more tablehack!!

so i've been working on http://ecmhack.com/tablehack/ tablehack a bit more ( spam spam spam) and the most recent feature kind of sends it to the next level if you are really into log analysis.

so the primary purpose of tablehack is to do table-matched analysis of data logs really quickly right. but i figured, what if your ECM doesn't have ready access to the log data you need..... what if you need to, for example, compare your open loop AFR target to your wideband, but..... there's no AFR target in your logs? what if you have ANY table of data but that isn't in your log, but you need to analyze it, filter with it, whatever?

well, if you have your AFR target table itself, and a pretty good idea how it works math-wise, and the map/rpm data you need to feed it IS in the log......

so i built a feature into tablehack to calculate those additional columns based on other log data. i call it a dynamic column, which is actually a formula you can define yourself. people do this with spreadsheets all the time but this is 1000x easier, especially considering you can have multiple lookup tables in the same formula. yes, lookup tables.

so for our AFR target table example, i paste the table data into a dynamic column, and it is automatically assigned a variable A. the formula could then be A and the column would be the result of a lookup on the table.

or, i could add the wideband data column (B), and go A / B and call the column "AFR target error"

you could use any formula really, you could create a column of AFR / 14.7 called LAMBDA or whatever and it would just automatically show up in your logs.

real world example, i have a boost target table (boost target is not logged on my subaru) that is TPS vs RPM. now i can easily add that table from my bin and it becomes part of my log, automatically populating with TPS and RPM data and presuming boost target. now i can run an analysis of MAP vs boost target and see how close to my target i am.

attached is a proof of concept picture of tablehack's calculated spark (from two timing tables and a modifier table) and the actual ECM spark output. so basically if you did not have spark advance in your log, but you had your spark table, you could just throw it in there.

this could teach you stuff too, like if the data differs, you could go 'oh, damn, there is a modifier table somewhere i'm missing'

get the idea?

simple use case in case the above is confusing. lets say your log has LBLM and RBLM in it.

lets say you define a dynamic column like

A=LBLM
B=RBLM

then a formula like

( (A + B) / 2 ) / 128

and call it BLM%

every log you load that contains LBLM and RBLM will now have a BLM% column added to it like that

even simpler example, unit conversion. your log has MAP KPA but you like INCHES OF VACUUM

create a dynamic column called vacuum and A=MAP, then the formula A/6.895. all the logs you use in tablehack now have inches of vacuum in there.

Great work steveo, and amazing idea. something like table lookup as the pcm sees it. and the lookup value is added to the log. Some call it data tracing but this goes to a whole new level.

I am experimenting with tablehack today. Setting up a simple VE histogram. I get an analysis error: Could not refactor log logfilename
I used an XDF to load the table and I am using a dynamic column as the data. What am I missing...


Edit: Duh nevermind I figured it out. I did not select the proper columns after I loaded the table from the xdf.

its a total monster of a program with lots of bugs that you get used to

like please for the love god do not make any 2d tables