EEVblog Electronics Community Forum
Electronics => Repair => Topic started by: oldmannewstuff on January 25, 2017, 08:11:07 am
-
Hello everyone! :)
My Zyxel ES1100 PoE router got smacked by a power surge. The router consists of a smps and the actual board that is responsible for the networking.
The networking board is still ok. I've tested it with another power supply. The reason why I'm trying to fix the original psu, is that my replacement psu is too big to fit inside the router case - and I thought it would be fun and educational to fix the original one :D
The original psu is 48Vdc and 1.56A. Input voltage is 240Vac.
The power surge took out atleast the fuse, MOV, coil, resistor (R16) and obliterated the mosfet.
Here's what I've replaced so far:
- the bridge rectifier
- MOV and NTC
- all capacitors (except smd capacitors) on the primary side
- original mosfet SMK1260 with 11N60C3
- resistor R16
- original two coils with one made from slightly thicker wire. I took the replacement from computer psu.
I've also checked the diode D2 which is next to resistor R16, and it seems to be ok.
After replacing all the parts I turned the device on only to blow the fuse. After releasing the magic smoke, I realized that my replacement mosfet has a metal back, whereas the original has a plastic one. So the mosfets drain was connected straight to GND via the heatsink.
Removing the heatsink GND connection solved this problem and the psu turned on without blowing a fuse. But there was no voltage arriving to the primary side transformer windings (shown in red).
The PWM controller chip is OB2269. Pinout is as follows:
1. GND
2. FB
3. VIN
4. RI
5. RT
6. Sense
7. VDD
8. Gate
Voltage arriving at pin 3 (through R3, R4 and R5) is 120Vdc! Datasheet shows VIN max voltage should be 36V. :wtf: The resistors R3-R5 all measure 0.324Mohm.
And to top it all off, while having my multimeter positive lead connected to rectifiers + side, and trying to measure gate voltage with the negative lead I've managed to blow the fuse - again :palm: And ideas what went wrong? I've always thought that one can measure voltages anyway round with a handheld multimeter. I'm 99% sure I didn't short anything out while doing the measurement ::)
So now I have a dead mosfet, resistors R13 and R14 shorted and probably the IC is toast as well :-/O
Is it possible that the IC was toast in the first place and because of that couldn't drive the mosfet properly thus leaving it in some intermittent state? But that does not explain why there is 120Vdc arriving to the IC.
Anyway, thanks to those who made it through my wall-of-text. Even if this simple circuit beats me, I'm going to call myself a winner for learning how switched mode power supplies work - in theory :-DD
-
That took quite a hit! I'm surprised the networking board still works after that.
I would investigate the cause of the inpressive looking black mark and crack/delamination on the bottom of the board to the left of R10. Something dramatic happened topside, I can't see for sure but it looks as if it under the transformer. If so, it could be that the transformer primary insulation broke down and arced.
P.S. Welcome to the forum. :)
Edit: Actually, no, it's on the same side of the board as the rectifier isn't it. It looks as if there was a discharge inside the laminations of the PCB. I would certainly check the transformer though, due to its circuit proximity with the Mosfet.
-
Thanks for your reply :)
I forgot to mention that I've desoldered the transformer primary side windings. Both windings have continuity - but not with each other so I guess the transformer is ok. I've also checked the optocoupler and it is fine also.
The bridge rectifier is GBU600 which I replaced with GBU6J.
-
-Perhaps get/use an isolation transformer (makes everything much safer and easier to measure).
-A lightbulb (50-150W) in series with the PSU works great in minimizing the damage when testing (lightbulb works as cheap and powerful PTC and will simply light up if there is an error in the circuit)
-Lowside feedback can sometimes tested by injecting voltage on the low side with lab PSU. Looks like that first blue ceramic capacitor couples the feedack signal onto the high side. (over to C5 and JP1).
-If that works, you can look at datasheet of SMPS controller chip and also feed it with external power and see whether it works.
-If everything works so far, you can use that lightbulb method and see if it works as complete unit.
-
Thanks for the tips! :-+
I'll definitely get that isolation transformer. Infact I feel a little bit foolish for doing those measurements without such a device. :-[
While waiting for replacement parts to arrive, I will build that light bulb thing you mentioned.
Would it be better that the light bulb total wattage be greater than the DUT would normally use? I mean the psu is rated 75W, so should I put a 100W light bulb in series with it or will 60W do?
Anyone have any ideas what could have gone wrong when I was trying to measure gate voltage as mentioned in the first post? I'd like to learn from these kinds of events. I'm still very sure I didn't short any rails with the multimeter tip ???
-
Well, I've managed to repair the darn thing :D
I replaced the following
- capacitor C6 which was probably busted as it didn't show anything on my ESR-meter.
- resistors R13 and R14 which showed abnormal values
- PWM controller chip
It didn't come out pretty, but function over form as the say ::)
Repairing didn't make sense financially, but I learned a lot, and wasn't electrocuted :-/O :box:
-
The power of perseverence! :-+
As you say, a good learning experience even if it took a while.
Thanks for reporting back, it's always nice to know how one of these stories turned out, especially when it was saved from going in the bin. :)