Power Supply Repair

[patrickoneal]

I have an automotive wheel balancer that is throwing error codes for low vcc very often. I tried to balance a wheel on it, and sometime during the balancing process the machine always craps out and gives a low vcc code, usually after the 230v electric motor runs, but sometimes when the machine is just sitting idle. I don't have an ESR meter, but I did check the motor start and run caps with my fluke 87 and the capacitance is is dead on for those, so I don't think it's some huge inrush from the electric motor causing issues. I've also seen a high VCC code just once. The code for VCC too low is for voltage under 5.10V, and too high is over 5.35V.

The motherboard for the machine is situated on a larger power circuit board. This board contains controls to start and stop the electric motor and what looks to be a switching power supply to supply VCC to the motherboard. None of the capacitors outwardly look bad, but they're 15 years old, so I'm suspect of the caps in the 5V power supply section.

I physically measured the caps and noted the values and ordered replacements from digikey. Before I ordered, I read that low esr caps are a good thing in a power supply. Out of the 6 capacitors, I could only find the 3 on the right side of the photo in the low esr variety at digikey(1200uf x2, and the 560uf). I believe these are the output filtering caps. For the remaining three I could only find general purpose caps and the quoted ESR is quite high. I don't think that's an issue on the large 120uf 400V cap, but I'm not sure of the other two(the 33uf and the 47uf). The 47uf might have been a lower voltage than 63, but that was the closest I could find.

Two questions:

Do you think I'm on the right track to stabilize the power supply?

Do you think the general purpose caps are fine for the 120uf, 33uf, and 47uf?

Thanks for any help.

[patrickoneal]

The closest component to the 47uf cap, under my red arrow, is a TI TL431C: http://www.ti.com/lit/ds/symlink/tl431.pdf

The component under the black heatsink near the right side is a B1545 Schottky rectifier.

The component under the silver heatsink is unknown, because the heatsink is soldered to the board.

There is an optocoupler near what I assume is the transformer.

[t1d]

As for suspecting the power supply... Given the errors, sure, that's not a bad place to start looking.

I take it that the controls are built directly onto the balancer. If so, it is subject to a lot of vibration.

Vibration doesn't have to have a large amplitude, to be destructive. It just has to hit the wrong frequency. At that frequency, the shaking is most intense.

So why talk about vibration? Because I might suspect a broken solder joint, or a cracked PCB. Do a close, visual inspection, with magnification and light. Use a non-conductive stick, to gentle push on things... Reflowing all the joints, just because, may be just as easy.

As for caps... Suspecting 15 year old caps is certainly prudent. Whether, or not, they are the problem, replacing them, while you have things apart, makes sense.

Looking for visual indications that a cap has failed is the first place to start. But, caps can, and do, fail, with no visual indication. They dry out, or their ESR value fails.

This type of cheap ESR meter will find hidden cap failures and actually has a decent reputation . I have one and I love it. Look for the newest technology and firmware and the best price. https://www.ebay.com/itm/LCR-T4-Mega328-Transistor-Tester-Diode-Triode-Capacitance-ESR-Meter-W-Shell-US/302739677139?epid=15003944572&hash=item467cb0dfd3:rk:5:pf:0

As for low ESR caps... If Digikey doesn't have what you want/need, then don't be afraid to go to multiple vendors, to get the right stuff. I like Mouser.com.

Yes, you will pay multiple shipping costs. But, you can write that off, as a cost of repair and for the sake of the value of the machine. In other words, it is just plain worth it.

Good job, on getting the cap measurements. When that step gets missed, things don't fit.

Is anything getting warm? Apply isopropyl alcohol and watch the shine, to see what dries out first. This is will be a hot spot. Heat can be an indication, of a problem.

[FriedMule]

Is it possible to desolder the component under the  the silver heatsink without removing it, so you can test what component it is?
I would choose some good general purpose caps like Panasonic and when mounting them, give them a "bath" in hot glue, to minimize vibration damage in the future.

[patrickoneal]

Thanks guys.

The 3 low ESR caps I was able to order from digikey for what I believe are for the output of the power supply are from United Chemi-Con and they're automotive grade. I figured rightly or wrongly that they would be good on a machine that moves around quite a bit. I think low ESR is appropriate for the application.

