Repair of an old high power switching lab power supply (Horizon Titan DTPS6012)

[toli]

Hi All,

Finally got my new work bench ready, so I can jump right to the action. I've recently picked up a cheap (and faulty) adjustable power supply. I don't know its exact age, but without a doubt over 20 years old. It a fairly capable unit with a dual 60V 12A output. With such a high power output, you've guessed right, its a switching power supply. Seeing as I have limited experience with similar supplies, figured I'd share here on the forum and ask for the help/guidance of people more experienced with similar circuits.

The problem I'm seeing with this unit, is that right after power up the voltage on channel A start slowly rising (as in 5-10V per second), when it reaches about 10% over the maximum rated voltage of 60V, the unit enters protection mode and shuts off the main part (front panel display is still active as its powered independently). I tried loading it to see what happens then - when loaded with >0.1A it behaves well and the voltage/current limits can be set and are within ~2% of the value read by my DMM. The second channel seems to behave just fine, but its also turned off when the first enters protection mode.

I've had a quick look inside (see pictures attached below) to see if there is something obvious (perhaps minimum load resistors that has gone wrong). I see a couple of 120R power resistors which are connected to the outputs, but they both measure as 120R so they aren't to blame. Other than that I haven't really measured anything just yet. Visual inspection didn't reveal any obvious suspects yet (unless you see something in the pictures which I've missed ).
While at it I took a few pictures of the working boards too, just as a mini tear-down in case its of any interest to anyone.

There are a couple of (identical, one per channel) boards on the side I haven't taken out yet (one annoying screw that won't budge, I will get it out eventually ). I have taken pictures of these to the best of my ability in their current position. These boards seem to be the heart of the regulator, with a couple of OP400FY's, some 7915/7805/IRF640/BUK436 power MOS and some other devices I can't read the label off.

Let me know if you see anything suspicious or you have some insight as to what I should check first.
BTW, I was unable to find any info on this model on the WWW, but seeing as this is a local Israeli manufacturer of power supply's, I might try shooting them an email just in case - perhaps they could share some info about this unit. I've had used linear supply's from Horizon in the past, like the DHR40-1 which was the first bench power supply in the past. While not as sexy as some other gear, I was always happy with it - it seems like a reliable tool which I can use safely. so I really do hope I'll be able to get it running without keeping a constant load on that faulty channel 
Otherwise I'll probably start probing a few points to try and understand where the problem is coming from. "Lucky" for me, the problematic channel has a "grey zone" where for currents of ~0.1A it won't rise indefinitely, but won't go below ~2.5V. This could be used to compare voltages at selected nodes between the working and the problematic channel without a "time limit" of going into protection mode.

[coromonadalix]

woah  thats a cramped space psu loll     maybe some scr phase pwm  controlled psu, it must be it to save space vs power / wattage

Like you did not find anything on it : https://www.eagle.co.za/display_product_816.htm

Try to sent them an email  :  under an non disclosure agreement  maybe they will send you a schematic ??

[toli]

Horizon were quite quick to reply. Unfortunately they say the don't have a schematic for this old supply so they cannot share it with me.
They've offered I send it for a quotation on a repair.

I obviously prefer to try and fix it myself, so for the time being looks like I'll have to try and figure this one out without the schematic 

[coromonadalix]

I do think its kinda at least scr based  for the given power it has ...  thats why you seem to have many coils in it for the power sections

You could check some sorensen dcr b series schematics to have some ideas of the working principles

Or this attached photo

[patteck]

Most switched mode power supplies requires a minimum output current  to prevent the output inductance from dry running. IE buck topology.  Now if the output capacitor looses its capacitance due to age, the output will rise and  the unit will eventually enter into protection mode.(loss of capacitance causes instability in the loop and oscillations on the output that the loop cannot compensate)   If the active switch (Fet, transistor, SCR etc) becomes leaky, and is providing more currrent  than the required minimum value needed to prevent dry running of the output inductance, the output of the supply will rise.
I would scope probe the output of the supply and see if the voltage is at DC or if it is oscillating.
 

[toli]

Thank you both for the comments, this is very helpful.
Since I have limited experience with similar power supplies, this is an opportunity to gain some more experience.
I will take a look and report back when I have something to share. This might take a few days as I typically only have time to work on my hobby stuff over the weekend.
BTW, from the quick look I've had over the last weekend, the active device seems to be the BUK436-100A I've mentioned in one of the previous posts.

 

[PKTKS]

Due to your "limited" experience I think is worthy to mention 2 cents of caution

- THIS DEVICE IS DANGEROUS
- It uses high power density crafted in a minimum space
- IT DOES HAVE HIGH VOLTAGE SURGES INSIDE
- IT DOES HAVE HIGH AMOUNTS OF STORAGE ENERGY (CAPs)

... that said..  you may find prudent
- first try to have some clue about the SCHEMATIC..
- if not possible - dismantle the parts in blocks and figure out topology
- INSPECT WITH GREAT CARE ALL SOLDER JOINTS - you may have cracks
- you may also have visible cold joints of leaking caps
- INSPECT RELIABILITY ALL TERMINALS due to high power amperage and vibration
  that includes HIGH POWER BRIDGE, HIGH POWER TRANSFORMER and CAPs
  and the current sense resistors.

After extensive looking you will be able to devise a proper layout to start
some sort of testing with the device properly open

e.g A bench test GIG for high power test

Proper to say that this requires some road mileage...

The device is not entry level stuff

Paul

[toli]

Hi Paul,

Thanks, I'm aware of this. I've built and repaired more than a few instruments in the past, including some HV tube amplifiers. I'm aware of the danger, and work in an organized and careful manner.
What I meant by limited experience has to do with practical experience with switching regulators, not circuit work in general.
Having the actually schematic would have been great, but as I've mentioned earlier, this isn't possible - the manufacturer replied to my email saying the schematic for this old design isn't available.

