The Troubleshooting and Repair for On-board Charger (OBC) Thread

[electronpusher]

Additionally, I had a look at the riser board, there is no way to mount the caps on there, however I was thinking of designing a new riser board that will accept through the hole caps, thus making future repairs easier. Basically you would remove the potting once, replace the riser or through the hole caps that were there with the new riser, and solder the caps to the new riser board that supports through the hole caps. Any future replacement of the caps would only require the damaged ones desoldered and new caps soldered in, bypassing the need to remove the waffle plate.

Thoughts?

 

[coulomb]

I think you'd want to re-pot the (new) vertical PCB, so I don't see much advantage unless the capacitors were up high out of the main potting. The extra lead length might introduce too much inductance. Though the existing PCB would have nearly as much inductance.

Perhaps really thick tracks, perhaps made from double thickness copper, would be equivalent. I'd want to secure the capacitors to the vertical PCB though.

 

[electronpusher]

I think you'd want to re-pot the (new) vertical PCB, so I don't see much advantage unless the capacitors were up high out of the main potting. The extra lead length might introduce too much inductance. Though the existing PCB would have nearly as much inductance.

Perhaps really thick tracks, perhaps made from double thickness copper, would be equivalent. I'd want to secure the capacitors to the vertical PCB though.

True, but I figured it would not be any worse that what was already there, and would make future repairs easier. I would definitely repot it if I put in a riser. I was thinking of designing and sending for a professionally produced PCB with screen printing, not something I etched myself.

 

[electronpusher]

Ok, removed the potting around the riser, and had a go with some solder wick to remove the riser board, without success. I think the waffle has to come off. Thoughts? Any suggestions on how to remove the riser without removing the waffle?

If the waffle has to come off, I will take it to uni where I have access to a re-work station, something I do not possess.

BTW, I am video recording everything I am doing, so that we can create a how to repair video.

 

[skylogger]

Electronpusher: Probably the reason you are having difficulty with desoldering the riser PCB is that you have to remove solder in 8 places, not just 4. Out of each of the 4 holes coming out of the main PCB, there are 8 pins, 2 pins come out of each hole. One pin solders on top side of riser PCB and one pin solders to bottom side of riser PCB. Since the riser PCB was so badly damaged, I didn't have to baby it too much. You could even use wire cutters to chew at the riser pcb leaving just small sections tied to the pins, allowing to flex and torque them a bit to get the pins to free/desolder from the PCB one at a time. Then you will end with two pins coming out of each main PCB hold forming a fork. It is easy
to resolder a replacement through hole cap lead by lead forming and running the lead between the two pins of the fork then wraping it around. I just made sure I did not over heat the pins so they did not fall back down the hole when joining the new cap leads.

If you use KIEV's schematic of the waffle pins, you can use an ohm meter in diode check mode, you can check out 90% of the semi's on the waffle. If you run through the 2 x 4.7R resistor and relay check in the doghouse, you might run across same problem that I had.

 

[electronpusher]

Thanks @skylogger for the info.

I was hoping to save the riser, but I should be able to take the measurements I want off it first, before cutting it up for removal.

Finished up for the night, but tomorrow night I'll give it a go, and also check the semi's on the waffle and relay.

Still waiting on that fuse from Mitsu so I am not in too much of a hurry right now, I intend to order the solar fuse/holder tomorrow for testing, as I do not want to blow the 4 week wait Mitsu OEM fuse while testing the repairs.

Thanks for all the help.

 

[ChristopheFR]

Hello tout le monde,

I took pictures this week-end.

We see traces on the cover, I think it is not due to a short circuit, but just a significant heating of the material just below.

1 of the three capacitors is inflated on the top. Are the other two still in good condition?Do you think it is better to change the 3 pieces?

The two resistors 4.7 ohm seem in good condition, there is no trace of burn.
Ditto for the two blue capacitors

So I relized the “skylogger test" for the resistance of 4.7ohm
Is this a correct connection for my ohmmeter (green wires)?

and there, result: my ohmmeter reads 0L, no resistance.
Is it the larger resistor that fused? due to overcurrent or overvoltage? No apparent damage on this resistors!

