Mastech HY5020E 50v 20a Bench Supply: Economical to repair?

[richfiles]

So, I was given a blown 50 volt 20 amp single output bench power supply. It's a "Mastech" HY5020E.

opening it up, I can see that multiple parts have violently blown. The main board appears to be single sided, and I've not delved further into it to see how bad or how far charing reaches. The fuse on the rear is NOT blown, and I don't know if it was replaced or not after the initial magic smoke release. I have not attempted to power it up yet either.

If I can find service information, my question is, is this economical to repair? Obviously, if it's just a selection of transistors, diodes, resistors and capacitors blown, then maybe cleaning it up, rebuilding any vaporized traces, and replacing the bad parts might be worth while. 50 volts at 20 Amps isn't anything to scoff at, especially for a hobbyist on a budget. It's far overkill for anything I can think of doing, but if I could get it up and running again, it'd definitely be a nice supply... Also, the excuse of anything I can think of doing does not eliminate further proofing my shop.

If this level of damage is going to take a large investment though, I'd be better off putting it toward a dual tracking supply, which I'd find far more useful. I have a pair of Tektronix dual tracking supplies in a 500 series enclosure, but those are current limited to 400 mA, and I've already been held back by that no less than 4 times just this month! I need more power! 

So, seeing the multiple violently blown parts... Salvage parts or repair? What's more economical?

I must also say I was surprised at how small the toroidal transformer was, and theres'a a rectangular one as well. Am I possibly looking at a switching supply, vs a linear? Again, I've literally only popped the 4 cover screws, and not gone any further yet, so this is all pretty preliminary.

[SeanB]

Switch mode power supplt, and with the whole input side and power devices very well crisped. Repairable so long as you replace everything, resistors, transistors, diodes and capacitors, and get the charred board clean. Especially those Samwha input capacitors, looks like they dumped themselves on the board and ripple did the rest. Probably best to replace with better ones, and the output side as well.

[richfiles]

Replacing capacitors? No such thing as buying cheapie crapola caps!
If I buy a replacement, I buy the rated value, and then limit search terms by manufacturer and highest available temp rating! 

Yeah, I figured it'd be a full rebuild. I'm curious if anyone's done a rebuild like this before, can mayby throw me a guestimate at what I'd be in for.

Fortunately, Mastech's website specifically says that if you want a schematic, just email them and they'll send you a PDF of your model. Nice! More manufacturers need to do this!

[richfiles]

Well, got the schematic. We'll see if I can figure it out then.

[Shock]

I'd say max about $15-$30 in components and about 3 hours work for a proper job removing the pcb, components, scrubbing and rinsing, testing for shorts and other components, identifying the parts and ordering them, soldering them in, testing and assembly. Any consumables you don't have on top of that. Even the dirtiest unethical repair with components on hand would be about 30mins.

Of course if you don't know what your doing it could take longer and get you electrocuted.

[edpalmer42]

When it comes time to power it up for the first time, I'd strongly recommend that you use the series light bulb trick to limit the current and hopefully prevent more smoke release!

Ed

[MrSlack]

Pah wire and fire!

More seriously, determining the actual cause rather than replacing the components and hoping for the best is something that should be done. The cause might still be hiding at which point you're going to light another fire.

[richfiles]

You know...I really ought to build a DBT (Dim Bulb Tester). I like working with vintage calculators, so it'd be nice to current limit those when doing power on tests. I have a variac, but that just varies voltage. Shorts or over currents depend on the fuse blowing. Good advice! 

Last year I rebuilt a power feed for an old Bridgeport mill. The outlet is a 4 prong split single phase 240 volt outlet in a H1--N--H2 + Earth configuration. The building it had been installed at had the outlet wired wrong, and instead of fixing the outlet, they "fixed" the cord providing the 240 volt wiring. When my boss bought it, he discovered catastrophically that the power feed used 120 volts and pulled that from what SHOULD have been one of the hots and the neutral. Instead, it got hit by both hots (ouch). 

I did succeed in cleaning the charing off that board, and rebuilding it. given the DOUBLE over voltage, I chose a very thorough rebuild. All semiconductors, all caps, and any obviously bad components. That job cost about $45 in parts, mainly cause of really spendy bolt ended diodes and drivers. No TO packages there! I took no chances though, and it was a rather simple op-amp based circuit. I could afford to spam! Glorious SPAM! It wasn't my money!

