1
\$\begingroup\$

I have a cheap benchtop dual-output adjustable DC power supply. One of the outputs is working alright, the other one "burned out" when it was left loaded somewhere near 80% of its maximum for 24 hours.

After opening it up and taking a look, I found a few burned power resistors (0.5Ω rated for 5W) plus some resistors which were surprisingly still functional (though burned beyond recognition) sandwiched between/around them, all of which I replaced:

Before: Burned power resistors

After (with 3 resistors in parallel to give me an equivalent circuit until the replacement power resistor arrives - they haven't burned out in testing, so I'm good for now):

enter image description here

When these power resistors went out, they took a few more resistors nearer to the middle of the board out with them, and I replaced those as well. Powering up the board burned the regular resistors in the middle again, so I replaced them once more and tested all the other components on the board:

Burned resistors in the middle of board

Luckily for me, there are two identical halves of the device, and I was able to quickly compare resistance across components leads (disconnecting some where I wasn't sure, but not all of them) and don't seem to have any remaining bad components. None of the capacitors are visibly damaged and their basic functionality checks out (though I'm unable to test their compliance to their rated capacitance as at 470uF they are out of my DMM's range and I don't have an ESR). The diodes seem to check out (though I'm having trouble positively testing some of the high-voltage protection diodes). The rectifiers have tested OK, and none of the transistors are shorted out so far as I can tell.

Crucially, I'm not sure what failed components would result in this behavior. The power supply is actually advertised as being 0-30V (and 0-5A), the 48V it's giving out now are well out of spec.

The problem is much complicated by the fact that the PCB design is an absolute mess. A number of traces appear to have been (by comparison to the other, functioning board) purposely joined together via globs of solder. The solder obviously reflowed around the burned power resistors and possibly in a few other places, but tracing it as best as I could and comparing it to the other board, it seems that everything is fine (and the burned traces are still functional).

The unit is still stuck at 48 volts no matter how I adjust the pots (and I tested them, just in case it was ironically just a control problem after everything I did and tested - I also disconnected the control and output boards entirely to no effect), and I'm not sure what else to try. I'm not very well-versed in analog, though I've tried isolating the problem as much as I could only to always end up disproving my hypotheses.

Here's the board now: Front of the non-functional board Back of the non-functional board

The unit itself: enter image description here enter image description here

I found a few schematics online for similar Chinese units, but I haven't been able to find any that match one-for-one my unit.

EDIT

The power transistors up close:

enter image description here enter image description here

\$\endgroup\$
  • 2
    \$\begingroup\$ Check the pass transistor (main power transistor) for that channel, make sure it hasn't failed short. \$\endgroup\$ – ThreePhaseEel Apr 23 '16 at 20:29
  • 1
    \$\begingroup\$ The main power transistors and the smaller ones providing the base current for them should be the main suspects. Take them out and check them each. \$\endgroup\$ – Laszlo Valko Apr 23 '16 at 21:26
  • \$\begingroup\$ Thanks, guys - you're both all kinds of awesome. The biggest transistor (seems to be a genuine Fairchild D880-Y, one for each unit) was fine, but one of the medium-sized ones (no-name, generic D1047 NPNs) mounted on a common heatsink is fried. It didn't show in earlier testing because it's not totally gone, but the voltage drop across the BC and BE is 0.03V instead of ~0.6V. I don't have any power transistors rated for that kind of voltage (200V according to the datasheet), so I'll have to wait for a replacement part to come in. \$\endgroup\$ – Mahmoud Al-Qudsi Apr 23 '16 at 21:31
  • \$\begingroup\$ I guess if there is anything you can do to improve airflow it might help avoid future problems. And/or keep an eye on the temperature of the vulnerable components. Good question. \$\endgroup\$ – mkeith Apr 24 '16 at 3:55
  • \$\begingroup\$ The D880 is a TO220, the D1047 should be TO3. The D1047s are the power transistors (that's why they are mounted onto the large heat sink), the D880 is most likely the driving transistor for them. If there are multiple power transistors for a single channel (they are connected in parallel), you should not replace a single one of them from a different model/vendor. Either get the very same model, or if that's not possible, replace all of them with a single model (otherwise current sharing may probably not work, and then "the weakest point in the chain" will break first). \$\endgroup\$ – Laszlo Valko Apr 24 '16 at 10:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.