[DeTomaso] EFI

[Jørn C. Olsen]

I have installed IR EFI on my 408 stroker with Pectel SQ6 engine management
system and will never go back to carburator.
The drivability and response on my engine was an huge improvement.

In my opinion is the trick to find a tuner who really knows the system you
need help to map.

J.C.  

-----Opprinnelig melding-----
Fra: detomaso-bounces at realbig.com [mailto:detomaso-bounces at realbig.com] På
vegne av Daniel C Jones
Sendt: 5. oktober 2011 17:52
Til: Pantera REALBIG forum
Emne: Re: [DeTomaso] EFI

> Question: Are any of those EFIs closed loop? That's how OEM works, and 
> if it is, then tuning should be unnecessary

OEM EFI are combined open and closed loop systems and, even if they were
completely closed loop, would need tuning.  The gains (and limits,
deadbands, etc.) in the forward and feedback paths are scheduled with
various parameters and tuning is required to determine the values of those
gains.  The reason open loop tables are used by the OEMs is the sensors
(mass air flow, O2, etc) don't have the bandwidth required (are too slow to
react to changes like throttle transients).

There are attempts at automated tuning, either model based or target based,
though I don't know how successful they are.  Model based systems attempt to
model the engine to determine the fuel (and spark) needs.  I know some of
the aftermarket systems use a model to come up with an initial set of
gains/tables that will at least start the engine and serve as a basis from
which a tuner can work.  In target based systems, you defined desired
air-fuel ratio tables (the targets:
ratios versus
RPM and load as determined by throttle position and possibly manifold
pressure) and
the computer adds or subtracts fuel (and spark advance) to iterate to
towards the targets over time.

I'm a flight controls engineer by trade.  There are all kinds of controls
schemes.
Implicit and explicit model following, neural networks, etc.  I've even seen
a Darwinian gain scheduling approach in which you define populations of
gains and success criteria then run simulations to determine the best
(survival of the fittest) of the gain sets then introduce random variations
(mutations) in the gains are recompete the resulting sets until the best set
is found that satisfies the success criteria.

One of the problems guys have with aftermarket EFI is picking the wrong
controls approach for their specific engine.  I've made a lengthy previous
post which detailed the three basic engine control system schemes (mass air
flow, speed density and Alpha-N) and which ones are best suited for manifold
type (common plenum or IR) and cam overlap.
If you pick the wrong control approach for your engine, all the tuning in
the world won't make it work properly.

Dan Jones
_______________________________________________

Detomaso Forum Managed by POCA

Archive Search Engine Now Available at http://www.realbig.com/detomaso/

DeTomaso mailing list
DeTomaso at list.realbig.com
http://list.realbig.com/mailman/listinfo/detomaso