Engine won't accelerate until check engine light comes on!

[AndyNW66]

I'm looking for advice regarding my NA, which I've had from new.
For some years now the check engine light has come on after the engine has warmed up for a few minutes. This seemed to have little effect on the running of the car and isn't my primary problem, but I haven't been able to fix it. It may become critical if it is an MoT failure issue rather than just a note. The more significant problem appeared last year. After some initial difficulties starting the engine it now starts easily BUT won't accelerate until the check engine light has come on! It will respond to light blipping of the throttle but just dies when trying to drive off. Once the check engine light has come on the engine responds normally to the throttle at which point I feel safe to drive off. However, often after a short distance the light will go off again and the engine will die. A few seconds or minutes at idle and the light will come on again and I can drive off again.
A spark tester shows good sparking as the engine is dying, suggesting a fuel issue. The fuel filter is OK.
All suggestions gratefully received.

[matts1972]

Just a wild guess: this sounds like a sensor fault (potentially TPS or MAP).
A sensor might give erratic data until the ECU notices the fault and ignores this sensor altogether.
If you cannot get an error code to pinpoint the sensor in question, I would try to unplug one of the sensors at a time.
If you pick the right one, you should get the CEL but the engine should run fine.

This is a shot in the dark - I have not tried it on the Elan but on other cars.

Matthias

[AndyNW66]

Thanks Matthias, I'll try that. I'd already replaced the MAP without success, but I'll have a look at the TPS -as soon as I can work out how to separate the connector. I might also be able to check the voltages at the sensor. No fault is shown by the paper clip test. And it seems Elanscan doesn't work on NAs.
I'm also in the process of removing the fuel pump to check for grunge, but it seems you need to remove a load of trim first, and then the hood panel catch.

[matts1972]

No fault is shown by the paper clip test.

Hi Andy, I guess you did the paper clip test while the CEL was lit? Then you really should go through all suspect sensors. I cannot imagine how a fuel problem could disappear with CEL lighting up.
In regards to sensors I would assume TPS and MAP have the most immediate effect, but also cooling fluid and air temperature could have a very negative impact on the mixture if they are way off. It’s a shame that ElanScan does not work on the NA.
Good luck with finding the culprit
Matthias

PS: If I remember correctly, the connector for the TPS is not directly on it but a few inches down on the cabling. I also tried to unplug it at the wrong place (this is from memory, I am not at my car)

[Jamie N]

I'm also in the process of removing the fuel pump to check for grunge,

You reckon there might be some Nirvana in there Andy?.

[AndyNW66]

Matthias, you are right, the tps connector is about 6inches down the line, but it doesn't come apart easily. The paper clip test just produces 1-2 continually. The voltages at the tps sensor seem to be correct and respond properly to the throttle movement, as far as I can tell from a voltmeter.
Interestingly I could hear some solenoid clicking which I've traced to the egr solenoid so I've tried disconnecting that. The result seemed to be reasonable engine running and acceleration with the cel on, but it may just be that the engine had warmed up. I'll try again the morning.

Strangely the cel has now taken to flashing 1-2 continually with the ignition on but the engine not running, with no paper clip! With the engine running the cell is flashing randomly, sometimes. So the ecm seems to have put itself into diagnostic mode. Maybe this suggests a short somewhere.

[matts1972]

The CEL behaving strangely sure does not help. Does disconnecting the battery clean this up?
That being said, the situation has not become easier with the TPS (and MAP) checking out fine.
I still believe the ECU notices a defect of something directly connected to it. I would not have thought this to be the EGR solenoid, though, but hopefully you found it.

[rip]

I can offer some information which I think is useless, but hopefully it may mean something.
CEL 1-2 is the code it uses to signify error reading mode.
If you have error 34 (Just an example, I have no idea what it is, if indeed it is valid), then you will see 1-2, 1-2, 1-2, 3-4, 3-4, 3-4, 1-2, 1-2, 1-2.
If you also have error 48, then you should see 1-2, 1-2, 1-2, 3-4, 3-4, 3-4, 4-8, 4-8, 4-8, 1-2, 1-2, 1-2.

The engine running ok only when the CEL is lit sounds strange but there is a logic to it.
The CEL shows when the ECU acknowledges a bad reading from a sensor, so it could well be programmed to ignore it & use a default mapping at this time. With no CEL, the ECU understands all sensors to be working normally, so it will react as if the sensor is correct.
An example to this is the coolant temperature sensor. If it is designed to work between -40c & +140c, then the ECU will have a mapping for it at -38c. It relies solely on this sender so will fuel for a mapping it has for -38c...but if you are driving with the roof down wearing just a T-shirt, this reading is certainly wrong. If the reading then goes further out, to -41c, the ECU is programmed to understand this to be an error, show the CEL & switch to a default mapping. This could in theory be a complete shutdown but is more likely to be a conservative set of parameters
I am not suggesting that the coolant temp sensor is your issue, or that these temperatures examples are what the car uses, but it is my understanding of how the ECU works.

[AndyNW66]

Thanks for your response Rip. That agrees with my understanding.
Using e.g. a paperclip to short two pins on the ALDL causes the ecu to enter a diagnostic mode and flash the cel. Among the many things puzzling me at the moment is how the ecu has apparently entered diagnostic mode without a short in place across the two pins.
Another is why is the egr solenoid clicking when the ignition is on, but the engine not running.
I can also see the secondary throttle actuator twitching when the engine is idling. Surely it should only move at high rpm (5000+?).