As for the two small caps, I ordered these, but didn't notice that the high ESR was rated at 120Hz rather than 120KHz. Does ESR rise or fall with frequency? These are the two parts, no big loss if they go in the trash:

https://www.digikey.com/product-detail/en/illinois-capacitor/336CKE063M/1572-1125-ND/5344055
https://www.digikey.com/product-detail/en/illinois-capacitor/476CKH063M/1572-1476-ND/5412774

For the 400v cap, which I'm guessing is the input filter, I order this one, and I believe low ESR isn't a concern for this application, so hopefully it's fine:

https://www.digikey.com/product-detail/en/panasonic-electronic-components/ECO-S2GB121BA/P10138-ND/262200

Also, thanks to t1d for the isopropyl suggestion, I'd never heard of doing that.

I'll definitely checked for cracked or melted solder joints as well. I've ran into that before and it can definitely cause some serious headaches.

[patrickoneal]

Is it possible to desolder the component under the  the silver heatsink without removing it, so you can test what component it is?
I would choose some good general purpose caps like Panasonic and when mounting them, give them a "bath" in hot glue, to minimize vibration damage in the future.

I think I'd have to desolder the heatsink, I'll have to look again and see the degree of difficulty on that, I don't know if the legs are twisted to lock into the board.

[patrickoneal]

I replaced all the electrolytic caps in the switching power supply section, and the x2 safety cap on the power input as well. I used low esr on the three caps that looked like the output caps, and general purpose for the rest. I also touched up one leg of the inductor on the output just because the hole wasn't completely full of solder, but I think it was likely fine. I fired up the machine and let it try to balance a tire. I did this multiple times and it works great now with no VCC warnings at all.  The interesting thing is that not a single cap had outward signs of failure and the capacitance was correct for each of them on my fluke. I'm going to get an ESR meter because I'm dying to know if the problem shows up there.

The only other problem on the machine is what I assume is a high tolerance potentiometer. There is an arm that pulls out of the machine that you touch the car wheel with. I has one pot that it rotates to determine diameter, and another that rotates via a string around a pulley that lets the machine know the distance from itself to the wheel. When I first tried the machine it threw error codes for the distance pot, so I went into the calibration mode and read the pot values. The diameter pot was working, but the distance pot was "stuck" on one value. Turns out the shaft had pulled out of the pot. I pushed it back in, and adjusted it to spec. This function of the machine now works, but the pot shaft can probably fall out pretty easily.

The end of the potentiometer shaft is splined. Does superglue or loctite sound like a good solution to permanently fix the shaft back in place?

I would get some better pics of the pot, but I don't really want to pull it out as it's currently working. I'm hoping there's a part number on it, but it may be expensive to replace as the machine is from Germany and the distance meter reads to a 1mm resolution.

Thanks in advance for any advice. I'm really happy that it's turned out to be a simple fix for the VCC issue.

[bitseeker]

Patrick, welcome to the forum! Looks like you've got things well under control and back in business. Just wanted to add that the component under the silver heatsink is probably the power MOSFET that generates the high-frequency pulses that get fed into the transformer.

Regarding your question about ESR vs. frequency, it depends on the type of capacitor and frequency range. Typically, it drops as frequency increases, but with some caps (e.g., ceramic) it could rise again above a certain frequency.

Also note that low-ESR is usually good, especially for filter caps. However, it isn't always desirable as ESR can be utilized to stabilize a circuit (eliminating unwanted oscillations) or to achieve a particular rate of oscillation. So, if you're not sure the role that a capacitor plays, replace it with one as similar as possible to the original.

[RobertHolcombe]

For the splined shaft I'd personally try shimming the shaft with aluminium foil (or similar) over using an adhesive, at least as a test.

[tpowell1830]

Nice post, good repair. Wish we had more like this.

[t1d]

Here's several more;-)

There are strong glues that are also intended, and designed, to be removable. Check into those. You need an industrial glue guy. Maybe call a manufacturer. Companies like to help folks. Maybe 3M?

As an example, but not necessarily for this particular use, is hot glue. It is easily defeated by isopropyl alcohol. It shocks (not electrically) the hot glue's chemistry, somehow, and the glue instantly becomes brittle. You just chip it away, in chunks. The effect only lasts, until the alcohol dries, but you just have to reapply it.

