OBD Fuel Trim vs ECUWorx DME Utility LTFT

[Shonky]

I'm chasing a bit of a minor issue. My LTFTs tend somewhat negative usually ending up around -8 to -9% (i.e. a bit rich but not by heaps). When cold / open loop it seems to run fairly rich but otherwise runs fine.

I reset the adaptations using the ECUWorx tool and they read back as 0.00 as expect (edit: I think although perhaps that was 1.00 based on my realisation below). Drove the car to work (about 15km)

Using the Android App "Car Scanner" by the end of the drive you could see the LTFTs start at 0 as expected after the reset and slowly drift negative ending up at -6.25 and -5.47

I then tried the ECUWorx DME tool again and it gave 0.94 and 0.95. Before the reset, they read as 0.92 and 0.92.

Ah I think I figured it out whilst writing this.... Do ECUWorx fuel trims show as a scaling factor i.e. 0.94 means -6.00% and it just loses some precision compared to the normal OBD tools?

[Martyn]

STFT / Additive figures are in are pulsewidth in ms.
LTFT / Multiplicative figures are percentage / factor modification of injector pulsewidths.

[Shonky]

Thanks for confirming.

[sapote]

STFT / Additive figures are in are pulsewidth in ms.
LTFT / Multiplicative figures are percentage / factor modification of injector pulsewidths.

In this case the ECUWorx shows STFT1=-0.07, LTFT1 = 0.94.
Does this mean the DME reduces 0.07ms of the calculated injector pulse width and add 0.94% to the calculated pulse width. However OP said the OBD2 app reported LTFT at -6.25%, and so I don't see the equivalent of ECUWorx LTFT of +0.94% and standard OBD2 LTFT of -6.25%. Are we comparing apple to apple here?
Can someone explain how to read ECUWORx LTFT that makes sense when compared with OBD2 LTFT?

[Shonky]

LTFT is not a percentage - his response is little confusing. The number is a multiplier. So (value x 0.94) is the same as (value - 6%).

Im not 100% sure about the STFT values.

[sapote]

How do you get LTFT 0.94 is the same as -6%?

[heinzboehmer]

How do you get LTFT 0.94 is the same as -6%?

0.94 * value = (1 - 0.06) * value

[eacmen]

This is super confusing. The rest of the world uses percentage for this value. ecuworx should do the same.

[repoman89]

Must be an engineer thing. My first instinct was to assume it’s a multiplicative factor.

[sapote]

0.94 * value = (1 - 0.06) * value

Oh man, I never see this. Thanks. And if the fuel trim = 1.06 then it is +6%.
They should just stick with the OBD standard of fuel trim data.

[heinzboehmer]

I have a feeling the additive vs multiplicative factor is a DME thing, not an ecuworx tool thing. This approach makes sense if the engineers wanted more resolution on the STFT than on the LTFT, while constraining both to a float of a certain bit width.

Pure speculation though

[sapote]

In this case the ECUWorx shows STFT1=-0.07, LTFT1 = 0.94.

So what it means for STFT=-0.07? It makes no sense if it means 0.07ms subtracted from the base calculated/table pulse width as we don't know what the base value is.

[heinzboehmer]

So what it means for STFT=-0.07? It makes no sense if it means 0.07ms subtracted from the base calculated/table pulse width as we don't know what the base value is.

Base value would be (some constant grabbed from a table in ms) * LTFT, no?

[Cubieman]

What is the "value" that is multiplied by .94 (in this case) to get to -6%?

[sapote]

What is the "value" that is multiplied by .94 (in this case) to get to -6%?

We don't need to know the "value" (which is based on the air mass, rpm, temp, etc.) but we know that the DME reduced that value by 6% as the mixture was a little too rich. 6% is not too bad, but more than 8% is problematic.

0.94 is 94% of the "value" is injected instead of 100% the value, so this means the injected fuel is 6% reduction from the idea value.

In ECUWORx world, LTFT = 1 is perfect as no modification to the "value" is needed. It's unnecessary caused confusion, and they shouldn't use LTFT or STFT for their data format.