How do you deal with you're first 'Beyond Economical Repair' / failure

[darky8]

Hi Guy's,

1st post But also a fist time experience for me.

I have (up until now) prided myself on having a 100% customer satisfaction rate, based mainly on having fixed everything I have been given. Just general electrical stuff, I have a knack (logic) of making things work again. But if I know it's beyond my knowledge, I stop and admit it, and just hand it back to them. I know my limits, a young player as DJ would say. However, I do want to extended my knowledge and I do that via helpful you tube videos and just messing around.

So guy called me up about a Fluke 2813 System Power supply that had the left hand side of it's LCD blank. He couldn't get anything out of it because he couldn't make out the options. Easy..?

A few plastic clips, cleaned contacts and realigned zebra strips later.. a full display

However, after a more detailed general inspection. I noticed this piece of kit had sustained a substation whack to the left had chassis, but not a mark on the case!?

Even closer inspection revealed no output on any of the 3 power modules, even further each input fuses blown. And they blew again with a new fuse, shorted FET. So I thought I would get at least one channel going, but after parts etc it's show no sign of life (And without a isolation transformer, nor would I if I continued).

The power modules, and PSU module were made by Mitra Power Systems on behalf of Phillips/Fluke. There are no schematics for these modules, and no silkscreen on the PCB either! Except for a hand + - on the cap holes and a random resistor!

I then found out the primary FET heat sink was at -150v even though the FET was isolated from it. I got on channel of it no to blow it's fuse, but this thing is shafted. Whatever took all three PSU's out at once has clearly caused a whole load more damage.

So I find myself in unfamiliar territory. I've never really given up on anything before

It's basically a nasty piece of work and I don't wan't to touch it. If it's shorted primary ST BUV45A's, I thought I would replace one and associated caps and 1 reg but there is at least another 12 semi's that could also be screwed and then x 3. Hey it never blew the fuse afterwards... :/

This must be a noob thing, and I probably should have stopped long before I did. My instinct tells me fix it, don't stop until it works. But my brains says, this is someone who's bought a Fluke made by Phillips, but with the main PSU, and all the output PSU's made by someone else (Mitra Power Systems, hence no schematic, with no silkscreen or numbered components, hunting around in the dark with -150v heatsinks... (why would you leave a heatsink on a -150v rail if the FET stuck to it was isolated!!).

I'm sorry this is so long, but for the first time I think I'm justified in handing this thing back to the guy and wishing him luck! I might have to nurse my ego a bit, it hurts as my real first fail to fix. Do you remember going through something similar on a first beyond-economic-repair??

[ebastler]

I would agree with your rational assessment -- call it a day, and hand it back to the owner.

It is easy to get sucked into the "I was getting so close, just a bit more effort and I will get it fixed" trap. Especially in a situation like this, where you had initial success fixing the display, and were already hoping you had completed the job. But also in general, it is never easy to admit defeat, and to realize that you have to write off a couple hours of time which you have already spent.

In this particular case, who knows how much damage the unit has sustained from the whack it apparently received? It is not unlikely that, even if you had schematics, you would end up peeling back several layers of damage, and spend a lot of time on this unit.

I think you can return it to the owner without bad feelings. You took on the job assuming an isolated failure of the display, and (justifiably) expecting a manageable repair. Now it has turned out that the underlying damage is more complex, and potentially more severe. So, nothing wrong with concluding that this repair does not justify the effort, in my book.

[daveyk]

The picture of the FET. There is a second pic of FET after replacing? Why is that Electrolitic cap to it left, look replaced, but in a different direction?


Sent from my iPhone using Tapatalk

[Seekonk]

Good eye.    I sometimes get feeling sorry for people and think if I don't fix it no one else will even look at it.  You have to look at it the way the rest of the world does, it's not my problem.

[ebastler]

The picture of the FET. There is a second pic of FET after replacing? Why is that Electrolitic cap to it left, look replaced, but in a different direction?

Well spotted! But probably not the cause of the lack of repair success, as long as the capacitor has not blown up?

[darky8]

Noticed that in bed at 23:40 so put it right in the morning, didn't help it. It's amazing no silk screen on the boards at all, yet I miss a + - !! lol, oh hang on they did mark the only VR as R'whatever, but other that nothing either side. No schematic either, call it a day. Hey, it doesn't blow the fuse any more! What ever happened to the primary side of all three power boards +155v, GND, -155v could have only come from one source, and that's the main PSU board. All three boards primary side fuses blown and FETS shorted? Must have come from PSU board issue, plus really quite large force to dent 1 or 1.2mm chassis steel. I'm not sure how you dent something like this without marking the case. Even with it off, I had to check bolt pitches on 19" rack mount and it's clearly deformed.