[lotusflasherman]

Thanks for your response Rip. That agrees with my understanding.
Using e.g. a paperclip to short two pins on the ALDL causes the ecu to enter a diagnostic mode and flash the cel. Among the many things puzzling me at the moment is how the ecu has apparently entered diagnostic mode without a short in place across the two pins.
Another is why is the egr solenoid clicking when the ignition is on, but the engine not running.
I can also see the secondary throttle actuator twitching when the engine is idling. Surely it should only move at high rpm (5000+?).

Correct about the secondary throttle - on the SE it's 4,700 rpm so I suspect NA is in the same area. Secondary throttle is operated by a vacuum actuator fed from a vacuum solenoid valve controlled by the ECU. Vacuum comes from the vacuum reservoir which is fed from primary throttle. EGR solenoid is another vacuum operated device. - I'd be checking all your vacuum pipes for leaks ..

[AndyNW66]

Thanks Phil, it's certainly worth checking the vacuum pipes but I'm not convinced. It seems to me a bit like putting the chicken before the egg! Or some such expression. The egr solenoid, and the secondary throttle solenoid, should be driven by a signal from the ecm and it shouldn't matter whether the vacuum reaches the vsv or not. The egr solenoid should operate when:-
Coolant temp is above 30C,
TPS is above idle,
RPM above 1000 and
MAP indicating the engine is under load.
With the engine not running but the ignition on only one of those should give a .YES., if the coolant is warm. And yet my solenoid is clicking away. The other signals might be erroneous on my car.

I checked the egr connector with a multimeter. It should show open circuit with ignition on, engine off, unless the diagnostic pin is grounded, but it was indicating a finite, and varying, resistance to earth. Again this points me at a possible short circuit or erroneous signals to the ecm, or from the ecm.
My next job is to see if I can check the pin voltages on the ecm. The EML Service Notes will tell me what to expect.

[matts1972]

I checked the egr connector with a multimeter. It should show open circuit with ignition on, engine off, unless the diagnostic pin is grounded, but it was indicating a finite, and varying, resistance to earth. Again this points me at a possible short circuit or erroneous signals to the ecm, or from the ecm.
My next job is to see if I can check the pin voltages on the ecm. The EML Service Notes will tell me what to expect.

I agree with this - your ECU currently is stuck in service mode which might explain the behaviour of the EGR.
I also would concentrate on why the diagnostic pin seems to be grounded.

[GeoffSmith]

Measure the voltage on the diagnostic pin. If it isn't close to 5V then it's being dragged down by something.

[AndyNW66]

Not much success today measuring pin voltages. The ecm is on a short cable length which means it's difficult to get a voltmeter probe at back of the pins. My plan is to use a safety pin on each pin in turn. The 24 pin and 32 pin connector blocks come with plastic clips that can be removed to allow better access. Note to self:- don't poke things at the 12volt supply pin. I thought I'd blown the fuel pump relay, a maxi fuse behind the switches panel or the ecm itself. Actually it was just the ecm fuse (10amps).
The diagnostic pin on the ALDL shows 4.8 volts, which I'm thinking is close enough to 5.0.
I've adjusted the throttle stop back so as to reduce idle rpm and bring the tps voltage down to 0.4+/-.0.02. It was about 0.48, i.e. above idle as regards the ecm.
The secondary throttle actuator is continuing to show random movements while the engine is idling. The egr solenoid clicks, ignition on, engine warm.
Day off tomorrow, golf!

[AndyNW66]

Good progress this afternoon. I've managed to check most of the pin voltages at the ecm, except the co pot, which somehow I missed. I'll check it tomorrow.
The output voltages are out of spec on the egr control and the secondary throttle, which matches the symptoms seen. Also the electronic spark timing voltage is well out of range with engine running.
The only incorrect input is the coolant temp sensor which is out of spec at idle and showing variable values with key on.
According to my EML service notes the CTS affects fuel delivery, idle speed, spark timing, egr control, secondary throttles and CEL. (And air con, but I don't have any of that). In other words, everything. Bingo!
Fluctuations of the CTS voltage coincide with the noisy clicking of the egr solenoid. Cause and effect?
Googling elan m100 coolant temperature sensor leads me to an LEC article!
Clearly I need to check the electrical contacts and possibly replace the sensor. I'm not yet sure that's 100% of my problems, but it's got to be a huge step forward.

[GeoffSmith]

You could always measure the resistance at the CTS and from the ECU connector. Resistance values at spot temperatures are in the Wiki.

[AndyNW66]

That's a good idea, thanks Geoff. That should tell me whether I have a faulty sensor or a wiring/connector issue.

Just for the benefit of any NA owner reading this in future, the thermostat housing is different on NAs. I'm not sure yet whether that makes it easier or harder to remove the CTS.

[AndyNW66]

A disappointing lack of success since the last posting. I replaced the cts but it seems to have had no significant effect. I now see from the EML manual that I should be using a high impedance voltmeter, otherwise I can't trust the readings I'm getting.
I still think that the clicking egr solenoid is telling something significant, I'm just not sure what. I feel I ought to be checking the idle air control valve, but I can't get at it. It seems to be really inaccessible.
I'm tempted to suspect the ecu itself but I can't see why my problem would be temperature-sensitive.