Re: ur-q EFI - doing it again.....

[Robert Houk - SMCC WorkGroup Server Firmware <Robert.Houk@East.Sun.COM>]

   It's doing it again. (big surprise - few mechanical devices actually =
   heal themselves)

Yeah, but we keep hoping, eh?

                    After about 1 hour of normal closed-loop operation the =
   ur-q CIS gets funky. I had the % duty-cycle meter on the CIS FV this =
   time. In normal operation the duty-cycle runs about 45-60% and slowly =
   increases from the low end to the high end of this scale as the throttle =
   is slowly opened and then jumps to about 80% and steady open-loop under =
   boost, all very well and good.=20

   What happened with this observed failure is the duty-cycle slowly =
   increases from 60% to 70-80-90 and then up to 100% duty-cycle and, of =
   course, the O2 sensor reading goes to full rich. It looks like the FV is =
   driving the CIS to full rich when it should not be, but only after =
   0.5-1.0 hours of normal operation. This failure is intermittent and can =
   come and go after it starts to happen. The coolant temp sensor is a =
   prime suspect, but is new. In previous observed failures without the =
   benefit of the % duty-cycle meter on the FV the O2 meter would indicate =
   full rich and then, for short periods like 5-20 seconds it would =
   indicate full lean. I wonder if this was due to the FV overheating due =
   to running at 100% duty-cycle and shutting down completely and then once =
   cooled down a bit starting up again and kicking the mixture back up to =
   full rich. I have not duplicate this full rich/full lean problem (yet) =
   with the duty-cycle meter connected to the FV, but I bet I =
   will...........

Idle, on the road, or ???

If you see the duty cycle go *up* and O2 sensor read rich, then it sounds
like the ECU is the focal point. I could never fool mine into going much
past 90% (max-cold engine, etc.), so that 100% number is suspect.

I think the best operational test would be (assuming this happens at
idle...and sustains itself long enough...) would be to manually close
the full-throttle switch -- the duty cycle should lock at 50%. This
should at least tell you whether or not the ECU can still exercise
control over the FV. (Of course you might want to verify this operation
works normally when the system is "working normally" first...).

If that works, I'd start to suspect wiring leading to the ECU. E.g.,
the O2 sensor [voltage] is not making its way into the ECU proper; or
maybe the block temp sensor [voltage] is not making it; or ???
At least the ECU uses a nice widely-spaced connector pinout, so it's
relatively trivial to get to the pins from the inside of the ECU (I
drove around for a year with my ECU dangling (well, resting on the
passenger side floor/wall) with assorted probes hanging off of various
pins...).

My first bet would be bad O2 signal, followed by bad block temp signal
(although I don't believe that can force open-loop past the 80% point
or so -- I'd hafta go back and lookup max open-loop duty cycle -- and
assuming your ECU hasn't been modified...). Although, if you've re-
cently replace the temp sensor, that in itself is pretty suspicious,
try putting the old one back in (or just verify voltages at the pins
inside the ECU).

Hmmm...another interesting experiment would be to disconnect the O2
sensor -- I don't remember what the ECU's behavior is in this state,
but I seem to remember that it does "react" to no O2 sensor in some
definite manner (try it working and not-working!).

I don't think any other sensor will affect the FV duty cycle. The
intake manifold air-temp sensor only diddles the timing, and then
only under boost.

Unfortunately, I think I've long-ago lost most of my "voltage-at-the-
pins" notes (one of the drawbacks of moving and "cleaning up"!), so
I can't tell you offhand what "good" values are.

One other possiblity would be a thermally-sensitive compenent inside
your ECU, see if the problem goes away with the ECU out/open with
the cover off (get one of those little car fans and keep a good cooling
airflow on the ECU; should run substantially cooler than closed up
and hiding beside the glove box).

I hate erratic intermittant problems!

Good luck!

					-RDH