math to calculate mpg from injector pulse width
 

i'm making some assumptions with this math. can anyone here confirm if what i'm coming up with is accurate or if my assumptions are correct?

this is specific to honda/acura OBD1 ecus


for this example:
rpm=3250
mph=59
pulse width as taken from ECU map based on load=.127ms=.000127sec
injectors= 240cc
1cc=.000264gal


my major assumption is that injectors pulse once per revolution. is that correct?


so, here we go:

cc/pulse=.03048
gal/pulse=.00000805196
gal/min=.026168
gal/mile=.0257327
mile/gal=38.86


based on the reverse math from known mpg based on my weekly fill-ups, the gal/mile calc above is spot on.


i'd like to use the above logic to request a change to FreeLog (popular honda tuning datalogging software). if the math is right and accurate, a free OBD1 honda realtime MPG display could be developed within freelog.

But do you have a vacuum biased FPR? in which case you'll have to read MAP data and adjust for actual injection pressure.

yeah, trudat. good point. hmm... let me see what data i have.

to be truly accurate i'd have to account for IAT compensation, too.

eh: "f the math is right and accurate, a free OBD1 honda realtime MPG display could be developed within freelog."

I thought freelog depended on a connection via something like a Hulog (https://www.moates.net/documentation....ntation_id=25). Or is there a simpler way to use freelog? I don't want to hack my ECU, I just want to monitor some data in a simple way.

"pulse width as taken from ECU map based on load=.127ms=.000127sec"

When you say you got that data from an ECU map, do you mean via something like a Hulog? Or is there another way to do it?

Sorry if that's a dumb question!

yes, freelog connects via the logging header on the ECU. a hulog or DKU-5 nokia cellphone cable is required to interface with the ECU. at a minimum a 4pin header will need to be soldered to the board to log.

no, i opened up an OEM .bin for my ECU in CROME (free honda ROM editor) which is used for tuning fuel/timing. similar honda editors are hondata, uberdata, neptune, ecTune

"i opened up an OEM .bin for my ECU in CROME"

OK, thanks for explaining.

I'm dimly familiar with that world, mostly because of hanging out at pgmfi.org a little bit. I just stumbled across a page (https://www.phearable.net/tech/efibeginnerguide.html) that does a good job of explaining this business of hacking a Honda ECU.

For various reasons you can probably guess, it would be nice if I could find a .bin for my car (VX). But such a thing probably does not exist (outside of a lab inside Honda), because my ECU is not chippable. Right?

This is something I've wondered about, and I have a feeling you know the answer.

actually, i'm not familiar with the VX ecu board.

if i could see a picture of the board, i could tell you if it was. you're assumption about the .bin for it not being available is probably true. 99.99% of the individuals socketing/tuning ECUs aren't going for FE and aren't tuning a d15z1

"if i could see a picture of the board"

I wish I had one, but I don't.

I've seen lists of Honda ECUs that are considered chippable, and the d15z1 ECU is definitely absent from that list. So I'm going to assume that means I'm not going to find a .bin floating around somewhere.

Anyway, the numbers you posted are interesting and helpful, so I'll be paying attention as you continue your efforts. I'm also very interested in the accuracy issues that were mentioned (e.g., IAT compensation).

Actually no, you can find a P07 bin. I have one on my computer. PM me if you want me to email it to you.

i'm simply trying to nail down all the formulas so that my findings could be presented to the freelog programmers. IAT compensation is super easy to handle. the main trouble now is coming up with the appropriate scalar for final fuel injector flow based on changes in fuel pressure. that actually has a major affect on the accuracy of the calcs.