Another glue trick I use is "Shoo Glue." It is made for putting the sole back on a shoe. Tough stuff... But, it can be removed, by pulling it off. Stretching it causes it to relieve. It takes a lot of careful work, though.

Also, you need to find out what type of plastic the wheel is made out of. Some glues won't stick. Super glue, for instance, doesn't stick to some plastics. And, some types might hurt the plastic... Think about how model airplane glue sort of melts the plastic. You can't have that happen.

Oh, here's an easy idea... Use a hose clamp, over the base of the pulley. Get the kind of clamp that will unscrew the track, completely, to open it. Some won't open that far. Slide it over the hub and squeeze the hub down, onto the stem.

Here's another idea... Cut a keyway, into the stem of the pot and the hub of the pulley. Install a key, to mate the two.

Yet another... Cut the stem, of the pot, down the center. Drive a wedge, into the cut, to spread the stem.

Maybe try a Dremel rotary tool, with a cutting wheel. Because the pot is not in the best of shape, be careful, with vibration, and seal it with removable tape, to keep the dust out of it.

But, to be honest, I bet that pot is a common pot. Replace it, with one that has the right value, body size and pick a stem that will work to repair the slippage.

[patrickoneal]

Here's several more;-)

There are strong glues that are also intended, and designed, to be removable. Check into those. You need an industrial glue guy. Maybe call a manufacturer. Companies like to help folks. Maybe 3M?

As an example, but not necessarily for this particular use, is hot glue. It is easily defeated by isopropyl alcohol. It shocks (not electrically) the hot glue's chemistry, somehow, and the glue instantly becomes brittle. You just chip it away, in chunks. The effect only lasts, until the alcohol dries, but you just have to reapply it.

Another glue trick I use is "Shoo Glue." It is made for putting the sole back on a shoe. Tough stuff... But, it can be removed, by pulling it off. Stretching it causes it to relieve. It takes a lot of careful work, though.

Also, you need to find out what type of plastic the wheel is made out of. Some glues won't stick. Super glue, for instance, doesn't stick to some plastics. And, some types might hurt the plastic... Think about how model airplane glue sort of melts the plastic. You can't have that happen.

Oh, here's an easy idea... Use a hose clamp, over the base of the pulley. Get the kind of clamp that will unscrew the track, completely, to open it. Some won't open that far. Slide it over the hub and squeeze the hub down, onto the stem.

Here's another idea... Cut a keyway, into the stem of the pot and the hub of the pulley. Install a key, to mate the two.

Yet another... Cut the stem, of the pot, down the center. Drive a wedge, into the cut, to spread the stem.

Maybe try a Dremel rotary tool, with a cutting wheel. Because the pot is not in the best of shape, be careful, with vibration, and seal it with removable tape, to keep the dust out of it.

But, to be honest, I bet that pot is a common pot. Replace it, with one that has the right value, body size and pick a stem that will work to repair the slippage.

I don't think I was completely clear on the issue. The pulley has a collet that locks down on the pot shaft. That pulley and collet is fine and grips as it should. The actual shaft of the pot can be completely pulled out of the pot itself, with the shaft still attached to the pulley. It's been working fine since I pushed the shaft back in, but I might pull the machine back apart, check for a part number on the pot and see if it's anything special, and glue the shaft back into it. I do appreciate the advice though.

[tautech]

I don't think I was completely clear on the issue. The pulley has a collet that locks down on the pot shaft. That pulley and collet is fine and grips as it should. The actual shaft of the pot can be completely pulled out of the pot itself, with the shaft still attached to the pulley. It's been working fine since I pushed the shaft back in, but I might pull the machine back apart, check for a part number on the pot and see if it's anything special, and glue the shaft back into it. I do appreciate the advice though.

I'd guess the pot was a sealed preset type, in that the shaft is in the form of an X and just slides into the X in the pot center axis and the string tension on the pulley stops the pulley and shaft from disengaging....or it's supposed to.
If the pot's operation is OK and just the shaft keeps wandering out, why not just form a long metal finger to keep the pulley center held towards the pot. Bolt it onto the flat surface below the pulley or from the side of the box somewhere handy.

Finding a pot that can accommodate a sliding shaft, well not impossible but it might be easier to do a simple fix.
OTOH, was the shaft originally held into the pot with a circlip...well an E clip ?

