AFR gets leaner with higher IAT
 

My AFRs get substantially leaner as IAT goes up. For example, I can go out for a drive do a bunch of tuning, think I have things dialed in. Then I park the car for like an hour, IAT goes from about 80 up to about 120, and CLT goes from about 205 down to 175 (not sure my sensor is calibrated properly). Then I restart the car and suddenly my idle and low-speed (10-15mph) AFRs are really lean until the IAT comes down again.

The only two things I can think of that would be affecting it are IAT-related timing (because timing can affect AFR), and maybe an incorrect Req Fuel that is causing the ECU to make too large of a pulse width change for a given IAT change. My Req Fuel is 13.0, and I'm using 2000 injectors (240cc according to one of the FAQs here, haven't verified). Bigger injector = smaller Req Fuel. Smaller Injector = larger pulse width change to achieve a given fuel delivery change, right?

Any ideas?

-Mike

check your warm up compensation maybe the values on that are to large on top or to small down low.

It's not warmup or after start. The coolant is still hot enough to be beyond the warmup settings.

-Mike

Where is your IAT located? Sounds like it is heatsoaking. If your sensor is reading a higher temp than is actually being ingested, you will run lean. Hot air is less dense, so less fuel is required.

It's in the plastic tube that feeds the throttle body, right in front of the throttle body. It's an open-element GM sensor.

There aren't any IAT-based fuel corrections on the megasquirt, are there? Only the spark adjustments?

-Mike

Lex was talking about this in Reverant's "Occasional rough idle" thread. In MS2 I can change IAT based enrichment/enleanment. I'm pretty sure you can do this in MS-SnE, too. I added 1% at 100 degrees, 7% at 130 and 12% at 160. Those numbers are crude and could stand some tuning, though, and I haven't seen intake temps higher than 120 so they're just a guess beyond that.

Edit the airdenfactor.inc file so that you don't get as lean as you describe when the "indicated" IAT goes up. You will have to recompile the firmware afterwards.

Jim

What does that file do? It sounds related to what I was wondering on the Req Fuel setting. One of my suspicions was that I have it set to high, basically telling the MS that my injectors are smaller than they actually are, causing it make a larger PW change than it should for a given IAT change.

I found the IAT adjustments, though. They're buried deep under "Coolant-Related Air Density". Then you hit settings "Coolant-Related Air Density", change "Air Density Correction" to "Corrected", and then change "Correction Based on" to IAT corrected.

I'll give that a whirl at some point.

Is it normal to have VEs up around 115 at high load and high RPM at about 12.5:1?

-Mike

yes it is you have a turbo after all 200% would not be impossible at some point. Trust me if the sensor is heat soaking then it will keep you lean on inittial start up as you wont have much compensation in that coolant range there is adjustment all the way up to 74 degrees celcius for that.

Sorry, meant to say NA motor. :) Haven't turbo'd yet.

-Mike

It defines how much (in %) the MS should squirt relative to the air temperature. Ie:

50F 101%
70F 100%
80F 99%
90F 98%
...
And that's for each of the 255 discreet values defined in the table.

Jim

I finally got my 440cc injectors in, and have the car running again, but I still see this problem where the AFR leans as IAT goes up, and I also need about 117 VE to get about 13.0:1, which seems way to high for an NA motor.

Can anyone share all their misc .inc files? Also, what's the difference between airdenfactor.inc and matfactor.inc?

-Mike

dude you should be way richer than 13-1 what is your DC? I would recalibrate my 02 sensor and then try my afr table. Attachment 216329 GL Just becouse you hit 117 VE doesnt mean you are overcompensating if the DC and everything else is normal then you migth just be making some nice power. try theses AF's in autotune and see how they work for ya.

I can't share mine since I have a 2002 with the stock sensor. matfactor.inc tells the MS how much temperature (in degrees F,C) is each voltage "step" of the sensor. airdenfactor.inc tells the MS how much more/less it should open the injectors, relative to the air temperature. THAT is want you want to change. Which air temperature sensor do you have?

Jim

I'm running a GM open element sensor.

My airdenfactor.inc looks like this:

Which column is the percentage I'd want to change?

-Mike

The second column, the one that says 100T, 99T, 98T. That number is the actual percentage of the time to squirt. Ie:

For 73.36 degrees F, squirt 99% of the calculated pulse width.
For 323.23 degrees F, squirt 67% of the calculated pulse width.

To find item your ideal temperature, start a cold datalog and let the car warm up. Watch the AFR as the car warms up. When the AFR starts to lean out, write down the temperature. Let the car idle until the IAT rises no more - ie its fully heatsoaked. Give it some revs if you feel you need to change more areas. Once you have the datalog, open it up and notice at say, every 10degF increments how much off is the AFR. Calculate how much more you need to open up the injectors in % (dont' forget the 1ms opening time) and note the numbers down. Open up airdenfactor.inc, and for each 10degF increment, add the percentage to the second column, taking care NOT to the delete the capital T character. Once you are ready, recompile the code (run 'compile.bat') and you will end up with a file inside the "src" directory called "msns-extra.s19". Copy it to the parent folder (there will already be one there, overwrite it) and download the newly created firmware to your Megasquirt (connect the laptop to your MS and run "download-firmware.bat").

You may need to fine-tune the figures a second time using the same method, but you should be pretty close the first time around. Hope that helps.

Jim

Thanks for the detailed explanation. You answered a bunch of questions I hadn't asked yet. :) Should I turn off warmup enrichment while the car warms up for this test?

I think I'll upgrade to 29v at the same time, since I don't have the source for 29t (I just got the .msq).

Do you realistically expect this to bring my high load high RPM cells down to something more reasonable? I looked at your .msq (don't remember where I got it), and your highest VEs are in the low 90s, which makes more sense for an NA motor.

-Mike

1) Don't turn off warmup enrichments.
2) Yes, airdenfactor.inc affects the entire RPM range so it should bring them down. Mine (VE) reaches 95% on the max torque cell.

Jim

Reverant, can you share your .inc file with me? I want to compare it against mine and the one that comes with 029v. Mine is extremely close to 029v (within about .5%), which is set up for a (the?) GM sensor. So I'm wondering if I have something else going on, like incorrect bias resistors or something. I've got a few mods to do to my board (TPS, Boost Control, Fan Control) so I'm going to poke around with the multi-meter a bit before I do a reflash.

-Mike

I'm running stock sensors (2002 Mazda CLT & IAT) so I don't know if it makes sense. No problem if you still want them though.

Jim