633csi starts then stalls, nothing seems out of the ordinary.

[EmperorNero]

I have a 633csi, build date 11/81. I replaced the plugs, distributor cap, and distributor rotor recently due to a misfire. It cleared the misfire up, and the engine was running smoothly. After driving around for about 20 minutes, I noticed the car was beginning to idle extremely high, around 3000 rpm. After a few more minutes, the car stalled out. Since then, the engine has started every time I crank it but immediately stalled. Whether I have the throttle closed or open doesn't change anything. Fuel pressure is 38 psi at the rail. Smoke tested intake for vacuum leaks, found a few small ones and fixed them but it didn't correct the issue. Speed and reference sensors have proper resistance, car won't start at all with them unplugged so it seems they are good. Throttle position sensor was out of adjustment, corrected it, resistance tests fine, didn't fix the issue. AFM resistance tests fine, door isn't binding. IAC resistance tests fine. I've seen many people say the car could be firing on the cold start valve but not the main injectors, however I disconnected the fuel hose from the cold start valve in order to test fuel pressure and it still starts up then dies. I've tried swapping both the main relay and the fuel pump relay with ones from my e36, known good, and it didn't correct the issue.

Does anyone have any ideas? I have yet to test the wiring harness for any opens or high resistance, but other than that, does anyone have any ideas as to what could be causing this? The excessively high idle really makes me think it's a vacuum leak, but I've checked and double checked and can't find any others.

Been lurking this forum for quite some time, any help is greatly appreciated!

[bpoliakoff]

Before anyone starts throwing things against the wall, what injection system? Ljet or Motronic? if Motoronic 1.0 or 1.1? You are right on the cusp. I am guessing it is a US manufactured car?

[EmperorNero]

Before anyone starts throwing things against the wall, what injection system? Ljet or Motronic? if Motoronic 1.0 or 1.1? You are right on the cusp. I am guessing it is a US manufactured car?

It's a US car, Motronic. I believe 1.0, ECU part number is 0 261 200 008.

[Brucey]

you were right to look for air leaks; there is only one way the idle speed can be too high and that is if there is too much air getting into the engine. The fuel pressure should reduce when the engine is running at idle; it should be constant (2.5 bar or 3.0bar depending on model) w.r.t. the inlet manifold, which sees a lower pressure at idle.

US-spec cars run under closed loop control but only once the engine is hot which means that the running characteristics may be very different between a cold and a hot engine, when there is a fault of some kind. When the engine is cold the Motronic will be running 'open loop'.

Under open loop control, if the idle contact in the TPS (throttle position switch) is engaged, fuel is cut altogether above about 1200 rpm and is only reinstated at about 750rpm. Fuel is then added in proportion to the AFM signal. [Things are similar under closed loop control but the fuel is added such that the mixture strength remains stoichiometric; if there is a fuelling fault it can be masked by the closed loop control.] The control system attempts to vary the air intake (via the ICV (Idle Control Valve)) such that the idle speed is maintained at ~750rpm (exact speed varies with model).

There are various things that control the mixture strength under open loop conditions and IME these things can be well out of kilter on a car that normally runs closed loop (everything sorts itself out more or less under closed loop control) and quite a lot of people will just put up with weird running when the engine is cold, right up until the car becomes completely undrivable. By contrast the same faults on a car that normally runs open loop are usually attended to much sooner; the car more quickly becomes undrivable when there is a fault of any kind.


A very common problem is that the ICV is driven full open on a cold start (because the thottle linkage wears such that the throttle butterfly slowly closes itself at idle) and this can even fry the Idle control circuit . However more commonly the ICV simply jams open. Often cleaning the ICV helps to stop it from becoming jammed and resetting the throttle butterfly (butterfly first, TPS to match butterfly setting after) helps to prevent a recurrence. However the ICV can also jam shut or remain shut because the control electronics are fried.

So I'd suggest that you check the ICV/ICV control and perhaps try running the car with the ICV pipes blanked off; this will soon tell you if excess air is getting in that way or not. The root cause may be a combination of a TB that needs to be reset and a sticky ICV.

cheers

[EmperorNero]

you were right to look for air leaks; there is only one way the idle speed can be too high and that is if there is too much air getting into the engine. The fuel pressure should reduce when the engine is running at idle; it should be constant (2.5 bar or 3.0bar depending on model) w.r.t. the inlet manifold, which sees a lower pressure at idle.

US-spec cars run under closed loop control but only once the engine is hot which means that the running characteristics may be very different between a cold and a hot engine, when there is a fault of some kind. When the engine is cold the Motronic will be running 'open loop'.

Under open loop control, if the idle contact in the TPS (throttle position switch) is engaged, fuel is cut altogether above about 1200 rpm and is only reinstated at about 750rpm. Fuel is then added in proportion to the AFM signal. [Things are similar under closed loop control but the fuel is added such that the mixture strength remains stoichiometric; if there is a fuelling fault it can be masked by the closed loop control.] The control system attempts to vary the air intake (via the ICV (Idle Control Valve)) such that the idle speed is maintained at ~750rpm (exact speed varies with model).

