DIY Electric Car Forums banner

141 - 160 of 353 Posts

·
Registered
Joined
·
1,148 Posts
I must be the only schmuck attempting to repair these chargers at the moment. I got all the way to testing with a real battery recently, when I decided to pull the plug before it connected (last time when it turned on the output relay was when the MOSFETs blew again). I did this because I heard the soft cracking sound that my friend (who is kindly loaning the battery, part of his EV pack) thought he heard last time. Last time, I wasn't sure if it was just the input relay doing its normal thing, but this time we both agreed that the crackling sound was back.

He suggested we do a "binary search" for the cause of the sound. We disconnected the battery, since this would allow us to leave it turned in without actually connecting and possibly failing again. Using a piece of tubing as a crude stethoscope, we listened in one half, then the other. We were able to determine that it was on the PFC (mains) side, not the output side, before it stopped happening. My friend thought it might have been coming from the input relay.

I formed a theory about how the DC bus voltage might be too high, so the PFC chip might be shutting down due to over-voltage (there is a divider chain dedicated to this purpose, which seems to trigger at 425 VDC design center). This was aided by me misremembering the DC bus measurement I did at home; I thought it was 460 V. (The bus capacitors are all rated at 400 VDC, so that would be bad long term, not too bad short term.)

But I re-measured, and it's around 392 VDC with either 240 V mains, or with 48 VDC. That's 98% of the 400 V rating of the capacitors, but I believe that they are designed to run at 385 VDC (when running off 220-240 VAC mains), which is over 96% of their rated voltage anyway. The MOSFETs are rated at 550 VDC, so the extra 7 V isn't going to cause the MOSFETs to blow.

I decided to take a closer look at the input relay itself. As you can see, there is a small darkened spot near where the contacts are:



The yellow gunk is from the manufacturer, to keep the tall parts from being affected too badly by vibration. There is also printed circuit lacquer from my having conformal coated the whole PCB. The darkish spot circled in red is hard to see at most angles; that photo happened to capture it pretty well. So is it significant?

This is inside the relay cover:



It's starting to look a bit more serious, but maybe relays always do that.

Here is a closeup of one of the contacts:



You can see that it's taken a bit of a blast at some point. It would have opened under load twice when the MOSFETs blew; there may well have been significant over-current due to the shorted DC bus each time. The input relay always sees AC (unless you run the charger from DC; my testing at 50 VDC was always at quite low current, 2 A maximum). The relay contacts are rated at 250 VAC and 16 A, or 30 VDC and 16 A.

Any opinions on whether the contacts are seriously damaged? They seem to measure low resistance and with a reliable connection, as judged by a cheap multimeter on the beeping continuity setting. (I have a Fluke that removes the "scratchy" sound you get from cheap multimeters when lightly rubbing the probes together; in this instance, this feature is unwanted).

In particular, should I carefully and lightly sand the contacts with very fine sandpaper? This could remove some of the blackness, but may also remove some of the miracle coating that the contacts may have.

The input relay is an Omron G2R-1-E with 12VDC coil.

The way to remove the relay cover (after removing as much as possible of the yellow gunk) seems to be to grasp it firmly with a large pair of pliers. This has the effect of making it bulge slightly, which releases the two clips at the bottom (which are difficult to reach due to the tall components nearby). I suggest replacing the yellow gunk with silicone when all is done.
 

Attachments

·
Registered
Joined
·
44 Posts
I'm the friend Coulomb mentions in the previous post. Well spotted, Coulomb. I can well believe that subtle arcing of that burned contact when closed, could have been the source of the sound I heard.

I remind you that, although the sound went away suddenly while we were still trying to narrow it down, that relay was in the quadrant that we'd narrowed it to, and when I listened right at the relay (and only there) I could hear a very quiet sound that was of the same character as the louder crackly sound we had both heard.

I suggest you get your daughter (with younger ears having better high frequency response) to have a listen to that sound, through a plastic-tubing "stethoscope", assuming it's still there. Then either clean the contacts, or better-still, replace the relay, then ask her if the sound is still there.

