From: John De Armond
X-Source: The Hotrod Mailing list
Date: Feb 1992
Subject: Re: More on Legalities

>You should be skeptical.  The "instantaneous MPG meter" is probably
>just a manifold vacuum gauge, with markings just as likely calibrated

No.  It's a real MPG gauge.

>Unless this is some kind of gizmo hooked up to the engine computer

Yep.

>Correct me if I am wrong; if they are calculating actual fuel used per
>actual distance, I will be duly impressed.

Be impressed.  Actually it is very simple to do.  Given that the 
fuel system maintains a constant pressure across the injectors (all
port systems do), measuring fuel consumption is a simple matter of 
totalizing the number of seconds of open injector time, multiplying
it by the flow rate then multiplying by an injection correction
factor that compensates for fuel and pintle inerta.  You've got 
distance info for the pinball dashboard speedo display so MPG is a 
simple matter of division.

This is not new.  BFI (before fuel injection) I had a commercially 
manufactured MPG gauge that did measure actual fuel flow.  When I 
bought my 75 Zcar, I missed this capability so I built a device (using
discrete logic, remember this was the era of the 8080) that did
exactly what I described above.  Worked well.

John

From: John De Armond
Subject: Re: Looking for aftermarket fuel consumption meter
Date: Fri, 15 Oct 1999 13:46:10 EDT
Newsgroups: rec.autos.tech

Robert Hiebert wrote:
> 
> Hi Steve,
>      It seems that way, but you'd think with the cheap electronics
> available, and variety of sensors already on  EFI vehicles, someone
> would be tapping in to this market. I know they are becoming fairly
> common as an OEM option on many vehicles.
>      I've been able to find some that display fuel flow per hour, with
> applications for light aircraft, but all the ones I've found won't
> work with a fuel return system. I also found one for a BMW K series
> motorcycle, but it was specific to that bike. FloScan makes a flow per
> hour unit for marine use, using 2 flow transducers, pressure and
> return, to calculate fuel burn, but again, it's fuel per hour, and
> very pricey to boot. I keep hoping there is something out there, as
> building one may be beyond my ability.

The problem with trying to measure the fuel flow is that you face
the classic problem of trying to compute the small difference
between two large numbers.  The flow meter error will kill any
possibility of the calculation being meaningful.

The only way to compute fuel consumption on an EFI engine, and the
way the OEMs do it is to totalize the injector open time and,
knowing the injector dynamic characteristics, compute total flow. 
It is not as simple as just knowing the flow characteristics of the
injectors.  One must also know the dynamics of the injector since
under normal conditions, the injector spends so much time opening
and closing.  A typical injector can open in a millisecond.  It
takes a similar amount of time to close.  During the opening
process, the flow goes from zero to full flow in a very non-linear
fashion.  High speed photographic studies I've done on the old style
Pintle injector (where the valve pintle is visible) shows that the
flow doesn't start until about 750 microseconds into the open
event.  The closing process is also nonlinear but of a different
shape.  During idle and cruise, the injector is running with an open
period of from about 0.5 ms to perhaps 1.5 ms.  Under these
conditions, the injector is continuously either opening or closing,
never reaching the full open state (or just barely doing so.)  The
OEM has characterized the injector and driver and knows the transfer
function. The OEM also knows the characteristics of his fuel rail 
pressure regulator, another vital input parameter. Without
specialized test equipment, you can't.  And since the transfer
function is unique to each engine combo, a universal solution cannot
exist. 

I have designed and built a specialized fuel consumption instrument
for use in engine management development work.  I have also designed
and built a fuel injector flow bench that will characterize the
injectors.  This process is quite accurate but is impractical for
end-user applications.

In summary, if the OEM didn't provide a fuel consumption readout,
it's not practical to add it as an aftermarket item unless you're
willing to do some engineering.

John