In 2009, I made a wonderful mistake: I bought a used 2006 Saab 9-3 Aero. It's an incredibly comfortable car, and is a kind of quirky I found enjoyable. At the time, I really hated driving, and so I bought a car that seemed like it could make driving fun. It did.
Of course, I quickly learned like most Saab owners eventually do that there are things about Saabs that aren't fun. It's from the GM Saab era, so it should have cheap parts, right? Yes, except for the places where Saab did something weird despite promising GM they wouldn't. That's Saab for you. Well, and, except for the parts which are identical between a Cadillac and a Saab, but are electronically restricted. In those cases, the new GM part will be a quarter of the price of a used, questionable Saab part. Good times.
By 2017, I was well aware that any issue means spending a kilobuck, sometimes more. I'd had to replace all kinds of things that were, well, several kilobucks. I'd overpaid in the first place, and along the way I put the same amount again into it. That Spring, on a weekend up in Victoria, British Columbia, from Olympia, Washington, I experienced my first clear-cut no-start condition. It wasn't the battery.
The awesome independent shop in Victoria (very much like our preferred awesome independent shop in Olympia) were helpful, sympathetic, thorough, and quickly hit a wall. They referred us to the local GM dealer who does Saab service for Vancouver Island, who were awful, charged us a bunch of money, and made a bad suggestion.
Several more no-starts followed, which meant all sorts of shenanigans to try to avoid taking it on a ferry back to Olympia, but eventually we had to. That was awful. I do not suggest taking a car that you aren't sure will start on a ferry.
It didn't happen often, in fact; more on hot days. The ferry ride was fine, as was a bunch of driving that followed. Until, that is, it wasn't.
So the car and I got to visit Swedish Automotive, who identified a straightforward issue and fixed it. I picked it up from them, and it wouldn't start right there in the lot. It started after a few minutes, as it always did, and they suggested replacing another part that the Saab part dealer across town had in-stock. I raced across town, and the dealer even installed it at no cost, while they were doing some recall fixes. They squeezed me in and did all of that. That was Park Place; I adored them, once I finally found their service entrance, anyway!
A few days later, I was driving down to a work conference when, in the middle of nowhere, the car wouldn't start after a brief stop. Yikes.
So it went back to Swedish, who held it for almost two months, trying to track down the issue. I picked it up once they'd, with good spirit, total transparency, and general wonderfulness, helped me fix a few more things that might've been broken.
Their loaner program is great; I had their loaner for basically no cost for all that time. I returned the loaner the day before my ordination to the priesthood, after picking my sister up from the airport, to go with family and friends to a big rented house that could fit all of us as we awaited and then celebrated my ordination. We got dinner after I got priested, and the car wouldn't start for twenty minutes. Saabs tho.
With more data points on when it was failing and what it was doing, it went back to Swedish. They fixed some things that, in hindsight, it's hard to say whether they were really failing, or if they were secondary. They were wonderful overall. They fixed everything they could, and then decided it was the Engine Control Module, failing when the car is heat-soaked.
Saab sure loved one-off approaches to things. In 2006, they used a Bosch ME9.6 for the ECU/ECM (hereafter ECM) in the V6 Saab 9-3, i.e. the Aero in the United States market. That ECM wasn't like Saab ECMs previously (the Trionic series), and wasn't on otherwise-identical 9-3s that year. It was used by GM in their cars that used the same, and larger versions of the same, engine.
There's a lot of tuners, mechanics, etc., out there who hate the thing. It doesn't work in the same way as the Trionic ECMs, and you don't hack it the same way, either. It's peculiar, an oddity. In the years following, it got revamped and used more widely, so on top of it all, the 2006 V6 ECM is especially quirky.
Swedish were certain that you couldn't take a used ECM of the same kind and use it. Even if you could marry it to the car, they said, there's car-specific stuff that has to be programmed at the factory.
Of course, that's not how Saab did it at first, or so the lore goes. In 2006, they were shipping unprogrammed ECMs and then they'd be programmed at the time they were married to the car. But maybe it's write-once, right? Except there's a lot of people out there, rather less visible, who are using junkyard ECMs all the time. For some of them it goes wrong, but the advice is, for the sake of superstition if nothing else, get one from as similar of a car as you can.
I'm convinced that you should be able to flash a GM ECM of the same underlying hardware, but I haven't seen it done. It may require a BDM flash rather than being flashed over CAN, I'm not sure. I'm not a car person, really.
Swedish told me the only way forward was a new ECM, which they couldn't source. I found someone who swore up and down that he could source one, and had a delivery date and everything. He was wrong.
A new ECM is a kilobuck, and fitting it will cost you half that again. That would be tolerable, in Saab terms, except that it was going to be at least 3 months, and possibly never.
