Troubleshooting Poor Cold Start on a 3.2

[mysocal911]

Originally Posted by et cetera

I took the ECU out to hopefully test it in a friends car soon. Figured I would open it up to see what chip is in it. Clearly it had been opened before. Chip has a sticker on it, PAE i8. I did a search and couldn’t find a reference for that.

I had also opened up the AFM and noticed the spring was moved one notch towards the lean side.

It's not a stock (original) memory chip (PAE). Who knows what effects it has? Besides, the processor chip has been replaced too. The original DME ECM didn't use a 8031 chip.

[mysocal911]

Originally Posted by et cetera

I agree, this is much more complicated than I originally hoped, or had described to me.

I posted the chip and AMF spring settings.

To my understanding these are OEM injectors that had been cleaned and refreshed.

It does seem like it might be a fuel pressure issue. I took the vacuum lines going to the FP damper and regulator off, they weren’t wet, but they smelled of fuel.

The 1st is only resolved by actual part numbers, and the 2nd by using a pressure gauge.

[pmax]

Originally Posted by mysocal911

He assumes this!

Now on that point I'm in complete agreement.

Needs to be measured and confirmed as already emphasized.

Originally Posted by mysocal911

Who knows what kind of DME ECM chip is installed? It's total confusion at this point. No more guessing.

As I said up-thread, time for another DME ECM!

How did you come to the conclusion that the OP needs another DME ?

Describe your line of reasoning for our enlightenment.

Thanks

[mysocal911]

Originally Posted by pmax

Now on that point I'm in complete agreement.

Needs to be measured and confirmed as already emphasized.




How did you come to the conclusion that the OP needs another DME ?

Describe your line of reasoning for our enlightenment.

Thanks

That hasn't been concluded yet. Since it's been modified, it needs to be fully tested, hopefully on another car and eliminated.
It can also be bench tested!

[pmax]

Originally Posted by mysocal911

That hasn't been concluded yet. Since it's been modified, it needs to be fully tested, hopefully on another car and eliminated.
It can also be bench tested!

We are on the same page wrt testing and measuring and all the good stuff during the process of elimination. You might be right that the engine has a perfect vacuum seal but why just assume ?

Did you read the OP's post about the vacuum ports being plugged ? Would you know if that requires tweaking the ECU to compensate ?

[mysocal911]

Originally Posted by pmax

We are on the same page wrt testing and measuring and all the good stuff during the process of elimination. You might be right that the engine has a perfect vacuum seal but why just assume ?

Most lack a correct way of testing for a vacuum leak, i.e. proper equipment, so other tests should be done first. An EFI system is not like a CIS where vacuum is critical.
Besides, vacuum leaks are rare on the 911 3.2. If a vacuum leak is of concern, one can always carefully use carb cleaner to find a leak.

Originally Posted by pmax

Did you read the OP's post about the vacuum ports being plugged ? Would you know if that requires tweaking the ECU to compensate ?

[et cetera]

I’m totally with you guys. Going to be putting the ECU into a known running car tomorrow.

After that I’ll be moving on to testing the fuel system.

I appreciate all the help!

[RobFrost]

Originally Posted by mysocal911

It still won't explain why the engine is running TOO rich!

Yes, it could do, because somebody could respond to a vacuum leak with some further maladaption. It's moot, though, because I'm not taking "too rich" as gospel yet.

Sent from my SM-G988B using Tapatalk

[et cetera]

For kicks, I also tried the AFM at a few different spring settings to see what would happen. There was no change. I think I can rule that out as a primary suspect.

[RobFrost]

Please, check the fuel pressure is responding to vacuum. Ever since you said engine revs don't momentarily respond to lifting the oil cap, all roads point to that. Even if there are other problems, you'll never get them fixed until you've got the fuel pressure adjustment working.

The vacuum lines smelling of fuel is a red flag for a leaking diaphragm. Try plugging them so fuel can't travel down them to the throttle body. Ideally you will have an AFR meter when you do so. See if plugging them leans out the mixture.

The modded chip probably worked fine when installed into a fully working system, else it would have been swapped back out for the original.

Sent from my SM-G988B using Tapatalk

[mysocal911]

Originally Posted by et cetera

For kicks, I also tried the AFM at a few different spring settings to see what would happen. There was no change. I think I can rule that out as a primary suspect.

Then that's a problem. The major way the injectors provide more fuel as the RPM/load changes is thru the AFM. Small changes in the AFM voltages result in significant changes to the AFRs.
You need to try to understand how an EFI engine operates.

NOT GOOD!

[mysocal911]

Originally Posted by RobFrost

Please, check the fuel pressure is responding to vacuum. Ever since you said engine revs don't momentarily respond to lifting the oil cap, all roads point to that. Even if there are other problems, you'll never get them fixed until you've got the fuel pressure adjustment working.

The vacuum lines smelling of fuel is a red flag for a leaking diaphragm. Try plugging them so fuel can't travel down them to the throttle body. Ideally you will have an AFR meter when you do so. See if plugging them leans out the mixture.

The modded chip probably worked fine when installed into a fully working system, else it would have been swapped back out for the original.

Sent from my SM-G988B using Tapatalk

Remember, vacuum lines lead to the intake manifold which smells like fuel.
Fuel actually in the vacuum line is a problem.

