pgmfi.org

Voltages are fine, C14 is fine. Left to check/replace is QM3 but is there anything else that causes a delay in the fuel pump prime? When the car is cold, it can take anywhere from 1 minute to 10 minutes before the check engine light turns off and the fuel pump primes. After that, the car usually starts without delay.

The check engine light and fuel pump prime should occur at the same time when power is applied to the ECU. If the check engine light is on and the fuel pump is not priming along with it, , the CEL never goes off, and you can't pull a code at all, it's possible the ECU is being reset over and over.

The prime/CEL is done in the code, so your processor itself or the oscillator crystal that controls processor timing may be to blame as well. Good luck!

So aside from the crystal, if an ECU does not prime the fuel pump immediately after powering up, there are no other components that can introduce a delay to this process which eventually completes? The fact that it is apparant when cold does not point to any component in particular?

Well think about it. If it's not the crystal, things contract with cold. If a circuit is not being completed because of a contraction in the cold and later warms up to complete the circuit... boom. HOWEVER, here's where the problem lies:

You're saying the fuel pump PRIME has a delay. Whether an output component was damaged or not the ECU would go through the cycle timed like it normally would be, it simply wouldn't activate the fuel pump because the component failed. The prime would not occur again once components warmed up, because it only occurs once at key on. So technically if it's simply an output component, you should still see the CEL follow the normal prime at key on for the normal duration, just not the fuel pump. That's not really what I understand is happening, from your description.

Now, if the ECU is being reset over and over it wouldn't get to the prime cycle until it stopped resetting, if the crystal was outputting anything other than it's proper cycle the ECU would not run at the speed it's supposed to be, and if there was no signal from the crystal the ECU would not run at all until there was one.

Ok, so I scoped the NMI and RESET lines on the processor and started the ECU. THe check engine light stays lit and the pump does not prime. The scope tells me the processor is being held "off" hence it's not executing any code. Suddenly the pump primes, the CE light goes off and the scope shows me the processor is running.

...so if the timing is in the code, and I have swapped/checked most of the components which control the processor including the TA8903 power supervisor chip, where is my problem?

Well after several hours of probing, scoping and testing, I figured it out and was able to repair the ECU. I've got 4-5 more with the same problem and will validate my fix with these.

so how did you fix it??

yup... share it with me also... cause I have a long P30 with the symptom close to this....
TQ

Ok, repair validated with the other ECUs!

The problem with ECUs that take long to activate the main relay & fuel pump is (drum roll.......)

The TA8903 Power Supervisor IC. Even though I had replaced it earlier, it seems I swapped in another defective one

...now where to find a place to source these as my contact in Malasia (yes, you e-tech!) told me he can't get them. I repaired the existing ECUs by removing the TA8903 from parts boards.

Ahhh yes. Low power could also make things run slower than normal.

Thanks.... guess i have to try harder to help those with the longtype ecus... i dun't get much of those here...mostly JDM square ecu.

BTW is there a way to check by multimeter if the TA8903 Power Supervisor IC really spoilted?

Thanks again for sharing....

I'm pretty sure the TA8903 is used on the JDM ECUs also. They are even used on the OBD2 ECUs so they should still be available

oh... man you r right... it is used in square ecu too...- IC17...
So at least I have 2 pcs spare from scrap brds.

Do you know the maker? Is it toshiba?

Yes, Toshiba - I sure hope you can find them! LMK asap via email. Thanks Eric!

A quick Google search and i found 2 places that say they have them.

http://www.iso-group.com/Public/PartDet ... 903SN.aspx

http://www.aes-online.com/ Just put TA8903 in the search box