It looks daunting, but I have a feeling that once I remove the PC board and unscrew the heatsinks, I'll see that it's probably manageable. I actually just snagged a nice 30 volt 2 amp dual tracing supply with an auxiliary 5 volt supply. It's more inline with the type of work I tend to do. Still, It'd be nice to fix this behemoth (to me anyway) of a power source!

[grumpydoc]

You know...I really ought to build a DBT (Dim Bulb Tester). I like working with vintage calculators, so it'd be nice to current limit those when doing power on tests. I have a variac, but that just varies voltage. Shorts or over currents depend on the fuse blowing. Good advice! 

I suspect a lot of designs would run off a 60V current limited DC supply, so if you have a 2-channel 30V supply you could use that to power the thing. Much safer than a "dim bulb" tester, if it works

I have one of a similar design in the repair pile, if I ever truely get to the end of the PM33[89]X repairs I have on the bench I'll go back to the PSU and try running it off bench supply to see how well the idea works (if at all).

[CJay]

When it comes time to power it up for the first time, I'd strongly recommend that you use the series light bulb trick to limit the current and hopefully prevent more smoke release!

Ed

With the dire warning that if it's got auto input voltage sensing then using it on 220V with a series light bulb could just trigger the rapid, violent release of more smoke.

If however it's got a nice switch to select between 120/220 on the back then yes, use a light bulb or, even better, a variac.

[richfiles]

With the dire warning that if it's got auto input voltage sensing then using it on 220V with a series light bulb could just trigger the rapid, violent release of more smoke.

If however it's got a nice switch to select between 120/220 on the back then yes, use a light bulb or, even better, a variac.

The real trick comes in using both. A variac will limit voltage, but doe not limit current any more than just popping the fuse under and overload. The DBT in front of a variac will auto current limit, while still giving you voltage control. Also, I presume the suggestion of using the current limited bench meter to power it, refers to attaching it passed the rectifier... Never really thought of doing that before.

My old bench supply is a Tektronix 500 series power supply module installed next to a DMM and a function gen module. It's limited to 400 mA. I'll have to learn to think entirely differently once that dual 30v/2A bench supply arrives! Not used to having anything good on my bench!

[CJay]

With the dire warning that if it's got auto input voltage sensing then using it on 220V with a series light bulb could just trigger the rapid, violent release of more smoke.

If however it's got a nice switch to select between 120/220 on the back then yes, use a light bulb or, even better, a variac.

The real trick comes in using both. A variac will limit voltage, but doe not limit current any more than just popping the fuse under and overload. The DBT in front of a variac will auto current limit, while still giving you voltage control. Also, I presume the suggestion of using the current limited bench meter to power it, refers to attaching it passed the rectifier... Never really thought of doing that before.

My old bench supply is a Tektronix 500 series power supply module installed next to a DMM and a function gen module. It's limited to 400 mA. I'll have to learn to think entirely differently once that dual 30v/2A bench supply arrives! Not used to having anything good on my bench!

The reason for mentioning a variac is that you can wind up the supply voltage and monitor the current as the supply starts up, but of course you're right, a variac won't limit the current in any sort of sensible way so yes, using both would be useful if you had no way to monitor current.

Doesn't matter if you bypass the rectifier or not, you'll just have to suffer the forward voltage drop of the diodes but at a hundred volts it's unlikely to be significant enough to affect the operation of the supply and will work (lots of switchers will happily run on DC applied to the input as long as the voltage is high enough).

[station240]

Oh good, it has transformer isolation between the destroyed mosfets and the driver ICs.
Looks fairly straightfoward to repair, new caps, mosfets and a few resistors.
I'd also replace C003 with a proper X class capacitor, given it's directly across the mains.

[richfiles]

All these part upgrades and repairs remind me of what it takes to get your average harbor Freight <Insert X Here> to be remotely usable! 

To the credit of this PSU though, it did do the job, for a time. I think it died when an entire bank of DC brush motors on a test rack got reversed while attached to it! Brutal!