Skylogger, do you think for the second test with 5 volts on the terminals of diode D301, that a 4.5volt battery like this one is enough to carry out the test? Is this correct for the wire connections?

what do you think of this failure? A problem due to a storm? A problem due to the capacitor that has inflated?

reminder: my dc-dc charger for the auxiliary battery still works, and I changed the auxiliary battery in 2015 when it lost power

 

[coulomb]

We see traces on the cover, I think it is not due to a short circuit, but just a significant heating of the material just below.

It looks like this discoloration is common, sadly.

1 of the three capacitors is inflated on the top. Are the other two still in good condition?Do you think it is better to change the 3 pieces?

Difficult to see in your photo. Presuming that one really is bad, then yes I would replace all three.

The two resistors 4.7 ohm seem in good condition, there is no trace of burn.
Ditto for the two blue capacitors

Yes, unlike many charrgers so far. But the 4.7 Ω resistors may still be bad, see below.

Is this a correct connection for my ohmmeter (green wires)?

It looks right to me, but others please chime in.

and there, result: my ohmmeter reads 0L, no resistance.

Actually, "OL" represents OverLoad, which is infinite resistance, i.e. zero conductivity. Either one of the 4.7 Ω resistors is open circuit, or the test isn't valid. Energising the relay will hopefully show which.

Is it the larger resistor that fused?

They are in series. So it's not possible to tell which one failed, or possibly both.

due to overcurrent or overvoltage? No apparent damage on this resistors!

It's certainly unusual. Unless the resistors somehow saw 240 V across them; in that case, the fuse inside one of them would blow quickly, before the resistors overheated and discolored.

Skylogger, do you think for the second test with 5 volts on the terminals of diode D301, that a 4.5volt battery like this one is enough to carry out the test?

That battery should be fine.

Is this correct for the wire connections?

NO! The stripe (on the diode and also on the silk screen overlay of the printed circuit board) is for the cathode of the diode, which is the negative end if you want the diode to conduct. But in this case, with no current limiting, we don't want the diode to conduct, except briefly as the power is removed from the coil. This gives a path for the current to continue flowing after the battery is disconnected, so there isn't a large voltage spike that could damage the other electronics (though the other electronics on the upper board should be disconnected for this test). [ Edit: so you need to reverse the battery compared to your photo; battery positive to stripe on diode; battery negative to non-stripe end of diode. ]

With the relay energized, you should see much less resistance across your green wires, reading "0.0" or perhaps "0.2" due to the resistance of the multimeter leads and green wires.

If you don't see this, and it remains at "OL", then either one of your resistors is open circuit and the relay contacts are open circuit as well, or your green wires aren't right. Another possibility is blown PCB tracks or badly soldered quick connects (etc), but these seem very unlikely.

what do you think of this failure? A problem due to a storm?

I think it's too early to tell.

A problem due to the capacitor that has inflated?

I don't think so. Those three large capacitors only work at 100 or 120 Hz, not at several kilohertz switching frequency. So they can degrade a little and still perform adequately. If they went low capacitance, there would be higher than normal 100 or 120 Hz ripple on the ~400 V DC bus, but I don't see that causing the pre-charge resistors to open circuit.

 

[redcane]

An update on failure #14 - our charger was replaced as a recall in January 2017, so it's been in service less than 2 years. It's dated 2015 manufacture.

 

[kiev]

The blue 1000pF snubber caps are rated 2kV and are ~9.3mm diameter. i have a bag full of some 3kV @ 12.0mm, and 6kV @ 17.3mm, that would provide a bit of HV margin--will make these available in a repair kit if you can't find a local source.

 

[skylogger]

ChristoferFR:
It's a good thing Coulomb pointed out that you have the battery polarity wrong. Connecting it the way you show, will blow the diode, and that particular diode is a bit hard to identify. I think we might suggest putting a resistor in series with the battery in future documentation on this test, so if someone does get the polarity wrong, it will current limit and not blow up the diode. I can see people naturally think to connect battery for forward bias, but for this test, we want to make sure the diode is reverse bias during the test.