[PKTKS]

This device is very challenging. 

To turn ON the device  on the bench you will need to
safely position parts to allow some tests.

And by the amount of power you will need to test a
proper LOAD should be arranged.

Judging by the pictures itself -  I have some SIMILAR CAPACITORS
on my shelf dating early 80s  (those yellow rounded...)

If this is the case ...
early 80s is a **LOT**  of thermal and aging stress..

RECAP in this case should be considered even  before tests...

Paul

[schmitt trigger]

The important thing here, is that it appears to be several floating supplies, whose common rail may not be connected to the main common rail.

So it may be a challenge deciding, for instance, where to attach your scope probe's grounding clip.

[PKTKS]

The important thing here, is that it appears to be several floating supplies, whose common rail may not be connected to the main common rail.

So it may be a challenge deciding, for instance, where to attach your scope probe's grounding clip.

That is where FAB support with schematics counts  a lot

I will not put my instruments on such device without the basic view of that .

When in doubt my last resource is to use the "DONUT"
probe  (e.g.  one wire turn wrapped to check what sort of signal we have...) 

Paul

[toli]

After a long delay (due to very limited free time, and a few other projects that needed attention), I have finally had an opportunity to take a look at the unit tonight.
I took apart the other side of the case, where the regulator boards resides. Working without a schematic is a PITA obviously, but some things are a must check after visual inspection. For instance, all the local regulators are working (+15,-15,+5). Additionally, it seems like there are 2 switching devices, IRF640 and BUK436. The IRF640 seems to be connected with a .51R ohm to ground. The larger of the two (BUK436-100A) has a 0.47R resistor at its source going to a negative ~3.3V. Checking the working channel, it has ~0.35V over this resistor. I expect this is the part of the circuit in-charge of keeping a minimum load current of ~700mA (under similar conditions the faulty channel behaves just fine - I've tested this before I  took it a part with a dummy load). On the faulty channel, this resistor has no current flowing through it, despite the resistor itself being the correct value. Next step would be to check if the fault is the transistor itself, or the control to its gate. This is already good news as I now have a fairly good estimate of the location of the fault, and its over on the low voltage side of the unit, so much easier and safer to debug.

Too late to keep working on this now, but I will keep posting the next time find a little time to work on this unit.

[toli]

Found a bit more time this evening to keep pushing forward with this. All I had to do is find the reason the current source isn't conducting.
The transistor measures fine, no short of G to S, yet its gate is at the negative rail. This means the problem is somewhere in the control circuit. Since MOSFET's vary quite a bit in parameters, its clear they will have it inside a feedback loop to keep the current at the desired value. I've followed the trace to the opamp controlling it, and checked its positive input terminal, this one is already too low. This voltage is set by a simple voltage divider of 2 resistors + a capacitor to filter it. All 3 of these passives seem to be ok, so the only suspect left is the opamp itself. These opamps (OP400FY) are placed in a socket, so I simply swapped the opamps between channel as an easy and quick way of checking it. Indeed, problem switched sides. Swapped them back to the original place, problem moved back with it. So the problem seems like the opamp indeed, and should be easy enough to fix,

I'll add it to my next Mouser order, so hopefully, I'll have it here and the unit will be running in a couple of weeks.

[schmitt trigger]

PMI Opamps....WOW!
It has been a while since I saw one.

And they are premium devices too. The IC package is frit-sealed.

[toli]

Yep, priced accordingly too 

[coromonadalix]

could you put an lm324 for the tests

an OP4177  could do the job  around 12 $  at digikey ??  outch

oh saw the price for the OP400FY  near 20$      re- outch

[TerraHertz]

Why are is there a DIP switch block on a power supply board?
Also, there are 9 switches. I didn't know they came in any sizes other than 4 and 8.

[toli]

I didn't check all of them, but judging by the 2 I did look at, seems like this board (like others probably) was shared between multiple designs/models (perhaps with different voltage/current rating). Some of these switches are used to bypass certain resistors on the board, probably to tune the full-scale current/voltage, and perhaps other parameters. None  of them were part of the traces I followed to the gate of the MOS. However, I did see that 3 out of the 9 are set differently between the two boards, so there is probably something more going on with these other than just V/I setting.

[Shock]

Why are is there a DIP switch block on a power supply board?
Also, there are 9 switches. I didn't know they came in any sizes other than 4 and 8.

Analog parity bit.

[toli]

Opamp will be here in a day or two, and hopefully all will work properly then.

One more little thing that must be purchased is another know as one of the 4 knobs is missing. After looking around a little, seems like this collet type knob is not too common (463-8457/020-3520 + cap + nut cover). Only problem is that with the local branch of RS here in Israel price is significantly higher than the RS in UK (which doesn't ship to Israel). So if I'll order these parts from them I'll end paying more for the knob than for the entire instrument 

Now I have to look for another replacement which probably won't look as nice, perhaps one without a collet, but regular screw instead.

[toli]

New opamp finally got here, with an AD logo obviously, no more PMI parts to find

Both channels are now happily idling at 0 load without voltage ramping up. I am still on the lookout for a replacement knob. I will either order a few extra (despite the price), or simply pick one up next time I'm at the local junk-yard. They should have plenty of faulty 80W supply's from Horizon. IIRC they use the same type of knob too.

So far this power-supply (and replacement part) set me back 60$, and roughly 2 hours worth of work. I think it will pay for this easily over time

[lordvader88]

did anyone mention that those funny caps with the white ceramic cases? They are from the 60' at least, and they go bad also.