delayed start if engine shut down 5 mins or longer

[rennbod]

A friend and I have been struggling with a strange problem with his 2006 997 4S with tiptronic transmission.

 

When you start his car you need to turn the key and let it crank on the starter motor for around 10 seconds before it will start, and when it does it seems to be a bit lumpy for the first second or two.

 

However, if you shut it down while it is running and restart it right away it fires straight up within a second.

 

We have pulled off a coil pack and put a new plug into it (earthed out with a jump lead) and observed that while it is cranking there is no spark, apart from maybe one every 3 seconds, and then rapid sparking at the plug as soon as it fires up.

 

We have also hooked up fuel pressure monitoring kit to the car, and the fuel pressure seems to be right, as I thought it may be the fuel pump check valve or pressure regulator bleeding off pressure, but then we discovered what was happening with the spark not being there until 10 or so seconds of continuous cranking.

 

We have hooked the car up to the PIWIS and PIWIS-2 that the local Porsche approved bodyshop loaned to us, and there are no fault codes anywhere.

 

We also had hoped to watch the live data while cranking, but it seems that communication disconnects during starter motor cranking, which we are told is normal.

 

The strange thing is that the car idles fine, runs fine, no hesitation, we have followed the Porsche test plan for checking the mass airflow sensor, we have used a smoke machine on the intake to look for vacuum leaks, we have even checked the various temperature sensors on the PIWIS and all seem to be reporting good temperatures.

 

While we had the PIWIS-2 we even reprogrammed the DME with the latest software, which it said was already loaded but we did it anyway in case something was corrupted.

 

We have even disconnected the battery over night and tried the various re-training adapting methods on the internet.

 

Fuel trims are all hovering around 900 and 1.1

 

We have even changed the crank sensor for a new one and no difference.

 

If it were not for the loss of spark, I would be tempted to say it was the fuel pressure bleeding off, because it is a little like it is trying to get fuel back up to the injectors when cranking, which would also explain the way it only has this problem 5 mins after shutting the engine down, but not less than 5 mins.

 

Does the 997 DME monitor oil pressure before it supplies a spark? My reason for asking is that this would also be something which would match up with the symptoms, once the oil pressure has bled off completely, it would take a few cranks to regain it, especially if the oil circuit had a fault... But when running the oil pressure is good, however it does seem to be that the oil pressure is at the bottom on the instruments when cranking and either it shoots up and then the engine runs, or maybe it just shoots up due to the engine running at a faster speed than starter motor cranking.

 

We even worried it may be cranking a little slower than the DME was happy with, so even replaced the dreaded alternator/starter motor/charging harness on the engine.. Which now has it cranking a fraction faster, but still the same problem.

 

Having compared the cranking speed to my Boxster and 996, it seems to be turning over at the same kind of speed.

 

Any ideas anyone??

 

[JFP in PA]

The oil pressure sensor system does not have an ignition interrupt.

 

When you check the fuel pressure, what value did you see before starting and then running?  And did you leave the system connected to see if it is bleeding off pressure?

[rennbod]

On the fuel pressure gauge it looked like it we were getting..

 

Engine stopped - Fuel pump activated by PIWIS - 3.7 Bar

Engine stopped - Fuel pump switched off on PIWIS - 3.7 Bar

Engine stopped - 20 mins after fuel pump shut down - 3.5 Bar

Engine stopped - 40 mins after fuel pump shut down - 3.2 Bar

Engine running - 3.7 Bar

During prolonged cranking - rapidly shoots up to 3.7 Bar from wherever it was within a second.

 

The interesting thing is the spark... No spark at all, other than a single pulse (on whichever coil we check) once every 3 seconds. Then as it starts, it has rapid sparks, which obviously continue while it is running (albeit missing a bit because one coil pack isn't hooked up to the spark plug in the cylinder)

 

This to me makes me think that the DME is looking for something and not seeing it until many seconds after initial cranking if the engine has been shut down for more than 5 mins... Or, the DME has a fault and is in a deep sleep until the engine has been cranking for several seconds.

 

Once the engine has run within the last 5 mins it will start perfectly within a second.

 

Part of me thinks it might be a component fault within the DME.

 

Another part of me thinks it could be either the DME is getting a delayed signal (via can bus?) that the engine is cranking? Other than the crank sensor, is it told by the ignition switch box or the security system that the key is in the start position?

