Roadster PEM failure

[warden_clyffe]

You can clean the PEM without disconnecting anything. With the car up in the air, you can remove the cooling hose and access both the inlet and outlet ports on the PEM. The cooling air just blows across the heat sinks on the bottom of the PEM, not the internal electronics, so you can get in there with a blow gun and not worry about damaging anything.

...hmm, this is a half done job then. the problem with dirt in the pem's is not the visible dirt (outlet), it's all that crap in the inlet where you don't get proper access without disassembling the megapole-block (black plastic base) and for that you surely have to take out the pem

 

[ML Auto]

...hmm, this is a half done job then. the problem with dirt in the pem's is not the visible dirt (outlet), it's all that crap in the inlet where you don't get proper access without disassembling the megapole-block (black plastic base) and for that you surely have to take out the pem

Once you remove the inlet hose you have a view of the inlet port. Blowing air into the outlets flushes the debris back out of the inlet. It is exactly what you do with the PEM removed, you just have to work around the motor to get at all the outlets. You would have to unbolt and raise the PEM a little if you want to get that little black cover off, but you can leave the connectors attached. I never found anything in there after I finished blowing air through it, so now I leave it alone.

 

[BridgetCK]

I agree most PEM failures seem to be due to damage of the connectors - ie sight misalignment when pushing back in - which can damage or push back pins - resulting in poor connections, overheating etc etc.

Reconnecting th e plugs is normally done 'blind' as they re under the PEM so these mistakes are easy to make.
The PEM only needs disconnecting to clean it - so obvious suggestion is to stop dirt / dust from entering in th 1st pace.
My solution to this is a large area filter in wheel arches (small area filters restrict flow and may add to any over heating IMHO).

.

How can we tell if the connectors are damaged. We just had all the batteries replaced and am instantly getting overheating related to the PEM.

 

[DanielFriederich]

What error Codes do you get in debug mode?

 

[csummers]

Silver Bullet #302 update. The car came back in June from Gruber and did not see any alerts or errors in first 5 days driving. Then at first home charge got the same fault as before. Contacted SC and Gruber again. More back and forth for a few weeks. SC said fuses were installed backwards, Gruber said they are multi-directional. then finally got sent back to Gruber to look at PEM again. Latest is that some of the micro switches are failing preventing charging. Now going on six months in the shop. Luckily have ICE backup car. Since the AC is so underwhelming not really missing it much in the TX heat

 

[Fuzzylogic]

Thanks to all who fixed their PEMs and explained what needed to be done!

My roadsters PEM also had crumbling insulation material, and had shorted to the heatsinks on phases 1 and 3, while charging the car.
three IGBT's had burn marks on them, i replaced them, although they appeared to be still functioning OK.

I opted for the same thickness as the stock material 0.5mm.
It lasted 9 years, so i'll check every couple of years, and replace it if it seems to be squeezed out from under the IGBT's.
This is what i ended up using:
Softtherm 86/600 200x120x0.5mm
You can buy them from Conrad. You'll need three films, one for each phase/megapole.

As @warden_clyffe said, you can just cut the film into 4 strips, each 30mm wide.
I removed all nuts from the threaded studs, and cleaned them as best i as could, so they would not interfere with the torquing of the nuts later on. Then i screwed the threaded studs to the heatsink.

After mounting the board onto the heatsink, i added the clamps, and washers / nuts. I tightened them all to 0.8Nm
Then i put a drop of superglue on top of the nuts. (it looked to me tesla used something similar)

Put it back together, installed it into the car, and it's charging/driving again!

 

[DanielFriederich]

Fuses for AC inlet (as electronic symbol states a diode is included => basic rectifying => correct direction necessary)

 

[slcasner]

Fuses for AC inlet (as electronic symbol states a diode is included => basic rectifying => correct direction necessary)

That's not correct. They are "semiconductor fuses", so named because they are very fast acting in order to protect semiconductors from hard short circuits. They are made of metal, not semiconductor material.

I am now carrying spare fuses and the necessary tools to replace them in my car. I tested one with my ohmmeter to verify that it conducts equally in both directions.

If the AC line fuses in the PEM were diodes then the input power would be cut in half (if they were installed in opposite directions) or blocked completely (if installed in the same direction).

 

[hcsharp]

I've always been curious why they had a diode symbol on them. I knew they weren't diodes for obvious reasons and testing them proved as much. Do they act like diodes in certain conditions?

 

[DanielFriederich]

That's what I misinterpreted too.
Data sheet doesn't mention anything about the diode symbol...

https://www.semic-shop.de/_obchody/semic.obchodak.net/prilohy/7387/fwh-150b-3251e6.pdf

 

[Fuzzylogic]

Here is a reference to these semiconductor fuses, from a model S thread:

DC - DC converter / AC compressor repair

 

[gfs]

I have #1362 (one of the last US v2.5 sport units made).
After many years of no issues (other than TPMS), I got code 905 and car would not go.
I rebuilt my PEM (new insulators and capacitors), while there I reinforced the back of the molex connectors with epoxy so none of the pins could back out.
The car runs fine with good acceleration, pem screen says about 37C max temp. I keep getting 1144 and 1146 codes. so I completely replaced the 4 pin connector, and used gold plated sockets.
Just drove it a few miles and am still getting 1144 and 1146 codes. when I stopped and looked in thru the back grill, I can see the pem/motor fan spinning (both sides), then it stops when cool.
-- these are "debug" messages - should I just ignore them (i.e. turn off debug mode) and keep monitoring the motor and PEM temperatures?

 

[PV-EV]

Did you replace the PCB mounted fan connector inside the PEM?

 

[gfs]

I did not replace that part. I reinforced the old one.
I was not getting any errors before the insulator crack issue, so I did not bother to change out that part.
it looked just fine, no burn marks or discoloration..
And the fan is actually spinning.
I am not sure what algorithm is used to determine there is a "problem"

 

[warden_clyffe]

that's your problem then. it's not only that the pins back out to the inside (then it's too late anyway), it's that they wear off, even they look fine.

b.t.w. only 1146 is a debug-massage, 1144 not.

before you go through the trouble of undoing your pem, replacing socket etc., please make sure your aircon is 100% ok.

 

[gfs]

what is an aircon?

 

[Austrianer]

What capacitors are used in the PEMs, anyone has info about V, uF, Dimensions, or a link to a direct replacement type?

 

[ML Auto]

what is an aircon?

Air conditioning

 

[gfs]

I replaced the caps with
epcos/TDK B43547A5477M000
470µF 450V Aluminum Electrolytic Capacitors Radial, Can - Snap-In
35mm diameter, 52mm tall

 

[gfs]

you will need a high wattage soldering iron as the pads have a huge amount of copper around them. I uses "resistance soldering" with a weller solder gun with the tip cut off