I was previously thinking it was the relay _coil_ acting as a speaker for a noisy power supply, but the idea that it is the contacts, has the benefit of a ready explanation of why it suddenly went much quieter, i.e. making better contact by semi-welding itself.

However, I have no idea how a noisy 240 Vac relay contact could possibly lead to MOSFETs blowing up. As you suggest, it might well be a _result_ of the blowups rather than a cause of them.
 

·
Registered
Joined
·
1,486 Posts
Discussion Starter #143
Howdy Mike, i'm busy with OEM bms cards right now, but crackling sounds from AC relays is not good. i use ultra fine paper for polishing metal, aka crocus cloth, for cleaning relay contacts.

AC has a way of eroding metal. In a sliding contact only the highest spots of the surface make connection and carry the current. When the current exceeds the point-contact thermal limit then a micro-fusing event occurs which vaporizes the metal and transfers the connection to the next highest spot on the face. Over time this erodes and discolors the contact surface and is the likely source of the crackling. If the contacts are held open by a small charred residue or debris from the fusing, then the ac will arc across in order to keep current flowing. This micro-arcing also erodes the surface and makes a crackling sound. On motorcycle regulator/rectifier boxes this occurs at the spade lug terminals from the alternator and usually melts/burns/chars the plastic connector housing--lots of fun to fix those too... cheers mate, kenny
 

·
Registered
Joined
·
1,148 Posts
I remind you that, although the sound went away suddenly while we were still trying to narrow it down, that relay was in the quadrant that we'd narrowed it to, and when I listened right at the relay (and only there) I could hear a very quiet sound that was of the same character as the louder crackly sound we had both heard.
Well, it turns out that the sound was not from the relay at all (contacts or coil), but from the 2.2 uF capacitor right next to it. It is best heard through a tube "stethoscope" aimed right at the middle of the largest face of the capacitor. This was vastly easier with the capacitor soldered under the PCB:



When we removed the old capacitor, we wondered about the discoloration in the white gunk (as opposed to the yellow gunk and the black gunk! :rolleyes:) under the capacitor:



But we decided that this wasn't a contributing factor.

There is another of these 2.2 uF MKP polypropylene capacitors across the mains input. We swapped the two capacitors, and found that the other capacitor made the same soft sound: a faint crackling. We left the charger running for an hour or two, and did not notice the louder version of the crackling. So the louder crackling seems to be an intermittent feature of the original PFC capacitor (the one between the relay and the bridge rectifier).

I ordered new Epcos replacements (now owned by TDK), and soldered them in place. I now could not hear any sound, but was this because it was so hard to get to the large face where the soft sound was coming from? I thought to solder one of the new ones underneath the PCB, but then it might be different because I had the new capacitor in place already. Removing the capacitors and especially cleaning the holes of solder is a considerable pain. So I soldered the *old* capacitor under the PCB first. I could hear a faint crackling, but it seemed much fainter than at Weber's place, where there wasn't a new capacitor soldered across it. Now I replaced it with a new capacitor under the PCB (still with the other new one on the top of the PCB), and found that the new capacitor also made the same faint crackling sound.

So these capacitor seem to be slightly acoustically active. I tried searching for this, and found that some ceramic capacitors act a bit like piezo electric sounders, and that there are ways of circumventing this problem. Of course, I found lots of pages about amplifiers causing crackling *through the speakers*, but that doesn't seem to be relevant.

Anyone else come across acoustically noisy capacitors before? (Apart from ones that have failed and oozed goo or the like, of course.)

Edit: Just to be clear, my tentative conclusion is that the faint crackling seems to be normal, though some brands or models may be noisier than others. The intermittent, louder crackling that could clearly be heard without the stethoscope, seems to be abnormal, and I hate to think what is happening when it comes on.

Edit 2: The more I think about it, the more I think that there wasn't anything wrong with the original capacitors, and that the intermittent loud crackling is due to intermittent behavior of the PFC stage. I wonder if it coincides with control tones on the mains, or if the PFC stage intermittently just goes mad. If the latter, it might explain the MOSFETs blowing up, through a higher than normal DC bus voltage.
 