The green wire on the left of your photo is in the correct place. It is hard to see from the photo if you have the green wire on the right connected to the proper NEUTRAL blade terminal or not. You can see the "N" marked in the silkscreen to make sure you have it connected to the correct blade terminal. The "OL" reading that you are getting means that either you are not at the correct test point, or that one of the two 4.7 resistors is blown open. If you accidently connected the green wire to the "Active" In stead of the "Neutral" then you would get the "OL" reading due to the wrong test point. If you have the correct test points, and BOTH resistors are good, you will see a value just slightly higher than 9.4 ohms. When you apply voltage (with correct polarity for reverse bias), The 9.4 ohms will change to close to zero, showing that the contacts of the relay are shorting out the resistor values. You should also hear the relay click.

4.5v should operate the relay. just make sure the battery is new, so you don't drop lower than that. When I tested mine, I used a 3 x 1.5v AA battery pack but used new fresh batteries. Make sure you can hear the relay click so you know it is activating the contacts.

If you find the resistors and/or the relay is faulty, then you will need to desolder all the pins along the two while strips, so you can remove the waffle from the Main PCB. Then you will have access to the bottom of the PCB to be able to rework/desolder the relay and resistors. I had a go at delsolder the waffle, and could not acomplish this with just solder wick. You will probably need to find someone with a proper solder sucking rework station to do this.

 

[ChristopheFR]

Hello tout le monde,
Thank you for ansers,
Mea culpa, I publish a corrected version of the relay test with good connections.

Question : good idea with the resistor, wich value ? and where in series?
On the side of the negative pole of the battery or on the side of the positive pole of the battery?

When I connected the ohmmeter with the white wire, I got OL. So one off the resistors is fused
When I connected the ohmmeter with the black wire, I got 8.5 Mohm . So my ohmmeter works!

Next time, I will realize the relay test.

if the relay is ok, that means that I could only change the resistors, right?

My idea: to weld the two new resistors on a small PCB, to fix them together in a little doghouse , and to connect them with big wires to both N of the main PCB (the same place as the first Skylogger’s test)
Advantage: no need to unsolder the waffle :lol:

Opinions on my idea? Are the PCB tracks thick enough to support amperage?

My problem at this moment : I have not found a seller for Tamura resistor references on Internet? Someone can help me?


Ah, I forgot: I am completely neophyte in electronics! you guessed it?

 

[skylogger]

ChristophFR: I'm still not sure that you are on the correct test point for your green wire on the right of the photo. I had a look at the photo again, and it looks like you have the green wire shoved into the female blade connector on the white wire, but its not clear what do you have the other end of the white wire going to? It looks like just below the white wire the male blade connector on the PCB looks like the white wire is not plugged into it. The male blade connector on the main PCB is the actual test point you want to use. Maybe you can send another photo a bit clearer. You need to connect the green wire to the male pin on the main PCB where the wire wire was connected. This is with the white wire unplugged and not connected to the top controller PCB. With the top controller PCB unplugged, If you measure between Neutral and ACTIVE "L" lines, you should see a "OL" or high reading. your last post said your reading was 8.5 mOhms which means there is a short between Active and Neutral, OR you are on wrong test point.

Your idea of connecting a resistor across the two Neutral testpoints to bypass the faulty resistor will not work. Have a look at the schematic on page 3 picture 9 of this thread. The resistors are in series with the donut inductor, so you can connect one point to one test point, but the other point is still on bottom of PCB where it connects to inductor.

After looking at trying to calculate a series resistor to be used during the relay test, It might be a problem. It needs to limit current enough to protect the diode if things get connected backwards by accident, but it also needs to allow enough current to still drive the relay. From the specs on the relay, I see the coil is rated at 400mw it has a resistance of around 68R . I don't have the specs on the diode to know what its forward contInious current max rating is. Might have to ask Coulomb if this is even possible with a 4.5v source, because a 100R value would protect the diode, but voltage drop would be too much to still be able to activate the relay. The relay needs a minimum voltage of 3.5v If you dropped resistor value to 20 ohms so the relay still has enough voltage/current to activate, it would be too much forward current for the diode if it was connected backwards, so it would not protect the diode.

 

[coulomb]

Your idea of connecting a resistor across the two Neutral testpoints to bypass the faulty resistor will not work. Have a look at the schematic on page 3 picture 9 of this thread. The resistors are in series with the donut inductor, so you can connect one point to one test point, but the other point is still on bottom of PCB where it connects to inductor.

Perhaps you can get to the inductor's neutral wire inside the doghouse, where it connects to the PCB. The neutral side seems to be the side nearest the two pre-charge resistors. You seem to need to access the end away from the edge of the PCB. But it's starting to get a bit complex.

Might have to ask Coulomb if this is even possible with a 4.5v source, because a 100R value would protect the diode, but voltage drop would be too much to still be able to activate the relay.

Yeah, not possible. I'd skip the series resistor, or use a higher voltage battery. If you had a 12 V battery, then a 100 Ω 1 W resistor in series with either the positive or negative leads to the battery (doesn't matter which) would be about right, assuming a 62.5 Ω relay coil. That resistor will get quite hot quite quickly, so perform the test quickly. Perhaps use two 220  Ω half watt resistors in parallel, or a pair of 47 Ω or 56 Ω resistors in series; whatever is handy. A 100 Ω 5 W ceramic resistor would be perfect.

The ideal tool for this is the current limited power supply. Set it for about an 80 mA current limit and 5 V voltage limit, and you can't go too far wrong. I have a digital one (it has digital display of both voltage and current), where it's quick and easy to set the limits accurately. But a lot of hobbyists don't have one, and can't justify the price tag for the once or twice a year that they would use it.

 

[electronpusher]

Hi All,

Question regarding potting compound and heat transfer compound.

Will the following work for our situation?

Potting Compound:
https://au.element14.com/robnor/el171c-566/resin-polyurethane-twin-pack-566g/dp/725973?st=potting
https://au.rs-online.com/web/p/potting-compounds/0851044/

Heat Transfer Compound
https://au.element14.com/electrolub...-silicone/dp/634920?st=heat transfer compound
https://au.rs-online.com/web/p/thermal-adhesives/0503357/
Thanks.

 

[coulomb]

Will the following work for our situation?
Potting Compound:
Heat Transfer Compound

I really don't know about the potting compound, and I think you might be supposed to buy other tools for mixing some of them.

I'd be really tempted to just use neutral cure silicone sealant from Bunnings. I know it's a lot softer than potting compound, but I think it would be find for supporting a pair of through-hole capacitors and a small printed circuit board.

The heat transfer compound you linked to seems expensive for what little you get. I've been using a 150g tube from Jaycar, branded Unick. I seem to recall it was of the order of AU$20. It seems to have been replaced by this, which seems just fine:

https://www.jaycar.com.au/heatsink-compound-150-gram-tube/p/NM2012

These brands are very Australia-centric. I'm sure equivalents can be found in other countries.

 

[skylogger]

Hi Electronpusher:

I agree with Coulomb on his suggestions on both the heat sink compound and the potting.
For the potting, I would definitely stay away from epoxy mix, as if you ever need to re-work again
in the future, there would be a lot more drama's digging through a hard epoxy rather than a
rubbery silicone type potting. Standard natural cure silicone would buffer vibration, non-conductive, and be a dust barrier.
For the heat sink compound, Both Jaycar and Altronics have smaller quantities of the general purpose white pinquin poo.
I'm not familiar with the other red metal oxide type so I'd stick with the standard white type.

 

[electronpusher]

@skylogger & @coulomb, thanks for the advice, we'll go with the JayCar heat transfer compound, and the natural cure silicone.

We got the fuse from Mitsubishi yesterday, I have solar fuses and fuse holders shipping which we should have tomorrow which I will use initially for testing as I do not want to wait another 4 weeks for a Mitsubishi fuse should it blow. I'll put the Mitsubishi fuse in when I am happy that the problem is fixed.

Hopefully she will be back on the road this weekend.

 

[electronpusher]

I removed the riser board, had to cut it, I managed to save the pins, and soldered the new caps directly to the pins (2 pins per leg), and made sure they were secure.

I am now testing the other components, about to test the semi's in the waffle board and the relay.

When testing the relay, would using a bench top power supply at 5V be ok?

 

[coulomb]

When testing the relay, would using a bench top power supply at 5V be ok?

Yes, ideal, especially if it has a current limit ability. Set it to about 200 mA if you can. Just in case you get the polarity wrong, to save the diode.