There are various things that control the mixture strength under open loop conditions and IME these things can be well out of kilter on a car that normally runs closed loop (everything sorts itself out more or less under closed loop control) and quite a lot of people will just put up with weird running when the engine is cold, right up until the car becomes completely undrivable. By contrast the same faults on a car that normally runs open loop are usually attended to much sooner; the car more quickly becomes undrivable when there is a fault of any kind.


A very common problem is that the ICV is driven full open on a cold start (because the thottle linkage wears such that the throttle butterfly slowly closes itself at idle) and this can even fry the Idle control circuit . However more commonly the ICV simply jams open. Often cleaning the ICV helps to stop it from becoming jammed and resetting the throttle butterfly (butterfly first, TPS to match butterfly setting after) helps to prevent a recurrence. However the ICV can also jam shut or remain shut because the control electronics are fried.

So I'd suggest that you check the ICV/ICV control and perhaps try running the car with the ICV pipes blanked off; this will soon tell you if excess air is getting in that way or not. The root cause may be a combination of a TB that needs to be reset and a sticky ICV.

cheers

I'll look into that, but the concern i have with this explanation is that I had been driving around for about 20 minutes, in other words in closed loop, when the problem first manifested and I stalled while driving. What is the function of the screw on the ICV?

[Brucey]

the big brass screw is an adjustable bypass. The idle control and bypass arrangements vary on early motronic from later ones.

BTW as I mentioned above, adjusting the TPS is a complete waste of time (actually, its worse than that it is counterproductive) unless you set the TB correctly beforehand to the factory specs.

The high idle speed has most likely got to be a massive air leak, TB jammed open, or a jammed/faulty ICV.

Have you checked the throttle cable to be sure that the cable isn't binding and making the throttle jam open?

cheers

[EmperorNero]

the big brass screw is an adjustable bypass. The idle control and bypass arrangements vary on early motronic from later ones.

BTW as I mentioned above, adjusting the TPS is a complete waste of time (actually, its worse than that it is counterproductive) unless you set the TB correctly beforehand to the factory specs.

The high idle speed has most likely got to be a massive air leak, TB jammed open, or a jammed/faulty ICV.

Have you checked the throttle cable to be sure that the cable isn't binding and making the throttle jam open?

cheers

Well I gapped throttle to .020 and readjusted sensor, throttle cable isn't binding. Cleaned ICV and the valve inside seems to be moving as normal. Nothing has changed. One of the vacuum leaks I had found previously was the joint where the vacuum hose connects to the brake booster. I disconnected the booster hose and capped off the hole on the intake manifold. Could blocking off a line like that cause issues?

[Brucey]

blocking off the brake servo vacuum pipe shouldn't cause the engine to run badly but of course it will stop the brakes from working properly.

If the idle is at 3000rpm, something is letting 3000rpm's worth of air into the engine. This is quite a big air leak into the inlet manifold or a fault in the various bypass hoses.

If you block/nip off the ICV hose, and this slows the motor, you have found your culprit. If it makes no difference then you should look elsewhere.

cheers

[bpoliakoff]

Unless I missed something reading through all of this, have you tried pulling the dipstick? If pulling it doesn't affect idle you have a problem there. A simple something that is often overlooked.

[EmperorNero]

blocking off the brake servo vacuum pipe shouldn't cause the engine to run badly but of course it will stop the brakes from working properly.

If the idle is at 3000rpm, something is letting 3000rpm's worth of air into the engine. This is quite a big air leak into the inlet manifold or a fault in the various bypass hoses.

If you block/nip off the ICV hose, and this slows the motor, you have found your culprit. If it makes no difference then you should look elsewhere.

cheers

Unfortunately I can't perform any tests with the engine running as the engine stalled after it began to idle at 3000 and has refused to run for more than a couple seconds since. I've followed all the lines through the engine bay to the charcoal canister, FPR, cabin temp sensor, and I can't find any leaks at all.

[Brucey]

if the engine won't run for 'more than a couple of seconds' then this is symptomatic of a no-start issue, and this isn't necessarily related to the high idle speed fault.

I'd therefore try a lot of the usual no start things before going much further. If the car fires on cranking and runs for a couple of seconds this sounds very much like you are getting cold start injector function and sparks during cranking, but are missing either sparks or main injector function in 'run'.

The most common cause of this particular fault is that the main relay is bad; it sees about x5 more current draw as soon as the main injectors want to fire, and this is enough to cause problems if the relay contact is bad. You might well still get sparks during cranking though. You can usually install a jumper in place of the main relay, as a test. [NB it isn't safe to drive the car like this, because in the event of a prang, the fuel pump may carry on running.]

cheers

[hornhospital]

I've seen many people say the car could be firing on the cold start valve but not the main injectors, however I disconnected the fuel hose from the cold start valve in order to test fuel pressure and it still starts up then dies.

