I was playing around with writing a sort of 'VE-Master' type code for post-processing and data decomposition/reduction, and had some thoughts I wanted to shake out with the crowd.

Regarding both MAF & MAP:
- Is the calculation: "Fraction Compensated" = BLM*INT/(128^2) reasonable and correct?
- Does doing this render 'cells' irrelevant and allow one to just look at the realtime fuel correction called for?
- Could this be preferable over looking at the BLMs alone? Seems it would. Save having to 'lock' BLMs?

Regarding MAP:
- Can the above value be accumulated into an array of RPM -vs- MAP kPa and then be used to correct VE tables?

Regarding MAF:
- Can you just take the above values (fuel corr -vs- g/s) into a vector of MAF g/s, and reverse-map the MAF 'counts' from the existing tables to generate new MAF g/s tables?
- Should you then remap the corrected g/s into a smooth function across the tables?

Anything else critical to consider with respect to fuel table corrections?

Hope this isn't too much questioning. I'm just trying to put together a tool which will make it easier to 'digest' acquired data and translate it into an improved burn...

 

I spent some good time making a VE table creater. I Left some really REALLY important things out and that is why I Never made it public.

1) That was too much correction for me. I think I finally settled on the equation
I usually used a value of 5 for X. It is a good way to control the scaling. This is more of a data thing, and it was something that fit for me when I had all my stuff sorted out. That is off the top of my head, I'll check files later.

2) If you ignore the cells like I did, then whenever you transition into a new Cell, the following BLMs still have to adjust. After a LOT of driving, this would even things out. But on a brief data capture where you aren't thinking constantly about keeping the throttle nice and steady, BLMs change quite a bit.

3) For reason #2 I gave above, I still ended up locking the BLMs in the end.

4) Hell yeah. That is the main reason to do it, right?

5) Only take data when learn control is on, at least 5 cells after BLM has been in that RPM/MAP position, you will have to rescale the equations you use to interpolate when shifting from the LOWER VE Table to the upper VE table, make sure the upper values in the lower table and the lower values in the upper table are the same (since the tables overlap), and by ALL means, incorperate a duty cylce check.

I ran my program time and time again. Only after spending a bunch of hours in it did I go back and check my DC. I had so much fun toying with my automatic VE table maker that I forgot to check it during normal burns and I needed to go back and change my injector constant and play with fuel pressure.

Those are some of the reasons that I never made it public. There were just too many things that *could* go wrong and I didn't want to have a bunch of beta testers come knocking on my door because they didn't verify that the changes should be made before they said "make the changes".

Good luck. If you need more MAP ideas re this, let me know. -Matt-

 

Here were the original instructions. Kinda gives a brief overview and you might be able to get the basic idea of things.

If nothing else, the small pics give a general idea of how to sort the BLMs and RPMs.

I disabled the link to the macro since it was flawed in more places than Quaker had oats.

http://expert.cc.purdue.edu/~msandine/Alpha730.doc

Should be anonymous entry if it asks for username/password.

 

HighHopes85,

Do you have a link to this that works? I would like to check it out if you dont mind. I am very interested in how this works in detail, & you seem to have a firm grip on it.

thanks
BW