wrote:
> Is there some conclusive test for an ECM failure?
> Is there something significant in the car's behavior that I described
> or are these units prone to failure?
>
> The following is originally from a vendor forum, I think.
>
> "...But, I have to ask, if you drove your car to the dealer with this
> problem what part of the computer failed? My experience with bad
> computers is really bad, as in you can hardly get anywhere with the
> vehicle. Or you just can't get anywhere. Car computers are real tough
> pieces of silicone. I've seen wiring harnesses melted together and
> after the repair the computer shook itself and went back to work...."
>
> Are there some other steps that I can go through to test the
> sub-components and rule out a simpler problem? What are the likely
> sources, given the symptoms or are there no other possible
> explanations?
>
> Thanks

I just happened on your new posts by chance. the ECM, you
basically have to verify that all inputs, outputs, powers and grounds
are functioning properly before you condemn it. If your Hyundai ECM is
the same build type and manufacture as similar year Mitsubishis, then I
would highly suspect it as the cause of nonsensical garbage codes. You
still need to check powers and grounds to the ECM, and ATLEAST make
sure that the input signals are sensible.

It is difficult to test an O2 sensor on a vehicle with a concurrent
funky ECM problem. You can force it to read rich by pinching off the
fuel pressure return hose, and you can force it to read lean by
introducing a large vacuum leak. Those two methods will tell you if it
is capable of reacting to known conditions by producing from 0 - 1 volt
accordingly. Chances are that your new sensor is OK, so you can move
on to other suspect areas for now.

The possible code 25 refers to a failure in the barometric circuit
inside the airflow meter. This could cause a large shift in fuel trim
if it was way off calibration, or shorted/open. Similar story: I had
a Land Cruiser in the shop that was setting a baro code. This
particular vehicle, unlike other Toyotas, inferred calculated Baro
readings from other sensors. IOW, it set a bizarre baro code even
though it didn't have a real baro sensor. It turns out that the ECM
was so far off in it's baro calculation that it ran extremely lean to
the point of near misfire and very high NOx (failed an emissions
inspection). A new ECM fixed the Land Cruiser.

If you are still suffering black smoke out the exaust, perhaps you
could find a junkyard airflow meter to try out. Your car uses a Karmen
Vortex design that outputs a frequency to the ECM based on airflow,
plus discrete air temp and baro inputs. There's plenty to go wrong
with this design that could send a false richening command to the ECM.
If the AFM is hard to find or too expensive to purchase as a quick
cheap parts swap, then it's back to more diagnosis. I am having too
difficult a time with deciphering these codes over usenet to be of much
use. I don't feel confident that you are getting consistent results
from the code retrieval process - and I blame Hyundai for that, not
you. Their service info seems pretty useless.

Toyota MDT in MO