Affordable Wideband ECUs - How to?

[Garfield Willis]

[I tweaked the subject in a blatant attempt to seque to a related
topic].

On Fri, 31 Mar 2000 11:15:14 -0500, Will McGonegal
<McGonegal.Will at etc.ec.gc.ca> wrote:

> I am eagerly waiting to find out what your
>product will be.  For now we'll hook our sensors up to the NTK box at work so we
>don't fry them.

Ahh, I was unsure until now whether you had their interface box as well
as their sensor. Watching the behavior of those NTK ($600) sensors
whilst under control by their own blue box is a perfectly safe way to
learn more about them. Just that you'll have to be cutting into some of
the leads/harness to measure currents. Others in your team may not
appreciate that, but hey, that's science. :)

>We were using the wide ratio O2 sensor yesterday on a vehicle with a
>programmable FI computer.  The car was on one of our chassis dynamometers.  I
>set up the dyno controller software to hold the dyno at steady speeds for the
>testing.  The throttle setting could be varied and the dyno would hold the speed
>constant.  The air-fuel-ratio was read off the NTK system and adjustments to the
>fuel injection computer were made to get desired AFR.  It really worked well. 
>You could see the AFR reading quickly respond as the FI settings were changed. 
>Handy tools to have.

I expect you must have been holding significant load on the engine via
the dyno, in order to get it into open-loop. This is the normal tuning
mode for non-wideband controllers.

What might be an interesting and fruitful thread to engage in, one that
we can all share in without product hangups, is to think about the "next
step", working out & discussing how one can affordably get to the point
of using the wideband O2 sensor in *closed-loop* at significant load (or
what is often referred to as WOT closed-loop).

There was some discussion early on, and then later some also in private
amongst a couple of us, about the possibility of programming an OEM
controller like the GM ECUs to force it to stay in closed-loop even at
high TPS and load, and trick the controller by mimicing the
stoich-crossing to occur at whatever AFR you chose. This was essentially
the idea of simulating the stoich-crossing input to the OEM controller,
and use an EGOR module (or an NTK box, or whatever) to allow "moving"
this simulated stoich-crossing to whatever AFR you commanded. Bruce
Plecan and some of his GN/GM buddies apparently looked into it, and
decided there likely wasn't enough "authority" in the tables that could
be reprogrammed, to keep the controller forced into closed-loop, and
gave the idea a miss, suggesting that you really needed to get at the
code to do it right. But coaxing those extremely affordable OEM
controllers to make the transition to WOT closed-loop controllers sure
sounds like a wholesome grail worth somemore pursuit, purely from an
economics standpoint.

Of course, this also raises the corollary question of what additional
coding algorithms would be required for the 332 project guys, for
example, once they're affordably able to read as input, the current AFR.

So there ya go, two mingled and possibly fruitful topics: (1) Q: any way
to move those useful and inexpensive OEM controllers into the realm of
WOT closed-loop ECU, once affordable AFR sensing becomes available, and
(2) Q: what kind of interesting controller phenom do you get into if
you're actually building your own controller from scratch, have access
to the code, and want to equip it to run closed-loop WOT via wideband O2
sensing.

Might also be revealing to hear from anyone with the wideband option on
a FelPro controller, explain any interesting programming additions (like
new modes or parameters), that are necessary for bringing up one of
these controllers on an engine. Anyone with a FP controller with the
wideband option?

Gar


----------------------------------------------------------------------------
To unsubscribe from diy_efi, send "unsubscribe diy_efi" (without the quotes)
in the body of a message (not the subject) to majordomo at lists.diy-efi.org