I'm going to stick my neck out here, and say that the deformation is no accident (a tool?). It's not a drop, the case is fine the front plastics were also ok. The fault with the left hand side LCD ties into the left corner impact. The main PSU board is up against that edge with it's connectors just shy of the corner. All power is just daisy chained from psu to module, 1-3.

I think there's a possibility that it's original owners may have attempted a controlled disposal of the unit. Only for it to pop up on Ebay a couple of weeks later good as new. As a lab power supply, it might have run without a case in a rack. Whatever it is, we'll get rid of it ASAP, and on to the next!

[daveyk]

Noticed that in bed at 23:40 so put it right in the morning, didn't help it. It's amazing no silk screen on the boards at all, yet I miss a + - !! lol, oh hang on they did mark the only VR as R'whatever, but other that nothing either side. No schematic either, call it a day.

Before pictures are vitally important sometimes. You did good by taking them.

Dave


Sent from my iPhone using Tapatalk

[StillTrying]

There's a service manual here, not that I think it helps much!
www.eevblog.com/files/philips_pm2811_pm2812_pm2813_sm.pdf

The -150V might be raw bridge-rectified live-neutral mains, the -Ve side of your rectified floating -Ve and +Ve drops down to -330V on every half mains cycle, and averages around -140V with respect to mains earth.

The thing does look horribly complicated to me.

[Electro Detective]

Sometimes a tough dog repair just isn't going to happen for a multitude of reasons, and you have to throw in the towel,
especially when the gear belongs to someone else with positive or negative expectations. 

Best to lose and be out of pocket/wasted time, and keep things sweet and straight up between both parties 

You don't have to like this seeming 'defeat' but that's how it goes sometimes,
and soon forgotten once you move on to the next dog..
that might be tougher   

[CatalinaWOW]

You might want to talk to the customer.  Tell him what you found and discuss with him (her?) what the cost might be to complete the repair.  This gear might have sentimental or other value changing the "beyond economic repair" equation.  The response also depends on what you told the customer when you accepted the task.  There is an enormous difference between "I will fix this for XXX", and "I will do what I can for XXX", and "I will look into it and tell you if repairs are looking spendy."

If your initial contract allows you to do so, and if continued work on this will interfere with other income activity you would be smart to chalk this one up to experience and move on.

[TiN]

You have to be honest with why you start fixing things in first place. If to catch an easy buck, sure, go for BER idea. If you fixing things to learn something, then BER is irrelevant, as you learning during fixing, and that have a price, just like when you paying for tutor to learn new subject.

Saying "it's all scary and unknown now" is not helping. No schematics? Make one, you surely can reverse schematics as you have real piece of HW in front of you, and a DMM with continuity mode to trace connections. No silk screen? Take a photo of the board and draw diagram and all orientations of all the parts before talking stuff apart. Cables, caps, connectors, whatever it takes. Also as free bonus - once you understand what every single part in that PSU does, you likely to know more about power supply design than some of EE graduates

[janoc]

Honestly, if I have discovered that the device I have on my bench has a much more serious problem than what it came in for repair with, I would stop working on it and call the client first. What good would it do to me if I have spent several days on it, lot of money in replacement parts (assuming they are obtainable) and then the client says they won't pay for all that because it is not worth it to them?

If you bring your car in for service, the garage will also not proceed to swap an engine if they discover that it is about to be busted and then present you with an eye watering bill for it. You would be completely right to refuse to pay for something you didn't order. They will get an approval of the work (and the projected repair cost) from you first instead.

If the client gives you a go ahead and approves the budget, by all means, go down the rabbit hole. If not, put it back together and hand it back to them. It is not a matter of professional pride but basic economics.

[darky8]

Thank you all for your valuable input.

I did contact the guy, and explained to him what I had discovered and it's repair would at this time be technically beyond me. So he asked me what I would do if it was mine. I said I would use it a learning tool and restore the thing, so he's donated it!!

So I'm now the owner of a FLUKE/Phillips/Mitra Prower Systems PM2813.

What I need to know, is what could have happened to blow the primary side of each power module? They have a 1.6A glass fuse blow, the big FET shorted and the primary side of the two optocouplers are blown. Everything else seems fine, the secondary side seems to be ok including all switching transistors etc. So it looks likes whatever has happened. has taken place on the primary side only.

So, logically thinking.. the only thing common to all three modules, is the internal power supply board. The voltages out of this are fine, but obviously I haven't been able to test is under any sort of load. The PSU board is also from Mitra, but doesn't really seem to have much going on. It uses an inverter to get the high voltage, the same big FET, but other than that there's nothing really going on.

So does any one have any ideas what could have happened? Power surge may be?

Thank you for all your comments, it's been really helpful.

Rowan

[Housedad]

The other boards the PSU feeds or the leads to them are not shorted in any  way?

[edpalmer42]

With blown inputs like that, my first thought would be misconfigured line voltage.  If it was set for 115V and you plugged it into 220V, it could be rather exciting!

Can you bypass/disconnect the input module and power the output module(s) from another source?  I'd hate to put in a lot of time and effort on the input module only to find later that there was also a fault on the output module.  If it turned out that there was a fault with the output module(s) I would call the unit cursed rather than BER and get it off my bench!

Ed

[darky8]

It was set correctly when it arrived, but there's a distinct possibility someone could have run this with it jumpered for 110v.

I'm in the UK and we are meant to be 230. or is it 240? or maybe bang on 250.5v as my meter says now!

I haven't had the PSU board out yet, but I'm going to. The outputs are +150v, GND, -150v then 24v, GND and 18v, GND and they test fine as that. The fuse (1.6A 250v) that blew on the modules was on the +150v which goes directly to the collector of the FET which shorted

I'm only working on one module, the others are disconnected. The other two have suffered the same fate (Fuse blown violently, and FET shorted), so they can sit there for now. I just don't want to fix this module and have it get blown again, because I would throw it out the window then.

Here's a couple of pics that may help (The voltage label is from the module not the PSU), but I'm going to take it out tomorrow and have a closer look.
 

[CatalinaWOW]

Don't know about your location, but a lightning induced power surge would be one thing I would think of.  Depending on what the owner of the instrument does there are other sources of large surges.  Welder?  Operator of large induction motors?

[edpalmer42]

Are you familiar with the light bulb trick for servicing switching power supplies?  Put an ordinary incandescent (or halogen) bulb in series with the supply.  Use one rated for the supply voltage - 240V in your case.  With no load, a working supply will draw very little current.  The bulb will stay cool and the resistance will be low.  If there's a short in the power supply, the current will rise, the bulb will heat up, the resistance rises, and the current is limited.  Hopefully, nothing else will smoke.  Start with a small bulb, 10, 20, 40 watts - whatever you've got.

Ed

[wasyoungonce]

The picture of the FET. There is a second pic of FET after replacing? Why is that Electrolitic cap to it left, look replaced, but in a different direction?...

Damn...you're good!  Well spotted!

[CJay]

OK, a few things, first, BUV45 is not a FET, it's a bipolar transistor (pedantic but important)

Heatsinks, they don't have to be at ground potential, it's nice if they are and much safer, but it's not that uncommon for them to be floating or at some 'ouchy' voltage so don't assume, always check and take precautions.

Make *certain* you know what the device is referencing as 0V, it may not be the same as ground and would explain your 150V reading on heatsinks.

The 'service' manual isn't quite useless, it lists voltages so disconnect the power modules and verify those are correct first, coincidentally there are 155V rails, so perhaps you've been measuring voltages with respect to a -155V rail?

It's rather nicely modular as I'm sure you've realised so you can easily disconnect all but one power module and work on that one but first things first, run your multimeter over *all* the discrete semiconductors on whichever module you're going to work on and see if the diode check/resistance readings make sense for the type of component (bearing in mind the not a FET comment in the first line, check part numbers to work out what parts are)

[Seekonk]

They are a nice supply, but a terrible display to read.  I have three of them, two still with problems.  I purchased them one at a time for about $25 shipped.  They are already in the realm  of not worth fixing. There is a little solder bridge over to change the addresses of the boards. Wish you could inspire me to fix the other two. Having never even used the working one that is not likely. If programmed in series they are a useful supply for the cost.

[darky8]

Hi Folks!

Well I know I was only working on one board at a time, but I just wanted to see if I replaced the input transistor and fuse on the other boards what would happen.

And guess what, that fixed them!!

So I might have fallen in the trap of making things a little more complicated than they actually are.

So hopefully the other channel will work once put back together.

Thanks for all your help

[darky8]

With blown inputs like that, my first thought would be misconfigured line voltage.

This is exactly what I think happened. You can also connect a spade connector incorrectly as well if you tried. The 220v jumper spade has no track behind it so it just stops that jumper flapping about.

Looking at the other channels, it seems it only damaged the main transistor, and blew the 1.6A 250v fuse. So I'm up and running. I won't stopping here though, I'm going to learn as much as I can from this unit.

[Lord of nothing]

sorry when i dont read everything here.

Some other point with self repair or repair for a customer is you get in trouble when something happen very bad.
For example my Dad and I have an small IT Company. On some products like Displays often small parts like a blown capassador cause an problem. So solder an new one in would not be any problem but... if something happen like the Display get burn up or cause some electrical damage you might know who get sued for? Right your Company. That would mean we have to prove that was not your fault. 
So that's why quite nobody try to repaid electronic stuff to avoid an trial.

[madires]

Actually the customer would have to prove that you've done something wrong. Don't you have an insurance for such cases? You can get reasonable and affordable "packages" for small companies.