You have already heard the main symptoms of the car – sudden (otherwise inexplicable) poor idle, rough running below 2,500 rpm, poor fuel consumption, and no real signs of trouble from the Rovacom except a split between the stored and current throttle voltages. Kevin had done some diagnostic work and found that the +5 volt supply to the Throttle Position Sensor had failed, and sent the ECU to be repaired – but when it was returned there was still no +5 volt supply when it was installed and it was assumed that the ECU was still faulty.
The problem turned out to be the pressure sensor in the fuel pump unit, which is also supplied by the same +5 volt drive from the ECU. This had failed to a short circuit and was effectively grounding the +5 volt line - as soon as the sensor was disconnected things returned to normal. This sensor was introduced in the 2002 model year with the installation of the “advanced evaps” (Advanced Evaporative Loss – AEL) system. This system depressurizes the fuel tank and uses the pressure sensor to check for leaks, and is (was) required for US emission compliance. This sensor failure in Henrietta required the replacement of the fuel pump assembly – which was not entirely unfortunate. GEMS cars (at least the 2002+ cars – perhaps the earlier ones too) - have some instances of the tank being contaminated by silicon sealer extruded from the pump/tank interface when it the pump is installed. This is then ingested into the valve at the foot of the small tank that surrounds the pump (and which is designed both to prevent the pump running dry and provide cooling at low fuel levels) and prevents it from retaining any fuel – and this, in turn, results in the pump overheating. Henrietta had a lot of this contamination in the tank and stuck in the valve. You can (usually) remove the pump by undoing the tank retaining bolts and moving it around until you can maneuver the pump out, but this proved to be difficult with Henrietta and so the tank was removed and the silicon vacuumed out.
So, as in all good stories, the heroine was restored to her former glory. I think that it is worth replacing the IACV and the TPS if there is the slightest sign of a problem; they are both mechanical (unlike the other sensors) and so have a finite life, even if it is often a long one. Richard may choose to have this done. The later IACVs also have a shorter time constant (they move faster) than the earlier ones. You are probably aware that if the time constant of the control system approaches that of the system being controlled the combination can become unstable, so an IACV with a faster response is additional insurance against problems in this regard.
So what are the lessons from all this? Here are some suggestions:
1.It is worth pulling the fuel pump on all GEMS cars and
checking for silicon contamination, cleaning the tank if there is silicon
present, and replacing the fuel pump (as a precaution) while the
thing is apart. The philosophy of “if it ain’t broke don’t fix it”
may be attractive, but chanting this mantra is not going to help when the
pump fails in the rain on the beltway at 3.00 AM.
2.Spontaneous failures are unlikely to be the ECU;
it has an excellent reputation and is very well engineered. Sensors
are of a more disposable nature, and many of them share common electrical
circuits (grounds, if not power, like the two culprits in the story).
So if you do have an odd failure it is worth
investigating to see if there are any common elements
that would cause complex results.
3.If one had written a manual for the GEMS system it might
be worth updating it in view of any developments that might have occurred
in the intervening
years. One might even consider publishing the GEMS
electrical schematics as a service to Morgan/GEMS owners.
4.Morgan/GEMS owners who feel that they would like the
GEMS electrical schematics before the manual is updated might like to contact
the writer of the
manual OG and ask for them, if he is not off swanning
around on some damn fool expedition to God knows where.