That wasn't an option.
My original ECM was labeled as a 55560279, a number that I now know like my own phone number. The service part, i.e. identical hardware from the same year, is 55560009.
People say that the 2006 V6 will take a 2007+ ECM which has a different service part number. They're easier to come by, cheaper, and even available new. I didn't get a chance to source a junkyard one, and it wasn't worth risking on a new one.
I went on eBay, and I looked at tons of pictures until I spotted one: 55560279, there alongside HFV6 and E55. Identical. Not from an identical car, so if there were major underlying differences, I'd be out of luck, but they'd come from the factory with the same software version, and the same general configuration. It turned out I could get one of those, access to what GM calls a "Techline PC" and a Tech2 and even buy an OBDLink SX so I could use TrionicCANFlasher if all else failed, to more directly poke at an ECM.
All of that cost less than the new, unobtainable ECM, to say nothing of getting one fitted.
Removing the hardware was pretty easy. I mean, it's tedious and awful. Still, I found a way.
I got the new ECM in, and I couldn't marry it to my car; the Tech2 insisted that there was an unsupported software version on the ECM.
It was suggested that with TrionicCANFlasher, I should be able to just change the VIN on the ECM, and have it work with my car as-is. In fact, that set me up for a world of pain down the line that I didn't see coming. Still, to avoid having to touch the software on the ECM, I gave it a shot. It didn't work. In retrospect, my actual issue here may have simply been that I had the wrong version of software on the Tech2 (version 148), as I had much greater success with having the ancient copy of GlobalTIS put its newest version (which was older than 148) on the Tech2, later on.
Changing the VIN was a mistake, I'd find later.
It seemed like there was nothing for it but to upgrade the ECM, by doing an SPS using GlobalTIS. I couldn't get my Tech2 to do pass-through, so I had to try to use the Tech2 itself to do SPS in remote mode. That didn't go well; it failed every time, flagging some error towards the end of the process. Something was wrong, broken, or incompatible.
And that was all after the point at which I realized we needed to be pushing power to the car continuously, the lack of which was to blame for my first few SPS failures.
I tried using TrionicCANFlasher's recovery mode to splat a full read of my original, kinda-working ECM onto the replacement ECM. That didn't improve things at all. I certainly was still a long way from plug-and-play.
At none of those points did a simple Tech2 add of the ECM work, either.
Eventually I downgraded the Tech2 to match the software version available in GlobalTIS, and suddenly using it as a pass-through worked. This meant that I could use GlobalTIS itself to do the SPS.
I did, using the reprogram mode, rather than the replace and program mode, because I wanted to leave the old ECM usable with the car. In retrospect, this was a mistake, too, but I wasn't confident I could add the old ECM to the car. When I got desperate later, I confirmed at length my ability to add and remove the old ECM.
I got all kinds of errors, the underlying cause of which I was able to diagnose based on experimentation with TrionicCANFlasher and a lot of patience. Nothing I tried worked.
I needed to SPS again, I figured, because some people online attributed errors I was seeing to a sort of split brain in the ECM: it was programmed for the car, but needed to be replaced since it was an ECM which had previously been married to another car, rather than just being reprogrammed like one would do for an ECM that was originally in the vehicle and being upgraded. That turned out to be difficult.
TrionicCANFlasher was a real lifesaver here, because you can't SPS something which already shows the same software version as the latest in GlobalTIS, without getting a VCI number from GM, which it sounds like is majorly annoying. After exhausting every other option, including removing and adding the CIM (which required putting back the old ECM to re-add, which was mildly terrifying for a few hours), I flashed my full read of my old ECM using TrionicCANFlasher. That didn't accomplish much, but it did screw up the calibration data enough that I could SPS again, this time in replace mode. It would only update the calibration data, not the rest of the ECM, but at least I had access to replace mode.
The replace mode SPS didn't change anything.
Eventually I realized that I needed to set the VIN of the ECM back to the VIN it had been programmed with originally. I had tried all kinds of things, from manually blanking the VIN, to trying to change it to a "generic" VIN (with zeroes, spaces, or NULs taking the place of the serial number digits at the end of the VIN), and none of them had worked. If there was a split-brain situation going on, where some security element in the ECM was tied to the VIN of its donor car, then maybe at least I could get it back to being a fresh junkyard ECM, ready to go.
I found a photo I'd taken with my phone (I can't stress enough the benefit of dumping as much data as you can along the way, and saving it, or at least photographing it), and I wrote "VIN orig" at the end of the notes sheet where I was tracking how different VIN settings broke in different ways. I used TrionicCANFlasher to put back the original VIN, after all of the SPS runs, and everything else.
The Tech2 added it without issue.
