-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Preventive eMMC replacement on MCU1
- Thread starter widodh
- Start date
Yes, if you check out the video in the startpost it's explaind.
This ROM (eMMC chip) is bound to fail at some point. Without it the car won't drive and not charge either.
Out of warranty the replacement is $3500/EUR3500 while the chip which fails is <50.
Read the rest of this topic and you'll learn more. But the lines above sum up the story
PC CMOS batteries are in a user replaceable holder in a user accessible location.
Ah... very interesting... and good to know. I don't know if their engineer ever tried swapping out the cap, although they did confirm it's only 0.1uf.
So with my DS300 reader, I get slightly different results in trying to read my chip.
With the SD adapter I got one block device and one character device showing up as a memory card (/dev/mmc*) under Linux. The block device was approximately the correct size for the chip (~7.5GB). I had expected to see one block device for the chip and then 4 subsequent partitions. I couldn't read from either device, which we know is problematic with the SD-card based All Socket reader.
With the DS300 I get two block devices, and two character devices showing up as drives (/dev/sd*). Attempting to read from /dev/sdb hangs indefinitely. Attampting to read from /dev/sdc immediately returns an error. Again, no 4 partitions.
Interestingly, this reader has some LED's for VCC and DATA. The manual says:
In my case I get these lights flashing when trying to access the chip. I've verified pin 1 is oriented correctly.
My assumption is that the errors are because the chip is toast. But I do wonder if there's any chance this reader ALSO has problems with this chip, despite what their engineer may have said. It would seem odd to design one programmer with one value capacitor, and a different reader with a different value cap for the same application...
So with my DS300 reader, I get slightly different results in trying to read my chip.
With the SD adapter I got one block device and one character device showing up as a memory card (/dev/mmc*) under Linux. The block device was approximately the correct size for the chip (~7.5GB). I had expected to see one block device for the chip and then 4 subsequent partitions. I couldn't read from either device, which we know is problematic with the SD-card based All Socket reader.
With the DS300 I get two block devices, and two character devices showing up as drives (/dev/sd*). Attempting to read from /dev/sdb hangs indefinitely. Attampting to read from /dev/sdc immediately returns an error. Again, no 4 partitions.
Interestingly, this reader has some LED's for VCC and DATA. The manual says:
In my case I get these lights flashing when trying to access the chip. I've verified pin 1 is oriented correctly.
My assumption is that the errors are because the chip is toast. But I do wonder if there's any chance this reader ALSO has problems with this chip, despite what their engineer may have said. It would seem odd to design one programmer with one value capacitor, and a different reader with a different value cap for the same application...
So since I haven't been able to read my flash chip, Plan C is in effect:
1) Order Fakra ethernet connector to build ethernet cable: Check
2) Source a copy of the boot firmware that matches what was previously installed on my car: Check
3) Obtain/rebuild as much of partition 3 (/var) as I can, omitting the car-specific stuff I can't read from my existing eMMC chip: Check
4) Write image/partitions to new flash: Check
5) Solder new chip back on MCU: Pending
6) Power MCU w. 12V and check for response from MCU over ethernet: Pending
7) Reinstall MCU: Pending
8) Source car-unique stuff (SSH keys, etc..), or see if Tesla will re-provision as a service visit: Pending
At least I'll get back on the road this way, albeit with some functionality caveats...
Thanks to @LuckyLuke for help on this...
1) Order Fakra ethernet connector to build ethernet cable: Check
2) Source a copy of the boot firmware that matches what was previously installed on my car: Check
3) Obtain/rebuild as much of partition 3 (/var) as I can, omitting the car-specific stuff I can't read from my existing eMMC chip: Check
4) Write image/partitions to new flash: Check
5) Solder new chip back on MCU: Pending
6) Power MCU w. 12V and check for response from MCU over ethernet: Pending
7) Reinstall MCU: Pending
8) Source car-unique stuff (SSH keys, etc..), or see if Tesla will re-provision as a service visit: Pending
At least I'll get back on the road this way, albeit with some functionality caveats...
Thanks to @LuckyLuke for help on this...
Last edited:
Well, my MCU just died 3 days ago. A new MCU was installed. I asked about upgrading to MCU version 2 -- as there were rumors they will do such now. However, I was told this is not the case and they don't do upgrades to MCU 2. Hopefully, this MCU will last at long enough for you guys to conquer this problem... Or maybe Tesla will come up with a better/cheaper fix...
Replacing the eMMC of my MCU (September 2013) didn't go that well this weekend.
