Request for firmware update. cranking pw / coolant temp

[gunni]

As it stands the cranking pw is nowhere near anything to be considered up to the task.

So , simple 2d table, coolant temp versus pw.
should sort alot of cold starting in places where temps get very cold.

[rob@vems.co.uk]

So that we could at least replicate something like the Cranking PW graph Cliff gave me?

[gunni]

Exactly what I was thinking.

can´t be that hard since the current situation is a linear setup -40C / 71C

[dnb]

It would certainly be nice to have.  I haven't really had much trouble with starting but the temperature here is never all that extreme - and if it were really cold, I wouldn't be using the TVR...

[gunni]

I think that is one of the reasons it hasn´t been done , cause it´s taken as a nice to have,
if vems is supposed to be serious, it has to be "right"

[GintsK]

I think VEMS do not support pulswidth longer than 25.5ms. At -20C it can be limiting factor.

Here is curve from some OE:


I vote for 4-point 2D graph. It must be enough.
70C- 100%
40C - 200%
5C - 400%
-40C - 2000%

Gints

[Sprocket]

I think VEMS do not support pulswidth longer than 25.5ms. At -20C it can be limiting factor.

Not really.

just use all banks rather than alternate at crank

[dnb]

Or bigger injectors...

[GintsK]

I think VEMS do not support pulswidth longer than 25.5ms. At -20C it can be limiting factor.

Not really.

just use all banks rather than alternate at crank

1.5..2ms is common cranking PW@>70C on n/a applications for me. It is when all banks fires at all ignition events. 1.5*20=30ms@-40C

[cliffb75]

That chart is real data from a production vehicle I did back in about 2002.

For OE we require a good clean consistent start with robustness to different fuel volatilities.

For VEMS, you could probably get an accurate enough curve by linear interpolation of 4 points as Gints has suggested. Sure some points will be a bit rich and others a bit lean, but I think given the constraints it would be good enough, and also a lot easier to calibrate for most users.

Alternatively, I found across a few vehicles that if you convert that into a multiplyer based on the fully warm pulse, the curve is pretty similar (though I can't guarantee that for all engines). You could therefore try something like a curve based on a rising rate calculation from a base value - that means only 2 calibratable values but a nice curve, but is probably worse from an ECU memory / calculation point of view.
 

[rob@vems.co.uk]

I spoke to Sambas about this and he said that no more features can be added before there are some proper tests done on 1.1.43alpha2

The following trigger types have been tested on the bench, but need some tests done on real cars...

• 6-1
• 12-1
• 18-1
• 36-1
• 60-2
• 10+1
• 24+1

If any of the signatories to the poll would like to try some cranking tests on their own engine it would hurry this change along

[gunni]

I´ll try and test my 60-2 today

[mattias]

It seems like this is the place to start asking about features like this one. I mentioned my demand for this ages ago in the Wiki, but I guess I'm the only one in the world who lives where it can be -30C on a cold morning ?
Living with the old hot/cold values has haunted me.

Sambas is doing most of the new features ? That's great!

[Sambas]

Ten point (similar to iac and warmup) cranking pw curve will be in next 1.1.x release. Don't know since when this was enabled but it allows to use upto 47.5ms pulsewidths.

Code: [Select]

allows configuration of higher
than 25.5 msec in 8 bit config value. Used for config.cwl, cwh, injpwmt, primep
Value 200 mean 20msec (megatune compatibility 100 usec step)
but 201 means 20.5 msec (changing to 500 usec step)
This allows 20 msec + 55 * 0.5 msec = 47.5 msec max pulsewidth


[gunni]

Awesome news Sambas.