3200 strange hot shutdown/start problem

[jluis]

I continue my quest to figure out why my 3200 AC doesn’t start once it’s hot and i have managed to at least consistently reproduce the problem and rule out a few things.
Starting with the obvious, the throttle body is refurbished and that made no difference.
The car always starts well when cold and up to the coolant 90C temperature.
When I shutdown the engine the throttle body becomes quiet, turn the key again. Throttle body hums and I can start the car.

Now, once the coolant is past 90C if I shutdown the car and remove the key, the throttle body will continue humming for a minute or so. If I try to start the engine it just cranks without starting.
Digging a little bit on the electric schematic and fiddling with the control relays I figured out that the spark plug coils control relay is energized when cranking the engine under normal conditions but when it’s in this hot state it simply will not energize, hence the no start condition.

It feels like the ECU initiated the engine shutdown procedure but it’s waiting on something, that only happens once hot, to complete until it times out and during that minute or two, I cannot start the engine.
And since the ECU is directly powered by the battery, removing the key will not make it reboot.

Being an IT guy this sounds a lot like a software bug but if that was the case then a lot of people would have the same issue.
Is there anything I’m missing that could be fooling the ECU into thinking the engine is still running?
OBS reports 0 RPMs

 

[Gecara8183]

I would replace crankshaft sensor if you haven't yet. As this is a similar diagnosis

Sent from my SM-G781B using Tapatalk

 

[Oneball]

it sounds like hardware to me.



I think it’s the other way to what you’re thinking. In that it doesn’t think the engine is turning over. Crank position sensor or camshaft position sensor is where I’d look.

 

[jluis]

I would replace crankshaft sensor if you haven't yet. As this is a similar diagnosis

Sent from my SM-G781B using Tapatalk

Did it a few years ago before putting it in storage because it was cheap and people were pointing to that and made no difference.

 

[jluis]

it sounds like hardware to me.



I think it’s the other way to what you’re thinking. In that it doesn’t think the engine is turning over. Crank position sensor or camshaft position sensor is where I’d look.

I thought of that and pulled out the camshaft sensor plug in case it was sending a bad signal to the ECU once hot but it continued humming.
The thing is, after that minute or so it will always start without issues. If I pull the fuse on the ECU to reset it then it will start right away

 

[jluis]

Am I right in saying it turns over on the starter motor but there is no spark?

Exactly, and checking with a multimeter on ignition coil relay I can see that the coils are not energized.

I did a quick test just now by disconnecting the cam shaft sensor and try to start the engine (warm) and it works so it seems the camshaft sensor is not always needed. Maybe it’s needed when hot and it is faulty after all. I will plug it to the oscilloscope later today

 

[Oneball]

I can’t see it only needing it when hot.

I think that the car will not let the starter spin over if it thinks the engine is running. Usually a modern car will need to know the crank has turned over more than 1 turn before it energises the ignition.

If you can get one I’d just try the crank sensor.

 

[davy83]

i take it you have tried replacing the relay that energises the ignition coils? The reason restarting the ECU makes it go is that some error condition is reset, I don't think the ECU is part of this problem. your experiment with the CAM sensor sounds like it might be worth replacing, it does get proper hot that one. do you get any error codes when it won't start? I would have thought a faulty cam or crank sensor would throw up an error? Same goes for the throttle body (which should be making that whining noise as soon as the ignition is on) if its faulty you will get an error code?

 

[jluis]

i take it you have tried replacing the relay that energises the ignition coils? The reason restarting the ECU makes it go is that some error condition is reset, I don't think the ECU is part of this problem. your experiment with the CAM sensor sounds like it might be worth replacing, it does get proper hot that one. do you get any error codes when it won't start? I would have thought a faulty cam or crank sensor would throw up an error? Same goes for the throttle body (which should be making that whining noise as soon as the ignition is on) if its faulty you will get an error code?

Yes I shuffled relays around and I’m measuring at the relay socket terminals when I crank and I see it’s not receiving the signal.
Starting without the camshaft sensor raises a misfire code, although no CEL, which I assume is the ECU guessing the engine phase in the absence of a camshaft sensor signal.
Other than that I never get any codes.

 

[allandwf]

If it's an AC it is OBDII compliant, so any generic, cheap, code reader will let you know any codes it is throwing. My bet is, as has been said above, the crank sensor. They fail when hot, but annoyingly start working again when sufficiently cooled.

 

[jluis]

If it's an AC it is OBDII compliant, so any generic, cheap, code reader will let you know any codes it is throwing. My bet is, as has been said above, the crank sensor. They fail when hot, but annoyingly start working again when sufficiently cooled.

Yes it’s OBD, I have been reading the sensors and the error codes. The problem is triggered by temperature but it’s not the heat itself. The problem will manifest itself as soon as the coolant hits 90C. With 89 it still starts fine. And if it’s above 90 and I reset the ECU by pulling the fuse it will start immediately even before cooling.

