[s-cars] Knock sensor failure?

[MLP]

As I recall, in the SAAB APC conversion / adaptation to other vehicles
posts there is some information about the difficulty one can experience
in trying to adapt the system, and the sensors to another vehicle.  One
check was to be sure the sensor is properly secured to the block.  By
properly secured as I recall, in addition to location, the sensors could
be somewhat finicky about being squeezed ~ not too loose, not too tight.

While I don't recall any Audi lister ever posting "My sensors got
loose...." in the Audi spirit of "My plugs keep moving, should I seize
or anti seize etc....,"  I guess it couldn't hurt to check them.  OTOH,
my guess would be this might be one part of the Audi system that doesn't
"fail safe" (but I really don't about know this) i.e., if one or more
knock sensors stopped working, wouldn't you get no signal at all,
therefore no knock sensor protection, or do the sensors by default put
out a signal which, if absent or "interrupted" begins a progressive
timing retard, go safe mode?  

Some one (Rod H.) on the Krummer list questioned the value of a knock
signal LED on the dash.  I suppose this would be a case where that could
prove helpful.  If you were getting regular knock sensor activation LED
activity while under little or no boost.....?  Last, does the VAG store
knock sensor interrupt codes?  Would something like Shadetree Software's
Palm OS allow running & monitoring the older system in real time to
watch for KS activity? 

mike

>-----Original Message-----
>From: s-car-list On Behalf Of Robert Myers

>Hi Y'all,
>
>I've been sitting here thinking...  (Miracles do sometimes happen.)
>
>I'm sure that knock sensors have failed from time to time but I would 
>expect that the normal failure mode would be that it simply 
>stop putting 
>out a signal.  However, has anyone experienced a knock sensor 
>failure in 
>which the knock sensor becomes overly sensitive and reports 
>knocking to the 
>ECU when none is actually happening?  Is there any direct way 
>to determine 
>if this might be the case or would it be just a "replace and 
>try" procedure.
>
>_____
>
>When Clinton lied no one died.
>
>Bob
>*****
>  Robert L. Myers  304-574-2372
>  Rt. 4, Box 57, Fayetteville, WV 25840 USA  WV tag Q SHIP
>  '95 urS6 Cashmere Grey - der Wunderwagen   ICQ 22170244
>  http://www.cob-net.org/church/pvcob.htm
>***** 
>