Attachments

·
Registered
Joined
·
1,148 Posts
It's possibly obvious, but since it just happened to me, I thought I'd point out a problem that can happen when reassembling the heat-sink clips:



If you managed to leave it in this condition, the heat-sink clip would short out all the MOSFET drain connections. This is considered harmful :eek:

The clip needs to end up on the black epoxy of the MOSFETs.
 

Attachments

·
Registered
Joined
·
383 Posts
It's possibly obvious, but since it just happened to me, I thought I'd point out a problem that can happen when reassembling the heat-sink clips:



If you managed to leave it in this condition, the heat-sink clip would short out all the MOSFET drain connections. This is considered harmful :eek:

The clip needs to end up on the black epoxy of the MOSFETs.
Oh crap! I never considered someone making that mistake or the implications of clipping them together. I noticed when I took them apart that the clip went on the plastic black part of the MOSFET's but I wasn't thinking this would connect the drains.

Did it blow up on you?
 

·
Registered
Joined
·
1,148 Posts
Did it blow up on you?
No, this was before I tightened down the screw, and fortunately I noticed it. Usually, there is little to no gap between the heat-sink clip and the heat-sink "wall", but this time there was about 1/8" (3 mm).

If I hadn't noticed, I might have screwed down the heatsink, which might have forced the MOSFETs down, possibly buckling their leads. The heat-sink clips obscure everything from above, so even this amount of mangling might go unnoticed to someone less familiar with these chargers. Hence the post.
 

·
Registered
Joined
·
1,148 Posts
I'm at Weber's house testing the repaired charger (so far so good).

But before charging, we decided to check for crackling sounds (using his far superior ears). Yes, the new after-bridge capacitor was making a very similar sound to what the original capacitor was, when it wasn't being very noisy.

But before that, at first switch-on, he yelled to turn off the power, he saw a puff of smoke! Arrgh - what now? All I had done since last test was to change the capacitors and also the relay. Oh wait - to get to the relay I removed the pre-charge resistors, and part of the pre-charge resistor always seems to come away with the yellow gunk. So I replaced them with new ones. New wire-wound ones, because I could never understand why the manufacturer stuck with carbon composition types* that burned up so badly whenever the power supply failed (often due to other failures). We suspected the new resistors, so we powered up again and noticed a puff of smoke from the resistors, but it didn't continue. Obviously, when the relay turns on, the pre-charge resistors are shorted, so they then dissipate no power. Sure enough, every time we started the charger, they would emit a little puff of smoke, but otherwise appear to be OK.

I then realized that each resistor is seeing the full 240 V at start-up, so that's initially a power of E^2/R = 240x240/150 = 384 W. Wow, that's a big overload for a 2 W resistor, just for a short period of time. Maybe the originals were 3 W or 4 W, but even so, it's obvious that high peak power types are required there. Ordinary wire-wound resistors are never going to achieve that.

Maybe the resistors can handle it; they seem to have the same resistance after a half dozen starts. But it seems like a bad idea.

I'm not sure what to replace them with. Perhaps two 5 W high pulse power resistors (these are wire wound, but with epoxy and a heat-sink). Or carbon composition resistors with a slow-blow fuse in series (say a 1 A 20x5 type). But it would take a lot of trial and error to get the fuse rating right, not causing nuisance blowing and yet still protecting against massive overload. Maybe a thermal fuse would be better than an electrical (current based) fuse.

* Edit: see post after next; it seems they aren't carbon composition types as I had guessed. Also, it seems that some wire-wound resistors are good at high pulse power too.
 

·
Registered
Joined
·
383 Posts
Yes, kennybobby and I did that calculation on the first couple we repaired. They all had these resistors burnt to a crisp. We puzzled over a way to power up the viper before the relay closes instead of going through these resistors but we never tried anything. It's a very common failure since any problem with the viper power and that relay opens and they get full power. Maybe they are meant to fuse but seems like the wrong component for that.
 