I think it's significant that it still starts and dies with no fuel to the CSV. It's getting a shot of fuel from SOMEWHERE or it wouldn't fire. And I agree 100% with Brucey about there being a very large vacuum leak somewhere, regardless of whether or not you found it doing a smoke test. Could be injector O rings....

[Brucey]

It occurs to me that if the CSV fuel hose is disconnected from the fuel rail, there will still be fuel in the line to the CSV. When the CSV opens (during cranking) there isn't anything like the same pressure across the injector as normal, but there will still be still some rather than none, so that fuel in the line will be delivered, just more slowly than normal. I'd expect the amount of fuel so delivered to cause the car to start (at least once) and then quit, if the main injectors are not firing.

cheers

[hornhospital]

I would like, for the sake of elimination, that the CSI fuel line be disconnected and the electrical connection to the valve be disconnected, then several starts attempted. If it fires more than once, it's not the CSI. Pressure after one start would be nonexistent, and fuel wouldn't get past the injector. I contend it would have a difficult time getting past it at all with only atmospheric pressure behind it.

[Brucey]

I would like, for the sake of elimination, that the CSI fuel line be disconnected and the electrical connection to the valve be disconnected, then several starts attempted. If it fires more than once, it's not the CSI. Pressure after one start would be nonexistent, and fuel wouldn't get past the injector.

I agree that (or something very similar) would be a good test.

I contend it would have a difficult time getting past it at all with only atmospheric pressure behind it.

When everything is normal on motronic systems there's only ever three atmospheres of pressure, tops, and on that particular model I think it may be 2.5? Note also that the flowrate through an injector is proportional to the square root of the pressure. This means that if you want to double the flowrate, you need four times the pressure, and if the driving pressure is only 1/4 normal, the flowrate is 'only' halved.

Remember that the normal CSI flowrate is complete overkill; IIRC it is about 20cc per second during cranking, and is intended to allow the car to start immediately even if the temperatures are about -20 degrees C and the petrol is barely wanting to produce any vapour at all. I have every reason to suppose that the car will fire up quite happily (at 'normal' ambient temperatures) even if the flowrate through the CSI is about half normal.

cheers

[EmperorNero]

i have started it about 10 times with the csv line disconnected. started and then died each time. checked resistance on all electrical connections to ecu, none more than 1.0 ohms.

[EmperorNero]

I would like, for the sake of elimination, that the CSI fuel line be disconnected and the electrical connection to the valve be disconnected, then several starts attempted. If it fires more than once, it's not the CSI. Pressure after one start would be nonexistent, and fuel wouldn't get past the injector.

I agree that (or something very similar) would be a good test.

I contend it would have a difficult time getting past it at all with only atmospheric pressure behind it.

When everything is normal on motronic systems there's only ever three atmospheres of pressure, tops, and on that particular model I think it may be 2.5? Note also that the flowrate through an injector is proportional to the square root of the pressure. This means that if you want to double the flowrate, you need four times the pressure, and if the driving pressure is only 1/4 normal, the flowrate is 'only' halved.

Remember that the normal CSI flowrate is complete overkill; IIRC it is about 20cc per second during cranking, and is intended to allow the car to start immediately even if the temperatures are about -20 degrees C and the petrol is barely wanting to produce any vapour at all. I have every reason to suppose that the car will fire up quite happily (at 'normal' ambient temperatures) even if the flowrate through the CSI is about half normal.

cheers

I've performed this test with no luck. My car has a green box mounted next to the ECU which is apparently the idle control module. I can't find any information whatsoever about this unit online, but was wondering if there was any known way to test it.

[Brucey]

I've performed this test with no luck.

meaning what exactly?

cheers

[hornhospital]

I guess it keeps starting and dying doing those tests. Question is, where's the fuel coming from if there's no fuel line to the CSI?

[EmperorNero]

I guess it keeps starting and dying doing those tests. Question is, where's the fuel coming from if there's no fuel line to the CSI?

Probably the regular injectors.

[EmperorNero]

I've performed this test with no luck.

meaning what exactly?

cheers

I've been away from this for a while. Been busy with other things. Even with the CSV disconnected from the fuel rail and the wiring harness, the engine would still start up then die, over and over. The previous owner said the injectors needed a rebuild, so I decided to replace them. With the new injectors installed, the engine no longer starts up at all unless the CSV is plugged in. It seems like the old injectors might have been leaking.

I got a noid light and plugged it into the wiring harness connectors, and it turns out the ECU is not firing the injectors. Is there anything that could cause this besides a bad main relay or ECU? Is there a conclusive way to test the main relay? Thanks.

[bpoliakoff]

Is it actually starting or just kind of knocking when you shut it off, like trying to start? If so it needs a valve job and is carbon build up in the combustion chamber

[EmperorNero]

Is it actually starting or just kind of knocking when you shut it off, like trying to start? If so it needs a valve job and is carbon build up in the combustion chamber

it was starting on the cold start valve, the ecu is not firing the main injectors