SOLVED: Identifying value of resistor in PSU (not Arduino related)

After plugging in a mobility scooter and hearing a "crack", the 24V output PSU no longer works - dead as a doornail. Might have to just buy a replacement, but can't resist attempting a repair.

I unscrewed it and was pleased to see that it is, in principle at least, quite repairable - nothing encased in epoxy or anything like that. No signs of anything major, no black marks or obviously damaged components. It is an offline regulator design based upon the UC3842B "high performance current mode PWM controller" IC. The IC appears to be powered by running rectified and smoothed 240V mains through a 300k resistor and then into the IC which internally incorporates a 34V Zener across its power pins. Seems a bit funky, but who am I to say.

After a bit of probing around with the multimeter, it seems that both the IC and the power N-MOSFET (AP09N70P, probably obsolete) were roasted on the primary side of the device (the half shown in the attached image). The FET's gate is shorted to source. I removed both the IC and the FET.

I've probed around the primary side some more and it seems like most of the components are showing sane values, bar R12, seen best in the close up. With the IC removed, it is only connected at one end and is reading 8.3k. It is connected to the "OUTPUT" pin of the IC and runs through a one-time thermal fuse to the gate of the FET.

The value of R12 is somewhat obscured. Actually, it was a little better before, but there was a raised bit that appeared to be flux residue or something and so I dutifully scraped it off - together with more of the white silk. Muppet me. That said, the middle digit was never really legible.

So after the TLDR, the question is: what is the likely value of R12?
I think the last digit may have been a 6 (as oriented in the picture), but that would make the resistor a very high value for driving a gate and would make the FET toasty warm. With the alternate orientation, the last digit would be 5, but likewise is a very high value. So maybe the middle digit is an R? That would make the value maybe 5R6, but that seems very low. What would the purpose be for such a low gate resistor?

I don't know the value of R12. I suggest you find the datasheet for a UC3842B and see if there is a reference design that matches the design of the board you have. If there is then you can work from that.

I occasionally repair switch mode power supplies, it is almost always the case that one of more electrolytic capacitors has failed. Unless you know what you are looking at it's not obvious, but they swell up. I would change all electrolytic capacitors whether you think they are good or not.

Ta Perry. I did already check the datasheet, but didn't see any significant reference design; just test circuits and partial circuits to demonstrate how to achieve particular design features.

You make a good point about the capacitors. The scooter is only two years old, hasn't seen much use in the last year (guess why, no prizes) and can't have been charged that often, and all the caps "look" fine. On the other hand, the PSU is original but isn't branded with the scooter manufacturer, so is probably a generic from a huge warehouse in China and may be have been sitting in said warehouse for some time - unlike a fine wine, caps don't improve with age, right? So maybe worth considering replacing them as you say. Just don't want to end up with a bunch of spare caps that I will struggle to use and will degrade.

The risk with the whole thing is that I replace all the bits that I know are bust, but miss something (maybe the something that blew the thing up in the first place), and within half a second of switch on it goes crack again. Still, no risk no fun...

arduarn:
The risk with the whole thing is that I replace all the bits that I know are bust, but miss something (maybe the something that blew the thing up in the first place), and within half a second of switch on it goes crack again. Still, no risk no fun...

That is the deal with SMPS repairs. :sunglasses:

5R6 seems very reasonable for a gate resistor, limiting rise/fall times to reduce EMI generation,
yet allowing pretty fast switching for use in a switch-mode supply.

With high voltage FETs is usually vital the gate has a very low ac impedance to ground so
that the high dV/dt values on the drain don't couple through the gate-drain capacitance and
cause the gate voltage to be affected too much. A low-value gate resistor helps with this.

Hi,
I found this in a data sheet; check page 16.

And this note;

Tom.. :slight_smile:

Thanks for the feedback folks, very useful.

MarkT:
5R6 seems very reasonable for a gate resistor, limiting rise/fall times to reduce EMI generation,
yet allowing pretty fast switching for use in a switch-mode supply.

Sounds good. I'll run with the 5R6 and take my chances.

The design does have other emission reducing measures. First the phase and neutral both go through what seem to be individual small chokes which float over a cutout in the PCB beneath the IEC connector; then through the small toroid choke in the bottom right of the picture; then through the bigger yellow choke. The heatsinks either side of the board are joined with another piece of folded alu that basically shields the unit on the four larger sides.

