knock sensor

[cohophysh]

90 leg 2.2, I just replaced the knock sensor, however, I did not put it back where it came. I found a bolt hole up and just towards the firewall that was easy access. So I mounted it there. Seems to me all the knock sensor needed was good contact to the block, so basically it could go anywhere the wire would reach. Is this right?

B

 

btw...knock sensor was in bad shape. split housing rusty...yuk

[Fairtax4me]

Yeah it's pretty common to relocate the knock sensor, I haven't heard of any ill effects from doing so.

[cohophysh]

here it is

[1 Lucky Texan]

pic of new location?

 

If I understand correctly, you could just add the new one to the new location and move the connector over right?

[porcupine73]

It can even be removed all together by replacing with about 560 kilo-ohm resistor. It's just then if there is knocking the ECU won't know about it so long term it could cause damage. I've done that on one of my soobs and it seems fine if premium fuel is run, but it will knock a bit at high load, low rpm's when it's hot, like going up a hill without downshifting.

[cohophysh]

I have someone telling me in another forum that the knock sensor shouldn't be moved because it won't sense knock as well, if at all and could cause damage. Any thoughts

[Cougar]

The location of the sensor on the engine could very well be critical to the sensor operation. The design engineers may have located it there for a reason and moving it may change the operation of the sensor.

[1 Lucky Texan]

good point.

 

if someone isn't planning on replacing it properly, cheaper to do the resister trick than buy a sensor and relocate it to where it's ineffective.

[cohophysh]

i will probably move it back where it belongs...

[Fairtax4me]

I'd hardly say relocating the sensor would make it ineffective.

All the sensor does is listen for a sound within a certain frequency range, and creates a small voltage pulse as the sensor element reacts to the vibration created by the sound. It's basically a microphone.

But it's listening to the block instead of open air. Sound travels much faster and much more efficiently through metal than it does through air. The sound that detonation (spark knock) makes even when it can't be heard by our ears, will still resonate through the entire structure of the block, and thus will be picked up by the sensor regardless of it's position.

 

If the sound of the knock is so small that a change in position of the sensor affects it's ability to react to that knock, the ECU would not pay attention to that knock anyway.

 

On a forced induction engine this may be a bit more critical, as even small knocks can mean catastrophic failure under high boost conditions. But for a normally aspirated engine it's not going to make a difference.

 

Also, if the sensor is replaced with a resistor, then you have absolutely NO reaction to knock at all. It will continue to knock and the ECU will NOT react to correct the knock because it doesn't know it's happening. With a relocated sensor, you still have the sensor. So when/if the engine begins to knock the ECU WILL react and will cut timing advance to try to stop it.

[jarl]

Isn't the knock sensor supposed to detect detonation on ANY cylinder? If so, the distance to the cylinders is not that critical (two of the cylinders are relatively far from the sensor), and using the alternate mounting hole should not be a biggie.

[presslab]

If you realize the work the design engineers have done to filter out resonances and amplify specific frequencies, you'd see that the location of the sensor is quite integral to it's design. "Listening" for engine knock is very tricky stuff, and the knock sensor will "hear" knock well before you or I can hear it with our ears.

 

I'd say leave it where it's meant to be, or else you will either cause the ECU to "hear" false knock, or the ECU won't be able to pick up knock until it's too loud and causing engine damage. The torque applied to the bolt is also important, so check your spec's.

 

In any case, is the risk of engine damage worth the time saved by not bolting it in the right location?

 

Some interesting reading:

http://www.diagnosticnews.com/tech/engine-knock-sensors-part-1/

http://www.diagnosticnews.com/tech/engine-knock-sensors-part-2/

 

And some nuts-and-bolts of knock detection implementation in a DSP:

http://www.ti.com/lit/an/spra039/spra039.pdf

the structural design of an engine and the mounting location of the knock sensor(s) affect which frequency modes are detectable by the sensor. Usually, the transfer function between the cylinder and the sensor is different for each cylinder. This causes both the relative and absolute magnitudes of the vibrational modes to be different for each cylinder. A good detection scheme should allow different calibrations for each cylinder.

[1 Lucky Texan]

nvrmd - well covered above

[jarl]

Presslab: interesting reading... goes straight to my tech doc "library"

 

There is a very good chance, however, that the resonant points/etc. were not the main issues when the position of the knock sensor was determined (what happens if the optimal position ends up being right under the intake manifold, as it probably is?). Replacing some components (i.e. gaskets), the tension on the bolts, etc. have some impact in the resonant frequencies of the engine block, and as such there's got to be some flexibility in the whole design, including the location of the sensor. It would be very nice to see an oscilloscope reading of pinging as seen from a sensor placed on either position, but I think whatever damping there might be between the two locations is probably not enough to reduce the amplitude in a meaningful way (i.e. compared to an old sensor mounted with a corroded bolt).

 

The short of it: if you can put the sensor back where it should go, by all means do it. If you can't, don't sweat it too much

[1 Lucky Texan]

acoustic coupling and damping is crazy stuff. I worked on some acoustic downhole instruments for a while (as a tech) in the wireline biz. We had to put some crazy bends in some tubing, add steel/lead cylinders of a certain dimension, couple things together with rubber bushing between everything - slide all that into a steel sleeve with slots cut out of it, all in an attempt to slow down and suppress the sound of the xmitter coming down the through the tool itself so formation travel time could be measured. fun stuff.

 

The point made about about the frequency being detected is good, as is concerns about mods/changes to the engi ne configuration.

 

With my luck though, if I moved it, the a/c clutch or some clunky brake/clutch linkage, or a relay would probably create false knock impulses. Or some worn out acc. belt bearing would mask the real knock sound or otherwise work to my disadvantage.

 

I'd put it back where 'they' had it.

[cohophysh]

okay, I put it back:clap:

[Quidam]

If you realize the work the design engineers have done to filter out resonances and amplify specific frequencies, you'd see that the location of the sensor is quite integral to it's design. "Listening" for engine knock is very tricky stuff, and the knock sensor will "hear" knock well before you or I can hear it with our ears.

 

I'd say leave it where it's meant to be, or else you will either cause the ECU to "hear" false knock, or the ECU won't be able to pick up knock until it's too loud and causing engine damage. The torque applied to the bolt is also important, so check your spec's.

 

In any case, is the risk of engine damage worth the time saved by not bolting it in the right location?

 

Some interesting reading:

http://www.diagnosticnews.com/tech/engine-knock-sensors-part-1/

http://www.diagnosticnews.com/tech/engine-knock-sensors-part-2/

 

And some nuts-and-bolts of knock detection implementation in a DSP:

http://www.ti.com/lit/an/spra039/spra039.pdf

 

Some interesting reading there, thanks for the TI link.

 

I'd say Subaru put the sensor close to the cylinder most likely to ping. On an old small block Chevy, that's cylinder 7 and 8. So, they put the knock sensor near the starter on those cylinders.

 

Doug

[cohophysh]

I miss the CEL coming on...NOT!