  if it sounds to good to be true....

In any case, it has a fixed 5.0V third channel whereas a number of others (e.g. have been also considering the Siglent SPD3303C) provide a number of voltages on CH3 including 3.3V. Some Thurlby PSUs seem to offer a selection of 4.0V, 4.5V, 5.0V, 5.5V and 6.0V. How difficult would it be to modify the 4.0V output to 3.3V? I found a service manual it didn't seem to have the relevant part of the circuit.

[WaveyDipole]

Well it looks like a solution may have presented itself. I found some NOS displays when I typed in the code MMN 59440 into the eBay search box. I had previously tried "Thurlby PL320" and "ESP 2404" but, more recently, after taking a photo of the back for reference to research the pinout details, I noticed another code printed on the PCB which I had previously not spotted. I found two listings by Spanish sellers with NOS displays available so I purchased a couple of units from one of them and am waiting for them to arrive.

I have no affiliation with either seller other than this purchase, but this one still has a few available for purchase:

https://www.ebay.co.uk/itm/MMN59440-LED-DISPLAY-GENERAL-ELECTRIC/152709861960?hash=item238e379248

They are not as cheap as similar far east products, but not unreasonably priced considering they are original NOS. I am posting the link in case it might help someone else who may have a PL320 with a faulty display.

[WaveyDipole]

The display arrived today, only 3 days after dispatch! The good news is that the PL320 is back in business.

The bad news is that after replacing it, I discovered another fault. After reassembling I noticed that the voltage was moving about a bit so I went to tighten the sense bars. I got a tingle when I toughed the left hand (negative) one. I soon discovered that this was because my hand caught the edge of the chassis at the same time. I checked the other unit and didn't get the same problem. I then measured with a DMM and got approximately 160vAC on the negative terminal. One can get that sort of thing with SMPSUs but with linear?

I did some further checking. The Earth to Line resistance checked out OK with the megger. With all secondaries disconnected the problem goes away, so it seems to be down to the TX. Measuring resistance with a DMM between the primary and the secondaries showed some interesting results. Between primary and 30V seconday I got about 8.1MΩ, however between the primary and 10.5-0.10.5 secondary I got only 0.65-0.70MΩ. Checking AC voltage between primary and secondary I got between 95VAC and 106VAC depending on which secondary terminal I checked.

Does this mean that the insulation in the transformer is breaking down? Both PSU's are in the same location and the other does not exhibit this problem so I suspect that the TX is breaking down but would like to confirm? If it is, then it looks like my display repair may unfortunately have been a waste of time and I will need to source another PSU after all!

[bd139]

Could be earth fault as well in your mains or you have a PME installation.  Check that first. Or alternatively haul the supply elsewhere and see if it does it.

Just got a good win on a PL330QMD the other day on eBay actually. £110 delivered. Seller did a terrible job of packing it and the case got bent, line switch and plug destroyed. Sent photos and the guy just refunded the entire amount no questions asked  . Panel bashed it back out and replaced the line switch and put a new plug on and it’s fine.

Anyway got a nice line up now 

[WaveyDipole]

That looks nice and clean! There were a couple which I had on my watch list for a little over a week but both now gone. Just the one damaged one left at £150 and another new ex Mod listing at £174. I held off because I was waiting for the display. I could probably use the bits from the faulty one to repair that £150 one but I think its a bit steep for one with damage. With the other one, for £40-£50 more I could get a new Siglent. I'm looking at a couple of Farnell ones going quite cheap too, but most don't seem to have the sense connectors and only up to 17V not 30V.

Regarding the earth fault, the other unit that I have, which is pretty much identical, does not exhibit this problem when plugged into the same socket.  I don't know whether I have a PME installation but I didn't think there should be any cross-over of voltage between primary and secondary, but I am not sufficiently knowledgeable about transformers.

[bd139]

It was quite clean. The front lamination has come off slightly but doesn’t really affect the operation.

Yeah I’d avoid damaged ones. You just never know what the deal is. Unless it gets damaged in transit which might be good news