TomGeorge:
I found this in a data sheet; check page 16.
https://www.onsemi.com/pub/Collateral/UC3842B-D.PDF

Oh, sweet. I was using the datasheet from ST, since it is the ST version: much thinner on detail. Maybe I should have looked for an AN.
Anyway, although that design on page 16 isn't going to be identical with the practical example in front of me (eg. doesn't have the optocoupler feedback from the secondary side), it should be close enough to give me a better understanding of how the diodes are connected up and allow for some more testing. Nice to see the 22R gate resistor.

I'll poke around a little more and then order up some parts. Maybe I'll have results in a week or so.
Thanks!

I got the replacement IC, FET, resistors and capacitors, banged them in and ... nothing, not a sausage. On the plus side, there was no loud noise, explosion, smoke etc, so that was positive. The new components didn't seem to take any harm.
Spent over a week then with the PSU sitting on my desk, giving it the occasional evil look and the odd probe.

Measuring across the IC's VIN and GND indicated about 13V, which is high enough for the IC to keep running but too low to exceed the 16V UVLO necessary for start-up. There was no 5V reference voltage output on pin 8 either.
I was able to power the low voltage part of the circuit with the IC using a bench power supply (connected to point A in the diagram). When the voltage was increased above 16V the IC sprang to life: 5V reference, waveform on output pin, new R12 5.6R resistor somewhat warm to the touch.
Output waveform:

bad_wave.png
Voltage not to scale, peak is around 14V.

Still couldn't figure out why it wasn't working, more probing. Eventually decided to reverse engineer the primary side of the device (excluding supply filtering) which didn't take as long as I thought; the following diagram was the result and is hopefully close to reality:

UPG 24BC2000T-2

The output waveform was showing a duty cycle of <1% and was about 6.5kHz. That seemed a very low frequency and why would the duty cycle be so low when the device has to power up and charge some big capacitors. So I worked through the diagram on the paths that control the PWM and feedback and measured across the load sense resistor R16 again...33k...? What? Took it out and it was indeed (visibly when removed) defective. Don't know how I missed that before because I did check it. Maybe I had checked the bands and expected 0.33, measured and saw 33 and ignored the rest. Very careless.

Had to order a replacement, with the best I could get being 10 4W way overkill wirewound ones (9 spares, LOL). Only just fitted in the holes.
Powered up again using the bench supply, wound up to 30V, then used a jumper to jump start the IC by applying a short burst of power to point A...and yoo hoo, it was running! R12 not just as warm as before, but still a little warm. Output waveform now 68kHz:

good_wave.png

Plugged into mains and now the PSU seems to be operating as normal again. Success at last.
Moral of the story: Sloppy measurements waste time.

Anyway, looks like the 5.6R was a good call. I am pretty sure that the reason the resistor was hard to read was a result of it failing and vapourising some of itself. All in all, U1, R12, Q1 and R16 were defective.

Thanks folks.

bad_wave.png

good_wave.png

Took it out and it was indeed (visibly when removed) defective.

Did you measure the resistance with a DMM ? What was the value ?
The resistor looks cooked. Were you not able to see it when you did your visual inspection ? (was it in a location
that made it difficult to see?) Just out of curiousity, did you use a bright flashlight for your visual ?

I am not known for being generous with compliments (unlike many members here) . In general I don’t think it’s
productive or helpful to pat someone on the back for breathing. I believe when it comes to electronics, someone
has to demonstrate at least a willingness to exercise some measure of discipline that demonstrates their commitment to learn and 'do the other things… not because they are easy but because they are hard…".
I don’t expect others to have the same commitment to electronics that I do , because for me, as a restaurant cook
making hamburgers 42 years ago, with no college education, (2 yrs jr college general ED doesn’t count), I chose
to learn electronics because I saw a future in it (I don’t think anyone realized 42 yrs ago how much progress we
would make) and remaining a restaurant cook was not an option, so I had to go all in. I later went back to school
to get a BS in Electronics Engineering Technology at the age of 43, with everyone else in the class being right out
of HS or slightly older. To this day , that degree still gets me jobs. That being said, what should be expected of the
weekend electronics hobbyist ? (everyone has their own idea). After reading about 17,000 posts, I would say that
your post is one of the best written ,organized and illustrated that I have ever seen… (I once repaired a 3/4" video recorder without a schematic but I got lucky).
Based on the fact that most of your last 1000 posts are in the “Programming Questions” forum topic, I’m guessing
that you are a programmer (and also German). I’m curious what scope you were using ?
So Kudos for a well written post and Karma for the noble effort and subsequent success. Keep up the good work…

raschemmel:
After reading about 17,000 posts, I would say that your post is one of the best written ,organized and illustrated that I have ever seen..

