Fixing a Porsche ECU

An old colleague and friend of mine contacted me recently as they had a customer at their Porsche business who had a car with a faulty ECU. The car was pretty old so the usual “we fix your ECU” people weren’t really interested, so (presumably) as a last resort he contacted me!

I think it is from a 911 (996), or a Boxster (986). It’s a 2000 Model Year, I know that much. I don’t know much about Porsches!

There’s nothing particularly special about an ECU – it’s just standard electronics. Although, having said that, there are some frustrations with them which I’ll discuss below…

The car in question simply didn’t respond to any diagnostics with the ECU connected, so he sensibly bought a donor ECU. When this ECU was connected to the car, the car responded to diagnostics fine, but naturally it was coded all wrong so you couldn’t actually start the car – makes sense so far.

My job – which I did accept – was to see if I could make the correctly-coded but broken ECU work, by harvesting anything necessary from the donor ECU.

Here she is:

Bosch M 5.2.2 2000MY 996 / 986 ECU top
Bosch M 5.2.2 2000MY 996 / 986 ECU bottom

Luckily my friend had the Bentley manual for the Porsche, so I had a wiring diagram which tells me what connects to what when the ECU is in the car. I flicked through the manual and found the correct variant, simply by comparing pin counts, numbering schemes, and by seeing what seemed the most likely diagram for the ECU in front of me.

This shows the ECU’s pinout and what it connects to in the car:

Bosch M 5.2.2 2000MY ECU wiring diagram (996/986)

The terminal we’re most interested in is called TERM 30. This is the name for the permanent +12V supply on a Porsche wiring diagram. It took a fair bit of staring at the diagram to work this out!

I hooked TERM 30 and a ground point to my bench supply and tentatively turned it on, with a 500mA current limit. It current-limited immediately, so I gave it a bit more current, and it continued to limit. So – there’s definitely a fault!

When this happens I turn to my trusty Thermal Camera. I own one of the cheapest and probably most rubbish thermal cameras there is, but it’s good enough for my needs – it’s a Hikmicro E01 – £180 on Amazon. The thermal camera showed this:

A thermal image of a faulty Bosch Porsche ECU

A single component much hotter than it’s surroundings. Clearly that is involved somehow in our fault. That component is this:

A faulty regulator on a Porsche Bosch ECU

Now this is where car ECUs are a pain. Due to the volumes involved, a lot of the parts are custom for each manufacturer (Bosch in this case), so it is sometimes very difficult to get data on the devices. Given the location of this device and looking at what it is connected to, I assumed this was probably some kind of voltage regulator.

You might think “ah ha! time to swap that component out and job done!”. But, not so fast! In my experience, components like this rarely fail. It is probably getting hot because of an excessive load on its output, forcing it to current-limit and sink a load of power, which can only be turned into heat.

Sure enough, I measured the resistance on the output of the regulator and it was pretty much a dead short. I started working my way through the components on the rail to see if any looked like they were misbehaving. A capacitor should never have DC continuity across it, and the very first ceramic capacitor I measured was a dead short!

Here’s the culprit, circled in red:

A ceramic capacitor gone short on a Porsche ECU

I swapped the capacitor from the donor module and as if by magic, the current consumption massively dropped and the ECU started behaving the same as the donor – it would take a few hundred milliamps, then put itself into a low-power sleep mode. Much better!

It’s also worth talking about TERM 15 on the wiring diagram. This is the terminal which goes to +12V when the ignition is turned on. On the donor I noted that when TERM 15 received 12V, the ECU went to a higher current state and the current started cycling – presumably normal operation.

Thankfully when I replaced the capacitor, the originally-broken ECU started behaving in this way too.

Hooray – nothing can beat the satisfaction of fixing something!

This post easily trips off the tongue, but of course this was several hours of faff, but enjoyable nonetheless.

I swapped the broken capacitor onto the donor ECU, but left it rotated on its pads so it wasn’t connected. Always worth keeping components that you take off!

Porsche 911 (996) / Boxster (986) ECU from 2000

My desk got a bit messy in the process:

My desk after fixing an ECU

You’ll note that I don’t have incredibly expensive or posh equipment. My microscope (just out of shot) was £200. I use WD-40 contact cleaner, £3.50 side cutters from Amazon and the cheapest tweezers I could find on RS – £7. My bench power supply cost £50 from Maplin back in the good old days when Maplin existed on the high street. I use a Hakko soldering iron and an (admittedly fairly posh) Fluke 179 Multimeter. I treated myself to both from eBay a little while ago – they were brand new in their box and a fraction of their new cost.

You can contact me if you are having issues with an ECU as a last resort, if you like.

Leave a Reply

Your email address will not be published. Required fields are marked *