how does a knock sensor work?

[Mark Agnello]

Here is an explanation I wrote for another mailing
list.  It is written so folks without an engineering
background can understand it.  The description is of a
circa 1988 knock sensor design.  Apologies if this is
not as sophisticated as some of you would like.  

Piezoelectric knock sensors are also known 
as accelerometers. The sensor consists of a piezo 
crystal that is mounted in compression inside a rigid 
housing. The crystal produces a voltage that is 
proportional to the force transmitted to it through 
the housing. This voltage is constantly changing in 
response to engine rpm and other factors. On an 
oscilloscope, the knock sensor signal will look like 
the kind of random ac signal produced by recorded 
music. This means that many frequecies are contained 
in the voltage signal from the sensor. The ECU will 
filter this voltage signal to measure specific signal 
strength at frequencies that indicate engine knock. 
On aircraft engines, the ECU actually matches the 
filter to the engine RPM and its harmonics. Apparently
auto manufacturers use similar strategies.  Texas
Instruments has an excellent application note on their
web site that describes a fairly sophisticated signal
processing approach to knock detection.  

When a sensor fails, it is usually because the crystal

is damaged due to mechanical stress or too much time 
at elevated temperatures. Another major failure mode 
is the wire connecting the sensor to the ECU. 
Remember how I described the sensor as having an 
output voltage proportional to force? I lied. 
Actually, the sensor output is in units of electrical 
charge called coulombs. In practical terms, the 
electrical charge output looks like a very weak 
voltage. It's very sensitive to the load it is 
connected to. Accelerometer cables are specially 
designed to provide an interface to the sensor that 
minimizes resistance and capacitance. If the cable is 
damaged or deformed by mechanical stresses, it will 
alter the sensor signal received by the ECU. If the 
cable is allowed to move too much or whip around in 
response to engine vibration, it will induce a false 
voltage due to the changing electrical parameters of 
the cable. It is very possible that cable failure will
occur on a 12-year old vehicle with high 
underhood temperatures like the Toyota Celica GT-4. 
The cable failed on my 1988 model.  I replaced it with
Mil-spec teflon insulated wire and it has worked well.
 

Mark Agnello
Washington Crossing, PA (USA)
88 Toyota Celica All Trac Turbo (restoration project)
92 Toyota Celica All Trac Turbo
96 Honda VFR
89 Honda Hawk GT (restoration project)


__________________________________________________
Do You Yahoo!?
Make international calls for as low as $.04/minute with Yahoo! Messenger
http://phonecard.yahoo.com/
----------------------------------------------------------------------------
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