 

And another theory is that the alarm/immobiliser/security system is not sending the immobiliser code immediately on ignition on, and there is a delay and perhaps if the DME has seen that immobiliser code within the last 5 mins it does not need it again?

 

I am a bit foxed to be honest.

 

 

[rennbod]

Another odd thing I have just realised in the wiring diagrams, which I have never noticed before...

 

The coil wiring has terminal A3 from each coil going into a black wire to a join where one black wire exits, it then goes to terminal A1 on a "condenser 3.3uF" and exits on terminal A2 of the same component as a brown wire to the ground point 10.

 

In the diagram the injectors have the same scheme with a condenser..

 

I have looked at the wiring on this car, and it seems that each bank has its own join for the coil grounds (x3 plus one exit) and no condenser to be found... 

 

I was hoping I was onto something here and perhaps the condenser needed to build up a charge as some kind of fault if left for 5 mins to discharge fully... But no condenser.

 

Ive not seen a condenser in the modern car since they had points in a distributor!!

 

 

 

[JFP in PA]

Coil on plug ignition systems do not need or use a condenser.

[Ahsai]

Have you tried new coils and plugs?

Remove the coil and plug on another cylinder and see if you get the same spark interruption. If you do, that means the DME is not firing the coils.

 

Next is to use a scope to check the CPS signal and the cam sensor signal on both banks (by back probing the DME pins). The DME picks up the TDC#1 from the CPS the but it needs the cam signal to tell between TDC #1 compression and TDC#1 overlap so it knows when to fire each cylinder. However, the DME might also just fire two opposing cylinders at the same time. One will fire up (at TDC compression) and the sparks will have no effect on the other one.

[rennbod]

1 hour ago, JFP in PA said:

Coil on plug ignition systems do not need or use a condenser.

 

Well exactly, but in this wiring diagram..

 

 

Which is from the Porsche PIWIS-TIS system shows the car should have a 3.3uF capacitor which it labels as a "condenser" on the earth wire of the coils.. over to the right you can see the same arrangement for the injectors... which is odd because I am more used to EFI cars having a common power and the DME fires the injectors by grounding the other side... I might somehow of pulled up the wrong wiring diagram, but it does seem to be the correct engine wiring diagram from a 2006 997 C4S as input to the system using the chassis number..

 

I have checked the harness and there is no component with two terminals between the coil pack grounds and the ground point, so unless like a resistor wire the capacitor/condenser is a special section of wire, it does not make much sense!

 

 

[rennbod]

1 hour ago, Ahsai said:

Have you tried new coils and plugs?

Remove the coil and plug on another cylinder and see if you get the same spark interruption. If you do, that means the DME is not firing the coils.

 

Next is to use a scope to check the CPS signal and the cam sensor signal on both banks (by back probing the DME pins). The DME picks up the TDC#1 from the CPS the but it needs the cam signal to tell between TDC #1 compression and TDC#1 overlap so it knows when to fire each cylinder. However, the DME might also just fire two opposing cylinders at the same time. One will fire up (at TDC compression) and the sparks will have no effect on the other one.

 

I have borrowed an automotive scope and am not entirely sure I am using it correctly, but it does seem that there is no signal at any of the coil packs when the engine is cranking until several seconds have passed.

 

The scope only has one channel so I can not duel trace both cam sensors or the coils to see their relationship side to side at the same time, but the crank sensor and both cam sensors seem to be generating a waveform which only changes in frequency as the engine fires and cranking, but is there as soon as you start cranking the engine.

 

I am beginning to suspect that either the security system is sending the immobiliser code to the DME late, or the DME has an internal fault.

 

I know that with my 996 the immobiliser/alarm sends the immobiliser code to the DME as soon as the ignition is switched on, and it is the DME that allows the starter block relay to close and allow the starter signal to get to the starter motor.. I am not sure if the 997 DME does the same thing? In which case, if it does, I can rule out the security system sending a delayed immobiliser code, because the car is cranking on the starter motor fine.

 

I even had a theory for a while that perhaps the DME was not getting a permanent live, so every time it was started it was being powered just off of an ignition live, which would maybe then cause the DME to perform its own housekeeping tasks, such as trying to initialise the throttle body adaption for the first few seconds (just as if the battery had just been connected) so may have been refusing send ignition signals to the coils.. But I have checked that it has both a permanent and ignition live..

