Mods and FCD inaccuracies. MPGuino?

[RobertSmalls]

No great surprise that the car still records fuel consumption even when I've got the fuel injector kill switch closed and my foot on the throttle. What does surprise me is that applying assist with MIMA makes the FCD register more fuel consumption. Full assist at 30mph = ~60mpg on the display, even at RPMs that would suggest DFCO. ~80% assist at 40mph = ~110mpg on the display, despite zero fuel consumed.

Has anyone here installed an MPGuino in an Insight? I suspect it might be a good way to predict how much more headroom you have within lean burn, so I've been thinking about building one. If you've used one, what do you think?

Btw, the MPGuino is an Arduino that reads VSS and FI signals and logs MPG. Seems kind of unnecessary if Honda's FCD is good enough for you, but the MPGuino will spit out accurate, precise numbers many times a second, unlike Honda's blocky bar graph FCD that refuses to jump around like your iMPG does.

[Mike Dabrowski 2000]

FAS opens stops the fuel injectors from operating only while the switch is pressed and the relays are held open.
When the engine quits, the fuel injector pulses stop, so no further fuel is used, and the ECM stops trying to fire the injectors. At best,2-5 seconds of injector pulsing may happen before the system gives up and decides that you have run out of gas.
During the 2-5 seconds, the car thinks it is using gas but it is not, so the actual MPG will be artificially reduced for 2-5 seconds. In my opinion, this "error" will fall into the noise, and even with hundreds of FAS events will not be something we would see.

[red1dr]

Originally Posted by Mike Dabrowski 2000

In my opinion, this "error" will fall into the noise, and even with hundreds of FAS events will not be something we would see.

I have MIMA with the FAS module, and I FAS a LOT. My calculated mpg dividing the tank mileage by the amount of gas pumped usually agrees within 1mpg with the average mpg displayed on the FCD for that tank. Tanks are typically 80-100mpg, so that's a bit over 1% maximum discrepancy. And if there is a consistent bias either way I can't see it in the data.

[RobertSmalls]

The error from engine kills should be very small, because the ECU is only trying to inject enough fuel to idle. Figure 4s * 0.11gph = 1.2*10^-4gal. i.e. 8200 engine kills per gallon.

It's the EV mode that creates meaningful errors. I've seen anywhere from 60mpg to 250mpg (lots of glides) displayed on >1mi EV mode trips with no actual fuel consumption.