[albert22]

Take care with the provided schematics. I spotted a small error and may be there are others.
The positive rail at R004 should connect with R207B and R105 (a dot is missing at the crossing of two lines)
Also I may be wrong but it seems that the switching supply is not working as a tracking pre-regulator so Q401 can get very hot and even fall outside its SOA. It is rated at 200W at 25degrees. And probably is not a MOSFET suited for linear operation.

Good luck with your repair.

[bitseeker]

When you've got the supply torn apart and clean, it'd be great to snap some high-res photos of the PCB for reference and fixing the schematics.

[richfiles]

I can do that. You should see the photos I've taken of my PC case build and my custom mechanical keyboard! 
When I start, I'll keep this thread updated.

[grumpydoc]

Also I may be wrong but it seems that the switching supply is not working as a tracking pre-regulator so Q401 can get very hot and even fall outside its SOA. It is rated at 200W at 25degrees. And probably is not a MOSFET suited for linear operation.

I have wondered about that in this design of PSU as well - but surely the thing would just blow up at high current/low voltage output settings?

[Shock]

With the dire warning that if it's got auto input voltage sensing then using it on 220V with a series light bulb could just trigger the rapid, violent release of more smoke.

I've not read much about this occurring, have you got any instances or references where this has happened and how a good SMPS failed from it? I can only seem to find it mentioned on the repairfaq.org site. Would the same occur using a variac?

[richfiles]

I'm 120v AC here (US), and that's what my input setting is set to. I do keep seeing a few references to 220v/240v here, so I thought I'd just mention that.
Man... I've never built a switching supply. Only ever worked with Linear ones. This will be my first switching supply repair, so be gentle! 

The only thing I know for certain about switching power supplies, is how not to build one and throw a fake CE logo on them! Lookin' at you China... 
 I know the basics of operation, the basics of how they work, etc. I just have not ever actually serviced one yet. I don't know what's typical of the controllers, and tend to follow schematics, instructions or guides to get through unknowns. I know well enough about upgrading to good high temp caps like Nichicons or Panasonics, and I can match a transistor or diode just fine. I used to do a lot of repairs as a profession, and got really good at tracking down parts lists and generating BoMs for them. These days, fewer professional repairs, but I still do this on occasion. More importantly, I will CERTAINLY take the best advice offered on more detail oriented suggestions, etc. I'll take into consideration the mention of C003 needing to be an X class cap, and I'll gladly take suggestions on Q401 This is the kind of gold I know I can find here, from people far more experienced than I am at this stuff. The whys behind everything... That's awesome too! Share why something should be different, why a specific change would be an upgrade, etc. I'll be listening. I wanna get this right. just know, it's all appreciated! 

[richfiles]

Just a dump of pictures. If you need any details enhanced, let me know.

Thing was covered in flux like some cheap Chinese junk yet had Nippon Chemicon caps. There were actually a few solder balls stuck down in flux. Really should have been cleaned, especially if they are gonna splurge for good caps! There's some smears cause I took a little isopropyl alcohol to a spot. Decided to deal with it later though. I didn't have time for more than the pictures today. There are two smaller module PC boards that plug in, but both are double sided boards using SMD parts. I can get more pics later.

A few pads are lifted, near the parts that blew, and in one spot where they soldered current handling wire onto the traces to beef them up. There is little to no charring on the solder side. I should have no trouble cleaning up the top side to prevent shorts, the bottom looks like it barely needs work!

[richfiles]

And a detail shot underneath some wires, as well as the larger of the three modules. I got a kick out of the fact that they used that big 2 row header for only 2 conductors!

[MrSlack]

That's actually considerably less terrible than I thought it was going to be. Thanks for posting the details.

[Richard Head]

I had a look at the schematic. They use a full bridge topology on the primary side and a single ended (forward topology) on the output. Un-F#cking-believable.

[grumpydoc]

I had a look at the schematic. They use a full bridge topology on the primary side and a single ended (forward topology) on the output. Un-F#cking-believable.

Could I pick your brains a bit for an explanation of which bits go with your comment above (well, I can see where the H-Bridge is) and why what they did was less than optimal.

As above I've always been puzzled why Q401 just looked to be a linear reg with no pre-regulation. Voltage feedback looks to go to the TL494 via R302. THat doesn't feel right because if you try to get 5V @ 20A out of this PSU (which should be well within its capability) the FET is going to try to disspiate 900W which doesn't add up.

Anything to help increase my understanding of the design would be great!