resistance between knock sensor harness and ground

[silverbox]

i am troubleshooting a check engine code #52, which is "knock control - malfunction" according to the internet or "no knock sensor signal" according to a chilton manual.

i removed the knock sensor, which is NOT THAT HARD, and i didn't have to remove the engine, or even cut up the floorboard, or even undo any motor mounts! ¡!MAZEL¡! i plan to document this soon.

with the sensor removed, i checked resistance between the single-wire knock sensor harness and ground, and measured a steady 3.3 kΩ. is this normal?

i will install the new sensor soon and see if the code goes away, but i also wanted to see if anybody knew about the expected resistance.

one other complication, my van was made before 8/1991, so it uses the 89615-12020 sensor, which is different from later previas, including the later 1991 ones.

thanks, and i'll post up the results too when the new sensor is in.

[silverbox]

i installed the new knock sensor (again, without cutting the floor or moving the engine) and then i cleared codes and drove a tiny bit before getting another code 52. so i will try to find a replacement wire harness for the knock sensor. looking for the part number for that harness for the early previas made before 8/1991. if i find it i'll edit this post

edit: i'll probably make my own harness and connect it directly to the ecu, bypassing the original wire. this is a low-current signal, susceptible to electrical noise, so i'll use shielded cable and find some kind of flexible heat-resistant conduit material to put it in. probably microphone cable, since it has an excellent braided jacket to shield the inner conductor from EMF, plus i have a lot of it. if anybody has a better idea, let me know.

still want to know if there's supposed to be any conductivity between the knock sensor wire and ground. mic cable will have basically zero resistance along the conductor, and infinite resistance from conductor to ground

[John Kaufmann]

Silverbox, my Previa, like yours is 91, and I did not know that the knock sensor (mine is 89615-12020, like yours) changed for later years. Still I would not expect that to be relevant to your question about resistance to ground at the harness, which surely should be in the range of megohms (probably tens of megohms), not 3.3K. Is that with the line connected to the ECU? If so, 3.3K sounds reasonable, because you are reading through the front-end signal conditioning circuitry (but I'm sure you knew that).

[silverbox]

yes it is connected to the ecu, i haven't actually measured it with the ecu disconnected, which would be a good idea. i imagined that this many ohms might be a feasible resistance to build into the ECU to convert a small current spike from the knock sensor to a nice digital high voltage signal to detect a knock.

really i haven't done a proper diagnosis of the wire yet, that is the sensible next step. i'll check resistance from one end of the conductor to the end with the ECU disconnected and also check resistance to ground with it disconnected. thanks for getting me thinking about that.

i did find some wire that might be a suitable replacement if needed: https://www.digikey.com/en/products/...603CY/12148479 it's a tiny 3' wire run and i could probably just ziptie this to the main harness, which runs above the engine to the driver's seat.

[silverbox]

update:

after disconnecting the ecu, i found that there is less than half an ohm resistance from one end of the single-wire knock sensor harness to the other end (good), and that the wire is an open circuit (infinite resistance) to ground. (also good, i assume)

it is possible that the grounded jacket (shield) of the knock sensor sub-harness has deteriorated over time, allowing EMF interference in and giving a bad signal, but i decided it was more likely that some grounding problem with the engine might be at fault- remember, it's a one-wire harness so the sensor is supposed to be grounded via the engine metal, which should be electrically tied to the chassis with grounding straps.

i found a grounding strap on the transmission (manual in my case), and one on the engine. i took them both off and refurbished all the surfaces involved, using sandpaper and stuff. shiny metal-on-metal contacts. also, i added another ground strap from the engine directly to the end of the negative battery cable, where it bolts onto some bracket. i also sanded the paint off of both sides of the motor mount bracket rod thingy that was part of the path to ground for one of those straps, it attaches to one of the motor mounts or transmission mounts whatever asldfjasdlfjkasdlfkj

in the meantime, my starter stopped working, and i think (gut feeling) that is really is the starter, not some other electrical issue. s i g h so i'll rebuild/replace/throwoffcliff that starter and get working on this knock sensor issue again.

that cable i was talking about came in + some automotive ecu pins that might be suitable, not exactly the same as factory but if i have to make my own replacement subharness for the toyota NLA one, i think i can. hopefully not necessary.

will post up more when i'm done defenestrating my starter and kicking various bumpers and things

[Jonny]

You probably know this, so this is just for general info, but surest way to test starter is set handbrake, tranny in park, (neutral, depress clutch if manual), turn ignition to on (not start), crawl under van, disconnect small wire from starter magnetic switch terminal and momentarily jump that terminal directly to the main terminal of the starter. This bypasses the ignition switch starting connection and the starter relay. Starter will operate if it is functional (unless the problem is within the positive wire from the battery, or grounding).
Edit: ignition on is for actually starting the car, leave ignition off for just testing.