T/A-Bob

I have been dealing with a cold start problem for the past two winters. Every possible item has been suggested as a cause. I posted once about it here last August:

https://www.camaroz28.com/forums/sho...d.php?t=631294

Fred made several posts in that thread, and based on what he said I decided to take a shot & change out the O2 sensors.

Viola!

Last night it went down to 5°F (-15°C) and just now when I went out to start the car it was 13°F (-11°C).. and it started right up, no problem

Thanks Fred.

Three very knowledgeable people I know here said there was no way the O2 sensors could possibly have anything to do with a cold start problem. Looks like these guys (two of them are professional mechanics) still have some stuff to learn.

bombebomb

I was unaware that the O2 sensors had anything to do with startup? I thought they did not become "on" until they was heated up (closed loop?)

T/A-Bob

My limited understanding of this is that information previously generated by the O2 sensors is stored in the computer and is used used during start-up -- even though the sensors are ignored during open loop operation.

So, if the O2 sensors were feeding bad data into the PCM before, that bad data is being used now when the engine is trying to start.

Angelis83LT

That would actually make sense if the computer quit accepting data upon the ignition being turned off, and if it was not cleared when the car is turned on. However one would think that either at the off or at the on, the thing would reset to the tune for the car?

Green96Z

What was your initial cold start problem? Continuous cranking with no start? Or a slight catch and then the engine dies? Because technically the O2s can't read anything on cranking because there are is no combustion.

T/A-Bob

In response to the two above posts... I think it's like this...

When you start the car, it has to use some sort of data. The only data it could have is from before you attempt to start the car.

This already-existing data is a function of what the O2 sensors were saying (in the past). So this data could be corrupted if the O2's were bad. Therefore, even though the O2's are ignored during the moment of start-up, their "legacy" - the bad prior data - still plays a role in start-up.

I'm not really sure this explanation is right. However, this is the explanation I used to convince myself to take a shot and change the O2 sensors. The result... the car started easily today, at a temperature that would have made it impossible last winter (as well as the winter before).


EDIT: Two years ago it was catching a bit, then dying. Last year, it was cranking with pretty much no response at all.

EDIT #2: I don't think the PCM re-sets itself when the engine is turned off. It stores past history data.

Green96Z

Hopefully Fred can post up how the PCM uses old O2 data in relation to fuel injector action during initial start-up, because I have no idea how it works. But at least you got your car running again! That's always a good feeling.

T/A-Bob

If you read thru Fred's posts in the thread I cited at the beginning, you'll see his explanation. I didn't completely follow it at the time (and I still don't), and you can see that the final post is me asking if bad O2s could contribute to a hard start problem. But somehow I got the sense that it might... so...

Yes, I'm very glad it starts. Tired of borrowing my father-in-law's car.

akafred

the ecm has a correction factor that it stores (blm) if your o2's are out of wack you may be running rich or lean at all times including startup. a scan would have shown that your computer was adding or subtracting a percentage of fuel due to faulty o2's

Injuneer

As indicated in several responses, the PCM uses the stored (learned) long term fuel corrections when it starts.

You've got it right

Turning off the engine, or restarting the engine does not clear the "learned" values, like the long tern fuel corrections, IAC start positon, low octane fuel flag, etc. That would not make much sense. The PCM learns how to run the engine correctly, then clears itself..... not rational. You can clear those values by pulling the power to the PCM, but there is a "PCM BAT" circuit that feeds power to the PCM when the ignition is in the "off" position. The "tune" is never reset, and never lost. Its stored in non-volatile memory.

A scan will tell you if the PCM is using the long terms to add or subtract fuel, but is does not tell you why its doing that. May be in response to "faulty o2's".... may be in response to other problems.

Faulty O2 sensors (or conditions that "fool" the O2 sensors, like exhaust leaks and misfires) really screw things up. But you have to look at both the long term fuel corrections (used all the time) and the short term fuel corrections (used only in closed loop). If the O2 sensors are far enough "off", the engine may still run decent in open loop when the only error is the "long term" error, but run very poor when the PCM is now using the "long term" error and adding a "short term" error. The long terms can only add 25% extra fuel, and subtract 15% of the normal fuel. The short terms are virtually unlimited (+/- 100%). The equation for determining the injector pulse width contains the product of the long terms X the short terms, so closed loop may make a bad situation even worse.

Angelis83LT

Thanks injuneer. I do not work alot wit Fuel injected cars, just with computers. I figured you would end up with it always back to where you started with the Stored stone of the ar as it was tuned by a shop or you. I didn't realize it used a fuzzy logic circuit in order to self correct to such a point it could cause it's self issues.

bombebomb

This is very interesting, ill definently remember this.