3

As part of a research project, we are doing accelerated ageing experiments on a set of Continental/BMW ECUs; part number seems to be TCU 1.5 MOST GSM US. The experiments involve powering up the devices inside a high-temperature high-humidity oven for a period of time.

We figured out the pins on the external header where we feed GND and 12V (simulating a battery-supplied voltage). It powers up, and consumes around 350mA ~ 400mA (as displayed by the power supply I'm using); however, after a minute or so, it shuts down (and it looks like it truly powers itself off — the power supply shows 0.000A current, while still supplying the 12V).

Any suggestions on how to keep this device powered up indefinitely? (the experiments will take in the order of several weeks, during which the device should be always powered up, inside the oven)

Notice that we're not allowed to wire power supply voltages into the board itself (say, soldering the wires into some points on the PCB's power rail to forcibly power-up the ICs) due to some of the constraints/requirements of the actual experiment we're conducting (I'm not at liberty to disclose too much detail; plus, it would not make much difference for this post/question)

2
  • 1
    If you are really supplying 12.00 volts and wish to study ECU temperature related to power dissipation, your results may be invalid. Real cars run on higher DC voltages (often over 14V) with significant continuous AC ripple from the alternator and occasional voltage spikes in excess of 30V from motors. Suggest you look into this and define your power parameters to either mimic the real world or to intentionally underpower and understress the ECU with clean 12V DC power.
    – MTA
    Commented Jun 14 at 15:46
  • How do you know an ecm doesn't need all or some its sensors providing input signals? Perhaps this ecm shutdown as a default in programming. Finding what parameters are needed to keep this ecm live, from BMW, may answer this mystery
    – F Dryer
    Commented Jun 14 at 19:32

2 Answers 2

3

The ECU may have multiple power ports to draw power, as one pin may not be enough to provide it what it needs to drive everything. Another reason there may be more than one pin is if there's one pin for the "on" state to provide power to everything, as well as a "maintenance" power to keep the memory valid over a period of time (think of how a radio is fed power to keep the stations programmed). Double check on this and providing power to all power pins might keep it alive. To my understanding, as long as the ECU has power, it should stay in an awake state.

3
  • @Cal-linux - One other thing which may be relevant ... how have you checked to ensure the ECU you're trying to utilize is good in the first place? Just a thought. Commented Jun 14 at 18:57
  • @Paulster2 ‒ we have multiple of these units; when we observed this unexpected behaviour, we tried a second one, and confirmed that the behaviour was the same on both. There are three adjacent pins that on the PCB look like tied together as 12V input (well, battery voltage, in any case); I'm only connecting to one of them, but with a consumption in the few hundred mA, I don't think that's a problem; you still have a good point, that there may be other pins (perhaps far away from the three pins that we identified as the power-in) necessary to provide the required total amount of power.
    – Cal-linux
    Commented Jun 14 at 20:31
  • It wouldn't hurt anything to try applying power to all power pins. You might also want to check your power supply to ensure it isn't just dumping power at some point. Commented Jun 14 at 22:05
0

Many automotive ECUs will have some kind of vehicle-wide power save features. Each manufacturer does it a different way and I don't know the specifics for BMW. You will likely need to provide some kind of signal on whichever protocol this ECU uses (possibly CAN, or if it's quite new possibly something like FlexRay or a similar ethernet-style architecture).

For this ECU it appears that if no multiplex signal is received within a time period it will shut itself off, this is quite normal. However, if you have a detailed wiring diagram, that will help. It may be that the power line you are using is a permanent 12v feed but another of the 12v pins will be for an ignition supply. This might be enough to keep it powered. As this appears to be a relatively simple and single-purpose infotainment ECU, you might be lucky and this might be the case.

I suggest you look for input from enthusiast/tuning forums to work out what sort of signals are needed. There are always hobbyists or aftermarket manufacturers reverse engineering aspects of vehicle architectures, especially infotainment networks.

The complexity of modern vehicles means that getting many ECUs running on the bench can be a non-trivial electronics engineering activity. It is possible or even likely that you will need to replicate specific messages and/or protocols in order to keep the thing awake.

In addition, depending on the kind of experimentation you are trying to do, you need to consider that an ECU which is simply powered on is little more than a random cluster of silicon chips. When the vehicle overall system is powered on and operating, and data is flowing, there will be much more happening and of course it will be a high noise electrical environment. The power supply, for instance, will not be anything like a clean 12 (or even 14) volts.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .