Cold Morning Strange No-start: 2 sec PW?

[CCS86]

Hi guys,

My car fires up beautifully every time. Cold start, hot start... until this morning. I'm guessing that it is a temperature related issue, as this was the coldest it has been since I started tuning it. 48*F

According to the log file, the injector PW is toggling between 2,000 ms and 0 ms. I don't know if that is true, or not. You can see that RPM barely flinches when it says the injectors are going wide open. Maybe they really are going wide open, because on the second crank I floor the pedal, which disables the injectors, and it eventually chugs to life. Interestingly, the injectors do turn back on while I am floored, but now at a much more reasonable PW.

Any ideas about which tables could cause this? Starts right up at 60*F.


cold-start.jpg

[Kcc0521]

Is the car on E85? I think I am having the same problem. There is a table to make the tune richer at certain temperatures.

[CCS86]

Nope, I run 93

[CCS86]

It might be a coincidence, but the behavior seems to occur only below: [ECM] 12111 - Open/Closed Loop Cold Start ECT which is set to 54*F

All the tables I have found that seem related appear to have good settings:

cold-tables.jpg

[bbrooks98]

Haha I ran into this last winter. I noticed the exact same thing with the below 60degree temps. I ended up moving the OL/CL coldstart ECT down to 0 and it seemed a lot better. I'm sure there is a better way with the lost fuel tables and base fuel but it worked in my case. I summed it up to crap injector data on my old injectors at the time.

[CCS86]

Okay, I tried lowering:

[ECM] 12111 - Open/Closed Loop Cold Start ECT: When ECT is lower than this it is considered a Cold start.

to 0*F, but the problem has repeated. I got it on a log again, and it's the same story. Mostly 0 ms PW while cranking with occasional 2,000 ms PW. Had to floor the pedal and crank a lot to get it going.

Help!!

[bbrooks98]

Try looking in the Lost fuel under BP. It seems to add 30 to 50% from lost fuel in the first few seconds in the cold startup temps. Maybe try zeroing and see if you still get such a large PW.

[CCS86]

Try looking in the Lost fuel under BP. It seems to add 30 to 50% from lost fuel in the first few seconds in the cold startup temps. Maybe try zeroing and see if you still get such a large PW.

Here's my table for that. It looks like a slightly diminished GT table.

Lost-fuel-BP.jpg


The description of the table doesn't seem to match. Without an axis relating to barometric pressure, how is it related?

At 35* and 0 sec, I have "0.44". Does this mean that if my base fuel was commanding EQ=0.97, this table is an EQ adder? ie 0.97 - .44 = 0.53 * 14.08 = 7.46. So fuel for a 7.46:1 AFR?

That is a substantial correction (that tapers pretty quickly with time), but it's hard to explain the 2000 ms PW. Normal idle PW is ~2 ms. Seeing a 1000x change seems to suggest something more severe. I just don't know what.

[bbrooks98]

Here's my table for that. It looks like a slightly diminished GT table.

Lost-fuel-BP.jpg


The description of the table doesn't seem to match. Without an axis relating to barometric pressure, how is it related?

At 35* and 0 sec, I have "0.44". Does this mean that if my base fuel was commanding EQ=0.97, this table is an EQ adder? ie 0.97 - .44 = 0.53 * 14.08 = 7.46. So fuel for a 7.46:1 AFR?

That is a substantial correction (that tapers pretty quickly with time), but it's hard to explain the 2000 ms PW. Normal idle PW is ~2 ms. Seeing a 1000x change seems to suggest something more severe. I just don't know what.

Yeah, that description is totally off. My take on it is it's adding .44 to your Base as you mentioned. I'm thinking maybe if the correction is large enough it hits an all-in(injector wide open) value and maybe that's where the 2000ms comes from?

[murfie]

Start up VCT changes with BP changes, just as the modes will enable at different load values for different BP's. The mode would be emissions reduction, basically get the cats warmed up as fast as possible so they can start working. The BP table is the same as the base decay table, but for the changes of the VCT that happens during the first 2 minutes after startup.