[mysocal911]

911 3.2 ('84-'89) Basic Troubleshooting A Running Problem

1. With Key-Off - DME ECM disconnected
a. check for ground on pins; 2, 5, 16, 17, 19
b. check for ground on pin 3 with throttle at full open

2. With Key-On and engine not running and cold - DME ECM connected
a. check for 12V on pins; 1, 14, 15, 18, 20, 35
b. check for about 5V on pin 9
c. check AFM input pin 7 for about .80V to 4.7V, idle to throttle at full open
d. check O2 input pin 24 for about .50V
e. check temp sensor input pin 13 for greater than 1.5V

3. With engine running and warm at idle;
a. check fuel pressure; 30-35 psi (DME relay output pin 20 < 1.5V)
b. check AFM pin 7 for about .80V to 1.2V
c. check temp sensor pin 13 for less than 1.0V
d. check pin 24 (O2 sensor input) for a varying signal (.30 - .80 volts)

Note: All test points are on the DME ECM 35 pin connector, back-probing with the connector shell
removed or with a breakout box.

[et cetera]

Went over to my buddies house today and tested out both my ECU and Relay on his car. Everything worked perfectly fine. Looks like I can rule those two out.

I do still want to check the fuel pressure, but I’m still a bit suspicious of the ICV. I know it worked when I bench tested it, It opened and closes when I apply a 9v battery to it, but I don’t suspect it was returning to a neutral position.

[mysocal911]

Originally Posted by et cetera

Went over to my buddies house today and tested out both my ECU and Relay on his car. Everything worked perfectly fine. Looks like I can rule those two out.

I do still want to check the fuel pressure, but I’m still a bit suspicious of the ICV. I know it worked when I bench tested it, It opened and closes when I apply a 9v battery to it, but I don’t suspect it was returning to a neutral position.

Read post #20 here; Bench testing Carrera Idle Control Valve

Of all of the problems you state in post #1, only one relates to a ICV;

Symptoms:

Starts when cold and immediately dies
When running it has short power cuts during hard acceleration
Wants to die when taking off from first gear, need to give it plenty of throttle to get going or it will die
The idle is just over 1100k RPMs and does tend to "bounce"
Occasionally, when I go to turn the car over the key feels manually locked out of the start position. After taking the key out and trying again it works after a few times.

Next, you need to follow post #53 above, key tests which you haven't done, or maybe it's time to visit a good Porsche repair shop.

[et cetera]

Originally Posted by mysocal911

911 3.2 ('84-'89) Basic Troubleshooting A Running Problem

1. With Key-Off
a. check for ground on pins; 2, 5, 16, 17, 19
b. check for ground on pin 3 with throttle at full open

1. With Key-On and engine not running and cold;
a. check for 12V on pins; 1, 14, 15, 18, 20, 35
b. check for about 5V on pin 9
c. check AFM input pin 7 for about .80V to 4.7V, idle to throttle at full open
d. check O2 input pin 24 for about .50V
e. check temp sensor input pin 13 for greater than 1.5V

2. With engine running and warm at idle;
a. check fuel pressure; 30-35 psi (DME relay output pin 20 < 1.5V)
b. check AFM pin 7 for about .80V to 1.2V
c. check temp sensor pin 13 for less than 1.0V
d. check pin 24 (O2 sensor input) for a varying signal (.30 - .80 volts)

Note: All test points are on the DME ECM 35 pin connector.

1: With Key Off
a. all ground tests pass
b. Seems to be grounded at both full throttle and off throttle

1. With Key-On and engine not running and cold
a. All connections read 12v
b. pin 9, 0 reading (also checked at the AFM harness and there did not appear to be power making it to the AFM)
c. pin 7, 0 reading
d. pin 24, 0 reading
e. pin 13, 0 reading

Looks to me like there may be a wiring problem. The shop who installed the motor did the wiring, so I need to research where power should be coming from for the AFM.

[mysocal911]

Originally Posted by et cetera

1: With Key Off
a. all ground tests pass
b. Seems to be grounded at both full throttle and off throttle

1. With Key-On and engine not running and cold
a. All connections read 12v
b. pin 9, 0 reading (also checked at the AFM harness and there did not appear to be power making it to the AFM)
c. pin 7, 0 reading
d. pin 24, 0 reading
e. pin 13, 0 reading

Looks to me like there may be a wiring problem. The shop who installed the motor did the wiring, so I need to research where power should be coming from for the AFM.

Great that you did the testing! We have good data now:

1. You're missing the ECU AFM power 9 output. Without this, there's no output on pin 7 to the ECU and no fuel mixture data. It's the KEY input to the ECU for proper running.
2. Pin 24 is the O2 sensor input. You may have a Euro ECU w/o O2.
3. Pin 13 is the temp sensor input. If pin 13 reads zero volts when the sensor is disconnected, then it's shorted to ground. This will make cold starting very difficult.

Nice work.

[et cetera]

Originally Posted by mysocal911

Great that you did the testing! We have good data now:

1. You're missing the ECU AFM power 9 output. Without this, there's no output on pin 7 to the ECU and no fuel mixture data. It's the KEY input to the ECU for proper running.
2. Pin 24 is the O2 sensor input. You may have a Euro ECU w/o O2.
3. Pin 13 is the temp sensor input. If pin 13 reads zero volts when the sensor is disconnected, then it's shorted to ground. This will make cold starting very difficult.

Nice work.

I was thinking a bit more about this.

1. Does power for the AFM come from the ECU? If so, my test would read zero because the tests were performed at the ECU harness connector. I want to connect the harness the ECU and test for power at the AFM connector.

3. For the CHT should I also be testing for ohms or just voltage?

[mysocal911]

Originally Posted by et cetera

I was thinking a bit more about this.

1. Does power for the AFM come from the ECU? If so, my test would read zero because the tests were performed at the ECU harness connector. I want to connect the harness the ECU and test for power at the AFM connector.

Correct.

Originally Posted by et cetera

3. For the CHT should I also be testing for ohms or just voltage?

Voltage when connected.

[et cetera]

Originally Posted by mysocal911

Correct.



Voltage when connected.

What would be the best way to check the voltage for the CHT when its plugged in? Open up the ECU and check from there or at the plug connector in the engine bay?