Page 1 of 1
knock sensor
Posted: Thu Dec 03, 2009 1:41 am
by vtxrider
cel code 22 knock sensor on 94 turbo 5speed. Have replaced the knock sensor with used ones and then finally with a new one. Have put in a ecu from a 92 turbo 5 speed and still have the cel with a 22.
Posted: Thu Dec 03, 2009 3:10 am
by SubiePrice88
Have you cleared the codes?
Search and you will find! I just dont have alot of time right now to explain.
Posted: Thu Dec 03, 2009 5:41 am
by vtxrider
Yes I have cleared the codes many times.
Posted: Thu Dec 03, 2009 5:42 am
by ericem
My bet is bad knock sensor wire. Might need to splice the wire back in order for it to work.
Posted: Fri Dec 04, 2009 3:57 am
by vtxrider
It could be wiring seeing that the engine compartment of a turbo gets hot. This is 16 years old and the wiring is brittle . It has been spliced but only back to the shielded cable. Can the shielded cable be take off farther back with out hurting it and the wire splice farther back?
Posted: Mon Dec 07, 2009 3:55 am
by Legacy777
The wire is a solid core wire. You'll probably be better off just running a new wire. You'll be fighting the stock one I think...
Posted: Mon Dec 07, 2009 9:20 am
by Dynamic Entry
Dumb question alert:
Does having a code 22 you can not get rid of NECESSARILY mean that the knock sensor is not doing it's job?
To clarify my meaning, consider that some times the component a sensor is monitoring can be in good order, and the only problem present is the MONITORING of that component, not the component itself. For instance, a coolant sensor float could break thus eliminating the ability to properly monitor coolant level and yet the coolant level could be fine.
So could a bad knock wire simply disallow the system to know if the knock sensor is working, or will it necessarily keep the knock sensor from working.
Oof.
Posted: Mon Dec 07, 2009 6:01 pm
by Legacy777
If the signal to the knock sensor is not there, the ECU goes into limp mode since it is not receiving that information.
If it's receiving the information, everything's working and you don't have a code.
I guess I may not understand the question....
Posted: Mon Dec 07, 2009 6:05 pm
by tahiti350
If the signal form the sensor is weak or no existant then the computer reads that as out of tolerance and sets the code. even when there is no knock there is still continuity on the wire that the computer should be reading.
So, to answer your question yes the wire can and does cuase the CEL for a knock code, even if there is no knock and a good sensor.
Other sensors and systems may or may not provide the same feedback, such as a failing MAF...
Posted: Tue Dec 08, 2009 7:56 am
by Dynamic Entry
I think you guys answered my question, which to rephrase would be put:
Can a motor without problematic knock and a working knock sensor still run badly because of a poor knock sensor cable? (as opposed to merely showing the CEL without any adverse affect to the motors power output)
And since you said the computer will go into limp mode without signal, the answer seems to be yes.
Posted: Tue Dec 08, 2009 7:16 pm
by Legacy777
You are correct.

Posted: Tue Dec 08, 2009 9:09 pm
by Dynamic Entry
but, if the cable IS problematic, then the CEL should light up ya?
I have a code 22 stored in my computer, that doesn't clear, but I never have a CEL...
Posted: Wed Dec 09, 2009 2:31 am
by Legacy777
Not necessarily.....it should show a CEL, but the ECU's sometimes don't.
Posted: Wed Dec 09, 2009 3:51 am
by magicmike
if the cable is somehow shorting to ground like at the bracket where the connector mounts to the back of the motor then perhaps.
Posted: Wed Dec 09, 2009 7:42 am
by BSOD2600
Run through the FSM knock sensor diagnostics. If all checks out, then the should be ok.

Posted: Wed Dec 09, 2009 9:15 am
by Dynamic Entry
my multimeter is effed, this is a good reason to get a new one
thanks for the replies guys, I need to get on this sometime. too bad I'm Cpt. Lazyass McPants
Posted: Wed Dec 09, 2009 4:26 pm
by magicmike
They get too technical with this garbage. The knock sensor is just a shock sensor basically.
Picture two cylinders with a very small gap between them. The inner cylinder is chassis ground (you can see it when you bolt the sensor to the block). That should be clean.
The outer cylinder is what the wire is connected to that goes to the ECU. The outer cylinder "floats" very close the inner cylinder.
If the engone shakes violently the two cylinders touch thus making the sensor wire go to ground and triggering the code.
got it?
Posted: Wed Dec 09, 2009 9:08 pm
by Dynamic Entry
good to know the theory behind it, I always like to understand the why, so thanks
but really my main concern is making my legacy less tempermental and more consistent, so, I'll check the function of the sensor/wire and see if it needs attention
Re: knock sensor
Posted: Thu Dec 10, 2009 7:27 am
by 93Leg-c
vtxrider wrote:cel code 22 knock sensor on 94 turbo 5speed. Have replaced the knock sensor with used ones and then finally with a new one. Have put in a ecu from a 92 turbo 5 speed and still have the cel with a 22.
You may also want to check the angle that the knock sensor is installed.
Some years ago I had a pinging and knocking problem with a new knock sensor. When I checked the FSM instructions, it says to install the knock sensor with the wire pointing at a 45 degree angle (if the passenger compartment direction is 12:00 and the direction toward the front of the car is the 6:00 position, the knock sensor wire should be pointing towards the 10:30 position).
That ended my pinging and knocking episodes along with the code 22 and CEL.
I don't know the technical reason why it should be installed that way but it solved the issue.
Re: knock sensor
Posted: Thu Dec 10, 2009 8:47 am
by BSOD2600
93Leg-c wrote:You may also want to check the angle that the knock sensor is installed.
Some years ago I had a pinging and knocking problem with a new knock sensor. When I checked the FSM instructions, it says to install the knock sensor with the wire pointing at a 45 degree angle (if the passenger compartment direction is 12:00 and the direction toward the front of the car is the 6:00 position, the knock sensor wire should be pointing towards the 10:30 position).
I've heard of this too, but never actually found a FSM scan which contained these instructions. Do you have a link to it per chance?
Re: knock sensor
Posted: Thu Dec 17, 2009 7:34 pm
by 93Leg-c
BSOD2600 wrote:93Leg-c wrote:You may also want to check the angle that the knock sensor is installed.
Some years ago I had a pinging and knocking problem with a new knock sensor. When I checked the FSM instructions, it says to install the knock sensor with the wire pointing at a 45 degree angle (if the passenger compartment direction is 12:00 and the direction toward the front of the car is the 6:00 position, the knock sensor wire should be pointing towards the 10:30 position).
I've heard of this too, but never actually found a FSM scan which contained these instructions. Do you have a link to it per chance?
No, sorry, I don't have a link to it. But I'll see if I can find it online somewhere. And my scanner broke some time ago...
Posted: Sun Dec 20, 2009 6:19 am
by log1call
The knock sensor is concentric and it doesn't make any difference to it's operation which way the wire comes out other than it needs to be secure so the wire doesn't shake around and, not be too tight.
Also, the sensor works on a piezo principle and does not earth or make a connection. It generates a small voltage.
If the sensor is overtightened to the block it can cause too much signal which can trigger a trouble code.
Bad fuel, heads that have been planned too many times, over-heating, oil burning, mechanical noises... they can all trigger knock detection and if the ecu can not stop the knock after it has retarded the timing as far as it's allowed(which is about twelve degrees.. varies a bit), then the knock sensor code gets set.
Posted: Thu Jan 07, 2010 5:08 am
by martimus919
Fixed it! when I tried soldering the bad wire before I had real crappy solder and gun, this time i spent more time on it and got the proper solder.