No crank, no start issue - need help!

[sapote]

Read key # 5 in the AK90 and it still is not showing the VIN or KM. This is also what INPA shows.

Should have read it before things failed.

[Rockies67]

Are you thinking that key # 5 is somehow corrupting the EWS? If so, if/when it begins starting again should I leave that key behind and switch to only using # 4?

[sapote]

Are you thinking that key # 5 is somehow corrupting the EWS? If so, if/when it begins starting again should I leave that key behind and switch to only using # 4?

I don't think the keys is the root cause, but the crazy EWS wiped out the key.
Next time when engine runs again, then shut off and read the key with AK90; read the key after it fails to crank and I think its data was wiped out.

[Rockies67]

Okay car started working again this evening. I had put the EWS into the freezer over night to test out a theory about potential microcracks in the EWS soldering. EWS went into the frezer not working and when it was put into the car it worked right away.

Took the opportunity to start the car with key # 5. Removed the key and read the key data with the AK90. The data is still blanked out (no VIN nor KM).

Also put key # 4 into the car and started the car to see if the KM data would get updated since I have not used this key during any of the test drives the other day. The KM was at 349648. Key was updated to 349654. Interestingly, the actual KM on the odometer is 349643. Not sure what to make of that?

Let the car run for a 10-15 minutes with a heater going to see if we could see the INPA actual key data (identification, code, or password) fail in real time. Eventually gave up on this as the data was really solid and not doing anything.

One thing we did notice though is that in comparison with previous times of watching the actual key data that the idenitication did not occassionally flicker from No to Yes today. I wonder if this is a clue to what is happening when the car fails to start. Any thoughts on what would cause the Idenitification to flicker between the two states?

[sapote]

“Key was updated to 349654. Interestingly, the actual KM on the odometer is 349643. Not sure what to make of that?”

it makes sense that the nonvolatile memory in the cluster not getting updated with odometer too often, and so its actual km is lagging compared to actual value in the EWS.


only use #4 key from now and see what happens

[Rockies67]

Car started and ran nicely for 7 separate starting sessions (spaced out over 30 hours) with key # 4 and then stopped on the 8th attempt - all the same symptoms. Checked AK90 data for key # 4 after the fail to start and it is all there, no problems.

[bmwfnatic]

“Key was updated to 349654. Interestingly, the actual KM on the odometer is 349643. Not sure what to make of that?”

it makes sense that the nonvolatile memory in the cluster not getting updated with odometer too often, and so its actually km is lagging compared to actual value in the EWS so

only use #4 key from now and see what happens

No, the clusters have a mileage offset that could be set once when they were new.

So a car that had say 11 kilometers like OP his car, when it was sold new the dealer would reset the odometer back to 0, and this value is stored and always subtracted from the displayed odometer.

The maximum value for this offset is 0xFF, so 255 kilometers.

[Rockies67]

Thanks for the info on how the odometer & keys work. That makes total sense to me.

[sapote]

Car started and ran nicely for 7 separate starting sessions (spaced out over 30 hours) with key # 4 and then stopped on the 8th attempt - all the same symptoms. Checked AK90 data for key # 4 after the fail to start and it is all there, no problems.

So this proves that key 5 KM and VIN memory locations are bad, an isolated issue.

[sapote]

So a car that had say 11 kilometers like OP his car, when it was sold new the dealer would reset the odometer back to 0, and this value is stored and always subtracted from the displayed odometer..

you meant Added to?

So the CPU reads the miles from the cluster and then adds (not subtracts) 11km offset before writing into the EWS? So the EWS has the actual mileage of the car?

[George Hill]

No, the clusters have a mileage offset that could be set once when they were new.

So a car that had say 11 kilometers like OP his car, when it was sold new the dealer would reset the odometer back to 0, and this value is stored and always subtracted from the displayed odometer.

The maximum value for this offset is 0xFF, so 255 kilometers.

I've PDI'd a bunch of E46s and we never reset the cluster mileage to 0. What it came with off the truck and after the PDI test drive is what it was sold with.

My experience is LCMs don't count past 300,000km.

[bmwfnatic]

I've PDI'd a bunch of E46s and we never reset the cluster mileage to 0. What it came with off the truck and after the PDI test drive is what it was sold with.

My experience is LCMs don't count past 300,000km.

Maybe only the miles driven on factory grounds/testing were reset this way, but the mechanism definitely exists, my personal one has 9 kilometers “hidden” via this system.

[George Hill]

Maybe only the miles driven on factory grounds/testing were reset this way, but the mechanism definitely exists, my personal one has 9 kilometers “hidden” via this system.

Maybe so, but it wasn't done at the dealer level.

Whenever I zero out a cluster eeprom they don't always read true zero, sometimes its positive a little and sometimes negative. Also, when we replaced instrument clusters (at the dealer) with brand new never been messed with cars and clusters they didn't always read the mileage that they came in, which I assumed was something to do with how the cluster and LCM sync AND stored their respective mileages. They almost always were close enough that after a test drive they would be higher than they came in with. But I do recall I had one that was significantly less after and this was a problem because the repair order system and key read were not equipped to show a negative out mileage and it was a bunch of work to get it passed through warranty with the mileage being less than before. So maybe that's related to what you are referring to as a new cluster wouldn't have this offset when the original did?

[Rockies67]

Update:

Three days ago I decided to "touch up" the bottom side solder on all of the through hole components on the EWS. The thinking here was that one of the lines in the EWS related to the INPA key data (identification, password, code) is glitching in/out and causing the no crank condition.

Since then the car has consistently started on over 25 separate times (note some times I just started the car and other times I drove it around) using key # 4 (note key # 5 has been left behind and will not be used on the car moving forward) with no instances of not starting. I also removed and reinstalled the EWS after about 20 starts to test it on the AK90+. All error codes were erased before the resoldering and no new error codes are present at this time.

I plan to continue to start and drive the car short distances until I feel confident that this is not a temporary phenomena--but so far, so good!

[sapote]

Update:

Three days ago I decided to "touch up" the bottom side solder on all of the through hole components on the EWS.

How did you clean up/remove the conformal coating on all soldering pads?