GM knock detection

garfield at pilgrimhouse.com garfield at pilgrimhouse.com
Tue May 19 21:13:47 GMT 1998


On Tue, 19 May 1998 11:59:05 -0400, "Bruce Plecan" <nacelp at bright.net>
wrote:

>Now if I have a new
>trigger for turning that 555 on (when ion eyes are shining), will
>I now just be able to feed that same acoustic signature into a 
>none esc type ecm, like a 730?...  There was some talk about 
>a different amount of resistance between new, and old style
>knock sensors.
>  Or would the sane approach be starting at square one.  Has
>anyone had both styles on an o-scope, and seen any difference
>in voltage generated?.  Any clues about v generated?.

Hey Dr. Pelican,

I don't know enough to understand half them questions you ask above, cuz
I am "ecm impaired", as you know, but I do have some thots on some
factors to consider in order to answer part of your questions. Whew, was
that convolved, or what?

Anyhoo, couple things to think about concerning ION. I expect we'll see
that she ends up with digital signals for fir/misfire indication (I know
you weren't askin bout that, but hey, I'm throwin it in for free!), and
adjustable thresholds for those. They may be in pairs ala Saab, or
individual per cylinder ala "who cares, the electronics don't cost
anything".

As for knock, I think we should follow Saab's lead and maybe do TWO
things. First, as per Saab, we should probably plan on an analog
ac-coupled 0-5V range signal that represents the actual measured
ionization current. This I would guess we will need to blank at LEAST
during spark time, but other than that, maybe just leave it available so
dataloggers and custom ECUs can sample it with an A/D, like Saab's ECU
does. The second thing we almost certainly will need to do, for cases
like you've brought up, is to have a digital signal again with a window
set to the ranges that Saab discloses in their patent (which they apply
in their ECU, BTW, not in the electics in the DI unit, as far as I can
tell) for watching for anomalous IGN events. These windows would be
generated by one-shots off the IGN signal and probably don't amount to
much hardware, cuz we'd have to do them on ION's board or some other
board (assuming you can't/don't change your stock ECM's code). At this
point, I have to allow for the possibility that not just a simple
adjustable threshold, but some additional signal conditioning/filtering
might be necessary to give us a reliable digital knock signal, so there
needs to be provision for that as well. And what's required isn't maybe
so predictable. Add to that, the stuff you need to "mimic" your previous
probably piezoelectric conventional knock sensor, at the tail end, so
you can feed the resulting signal directly to the std. ecm.

If you can see the drift above, it's beginning to look alot like drawing
the line around EGOR turned out. I would think that again, with ION,
what would be MOST universal would be (if Sandy's still in harness and
we can talk him into another "wee module" board design) that you'd have
ION's fire/misfire signals and the 0-5V scaled ionization analog signals
coming out, and that's it. Again, judging from parts count so far, I'd
bet it would be smaller than a 0.6" 40-pin DIP, maybe even a 28-pin
assuming surface mount. Then each person, whether they were making a
simple KNOCKnMISFIRE "meter" or hooking it up along with everything else
to a datalogger, and logging even the ionization signal during a dyno
run, or like yourself, hooking it up to a std. ECM; WHATEVER the
particular usage, the ION module stays the same, plugs in to your own
particular interface board that contains the OTHER circuits you need to
accomplish your particular application. In the case of a std. ECM, this
would amount to the window generators for determining WHEN to look for
KNOCK, as well as threshold comparator and maybe filtering if needed,
and finally the mimicry stuff.

That would keep the basic circuit small and sorta "universal".

Mind you, until we try this stuff out on our conventional inductive IGN
systems, and play a little, we won't really know what makes sense, but
that's my first take, such as it is. HTH.

Gar
(hand me some more plasma, willya EGOR; I feel like gettin ioneyes'd  B)




More information about the Diy_efi mailing list