Power supply 10-20V is stuck at 31.6V

[TehDucky]

Hello everyone, I'm newly registered, mostly lurker.

I have a problem with linear power supply, called Trione TR110, that is 10-20V and up to 10A, 6A constant but is now stuck at 31.6V even if I change the pots.

Power supply is used for ski service machine and is basicly connected to a motor with a grinding wheel. Since new original one is 446 EUR, I would like to fix this one.

From what I have figured out, is my machine still works at 31.6V, but spins ~60% faster and the power supply handles the current. This makes me believe that it is only problem with voltage regulation.

Capacitors and power resistors look fine. On main board there is L7805ABV, C33740  and a BD437. On big heatsinks outside are 3x 2N3055 which I believe should be fine.

Does anyone know what should I check first?

Thanks for all the help!

[madires]

When the output voltage is stuck at the maximum usually a pass transistor has a short. The first step is to check the 2N3055s for shorts.

[TehDucky]

Hello, thank you for suggestion.

I've desoldered all 3 2n3055 and they all appear fine.

[madires]

Have you checked the pn junctions B-E and B-C (diode check, around 0.6V drop)? And C-E (open)?

[Gyro]

If the pass transistors are fully intact then the next candidate would be the output driver transistor, which I assume is the BD437. You might be able to probe it in-circuit for direct shorts but not fully. It is only rated at 45V so it's possible that it might have suffered from an inductive kick from the motor, or just thermal fatigue.

As there is so little on the board, then you only have a couple more actives left to check at that point.

[tunk]

If you cannot fix it, I'd guess any 30V/10A bench PSU should work.
Or even cheaper, e.g. a 18V/20A PSU module from aliexpress/ebay
(20A to have headroom, aka. inflated specs). 

There are some trimpots - did you check them as well? Their wipers
could have bad contacts.

[TehDucky]

Have you checked the pn junctions B-E and B-C (diode check, around 0.6V drop)? And C-E (open)?

Hello, I've checked and all 3 had ~0.57V between B-C and B-E. Edit: and open C-E

If the pass transistors are fully intact then the next candidate would be the output driver transistor, which I assume is the BD437. You might be able to probe it in-circuit for direct shorts but not fully. It is only rated at 45V so it's possible that it might have suffered from an inductive kick from the motor, or just thermal fatigue.

As there is so little on the board, then you only have a couple more actives left to check at that point.

Will check it first thing tommorow. Will try to see if I can get something valuable from checking it on PCB. It is quite old, I believe 11 years now, so it should be thermal fatigue.

If you cannot fix it, I'd guess any 30V/10A bench PSU should work.
Or even cheaper, e.g. a 18V/20A PSU module from aliexpress/ebay
(20A to have headroom, aka. inflated specs). 

There are some trimpots - did you check them as well? Their wipers
could have bad contacts.

I have 60V/5A Aliex power supply that will have to do for now, it goes to around 1.5A at 20V in idle and when I put machine on ski it goes to 4.99A so it maxes, so will have to be careful with overheating it, I have quite a few skis waiting for tommorow

I have trouble reaching the tripots, will check rest first.

I called a few service places and they have too much backlog of things to finish (300 pieces of equipment to fix) and I would get it fixed after new year. The original costs 450 eur, and would send it to Italy for service, which would take 2-3 weeks too, and I'm in middle of winter season.  It is very heavy and have big heatsinks that is why I want to get it working, and it probably costs a few eur of material to fix since it is technicaly still working.

Thanks for all the help so far!

[DavidAlfa]

At least make it easy, do you expect us to guess how it's made?
You'd better remove that board and post some detailed pictures!

[p.larner]

did you remove the nut from the bd transistor fixing?

[TehDucky]

Sorry, managed to get some more pictures of pcb. The potentionmeters on pcb where display is are digital if I recognise this correctly, I measured 0 ohm on them.

I got BD437 and voltage regulator out of pcb and they are both ok. 0.6V on bd437 and 5V on output of voltage regulator.

Thanks for all the help so far.

[madires]

That's a classic LM723 based design. Since the pass and driver transistors are fine the next step is to check the LM723 and its surrounding circuitry. There's also an LM393 (dual comparator), One comparator seems to be tied into the LM723's frequency compensation.

[coromonadalix]

yes a basic  psu design, some fancy mcu for the voltage and current display

a classical lm723 design


lm723 may have died,   1st driver transistor  or the main power transistors

the 723 vcc limits are 40vdc,  it the rectified voltage had some swing  it may have busted it

check  in the voltage pot resistors,  you normally have a low value in the circuit who could have toasted ...  if toasted  the 723 may give a max value output

[Kleinstein]

If the power stages are OK, a point worth checking is the feedback signal. So the cable link from the front towards the regulator, especially if this goes via the flat band cable.

[DavidAlfa]

Some soldering joints in the LM723 are a bit suspicious, like surrounded by solder but not really soldered.
In general I don't like the soldering job in that board, I'd resolder everything!

Now I noticed a darkened zone under that diode... Or is it just a shadow?

[Haenk]

Do you need to change the voltage (aka rpm) or is a fixed voltage good enough?
If so, I'd suggest to dump the old PSU and get a new (switching mode) PSU @ 13,8V, like
https://www.ebay.de/itm/401718082434
They are usually tuneable within a certain voltage range.
Note this is just a sample, there are many offerings and different designs - just to get an idea.

Those PSUs are powerful, and cheap. There are even less fancy designs, and probably cheaper ones, when directly sourcing from China.
IMHO there is no reason to invest time and money in a vintage linear PSU, when the costs of labour certainly are higher than a new PSU.

[coromonadalix]

but where's the fun to repairs a simple circuit like the op psu,   learning curve ...


the best would be to do schematic, and do or add improvements  if needed

an smps psu is not always the best,  if you have big inductive / capacitive loads   etc ... some of them wont be happy

[TehDucky]

Thanks for all the help so far.

I've had loads of ski service these days and didn't have much time to answer and get going with the repair. I've bought this PSU https://www.aliexpress.com/item/1005003151034729.html?spm=a2g0o.order_detail.order_detail_item.3.6b7df19cIpe6Jf 30v-10A.

It worked perfectly for 3 days, I had to buy it asap, as it was going nowhere with my repair. Now while I was working it just made a loud bang, I checked that fuse is broken. So I changed it with same 5A fuse, and as soon as I turned on the PSU it stopped working again and fuse was broken again. Appears something got damaged.

So it looks like that linear PSU is the way to go. I was looking maybe buying Meanwell 24V 13.5A PSU as it goes down to 20V with adjustment, but I am scared since it is not linear, it too might have trouble soon. And on new PSU that died I saw that I was at 9-10A most of the time (for 10-15 seconds at time).

Regarding the picture near diodes that is black, its just shadow. It looks ok in person.

As I said, I was working on old linear PSU when it just stuck on 31.6V and it worked normally, except that the motor was spinning significantly faster.

Does anyone know what I should next?
I asked a guy that I know that services electronics, if he has some time to fix it, but haven't received answer yet.

If checked potentiometers, and they looked ok, the resistance was changing linearly on both V and A side.

[coromonadalix]

i would avoid any cheap priced smps  from ali or whatever place  you can buy them ...


not starting a war here,  linear has good and bad things,  smps too etc  ...  the choice is yours  but you have to know


just take some time to do the schematics  ...    check some designs on the web ... you will understand ...

but if you are time limited,   just find another good reputation linear / variable smps  brand and have fun ...

find ones with available schematics

[madires]

Make sure that the DC motor has a flyback diode to protect the PSU (some PSUs have one at the ouput). Meanwell is a good brand. Regarding the repair, download the LM723's datasheet and check for any strange things. Or try your luck and simply replace the LM723.

[p.larner]

i am a radio ham,switching power supplys are all but wiping out hf bands!.

[pienari]

Have you found the broken part?.

[Jwillis]

Check the voltage pot. Theirs nothing special about them because they have four pins. Some have a dual wiper and others have one dummy pin.If a wiper fails the supply can latch up to maximum regulated voltage.
Typically if the power transistors fails the voltage would go to the maximum collector voltage.
If the regulator fails it will either go to maximum input voltage of the regulator or to zero voltage at the output.The  later case being the most common.

[fmashockie]

I'd be curious to know what Vin is to the LM723.  Measure pins 7 and 12 on the LM723.  Linear regulators like this don't like a small difference between input and output voltage (in this case <3V).  If the input voltage is less than ~34-35V, I would definitely suspect that LM723.  Also check Vref of the LM723 (pins 6 and 7).  Should be between 6.8-7.5V.  The reference voltage is directly proportional to the output voltage.  If the IC's internal reference is shot, it will throw the output voltage out of wack.  You've already checked the series pass transistors.  Aside from the voltage adjust pot being funky, you've really narrowed down a lot of the potential issues for a fault like this. 

[xavier60]

I suspect it's an LDO type design which means that the LM723 can be powered by much less than the unregulated input voltage.

[fmashockie]

I suspect it's an LDO type design which means that the LM723 can be powered by much less than the unregulated input voltage.

I don't believe the LM723 is an LDO.  It has separate Vcc and Vin pins.  So it can be powered at less than the input voltage.  But it still needs headroom between Vin and Vout.