You would subtract the engine start tables from the other tables that are load vs ECT and arrive at a lost amount of lambda. This would be added to your desired lambda and the pulse width would be calculated from that. This is mainly to compensate for the lost fuel between piston and cylinder wall when they are cold. you wouldn't actually see a lambda like 8:1, the O2s would read near stoich, if they were actually warmed up and accurate at that time.

Getting to .61 lb/min maf at 610u period at 176RPM is like having 2.1 air load. you probably aren't getting twice barometric pressure in your cylinders on cranking. add in the lost fuel correction... Make sure your MAF curve has values of like .3 or lower on it. MAF temp compensation table is also good for this .01-.02 columns in the LB/s.

[CCS86]

Start up VCT changes with BP changes, just as the modes will enable at different load values for different BP's. The mode would be emissions reduction, basically get the cats warmed up as fast as possible so they can start working. The BP table is the same as the base decay table, but for the changes of the VCT that happens during the first 2 minutes after startup.

You would subtract the engine start tables from the other tables that are load vs ECT and arrive at a lost amount of lambda. This would be added to your desired lambda and the pulse width would be calculated from that. This is mainly to compensate for the lost fuel between piston and cylinder wall when they are cold. you wouldn't actually see a lambda like 8:1, the O2s would read near stoich, if they were actually warmed up and accurate at that time.

Getting to .61 lb/min maf at 610u period at 176RPM is like having 2.1 air load. you probably aren't getting twice barometric pressure in your cylinders on cranking. add in the lost fuel correction... Make sure your MAF curve has values of like .3 or lower on it. MAF temp compensation table is also good for this .01-.02 columns in the LB/s.

-I understand the concept of lost fuel. But without bogus values in one of the lost fuel compensation tables, this should obviously not be the result.

-We aren't talking about VCT here, as the cams are staying essentially parked. BP was mentioned because the table I pictured above is most likely mislabeled.

-The MAF curve looks nice and smooth to values below this. But if it was MAF curve related, I would most likely see issues cranking at all temps. This is definitely only at very low temp. The bottom 3 axis values are: 1400, 712, 662. The resting MAF signal is ~1400us. During cranking this jumps to ~660us.

-Right or wrong, the abs. load reported is about 57%. Calc load is 100%.


Did you take a look at the log or screenshot? This behavior does not look like normal "compensation" behavior. It looks like fail-safe. The injectors are acting like on/off switches. Cranking begins with 0 PW, despite showing a positive fuel flow rate value.

I am attaching a new no-start log and a comparison log of a normal start. I can't see any important differences, besides the injector behavior:


no-start-2.jpg

[CCS86]

I appear to have fixed it...

My [ECM] 44519 - Fuel Cranking vs. ECT: This is the commanded lambda value when the engine is cranking. table had been changed to blend smoothly from 1.0 at warm temps, to 0.86 at 35*F. My guess is that the other modifier tables mentioned don't play nice with a non-one value.

[Jn2]

I appear to have fixed it...

My [ECM] 44519 - Fuel Cranking vs. ECT: This is the commanded lambda value when the engine is cranking. table had been changed to blend smoothly from 1.0 at warm temps, to 0.86 at 35*F. My guess is that the other modifier tables mentioned don't play nice with a non-one value.

Yeah, I normally command .95 at most on e85 in cold temps, you will never need to crank on .86 lambda...

[CCS86]

Yeah, I normally command .95 at most on e85 in cold temps, you will never need to crank on .86 lambda...

This wasn't really an issue with cranking at .86 lambda. Incompatible settings were causing the injector to me off or wide open. With the cranking table set to all 1's, you will still command a richer startup mixture due to the lost fuel table.

[murfie]

Only one of your log uploaded larger than 3KB and is anything to look at.

I was going to suggest cranking fuel table ECM 1219 next, but glad you figured it out with the base table. Below 50* the mass values start to get really big.