It'* been a several months since I did this on our 2002 Bonneville and the repair has held up thus far, so it'* time to share. A new module goes for $140+. This repair costs less than $10, $25 if you need to buy a soldering iron, etc. Soldering experience is essential for this repair.

For those who are unfamiliar with the issue, symptoms of a failed rear integration module are a loss of electronic door locks, trunk release, reverse lights, rear defogger, remote horn, and probably a couple additional systems that I can not recall. The loss of all these functions due to a failing integration module can be sudden and complete or, as was the case for us, initially intermittent.

Repair

I determined that the cause of the failure in our module was a broken crystal oscillator. It makes sense that this would be a common failure as the module experiences a high degree of mechanical vibration and is mounted in a manner that exaggerates the mechanical shock. The oscillator is in a 6 pin surface mount package underneath the Faraday cage on the PCB; it'* labelled X1. This component can be difficult to remove without specialized equipment, but fortunately the two rows of pins on the chip are interconnected so that it is only necessary to heat up one side of the package. I used a small chisel bit (pictured) that is common to wood burning sets. Alternatively, an ordinary iron tip could be hammered flat and ground to shape to do the job.

I never got as far as finding the exact oscillator module for replacement. Instead, I opted to use components I had lying around. The oscillator is 8.00MHz, which is a fairly common value, and the layout lends itself very well for using 0805 bypass capacitors. I used a HC-49/U packaged 8.000MHz crystal. I'm not going to recommend the exact crystal I used since it'* an obsolete part, but DigiKey P/N X100-ND should work the same, in concert with two of 399-1112-1-ND (18pF caps, 0805 package). A surface mount HC49/US packaged crystal, such as P/N XC1271CT-ND, may mount better, but I haven't tried it.

I also replaced the resistor right next to the oscillator, R48, with an 0805 jumper (0 Ohm) resistor. It'* the now green one in the picture. This may not be absolutely necessary, but it'* how I did it and it worked. The old resistor can be removed using the same chisel bit used to remove the old oscillator. A suitable P/N from DigiKey for this is RMCF0805ZT0R00CT-ND, or a piece of wire could be used in which case the old resistor needn't even be removed.

As seen in the picture, the two caps can be mounted end to end on one side of the old oscillator'* land pattern. The new crystal'* leads can be soldered to the outer traces of the pattern. Assuming that the crystal is not a surface mount component it will also need to be secured in a way that its metal shell is isolated from any circuitry on the board. I used a dab of hot glue.

Preventative Maintenance

As mentioned earlier, mechanical stress is what causes the failure. Anything done to reduce the mechanical shock experienced by the module should help prevent future failures. What I did was move the module from the top of the seat back, where it would flap with every bump, and remounted it on the nearby cross member. I secured one side with a long zip tie. There'* a pre-existing hole on the other side that'* just the right size for one of the plastic rivets, or another zip tie could be used. I also put some padding between the module and the beam, although I doubt the padding does much. This complete mounting is shown in the last picture.

Very nice write up.

Thanks for taking the time to write-up and share!

I'm curious, how did you come to the conclusion that the crystal needed to be replaced?

And why did you replace the resistor with a jumper?

I tested the most common failure components first. The aluminum electrolytic capacitors appeared to be OK, so the next component I tried swapping out was the oscillator. Crystals are pretty common points of failure in electronics; they're very susceptible to shock. The crystal itself is usually cut extremely thin and is supported at only one or two points, making it very easy to shatter. One thing that makes them hard to diagnose though is that sometimes they'll still work even with a fracture. I once saw an article that went into great length describing crystal failures, and it even had a scope scan of a faulty one. I don't know where I saved it right now, but the scan was still periodic, but noisy looking. Here'* one article that I can find: How to test crystal with tester, checker and oscilloscope

In my own experience I've seen it before too, that'* how I got a 10MHz function generator for cheap. The thing looked trashed and had clearly been dropped. Other than cosmetic issues, guess what part failed? Works like a champ now. I've seen it in a couple wireless mice as well, except in those cases it usually is very difficult to find a suitable replacement due to the frequencies used.

As for the resistor, I couldn't tell why they had a resistor there (330Ohm IIRC) in the first place. I looked up the datasheet for the uC (MB8953xA) on the board, that is the big chip right next to the oscillator with which the oscillator interfaces exclusively, and it had IO'* for two different oscillators, a primary and secondary clock source. In the datasheet, it described a normal connection for the primary, that is a direct drive of the crystal with bypass capacitors to ground on each leg. For the secondary it additionally had a resistor in series with the crystal, that is between one leg of the crystal and the uC. The caps were shown as still tied directly to the crystal. That particular datasheet didn't get into detail as to why it needed this extra resistor, and I wasn't interested in hunting down better documentation at that point either because the oscillator was hooked up to the primary clock source! There is no secondary clock source inside the module. So, either the designer was as confused as I was looking at the layout and threw one in, or the old oscillator module was spec'd for one in the same way crystals require a particular load capacitance to run at their rated speed. Either way, the resistor doesn't apply with the substitute crystal. Still, it MAY still work with the resistor there, albeit the frequency will probably be somewhat off.

I was going to say that, but he beat me to it. Isn't it obvious?

Yeaaa, this is just a tad above my pay grade.