[6PTsocket]

I have not seen the pot so It is hard to make a suggestion but I can make a few points. Most types of Locktite EAT many types of plastic.I once rebuilt a fishtank diaphram pump. The nut on the diaphram had been locked with a dab of paint. I used a tiny drop of Loctite. The next time I looked, the whole side of the pump case was mush. I have used Goop for right duty plastic fastening and found no solvent damage. JB Weld works quite well on some plastics. I would look to see what was originally designed to hold the shaft and take it from there. A mechanical solution might work.

Sent from my SM-G900V using Tapatalk

[patrickoneal]

I finally got my DER EE DE-5000 LCR Meter. I tested all the caps I replaced, but none of them look bad to me. I'm confused as to why the machine works now, unless there were cracked solder joints I repaired in the process. None of the caps would produce a readout at 100khz on the meter, which I found a bit odd. I guess it's some limitation of the meter. Here are the higher frequency readings I got:

Electrolytic Caps:
Cap 1: 1200µf 35V measured at 1kHz: 1100µf 0.03 ohms
Cap 2: 1200µf 35V measured at 1kHz: 1084uf 0.03 ohms
Cap 3: 560µf 35V measured at 1kHz: 486µf 0.04 ohms
Cap 4: 47µf 63V measured at 10kHz: 43.9µf 0.53 ohms
Cap 5: 33µf 63V measured at 10kHz: 27.7µf 0.80 ohms
Cap 6: 120µf 400V measured at 10 kHz: 91µf 0.24 ohms

X2 Safety Cap(metalized film?):
0.1µf 275V This one is weird. It measured 11 Megaohms at 100 Hz, 9 Megaohms at 120Hz, 7.6 Ohms at 1 Khz, 1.6 Ohms at 10 kHz, and 0.30 Ohms at 100 kHz. I'm not really sure what it should measure, or how to further test it.

I got the meter to satisfy my curiosity first, and secondly to understand which of the parts I replaced had actually failed so I can apply that knowledge to future repairs. The machine works now, I'd just like to know why

Does anything jump out at anyone? Any thoughts on why the DE-5000 would just read OL when trying to read ESR at higher frequencies?

Thanks.

[cdev]

Those caps sound okay to me. Their ESR is getting very low at the higher frequency.

The DE-5000 is the best meter of its kind for under 4 times the price. 

Other people with more experience should weigh in but those capacitors sound to me to be not unlikely to have still been okay.

[patrickoneal]

I've already replaced them and now the machine works fine. That's why I'm confused about the apparent good readings on all of them.

[tautech]

I've already replaced them and now the machine works fine. That's why I'm confused about the apparent good readings on all of them.

Except these two:

Cap 4: 47µf 63V measured at 10kHz: 43.9µf 0.53 ohms
Cap 5: 33µf 63V measured at 10kHz: 27.7µf 0.80 ohms.....particularly this one.

They're related to proper operation of the SMPS and probably both across the Vcc of the controller circuitry.
If they're not right it won't kick off.

[bitseeker]

I've heard of caps that, despite seeming to be OK, would not work properly in a particular circuit. It could be that they were leaky or some other characteristic besides capacitance and ESR affected their behavior. It's also possible there was a bad solder joint, as you said.

[nickeevblog10]

  Not sure I have an idea here or not about the readings on meters for components when testing but, what may be a reason the op got multimeter readings that did not seem to be correct is perhaps because unless the meter is turned off after taking a reading and then turned back on before a second reading is taken the second reading may not appear to be correct because of stored power in the meter measuring process.
My UTI-61E automatically OL's but my first digital multi meter (NZ$20) seemed to not read consistently correctly and this seemed to be the reason I should get the new UTI-61E meter.
After I realized the Uni-T would OL automatically after each reading I began to use my $20 digital multi meter again but this time around turning it off after a single reading and readings came back to expected levels.
Lesson learned.

[W0MPM]

Bet you've got some Loctite thread locker goop nackumpucky in the shop somewhere. That should do it for the shaft. Loctite even makes a specific blend for shafts but I can't remember what. Plus hours of reading the data sheets and it still never was clear to me what, exactly, was the difference in some of their stuff. In that application and orientation even gasket maker would work and has the advantage that you could pull it apart need be.