·
Registered
Joined
·
1,148 Posts
... I could never understand why the manufacturer stuck with carbon composition types that burned up so badly whenever ...
It seems that my guess was wrong. Here is one of the original 150R pre-charge resistors split apart with a chisel:



I should have realized that they weren't carbon composition by the end caps**. The white material in the middle is totally non conducting, likely ceramic. You can see the spiral grooves that they use to trim the resistance of metal or carbon film resistors. By the dark color, and the smell when they burn up, I'd guess they are carbon film types, but I would not know.

So I think that the original pre-charge resistors aren't particularly specially chosen for their pulse power rating; perhaps they merely use coatings and paint that can tolerate the high temperature extremes from a pulse of power.

[ Edit: after a little reading, I think they may be metal oxide film resistors, which are good but not the best for high pulse power. ]

[ ** Edit 2: It seems that some composition resistor types also have end caps, especially ceramic composition types. So that's not the way to rule in or out whether a resistor is a composition type or not. ]
 

Attachments

·
Registered
Joined
·
383 Posts
That's good to know. I still think a fuse or some type of protection is needed there for when the relay fails to close to keep them from frying. Or another way to let the viper come up to full voltage before closing the input relay.
 

·
Registered
Joined
·
1,148 Posts
I finally settled on these replacements:

http://www.digikey.com/product-detail/en/ohmite/OX151KE/OX151KE-ND/823910



They're only 1 W continuous power, but that's because they're ceramic composition types (more or less the modern equivalent of carbon composition resistors), which use the whole body as the resistance. So they can't get rid of the heat from inside the core of the resistor as well, but that makes them ideal for absorbing energy pulses. At 250 V, the resistors need to absorb 50 J between them, and these can do 50 J per resistor. They are also the same color and about the same size as the originals; the only difference is the 10% tolerance compared to the originals at 5%. But exact resistance value is not important, as far as I can tell.

They're about 100x as expensive as the wire-wound types, but still only a few dollars for a single charger repair.

But I don't know what they are like for overload. They could well hang on for dear life. So I might investigate some thermal fuses as well. Ah, they could replace the longer leads of the pre-charge resistors, like the attachment.

The middle pads, which normally take the longer leads from the resistors, connect together. All three components could be siliconed into a blob for mechanical strength, insulation, and thermal connectivity.
 

Attachments

·
Registered
Joined
·
838 Posts
Is there some way to test what voltage my charger is set to prior to attaching a battery? I haven't hook up a computer to it yet, but I'm thinking it will say the same thing my multimeter did. It was 0.24V or something close to that at the Anderson connector. If I hook up the computer can it be set there?

I will have to do a lot more reading on setting this up and the right way to charge these batteries.
 

·
Registered
Joined
·
383 Posts
Is there some way to test what voltage my charger is set to prior to attaching a battery? I haven't hook up a computer to it yet, but I'm thinking it will say the same thing my multimeter did. It was 0.24V or something close to that at the Anderson connector. If I hook up the computer can it be set there?

I will have to do a lot more reading on setting this up and the right way to charge these batteries.
I assume you have an Elcon / TCCH charger as they are called sometimes.

There is usually a label on the side that has 10 curves to choose from like this:

The unit will flash the led the number of times to tell you the curve its using. To change it you hold down the button when powering up till it flashes the number for the curve you want then let go of the button.
 

Attachments

·
Registered
Joined
·
1,148 Posts
Is there some way to test what voltage my charger is set to prior to attaching a battery?
Occasionally, there won't be a list of settings on the side. If so, there is unfortunately no easy way to tell what it will do. 99% of the time, you will get a clue about either the nominal or the absolute maximum voltage the charger will put out from a separate label.

If you have the programming hardware per this post, then you might be lucky and not have the security bits set, so you can read the firmware. Someone like myself can tell you what the settings are from that. You don't need to open the charger for this, just remove a label.

If you have no idea what voltage level the charger is set for, you could open it up and hope that they have marked the high frequency transformer with the *nominal* voltage of the charger. For example, it might be marked "120 V" or "288 V 13:7:8". The former would be nominally 120 V, able to charge up to about 168 V. The latter will go to about 389 V; see the TC or Elcon charger web site for details. The transformer seems to be marked about 95% of the time.

