Tuning for cold start in cold weather.

[Sam Powell]

Now that the cold weather is here, the engine will not start unless I pump the gas. I know this is enriching the mixture with the TPS accel enrichment, but it should be possible to make it start without touching the gas. What numbers should I be upping here. After it starts, it runs lean for a few moments, and then starts to smooths out. Thanks. Sam

[Moonsteel]

Now that the cold weather is here, the engine will not start unless I pump the gas. I know this is enriching the mixture with the TPS accel enrichment, but it should be possible to make it start without touching the gas. What numbers should I be upping here. After it starts, it runs lean for a few moments, and then starts to smooths out. Thanks. Sam

If you can't get it started, you need to adjust your Cranking Pulsewidth.
Right after it starts car goes into AfterStart enrichment, used to keep car from stalling.
After a few seconds (10?) it switches to your Warmup enrichment.

Do you have a fast idle? I highly recommend it.

[zenon]

Now that the cold weather is here, the engine will not start unless I pump the gas. I know this is enriching the mixture with the TPS accel enrichment, but it should be possible to make it start without touching the gas. What numbers should I be upping here. After it starts, it runs lean for a few moments, and then starts to smooths out. Thanks. Sam

Does the accel enrichment really kick in when cranking? I thought not. My understanding is that the cranking pulsewidth remains the same regardless of MAP and throttle position, short of entering flood clear mode.

I found that my cold starts were similar - I had to open the throttle while cranking to get it to start. I took this to mean the engine wanted extra air because the mix was too rich. When I reduced the cranking pulsewidth, it then was able to start easily without touching the gas pedal.

I am using a Bosch aux-air regulator, btw - initial cold idle is about 1500RPM.

regards,
Zenon
'73 BMW 2002i
http://www.zeebuck.com/bimmers/tech/Meg ... quirt.html

[Sam Powell]

Why did the car start fine when the weather was warm if the mixture is now too rich. Would not the cold lean out the mixture, not richen it up? I guess the computer takes this into account with the air temp sensor, so maybe that is not good reasoning. You think it is just coincidence that it finally starts when I open the throttle? I guess I can try raising and lowering the cranking pulse witdth and just try it, but you don;t get many chances in a week to do that. What changes in the pulsewidth computer formula when the after start mode is over? Wht is the prevailing parameter up until then. Is the reqfuel ignored, or modified? Is the pulse width just kept constant at the selected cranking pulse width, or are other factors such as air temp, and coolant temp taken into account? Thanks. Sam

[zenon]

There are two boxes for setting cranking pulsewidth - one for low coolant temperature and one for near normal operating temp. The Megasquirt then draws a straight line between these two points to determine the PW for in-between temps. My guess is the cold PW setting is too large but warm was close. In my setup, I think my cold PW is set at 3.0ms, while warm is 2.0ms (and the warm setting is maybe still too rich...). The default was something like 12ms and 4ms, so it was originally way too rich.

[efahl]

Sam Powell wrote:

> Why did the car start fine when the weather was warm if the mixture is now too rich.

Sam,

To paraphrase Zenon, it's the slope that counts.

Let's pretend that MS uses coolant temperatures at 0 and 100 degrees to
compute the cranking pulsewidth used for starting. Say that you've
tuned the engine to start just fine with a coolant temp of 50 degrees,
which results in a pulsewidth of 10 milliseconds. There are a lot of
ways we could get 10 ms at 50 degrees, we could have our tables set up
like

CLT PW - Case 1
0 20
100 0
or

CLT PW - Case 2
0 10
100 10

or even

CLT PW - Case 3
0 0
100 20

and they'll all work just great at 50 degrees... But now, what happens
when the temperature drops to 0? Well, in the first case we might actually
get the thing to start, because it produces longer pulsewidths as the
coolant temps drop. Case 2 will probably be too lean to start and case
3 is a total disaster, we have zero pulsewidth at zero temperature!

Now raise the temperature to 100 degrees. In case 1 we have that same
problem that case 3 had when it got cold, no fuel at all. The other
end of the spectrum, case 3, that worked so well at 50 degrees is now
way to rich and we end up flooding on warm starts.

So, the trick is to find the slope of this line and set the end points
properly. Oh, and confounding it is probably the fact that it's not
really a line but more of a curve, and maybe not even a nice simple
curve! This is probably one of the hardest parts of tuning EFI, and
a bunch of us have spent a lot of time thinking about it and haven't yet
come to any good conclusions (but Matt Dupuis has come the closest in my
estimation, do a search of the forums for his posts on cranking pulsewidths
if you want to know lots more).

Eric

--
Eric Fahlgren http://www.not2fast.com/

[Sam Powell]

Thanks guys. That is big help. I don;t know where these values are now, but will check them out. Sam