Wow! High praise indeed. I don't think I ever remember a time when someone has accused me of writing well. Back in the day at school my grades for English Language and English Literature were, let's say, less than optimal. I'm not known for clarity and conciseness either. Nonetheless, compliments always welcome... :smiley: Thanks!
I'll keep your comments in mind for any future posts I make that would benefit from more extensive treatment.

The resistor measured as ~33k. I struggled at times to get good readings on the meter since the probe tips are not sharply pointed and, when not sliding from their target, would sometimes show a value which would then change with increased probe pressure - maybe flux residue or oxidation on the solder blobs.
A bright LED desk lamp was used during the inspection and the damage to the resistor would have been somewhat tricky to see looking down from above, but probably not impossible. What can I say, I missed it.
Those resistors seem to have quite a crumbly outer coating on them; it makes me wonder if they might be more susceptible to the slightly damp conditions of an unheated shed.

You're right, my background is in software, but I'm from Norn Iron. Scope is a R&S RTC1002.

I don't think I ever remember a time when someone has accused me of writing well. Back in the day at school my grades for English Language and English Literature were, let's say, less than optimal. I'm not known for clarity and conciseness either

Well to clarify, it was not so much your English compositional skills, (I excelled in that in school because I never
learned sports so I was a bookworm at 14 reading all the science fiction novels as well as Les Misérables and
pretty much all of Jack London and Arthur Conan Doyle) so much as it was the way you jumped right in without
wasting any time and got right to the meat of the matter, with little or no fluff. Pretty much everything you said
was relevant and your approach was quite comprehensive. Had it not been for the that resistor I would have pegged you as an electronics technician, but having been one myself , I have found that repairing a circuit board
is in many ways similar to a doctor treating a patient, in that there is little room for error so the bulk of your time
is spent doing boring things like measuring ALL the components , or at least all the relevant ones. When I was
an ICU Hospital Tech in the USAF during the Vietnam War I learned that all the things you have to do keep someone alive when they are in ICU are boring and redundant and sometimes not very much fun but if you
put them all together, it's the difference between life and death, so I learned to appreciate the value of the little
things or boring things because they are analogous to a detective pounding the pavement. I would have
been puzzled by the color coding on that resistor myself being as 5 is green and 6 is blue and the resistor bands
were both orange. The resistor wattage (physical size) is probably the only useful clue. It doesn't help that the
wattage wasn't on the schematic either. I don't know if I would have found the problem any quicker than you
did but when I know something burned up I focus more on an intense visual before getting into any schematic troubleshooting. When my landlord asked me to fix his 3/4" video tape recorder without a schematic I just put
my meter on diode scale and checked every power device on the board until I found the culprit. He couldn't
afford to pay me so he compensated by offering to do a custom charcoal sketch. (he was an artist, his name was Kieth Daniels)
I asked him to do a futuristic treatment of my Yamaha FJ-1100 motorcycle, but supersize it to star wars
mega size and give it rocket engines. The attached sketch shows the crew in the cockpit windows and the
spherical components of the rocket engine. It came out pretty nice. It took me 2 hours to fix his tape recorder
but it took him like 5 or 6 hours to draw the sketch. He was thrilled I fixed it because that meant he could
transfer all his reel videos to VHS tape, which he later did.
I worked with an Irish engineer for the last year until he left the company.
We got along great. He was such a gentleman. More like a gentleman's gentleman. (polite to a fault) I used to
tease him and say 'Gerry, you know I like you so much I'm going to upgrade you to VIP status and from now on,
whatever you want , you straight to the top of the cue !" (I was really just messing with him because my boss (his boss) told me to support him in any way necessary).
That is one fine scope. I'm jealous. I'm using old Tek scopes . One of them has a 5 1/4" floppy drive for saving screenshots that's how old it is. My previous three jobs they had top of the line everything. Apple and Google
had some pretty nice equipment. I got spoiled fast.

This topic was automatically closed 120 days after the last reply. New replies are no longer allowed.