Thread: Recovering a bricked NNN ecu

The chip quik arrived yesterday, and I got a chance today to try removing the eeprom.

I initially had a bit of trouble getting the chip quik to wet all the pins. The surface of the board is covered with a sticky coating which I guess protects against oil and water to a degree. After good work over with isopropyl alcohol and a tooth brush I was able to get the chip quik to wet and chip free from the board. Not quite as smooth sailing as the tutorial videos but fairly close. Considering I was envisaging cutting the leg and then desoldering them individually this much quicker and cleaner.

I'm pleased that I didn't damage the board and the eeprom has come off without any visible damage. None of the legs are bent or broken so I should be able to reuse.

The pic below shows the pads after desoldering the eeprom.


The pads look almost as good as the empty pads to the left which have never been soldered. I'm giving the chip quik a big thumbs up

The Willem programmer is due to turn up on Monday so with a bit of luck I'll be able to have the ECU up and running by mid week. I haven't been able to locate a original file for the NNN500030 as yet, but I have found one for a NNN000120. As Colin (BBS Guy) has detailed the OBD map upload overwrites the entire eeprom contents so all that is actually required is that the firmware boots the ECU to the point where a new map can be uploaded.

The willem turned up yesterday and looks to be a well made and packaged bit of kit. So I dragged home an old PC which had spent 4-5 years running an OpenBSD based mail server until about 2 years ago - it's has a ASUS P4 motherboard and has been totally reliable this far.

After a huge amount of mucking about I finally got XP installing using an app called Flash Boot to create a bootable USB version of the XP installer disc. It seems that after 7-8 years of sitting without being used isn't kind to cd drives. The situation hasn't been help by the fact that I use a Mac and I've had to do all the usb stick creation using an XP guest OS running under Virtual Box.

Update: As it turns out the saga is not done. After completing the inital part of the XP install the system rebooted to a "NTLDR is missing" error. A quick trip to the computer shop and $20 poorer I have a working cd/dvd drive and can boot from the install cd. hopefully this time I can actually get a working OS.

Brave project!

Will be watching with interest

I guess I should update on progress thus far.

After getting the new DVD drive the XP install was straight forward. If only I'd realised 2 days earlier.

The Willem seemed to be ok but I was having trouble getting it to read the chip id. This is apparently one of the key tests that you have the wilem correctly configured for the chip, and no other problems in your setup. The chip id read should return the manufacturer and the model of the chip - and AMD AM29F200BT in this case. To compound the problem I was having trouble getting the chip to seat in the PSOP44 adapter and was bending legs etc.

In the end I called it a day and decided to have a go with a fresh head.

The next day I spent a bit of time trying to get the chip to seat correctly. I managed to break off pin 1, which was fortunately a NC (not connected) pin - the only one on the chip. After a bit of fettling on some rough mouldings I got the chip to seat securely in the adapter. One issue down.

When I'd first set up the computer the pin holes for the parallel port wre filled with dust and grime which I'd vacuumed out. Given the random problems I was having I decide to clean each one out with a mini screwdriver which just fit into the hole. Just gave the screw driver a single turn to ensure there wasn't any muck left.

The final tweak was to change the LTP settings. Every "how to" I read suggests a different setting so I went with EPP rather than the ECP I'd used initially.

Not sure which change exactly did the trick but the ID check came back with AMD and AM29F200T. Erased the chip - success! Grabbed the NNN000120 binary file and hit program - success!! verified the write - again, success!!!

So this was looking pretty damn good.

I bit the bullet and solder the eeprom back into the ECU and went down to see if I'd managed to revive. Plugged it all in and tried to read faults - "communication with ECU failed". Bugger! Tried a few variation and permutations all with the same result.

This left me with two options - I'd done a really dodgy job on soldering which given the fact that my eyesight is definitely not good enough to check for solder whiskers on SMD chips was a fair possibility, or the NNN000120 firmware wasn't suitable for the NNN500030.

So today I finally gave in and bought a led illuminated magnifier - and saw what a shocker of a job I'd done on one side of the chip. Ooops.

So with the benefit of magnivision I desoldered the eeprom, straightened the bent legs and soldered it back with a degree of precision. I also took the time to jumper the stub of the broken pin to the board just in case.




Back down to the disco to test, and same result - no communication.

I'm at the point where I am fairly certain that the issue is the NNN000120 firmware is sufficiently different that it won't load on the NNN500030. I'm now hoping that I can either get a NNN500020 image from Td5Inside, or purchase an image for the NNN500030 from a Belgian mob. Pete at B-A-S has responded to a query I sent but is on holiday for another couple of weeks so can't assist.

As I was typing this an email came in from Wally at CTH Chiptuning with an original NNN500030 firmware file attached

Tomorrow it looks like another attempt to reprogram the eeprom. The pins on the current chip are starting to look a bit iffy so I'm not really confident that they will survive another round. With any luck the replacement chips will turn up next week so if all else fails I'll program a fresh chip and install that.

On interesting thing I discovered is that it is possible to use the larger AM29F400BT chips. These have twice the memory of the stock chip arranged in two banks. The two banks are addressed by grounding or floating pin 1. It is apparently possible to load 2 maps and use an external switch connected to pin one to select between them. Could be cool, but I'll keep focused on reviving as an initial goal

cheers
Paul

You know of course that for under $500 you can buy a generic, programmable ECU that can control multiple spark delivery and sequential EFI? and for a small amount more you can buy the TCU that can control most modern electronic auto boxes?

Just saying.

Originally Posted by jakeslouw

You know of course that for under $500 you can buy a generic, programmable ECU that can control multiple spark delivery and sequential EFI? and for a small amount more you can buy the TCU that can control most modern electronic auto boxes?

Just saying.

A Td5 ecu is around $300 and doesn't require a third party TCU. Just saying.

So the saga continues.

Checking in the cold light of day I found I had a couple of tiny solder bridges on the resistor packs which are located next to the eeprom. I removed those and thought I'd give that a go. Still no joy, so removed the eeprom again. This time the 5V+ pin broke off, so that puts paid to any more playing until the new chips turn up from HK.

While the chip was off I double checked the resistor packs to clean up any traces of ChipQuik. I can across one additional bridge and a couple of small fragments of loose chipquik. I've gone through and checked again for bridges between pins and also that there is continuity to the other side of the resistor packs. Hopefully that will ensure that there are minimal issues when I finally come to soldering in the new chip.


In the process of doing all this I've discovered the "right" way to reflash the eeprom. I haven't found a cheap way to do it as yet, but it doesn't require any soldering.

Following up on the previous post, I've discovered a cheap way to do the no solder repair.

I'd initially thought it would require $450 investment - and that is probably still the most user friendly way - but I've uncovered some free software that will work.

I'll post an update once I've had a chance to test this out.

Cheers
Paul