My MCU has the Hynix H26M42001FMR 8GB chip and after removing the chip from the MCU it wouldn't read.
There could be two reasons for this:
1: The H26M42001FMR chip needs a special reader/programmer. I doubt this
2: The eMMC died during the desoldering. The guy doing this for me took special car with the chip by desoldering at a maximum of 390C.
We wrote a new image with firmware 2019.20.2.1 (You need to know how to do this) on a new 64GB Swissbit eMMC and soldered that on the MCU.
The MCU booted with the new image as it's responding to Ethernet commands and the network came online. The MCU is being send back to me today and I'll put it back in the car tomorrow.
The homebrew Image is however lacking:
- VPN certificates for Tesla
- Some VIN specific information
We can hopefully recover this from the old eMMC with a different programmer, but otherwise a few features of the car like the app won't work anymore.
I'll update tomorrow with more information once the MCU is running again in the car.
My MCU has the Hynix H26M42001FMR 8GB chip and after removing the chip from the MCU it wouldn't read.
There could be two reasons for this:
1: The H26M42001FMR chip needs a special reader/programmer. I doubt this
2: The eMMC died during the desoldering. The guy doing this for me took special car with the chip by desoldering at a maximum of 390C.
We wrote a new image with firmware 2019.20.2.1 (You need to know how to do this) on a new 64GB Swissbit eMMC and soldered that on the MCU.
The MCU booted with the new image as it's responding to Ethernet commands and the network came online. The MCU is being send back to me today and I'll put it back in the car tomorrow.
The homebrew Image is however lacking:
- VPN certificates for Tesla
- Some VIN specific information
We can hopefully recover this from the old eMMC with a different programmer, but otherwise a few features of the car like the app won't work anymore.
I'll update tomorrow with more information once the MCU is running again in the car.
Last edited:
Jays200
Member
Gotta love Australian Consumer Law
"Fit for purpose"
It would be reasonable to expect the eMMc to function for the life of the product/car
https://www.google.com/url?sa=t&sou...FjACegQIAxAC&usg=AOvVaw2krahxiits_7qCQu00TfHg
"Fit for purpose"
It would be reasonable to expect the eMMc to function for the life of the product/car
https://www.google.com/url?sa=t&sou...FjACegQIAxAC&usg=AOvVaw2krahxiits_7qCQu00TfHg
@widodh : Thank you for the progress report!
Just a quick question; Isn't 390C way too hot.
It appears, and I am no expert, the good BGA rework stations use heating profiles which rarely peek over 250C (As I believe most BGA chips will take damage above this temperature).
Also, to avoid "popcorning", they "prebake" the boards.
Rework Profile Development - The BGA Rework Machine
I am about to get my chip de-soldered, but the phone repair guy I talked to said he just uses a hot air station, so I might need to find someone else to do this.
Seeing how many people have issues reading this chip, could it just be a case of improper de-soldering being very widespread?
Just a quick question; Isn't 390C way too hot.
It appears, and I am no expert, the good BGA rework stations use heating profiles which rarely peek over 250C (As I believe most BGA chips will take damage above this temperature).
Also, to avoid "popcorning", they "prebake" the boards.
Rework Profile Development - The BGA Rework Machine
I am about to get my chip de-soldered, but the phone repair guy I talked to said he just uses a hot air station, so I might need to find someone else to do this.
Seeing how many people have issues reading this chip, could it just be a case of improper de-soldering being very widespread?
I will.
And then try to convince Tesla to replace the MCU under warranty.
I have to trust this guy. He has been doing this kind of work on consoles, phones and other very sensitive PCBs for years now.
But it's very likely that the chip got toasted. The Hynix 42001 chip doesn't seem to be a very robust chip.
So the MCU would have to come with a lifetime warranty?
The problem with "lifetime" warranties is that the second something fails the life of the car has ended.
Jays200
Member
So the MCU would have to come with a lifetime warranty?
The problem with "lifetime" warranties is that the second something fails the life of the car has ended.
So, the way it works is, just because a manufacturer might only give a 12 or 24 month warranty, that doesn't limit what is "reasonable" insofar as it would be reasonable to expect a fridge or TV to last , say 5 years. We have strong consumer laws here that favour the consumer. You'd have to fight but you'd expect to win.
There is no way Tesla would get away with an AUD $200K car being rendered inoperable because of a small part with a known fail history failing. It just wouldn't happen.
Hope I'm not the first bloke to test it though
Similar threads