[Ahsai]

4 hours ago, rennbod said:

We have pulled off a coil pack and put a new plug into it (earthed out with a jump lead) and observed that while it is cranking there is no spark, apart from maybe one every 3 seconds, and then rapid sparking at the plug as soon as it fires up.

 

That is way too infrequent if the engine cranks at a normal speed (say 100-200rpm). Cylinder fires every 2 crank rotations => spark almost every second on EACH cylinder.

I'll ask again. Do the other cylinders exhibit the same firing interval? Have you tried new coils?

 

6 minutes ago, rennbod said:

but the crank sensor and both cam sensors seem to be generating a waveform which only changes in frequency as the engine fires and cranking, but is there as soon as you start cranking the engine

 

The crank sensor signal should also increase in magnitude as the engine speeds faster. i wonder if you have a weak CPS signal.

BTW, those small capacitors are there to suppress (electrical) noise produced by the coils and injectors. Absolutely nothing to do with the strength of the sparks you get.

[rennbod]

On 8/18/2017 at 6:53 PM, Ahsai said:

 

That is way too infrequent if the engine cranks at a normal speed (say 100-200rpm). Cylinder fires every 2 crank rotations => spark almost every second on EACH cylinder.

I'll ask again. Do the other cylinders exhibit the same firing interval? Have you tried new coils?

 

 

Absolutely, that is what I mean..

 

On 8/18/2017 at 6:53 PM, Ahsai said:

The crank sensor signal should also increase in magnitude as the engine speeds faster. i wonder if you have a weak CPS signal.

BTW, those small capacitors are there to suppress (electrical) noise produced by the coils and injectors. Absolutely nothing to do with the strength of the sparks you get.

 

Crank position sensor has been swapped with the new one, and the signal checked at the DME connector, all seems to be good.

 

Interestingly, if I go to start the car, but do not let it crank for more than a turn, then switch off the ignition and then back on and into the start position it always starts within a couple of revolutions of the engine... If though, when hot, I just try to start it, it cranks and cranks and cranks for about 10 seconds and then fires up.

 

It is almost as if the Porsche access system is delaying sending the immobiliser code to the DME for about 10 seconds after the key turns to the start position.. I know the 996 models alarm and DME package works in such a way as that the DME does not allow start (starter motor use) or run until it has the immobiliser code from the alarm control unit, so on a 996 this theory would not hold water.. But I am not sure if the DME in the 997 blocks the starter with a relay as well as run.. If it only blocks run, it is perhaps possible that the PAS is delaying in sending the immobiliser code and that is why the DME will not activate the coils or injectors until a few seconds have passed.

 

Other than this, I am at a bit of a loss.

 

 

Edited September 13, 2017 by rennbod

[Ahsai]

Very strange indeed. However, if the 10s delay is actually true, how to explain the ~once per every 3 seconds spark during the 10s? You would think the DME either wants to fire up the engine or cut the sparks completely but not something in between...

[rennbod]

--- UPDATE for the archives ---

 

Not been back here for a while.. But it seems the problem was solved.

 

Having loaned my friend the PIWIS for the last few months to be able to keep looking at live data and actual values and generally hack about with the electrics, the problem was found through a shot in the dark and false diagnosis by a main dealer!

 

He experienced problems with his remote key head which was only a couple of months old, so took the car into the local Bournemouth main dealer to resolve under warranty... They claimed it must have been a problem with the main security control unit, not the remote and installed a new one.

 

The bad news (for the main dealer) is that it did not solve the problem with the key remote, that turned out to be the key remote being faulty and replaced under warranty.

 

The good news (the the owner) is he left the main dealer with a new working key remote, and the new alarm control unit at no cost.. But the best news is this somehow also fixed his delayed warm start issue.

 

The last bit of bad news is he just dropped off my PIWIS back to me, now missing its main OBDII cable which he has lost somehow!!!

 

Anyway, for the archives, I guess the immobiliser was delaying telling the DME to "GO!" and for some reason the DME was omitting an occasional spark to the coil packs, but not a regular spark without that code.. or it is something else.. But the replacement of the Porsche alarm/access system fixed the problem!

[MartiniLinguini]

Immobilizer ignition can trip a lot of people up.  Glad it was resolved, thanks for the update.

 

For the record, there are coil-on-plug systems that utilize capacitors(condensers). The capacitors are used to stabilize the voltage to the coil-packs, and without them, you will get misfires.

Edited March 1, 2018 by MartiniLinguini