Some more tests I did in the meantime. The coil relay will only energize when the ECU detects engine movement, as suggested above. I tried disconnecting the crank sensor with the cool engine and measuring the coil relay socket terminals and they do not get energized.
Crank sensor has been replaced a few years ago before going into storage and did not solve the problem.

According to the manual, at 90C the fan is supposed to kick in which in my case it doesn’t. Only at 95C it starts at high speed
I tried pulling the fan relays in case one of them is shorting and somehow interfering with the ECU but no joy … still won’t start after 90C until after 1 minute shutdown or ECU reset.

Also, if I stop the engine at 90C and leave the car in the MAR position for a long time, it will not start even after it has cooled off. It needs to cycle the ECU in order to start again.

 

[Oneball]

“According to the manual, at 90C the fan is supposed to kick in which in my case it doesn’t. Only at 95C it starts at high speed”

Fix this. Forget about the starting issue. I’ll bet you a pint the start issue will go away when you get the fan working.

 

[jluis]

“According to the manual, at 90C the fan is supposed to kick in which in my case it doesn’t. Only at 95C it starts at high speed”

Fix this. Forget about the starting issue. I’ll bet you a pint the start issue will go away when you get the fan working.

Yes but looking at the electric diagram the fan relay is driven directly by the ECU, which means that if the relay doesn’t receive the signal from the ECU to start then I might have a damaged ECU …
I also noticed that when the car is in this condition the voltage at the crank sensor is lower than usual which might indicate an internal problem with the ECU, wiring or connectors

 

[Oneball]

Yes but looking at the electric diagram the fan relay is driven directly by the ECU, which means that if the relay doesn’t receive the signal from the ECU to start then I might have a damaged ECU …
I also noticed that when the car is in this condition the voltage at the crank sensor is lower than usual which might indicate an internal problem with the ECU, wiring or connectors

Maybe but we’ve kinda been here before and I can’t find the thread.

Someone had a hot start issue.

We went through all this checking testing and stuff.

They then said “the thermostat is stuck open”

I said “fix the problem you know”

After much procrastination they changed the thermostat and the hot start issue went away.

I can’t tell you why but it worked but “fix the problem you know”.

 

[Oneball]

Part 5 is what you need to check

 

[allandwf]

Yes it’s OBD, I have been reading the sensors and the error codes. The problem is triggered by temperature but it’s not the heat itself. The problem will manifest itself as soon as the coolant hits 90C. With 89 it still starts fine. And if it’s above 90 and I reset the ECU by pulling the fuse it will start immediately even before cooling.

Some more tests I did in the meantime. The coil relay will only energize when the ECU detects engine movement, as suggested above. I tried disconnecting the crank sensor with the cool engine and measuring the coil relay socket terminals and they do not get energized.
Crank sensor has been replaced a few years ago before going into storage and did not solve the problem.

According to the manual, at 90C the fan is supposed to kick in which in my case it doesn’t. Only at 95C it starts at high speed
I tried pulling the fan relays in case one of them is shorting and somehow interfering with the ECU but no joy … still won’t start after 90C until after 1 minute shutdown or ECU reset.

Also, if I stop the engine at 90C and leave the car in the MAR position for a long time, it will not start even after it has cooled off. It needs to cycle the ECU in order to start again.

Interesting. A pity we don't have a flow chart , or old style ladder logic, to show what needs to be clear or enabled to allow the engine to start. I'm not sure software could change in someway to cause just this fault, more likely a contact from somewhere. We need to ask why would an engine not start if it has a high temperature, above 90 on the gauge may be a red herring, wherever the contact goes it may think it is 150 degrees or similar, IE a trip, shut down condition, as you would have on industrial machinery, thus not allowing it to start until it has been reset by you powering off the ECU.

 

[jluis]

Part 5 is what you need to check

You may be onto something...
According to this manual page I can measure the resistor with a multimeter on terminal 87 of both relay sockets, which I did and returned infinity, therefore the resistor might be shot.

But how does the bloody computer know that, if the relays are directly connected to the fan motors and there doesn't seem to be any current sensing pin to the ECU anywhere?!
And if that's the case, why is there no error code?!

Anyway, time to get under the car to measure the resistor directly and order a new one if it's really broken and find out ...

 

[Potakas]

Well, I am not very experienced but had a similar issue and apparently everything initiated because the 1st fan wasn't working, in my case also the thermostat needed replacing as well.

This was the part that had failed, which is the same as Oneball said.
980000960

 

[jluis]

Well, I am not very experienced but had a similar issue and apparently everything initiated because the 1st fan wasn't working, in my case also the thermostat needed replacing as well.

This was the part that had failed, which is the same as Oneball said.
980000960

Mine is definitely failed. I pulled it out to find the internal fuse loose and the thing is impossible to solder.
Waiting on parts now ...

 

[jluis]

FYI, the radiator resistor is yet another NLA part from Maserati … bloody **** …
I think they are the same part that is used in some alfas.
For now I could source one from a German seller