The day after marrying the ECM to the car, I was seeing numerous DTCs related to the ECM's presence with various other components. Going through and adding each of those components allowed them to marry with the ECM, and you should expect to need to do that. Specifically, I needed to add the BCM and ESP+, but your car may vary. Annoyingly, this meant doing things like the yaw sensor calibration on an extremely uneven apartment building parking lot, with a car full of stuff. Luckily, such things can be redone at a later date.
I'm an impulsive, impatient sort of woman when faced with high-pressure, high-opacity technology things like this. I wish I'd been able to get better diagnostics along the way, and had more freedom. Still, I'd jumped to doing stupid things because I wasn't well-oriented. If I'd done the SPS properly first, and then done an add immediately after, I think it would have worked. If I had just put a compatible software onto the Tech2 and then done the add, rather than getting discouraged by the incompatible software version messages, it would have worked right away.
You can use a junkyard ECM, not properly divorced from its donor car, in a 2006 Saab 9-3 Aero, and marry it to the car using only a Tech2 with security access. Worst case, you might need to SPS it, but probably not.
Also, TrionicCANFlasher was very useful for checking things along the way, and for both making and fixing mistakes. It's a great and dangerous tool for an ignorant, impulsive, and careless hacker like me.
Most significantly, I believe I now know what several error messages mean that I found a few other lost souls struggling with on Google. Moreover, I know how to fix them, which it seems like the others didn't; I gather at least one car may have been parted out due to being unable to overcome one of these errors, even with direct GM help.
These are the errors I got while trying to add the ECM to my car with its VIN configuration in various states after SPSing. If you hit one of these errors, it's likely there's a previous VIN tied to your ECM that you could set it back to with TrionicCANFlasher. Good luck.
Error code <-7>: SSA sec not found / ECU id: 8450 / scode.c, line 5109
I got this error when attempting to use an empty or dummy VIN in the ECM.
I believe the issue here is that some internal security codes needed to be unlocked, basically, which were keyed to the original VIN; moreover, those VINs were not configured to the car.
This left the ECM in a state where the contents of the ECM neither belonged to the VIN of the car nor the VIN of the ECM.
If you imagine that the VIN was being used to encrypt and decrypt both new and stored marriage records, immobilizer codes, etc., then of course that would fail.
ME9 Motronic E9 / Negative Response: 7F AE E3 FE 05 / CPID Number: 7D / scode.c, line 5106
I got this error when using the VIN of the car in its entirety in the ECM.
I believe the issue here is something like this, similar to the above: the car tried to access things on the ECM which it should have been able to access, decrypt, or encrypt because the VIN of the ECM was the same as the VIN of the car, although those records were tied to another VIN altogether.
This left the ECM in a state where it could talk to the car, but not to itself, I think.
I hope that any of that proves useful to someone else. I don't know what I'm doing, and am certainly not a mechanic, and have no deep GM or Saab knowledge.
I do wish that the current Saab licensees, rather than dragging their heels on making another batch of these ECMs, would simply provide comprehensive flashing software and instructions for service centres to be able to use new GM ECMs for these cars. It should be doable. Customers deserve better than to have to go down rabbit holes such as this, and no company should leave its customers so lost in being able to continue using an essential and otherwise-maintainable product.
If someone out there manages to get new GM ECMs working in Saabs, they'll be doing everyone a huge favour. I thought I would make a more serious effort at that myself, but I believe I've hit my limit. I did do some work on that with a cheap Cadillac ECM I bought on eBay, but I took poor notes because it quickly ceased to be promising. I think that a new GM ECM, which has never been programmed to a VIN, may have fewer problems if you can get it to SPS. It's worth trying, but not for me right now, given the difficulty I had with the easier route.
TrionicCANFlasher and the Trionic Tuning message boards are wonderful.
It's terrifying to do things you only half-understand that might brick your daily driver.
If you need an ECM for one of these cars, you should be fine with a working ECM with the same part number on it, and maybe even the same service part number alone. In the former case, you can likely get away with just using a Tech2 with security access to add it to your car. In the latter case, you may want to SPS it, and probably should.
I left out of the narrative the part where my mostly-working original ECM got broken towards the end of the process, when I was out of ideas. I'd left it sitting in the engine bay, and in a fit of heat stroke, forgetfulness, and general irritability, I closed the hood to go inside for a break. The ECM was positioned perfectly to get jabbed through its protective casing and down onto the circuit board itself. It could have missed everything easily, but it didn't: a small resistor got broken off. Luckily, it was still working enough for me to re-add my CIM using it later, which it was necessary for, since I couldn't do any of the steps I needed to take to be able to successfully marry the ECM without the CIM being added first!
In any event, good luck to anyone who happens to find this: you're probably not having the best day of your life. I hope that my Saab story gives you properly-attenuated hope, and the courage to chart a path forward.