Sponsored

MG1 ECU Teardown | Attempting component-level repair on the Supra ECU

32bitsofGil

Well-Known Member
First Name
Gilbert
Joined
Nov 9, 2021
Threads
23
Messages
620
Reaction score
677
Location
Taylor, TX
Car(s)
2021 Supra 3.0
Hello fellow nerds,

I recently came into possession of a MG1 ECU that is "cloneable," the catch is this ECU was clearly involved in nasty wreck. I want to take this opportunity to do what I think has not been documented before. We are tearing open this enclosure and going to attempt to repair the ECU on a component level.

This thread will attempt to follow our journey, and maybe inspire someone smarter than me to pry these open and make a cool discovery or break the dreaded Bosch encryption for good. Our intent is to remove the enclosure, repair or replace any damaged components or the connectors, then seal it all back up and attempt an ECU flash over the currently non-working Can bus. If we are successful then we have gained one working "cloneable" ECU in this world, and maybe even given the courage to others to attempt similar repairs on other damaged MG1s

With the goal of flashing a working ECU in mind, we are taking precautions, but we have low expectations for this repair. I will do my best to stay chronological and detailed in my documentation of the process, but expect some jumping around as we are in very uncharted waters. The best case scenario is we get a fully working ECU again, the worst case we inspire someone else to fix a non-working ECU.









We started by following some decent videos detailing how to open the lid on other Bosch ECUs. It is very similar process and I will leave it to them to explain in detail how to separate the lid, and cut away the black silicone glue holding down the lid:



In the above video he used some "special lid extraction tool" or whatever he called it. Not at all needed. We used M5 screws and the trick is here that the OEM threads are smaller than that, but the hole on the lid of the enclosure is just big enough the M5 threads grab on to it and pry it up from the bottom half. We took it slowly, applying heat with a heat gun, slowly screwing down the M5 screw, then getting in there with a hobby knife to cut away at the silicone. Once we got all three sides that we could reach cut away from it's silicone bonds the lid just gently lifted away from the bottom half of the enclosure and we are left with this scene:

IMG_1374.jpg



Once we reached this stage we stopped to do some light analysis and inspection immediately we noticed a small component fall out from between the PCB and the bottom half of the enclosure. This signaled to us that the extent of the damage must only really be observable from the other side.

IMG_1385.webp


This made us conclude that we wouldn't be able to just reflow the solder from this side, and solidified the need to remove the bottom half of this enclosure. Now we begin to face some real challenges. After exploring some more we found lots of deposits of what appears to be a silicone based thermal paste (The white/gray looking stuff). This stuff is EVERYWHERE! And there is a HEALTHY amount under the PCB, adhering the PCB and the metal enclosure together. See some macro shots below:

IMG_1377.webp


IMG_1378.webp


IMG_1379.webp


This stuff is soft, and very very sticky. We hope not to destroy the PCB when trying to pry it out. We can also see some additional black silicone glue that is holding the top connectors (the plastic covers for them) directly to the exterior metal of the enclosure. This is going to be tough. We have a plan to try a two pronged method here.


This is where I will have to end my first post. Today our plan is to take this half clamshell and throw this bitch in the oven @ 170F for about 5 mins. The hope is the heat will weaken the bonds of the remaining glue, and we hope to gently lift the PCB out of the bottom half and access the side we need to repair. If this doesn't work our backup plan will be to hit this stuff with ISO 99% or acetone to see if we can get enough glue to dissolve and get some movement. This will all be much harder than the top half, since we have to be very gentle when prying on the bare PCB. If you have suggestions, or questions please we would love to hear your thoughts. Expect another update later tonight detailing our success or failure in removing the bottom half.

Thanks everyone! Happy Hacking!
Sponsored

 
OP
OP
32bitsofGil

32bitsofGil

Well-Known Member
First Name
Gilbert
Joined
Nov 9, 2021
Threads
23
Messages
620
Reaction score
677
Location
Taylor, TX
Car(s)
2021 Supra 3.0
Well we decided yesterday to give the ECU to the oven. This went about as well as expected.

IMG_1411.webp


We initially tried the lowest heat setting, doing our best to avoid heat exposing the plastics too much. This turned out to be not nearly enough. We moved on to using a 250F set temp, with about 10 minutes decided in the oven so it would get nice and hot.

vlcsnap-2022-04-20-10h59m44s338.webp


So we pull it out, blazing hot, and the black silicone along the outside edge is now very malleable, but the white/grey silicone all over the board and underneath won't budge a bit. Our plastic pry tools seemingly could do nothing to budge the PCB.

