"Andrew Stephenson" <ames@deltrak.demon.co.uk> wrote in message
news:1165689891snz@deltrak.demon.co.uk...[color=blue]
> In article <d3fdd$457af3d6$47c2b532$14646@msgid.meganewsservers.com>
> rokigawaATtristarassociatesDOTcom "Ray O" writes:
>[color=green]
>> The other thing that has me puzzled is the nature of the knock
>> sensor. Toyota uses a piezo-electric crystal inside the knock
>> sensor, and the crystal is struck or twisted, a static spark is
>> emitted. I don't know much about those crystals, but I can't
>> imagine how they can send a signal without a physical force,
>> i.e., knocking, exerted on them. [...][/color]
>
>
AFAIK, you are correct. The mechanical energy of deformation
> becomes electrical energy. No change in deformation, no spark.
>
> Possible useful data: IIRC, the source impedance of the piezo/e
> crystal is very high. A large serial resistance could probably
> be tolerated (guess: several tens of kilo-ohms); but it wouldn't
> take much shunt leakage (ie, across the wires carrying the pulse
> away) for the pulse to be crippled. Think of a battery with high
> internal resistance -- but _really_ high.
>
> Excessive shunt capacitance would also slug the pulse -- though I
> don't see how anyone could _accidentally_ introduce that into an
> established design. OTOH, if there were significant extra serial
> resistance, such from a manky connection, less shunt capacitance
> could slug the pulse. Shunt capacitance could be caused by using
> a non-standard cable, especially screened and excessively long.
> NB: This paragraph is speculative; it may also contain nuts.
> --
> Andrew Stephenson
>[/color]
This info seems to make it even less likely that the bad knock sensor would
make the ECM retard timing. In other words, a bad knock sensor is more
likely to keep timing too advanced, not too retarded.
--
Ray O
(correct punctuation to reply)