Edit: there is also the chance that it's a CAN bus version. In that case, the voltage and current are set externally.CAN bus models will have a small "dongle" to convert serial signals to CAN bus and back again.
 

·
Registered
Joined
·
1,148 Posts
It seems I never posted about the measurements possible without removing the PCB. I'll place them here.

It's hard to jam a multimeter probe at the bridge leads, at least while the heat-sink clamp is in place. Fortunately, the line end is accessible through the junction of the two long ends of the pre-charge resistors. So to measure the value of the pre-charge resistors, you can measure between there and the line input. This includes the resistance of two line chokes, but their resistance is negligible. It should measure 75 ohms, give or take about 10%.

Positive output from the bridge rectifier is easy to get to; it's the pin that's a larger distance from the others, towards the edge of the board, and accessible via a large pad.

Negative output from the bridge is essentially "GND"; there is a marked pad (it might be under some yellow gunk), and it also makes it to the output side of the board via the top side red jumper.
[ Edit: on the 2 kW chargers with three capacitors on the mains side, it is in about the same position relative to the large PFC inductor. In other words, the extra capacitor is nearer the yellow power supply transformer and power supply chip with the small copper heatsink. ]

The other AC input lead to the bridge rectifier connects essentially to the Neutral input.

So now the four pins of the bridge rectifier are accessible: L and N are the AC inputs, (the 75 Ω pre-charge resistors won't affect diode measurements much), negative output is "GND", and the positive output can be accessed by the large pad.

[ Edit: the positive output of the PFC stage, the "DC bus", is easiest to get to near the big capacitor on its own, where there are several vias, as shown. ]

2018/April: Finally, more of an observation than a measurement. If the processor is flashing its small red LED or the red/green LED, then you know that the 12 V power supply is working. If the input relay pulls in, then you know that the 15 V power supply is working (presuming non-welded contacts). Both of these tests can be done with a ~52 VDC power supply applied to the mains input.

[ Edit: Forgot to add the images; added comment re 2 kW chargers; easier access to DC bus plus; added comment re pre-charge resistors not affecting bridge diode measurements ]
 

Attachments

·
Registered
Joined
·
1,148 Posts
Now some repairs possible without removing the PCB. This one is replacing the 150 ohm pre-charge resistors, which often get burned up through various mishaps. In my case, I needed to replace the wire-wound resistors (from a previous repair, actually just getting clearance to replace the input relay) with ceramic composition types, because the wire-wound types were emitting a little puff of smoke with each switch-on of the charger. During the first 100 ms or so, these resistors are required to take up to 50 Joules between them, with a peak pulse power of nearly 200 W each.

I also wanted to try adding a slow-blow fuse, as posted recently. Back then I suggested a thermal fuse, but a slow-blow fuse of suitable rating should be about as good. I chose a 750 mA part; it seemed to be a good compromise between not nuisance blowing with the large surge of current, and blowing in 10 seconds or so if the resistors end up across the mains.

You can see the before and after photos in the attachments. I did the whole repair from above the PCB, but I found that after it was done, the fuse no longer connected to the bridge rectifier. So I was forced to take the PCB off after all :mad: It turns out I damaged the pad under the fuse. It looks like it would have been better to use the hole that is slightly further away from the edge (i.e. the one for R1, not the one for R23). It has a reasonable sized track attached to it, so it may survive the rigors of clearing solder from the pads better than the other pad. (Both middle pads connect to the same place; there is a small sliver of track between the two middle pads.)

You can also see one of the hassles of repairing from above the board: the shiny plastic gets in the way, and it's nearly impossible to avoid burning it with the soldering iron. (Hence the hole burned in it.)

Edit: the ceramic composition resistors seem to work well. With each switch on from 240 VAC, there is about a 5°C temperature rise, which dissipates over a minute or two. The 750 mA slow blow fuse also seems suitable, but really only time will tell if it will nuisance blow too soon.
 

Attachments

141 - 160 of 353 Posts
Top