What a bummer. This stuff might be some sort of silicone-based thermal paste in here. If that is true then the temperatures at which this will give way will be much higher than we want to heat the entire ECU. This leaves us with the remaining option of solvents.



I warn you the following images won't be easy for anyone who is looking for an ECU.

67211586318__53C8D761-80B0-43C6-9C31-A4299C83F6B1.webp


IMG_1427.webp



No we didn't do this lightly. We really think our only chance to get under here was to blow $80 in 99% ISO to get in there and dissolve the thermal paste. This will stay in it's container for 24 hours and I will drop another update if we get movement. Already the morning after I noticed some bubbles forming, so hoping for the best.

Please chime in if you think you know what this stuff is, or if you've cracked one of these open before.
 
Last edited:
OP
OP
32bitsofGil

32bitsofGil

Well-Known Member
First Name
Gilbert
Joined
Nov 9, 2021
Threads
23
Messages
620
Reaction score
677
Location
Taylor, TX
Car(s)
2021 Supra 3.0
Reserved again for final updates
 
OP
OP
32bitsofGil

32bitsofGil

Well-Known Member
First Name
Gilbert
Joined
Nov 9, 2021
Threads
23
Messages
620
Reaction score
677
Location
Taylor, TX
Car(s)
2021 Supra 3.0
New update above... things didn't go well last night. No progress was made, but we've set it in some ISO hoping for enough to break free we can lift it out of the enclosure.

Have you dealt with this crap before?
 

Loco38SUP

Well-Known Member
Joined
Feb 17, 2021
Threads
55
Messages
2,350
Reaction score
3,406
Location
Seattle
Car(s)
2021 GR Supra 3.0 (Renaissance Red)
That looks to be a conformal coating. They’ re popular for PCB assemblies. The key is to finding what type it is as there are many to chose from.

Not all coatings are permeable so removing them can be challenging. In some cases specialized solvents and ultrasonic baths are required.

-RJM
 
OP
OP
32bitsofGil

32bitsofGil

Well-Known Member
First Name
Gilbert
Joined
Nov 9, 2021
Threads
23
Messages
620
Reaction score
677
Location
Taylor, TX
Car(s)
2021 Supra 3.0
That looks to be a conformal coating. They’ re popular for PCB assemblies. The key is to finding what type it is as there are many to chose from.

Not all coatings are permeable so removing them can be challenging. In some cases specialized solvents and ultrasonic baths are required.

-RJM
It is some type of really interesting stuff. I can cut away at it after 48hours soaking in ISO. But the difficulty now is actually getting a blade underneath the PCB, between the enclosure.
I can’t seem to pry the PCB up enough to cut away this material, and I’m concerned going stronger with solvents will destroy the PCB unless I pick the magic one that is meant for work on these boards.

I’ve heard some people out there try solvent baths in ultrasonic tanks, but seemingly this will just rek the components in the process of dissolving this shit.

I sent some pictures over to a few ecu repair facilities that work on sub modern Bosch ECUs hoping for candid feedback, but I’m stuck in research mode until I can find a new attack angle here.

I’ve also considered nuclear options like cutting away the metal enclosure, doing my best to preserve the PCB, but this would require reverse engineering the enclosure, or sourcing another.


Probably not the right form for this, but since my post is now showing up in page one results for “MG1 tear down” I can hope to crowd source some ideas here.
 
OP
OP
32bitsofGil

32bitsofGil

Well-Known Member
First Name
Gilbert
Joined
Nov 9, 2021
Threads
23
Messages
620
Reaction score
677
Location
Taylor, TX
Car(s)
2021 Supra 3.0
Alright boys this weekend I’m getting desperate. First I am going to try a very high temp in the oven. 300F for at least 10 minutes.

if I can get it to pry up, then I’ll cut away what I can and try to free the PCB. If this doesn’t work I also have plans to boil the ecu in distilled water to see if I can get any movement.

beyond this I will likely call in a favor or phone a friend on this. Getting this guy out without butchering the PCB is tough.

anyone with connections feel free to PM me. I’ll pay for the service.
 
OP
OP
32bitsofGil

32bitsofGil

Well-Known Member
First Name
Gilbert
Joined
Nov 9, 2021
Threads
23
Messages
620
Reaction score
677
Location
Taylor, TX
Car(s)
2021 Supra 3.0
I've been told from a trusted source that the flash memory/CPU combo component on this ECU is likely dead. This process will yield nothing of value other than the information I have shared here.


Should I keep going? Even if I destroy the PCB trying to remove it now, it won't matter. Anyone here want to give this a shot now? :)
Sponsored

 
 








Top