I reviewed the Rohde & Schwarz HMP4040.04 back in October 2018, which was unfortunately damaged in transit. After almost a year without any resolution, I decided to crack it open and attempt a repair. The first time I opened it, I found a bunch of bad components but I couldn't narrow down the fault conclusively. After receiving a helpful comment which pointed me to the possibility of other bad components, I tested my switching MOSFETs and found them to be a dead short as well. This made it seem more promising that a successful repair might be on the cards. As a winner of an element14 shopping cart, I loaded up a list of components and hoped for the best. Unfortunately, when the element14 Repair and Restoration Project14 rolled around, the components had not all arrived yet so I had to shelve the plans.
In fact, the components had not all arrived even now due to a shipping discrepancy. Although I was assured it would be re-ordered for me, I still have not received it. Fortuitously, I had a reason to put in an order for element14 Australia just this week, so I decided to add the missing Littelfuse Nano^2 7A Fast Acting fuses to my order since there was local stock and I could get it next day. On their own, it wouldn't have been worth it, but seeing as I had a substantial basket of goods with free shipping, what's a few dollars?
That only means one thing - it's time to start up the soldering iron and see what happens.
Cracking Out the Soldering Iron
I'm no soldering genius especially when it comes to SMD components, but my soldering is passable enough that it shouldn't cause any new faults. So the first thing on the list was to get the two new Schottky diodes into place.
One of the diodes isn't quite on straight, but it should still be well connected. These aren't exactly the same model as the original, but they're equivalent in specifications and package. The blast from the heat gun has melted part of one of the toroidal filter base, but that hasn't harmed anything. The heavy copper and vias really suck the heat away from the board, making it a challenging job for a thin-tipped ~60W iron.
On the opposite side, I replaced the MOSFETs with better equivalents since the original Infineon parts could not be sourced. I put a bit too much solder on the right one and gave it a whack with the heat gun to get them to sit into place. The dirty looking flux still remains, but that's not a problem. Given that both MOSFETs had shorted out and the package on one was literally crumbling apart when desoldering, perhaps the Linear Technology switching controller above might also be damaged. While it can be sourced, it is a bit pricey and relatively fine-pitched and packed in with other components. I wouldn't be game to try replacing it - I would probably end up breaking it if I tried.
The surface mount capacitor was put back into place and the electrolytic was next. Because of the thin holes and plenty of copper, clearing out the remaining leaded solder was an absolute pain and I couldn't do it even with a solder sucker bulb. I don't have a desoldering gun so I resorted to bridging both pins with hot solder while "hot inserting" the component from the other side, then wicking away the excess. Unconventional and perhaps dangerous, but desperate times call for desperate measures. The 100 ohm SMD resistor went back in, but huh, I must have missed it before but that 100 ohm resistor next to it looks a bit smoky! Uh oh. That's not good news. Perhaps this means there's more to this fault than I thought, but then again, I know there's not much more I'd be game to try and replace - it's just too tight and finnicky. The transistor above however, is not damaged - that's just a bubble of flux on the package.
Finally, I decide to fit the two fuses, since they can be a bit heat sensitive. These two went back in just fine, although the bottom one is a bit crooked. Functionally, it should be just fine though.
Installing and Wiring the Module
The wires to the PCB were so heavy that I just decided to cut them off the board when removing the module. As a result, now, the wires had to be stripped to prepare for re-inserting into the module. This was possible as there was enough slack in the wire to allow for a few millimeters of wire to be consumed. The mains transformer has three windings for each module - in this case, I have marked the edge of the PCB with black permanent marker and one of the two equal-coloured wires with the same mark to indicate the polarity of the winding. This is because, in some devices, the polarity of the winding could be important although in this case, I doubt it. My biggest concern is getting all of it soldered back in.
The soldering job on the rear transformer inputs to the right side of the image was easy. The soldering of the DC outputs however, was a nightmare. The copper cables and the PCB sunk so much heat that dialling up the temperature to the maximum and holding the iron down for minutes with a ball of solder just didn't manage to flow it nicely.
While it looks ugly, I decided to give it an extended blast with the hot air gun to flow it as best as possible. The central two connections don't look well soldered, but it seems good enough to me since it seems the solder has permeated through the strands and it's more so the pooling of burnt flux that gives it the "shadow" around the wire. At this point, this is the best I could do, so it was back to screwing the whole unit back together. Luckily, even after all this time, I managed to reassemble it without any extra parts left over.
Testing the Power Supply
The moment of truth has arrived. I plug it in, turn it on, select all channels and hit the master output key. The relays click ...
NOOOO! The fourth channel is still as dead as it was before. Drats! It seems I have failed in my quest to repair the HMP4040.04 and perhaps there is a downstream fault in some of the smaller transistors that drive the pair of MOSFETs near the capacitor and resistors I was looking at before. But I know my limits and this repair was pretty close to it. Given the amount of time, effort (the unit is HEAVY!) and risk, I don't feel confident to go further. On the upside, I haven't broken anything more than it was broken before, so that is a win in my book.
Repair is often a bit of a tricky thing - it's rarely ever a guarantee. If you're like me, probing "in the dark" without a schematic, guessing at potential problems and swapping out components "shotgun" style, then success is rarely ever assured. If I had the skills of CuriousMarc for reverse engineering things and a bit more time to dedicate to it, perhaps I would have fared better. If I was willing to take more of a risk and start running the unit live and probing about its "companion" channels for comparison, that might have helped too. If I really wanted to make things hard and I dared, I could probably start swapping components between modules, but that's almost certainly going to result in some carnage based on prior experience.
But more often than not, repair can be a bit of a gamble. In this case, it cost me about AU$15 of parts and a few hours for the chance to revive the fourth channel of the unit. I didn't win this time, the channel is still "dead". But I could have come out of the gamble a lot worse - failing to reassemble the whole unit or collateral damage to other parts of the unit could have left me with a unit in worse shape than I had started with. Part of repair is knowing when the gamble doesn't pay off and quitting while you're ahead - in this case, I knew I wouldn't have much of a chance with the other components on the cramped PCB and fine pitches that I had made the decision that if this did not fix the problem, I wouldn't bother with another attempt. Now I know for sure that I couldn't have fixed it myself, but haven't really lost much in trying to fix it - perhaps a bit of time, a little bit of money and perhaps a tiny bit of pride having been defeated.
That doesn't mean it wasn't worth trying. That doesn't mean it's not worth repairing something in the future either. It's about picking the right battles, having the right skills and equipment, and sometimes having a little bit of luck and timing.