The LT trims are going to be quite difficult to track down because there is alot going on to make them work. There is a varying delay time for some reason. Even my car which has no trim problems will sometimes adjust the LT trims quite quickly, and other times take minutes at the same conditions to make the trims adjust.

Example. Yesterday while finding the ST trim function I made up a rom with the ST Trims set to lock at -15% and watched the LT Low trim. Within around 30 seconds it started to adjust into the negatives. After changing back to normal ST trim values it took almost an hour of driving and long periods of idling to get the LT Low trim to climb back up (or even move at all).

Driving... it would be completely random when the LT trims for mid and high would kick in and try to adjust. This is not going to be easy to find.

 

Your scaling should probably be used for the loggers now as well, instead of the popular (.1961*x)-25

Using (x-128)/5 will give more accurate results. A 0 trim will show as 0 with your scaling rather than .1008 with the commonly used scaling in the loggers.

 

I think I may be doing something wrong. I patched the definition into my 88590015 and into the base file for the scaling definition and when I open ecuflash- it's all good. when I open the new Ft tables, they don't open completely. I can only see the side with the temp and not the right side of both tables. Any ideas what I'm doing wrong?

 

C'mon man....I have faith in you.

 

Check the first post again. You forgot the scaling.

 

I just added this patch to my 88590015 rom and it worked out fine. Place the scaling at the top section of the base .xml. Doesnt have to be the first line, but the first section...

 

Do you have the entries for 96940011....please

 

Great find.

I was hitting the board up a month or two ago for this table hoping I could limit some of the correction the ECU does while running a VTA BOV.

Can't wait to try it out, still waiting for the TC and rear diff to show up though.

 

here is the correct scaling for trims
Or here when you want same values as OBDII shows.

But at that time nobody took me too seriously

 

Acamus, you are most likely (and usually) correct. I based my formula purely on the values I was seeing in the logs for trim in Evoscan and not from deeper disassembly. I figured someone else would find the true formula eventually.

I am so stupid!!! I could have just maxxed the trim to full positive and full negative and have a look at what it shows in the OBD. DOHHH!!!

I do remember seeing my OBD unit showing around 13 or 14% at one stage, so maybe it goes more than 10%. Tell you what, I am going out the to car now to check it!

 

Well well well, I think I have you beat this time Mr Acamus

I just did some checking and OBD shows +/-25% trims. Evoscan also will show raw 0 to 255 for the trims too, so it goes beyond 64 and 192. So far my simple formula seems to line up..... not bad for a guess! I just changed the formula in Evoscan to (x-128)/5 and it reads correctly and lines up with the OBD tool display.

I found something interesting. It will use the first setting in the map for its max or min for the first 20-30 seconds, then it instantly switches to the map setting. It had me baffled for a minute because I was pretty sure it was engine temp based but it followed another number for a short period. Something to note when tuning the maps.

 

I am more than happy to get corrected

If you look at http://en.wikipedia.org/wiki/OBD-II_PIDs
OBDII range for trims is -100% - 99.2%

So maybe I made mistake here

 

You guys crack me up.

So anyway, is it (x-128)/5 or (x-128)/6.4? one is just over the defunct value and the other is just under.

 

If Mattjin has max-ed and min-ed fuel trims and was seeing +-25% via OBDII protocol, then his scaling is correct.

25*(x-128)/128 = (x-128)/5.12 so approx. (x-128)/5 is correct.

 

I put in the scalings and it still is coming up with only half the graph- here's how I have it
<!-- *** SCALINGS *** -->

<scaling name="O2 Trim" units="%" toexpr="(x-128)/5" frexpr="(x*5)+128" format="%.1f" min="-25.6" max="25.4" inc="0.2" storagetype="uint8" endian="big"/>
<scaling name="AFR" units="AFR" toexpr="14.7*128/x" frexpr="14.7*128/x" format="%.1f" min="5" max="30" inc="0.1" storagetype="uint8" endian="big"/>
<scaling name="AFR16" units="AFR" toexpr="14.7*128/x" frexpr="14.7*128/x" format="%.1f" min="5" max="30" inc="0.1" storagetype="uint16" endian="big"/>

If I add remove this definition from the base file it doesn't matter how the table comes up in ecuflash. What am I doing wrong? I've never had a problem with patching. I've got 5.10, and a few other patches on my rom.