RE return to fuel tank

Greg Hermann bearbvd at sni.net
Thu Oct 22 22:20:02 GMT 1998


>Hi all
>
>Been reading this thread and can say that several european auto
>manufacturers are working on a returnless fuel system, where the fuel
>pump is controlled by the ECU using PW modulation so that only enough
>fuel is pumped into the fuel rail to meet the engines needs.
>
>The biggest reason behind this is emissions, passing warm fuel back to
>the tank causes fuel vapour to build up & pressurise the tank.  It
>cannot just be dumped to atmosphere anymore so the vapour is passed
>through a charcoal filter to sepperate any fuel before venting to
>atmosphere through a valve controlled by the ECU.

But--for really high performance--without quite so much worry about evap
emissions, I believe that it would be fairly easy to show the following
(with a bunch of differential equations):

The most stable fuel pressure control will be possible if you use a
variable drive fuel pump,  a pressure regulator,  feedback on what the
regulator is doing from a fuel pressure transducer, and a controlled
reference pressure on the back side of the fuel pressure regulator
diaphram, as well as a fuel temperature transducer to allow proper
correction for fuel density. Use the fuel pressure regulator as the primary
control with control over pump speed following what is going on so as to
minimize the amount that the regulator armature has to move in order to
keep the fuel pressure where the computer wants it. This also allows you to
vary fuel pressure to wherever you want it at different points on the
engine MAP/speed matrix so as to maximize the dynamic range of the
injectors.

With proper PID control over what the regulator and pump are doing, one
could get extremely stable, but variable fuel pressure control, and thus
much closer f/a ratio control over a wider dynamic range. Yes, there is
room here for a code guy to figure out how to put some learning blocks into
this loop so that the control loop could anticipate things a bit, and hold
things even closer.

Also, with the fuel pressure transducer input, the computer could catch any
pressure glitches on the fly, and correct pulse widths accordingly.
Injector pulse widths could also be corrected for fuel temperature in the
rails on the fly.

Including a float vent at the high point of each rail (piped back to the
low pressure side of the system) and mounting the fuel pressure regulator
at a lower point, with its outlet pointed downward, would (1) assure that
no bubbles ever got into the injectors to cause random lean mixture
anomalies, (2) make any hot soak vapor in the fuel rails situations and
problems into history, and (3) avoid any possibility of fuel pressure
anomalies (and resulting a/f mixture anomalies) due to the regulator
randomly passing fuel vapor instead of liquid fuel.
>
>To achieve this they fit a pressure sensor in the fuel line and by
>knowing pressure and injector duty cycle the pump duty cycle is
>calculated.
>
>>From experience on dynos I can say that for power you want hot fuel in
>the fuel rail

I strongly suspect that this finding would change if you were getting
better (finer) atomization of the fuel! Have any of your tests been run
using Bosch's "air shrouded" injectors and at part throttle (which is where
said injectors improve atomization)??

 and cold air because of its increased density.

No question about this, but let us not forget that fuel VAPORIZED (as
opposed to fuel ATOMIZED) before the intake valve closes displaces air and
affects VE negatively through this mechanism! And the lighter the molecular
weight of the fuel being used, the stronger this effect becomes.
>
>               Regards, Greg

I understand that I am discussing issues here that go WAY beyond normal
needs for street performance , although bits and pieces of it go straight
at some fairly common street problems. But HIGH performance is what
interests me most, and I would hazard a guess that a bunch of others on
list have interests in the same kind of stuff, so don't bother telling me
about the economics of it!!

Further, if you have doubts about the validity of the control theory I am
applying to fuel pressure control, I have a text on a shelf somewhere which
thoroughly describes the theory of accurate water level control in large
utility steam boilers (100 megawatts and on up in size) in more detail than
you could want. Fuel rail pressure control for an EFI system is a
reasonably analogous situation. Name and author of the book escapes me
right now, but if anyone wants to spend some money and wade through it in
an effort to learn, it IS a really good book, and I would post title and
author name later if there is interest.

-G.





More information about the Diy_efi mailing list