replacing a knock sensor

P

pinoyGT4

New member
BriinumsBo":2pzq1hna said:
yes, its 1-wire system, it should go to the ecu and grounding is via the block. maybe add some grounding wire from battery to the block to see if that helps.


i got a weird problem. i have some metallic vibration (havent checked yet, its freezing outside :D) at 1700-1800rpm.. at that point the knock sensor throws up the CEL :( i dont mind the safe mode, but i do care for fuel economy :D
vid:
http://www.youtube.com/watch?v=kHeSRCcBKPg
Click to expand...


aryt ill check on that one tom.. :D

hope it gets my KS prob done.. :D
 
T

timco

New member
Where does this wire plug into the at? I'm more than a noob with the wiring/ecu. With it being such a tight fit how do you manage to run the wire? Where do I plug it in? Just ordered a new KS(the ats gm one) because just recently ive been having this same issue, same codes as well. My last option(as far as i can tell) is to replace the KS and/or cable. Just wondering how difficult it is to locate the correct cable to replace or splice it. thanks
kris

edit: holy crap old thread
 
$

$1200alltrac

New member
Hey, I had the same problem. But I pulled my ks out and cleanedbthe flashing off of it, and it stayed off for like a day, now come back at 1800 rpm, and sometimes 2800 rpm. Then sometimes doesn't come on at all, what should I do? I think its the ks its self since I cleaned it and no more cel cam on for a day
 
C

ccsouls

New member
Did anyone ever successfully correct the CEL knock sensor issue at 1700-1800 RPMs?

If so, please post the solution.

I have the same problem now and have been chasing my tail for 3 months.

Thank you.
 
L

l0ch0w

New member
ccsouls":3g1hvlb2 said:
Did anyone ever successfully correct the CEL knock sensor issue at 1700-1800 RPMs?

If so, please post the solution.

I have the same problem now and have been chasing my tail for 3 months.

Thank you.
Click to expand...

I have heard of a few people fixing this issue by completely replacing the wire that leads to the KS. What usually happens is that the shielding on the wire becomes compromised or the internal wire has a short somewhere in it that causes it to ground out to the shielding.

I can make a small bypass harness for you if you send me your old KS connector. What is important is that you use well shielded wire (i use RG316U) and that you DO NOT splice the wires. Wire splices add mass to the unsheilded sections of wire and can act as antennas for EMI which will throw off your KS reading. I only use the OEM crimp at the ECU terminal and I solder the new wire to the old terminal in the KS connector if it is the round style, or I use a new terminal in the case of the square type plugs in the newer KS's. Here is what the patch harness looks like:

patchharness.jpg

By alochow at 2011-11-14

Basically all you do is remove the ECU pin terminal from the ECU for the Knock signal, cut a tiny hole in the harness boot on your firewall and feed it through. The new pin goes into the ECU, the grounding terminal goes to any bolt that has a good ground (IG ground is reccomended) and the connector plugs into the KS. Simple enough :)

I recently made the one pictured for $1200alltrac, and am waiting on him to finish up his clutch replacement to hear back as to weather or not this fixes his problem. I had him ohm out his old KS wiring and it appeared to be the wire. In which case this should solve his problem. If it does turn out to be the case, Ill be making a few more of these to sell.

PM me if you would like a quote.
 
T

tubasteve

New member
I would like to majorly chime in on this!!!

I hah this dreded code after getting some gas here about a year ago.(nothing to do with the gas station, but I did go back and ask...)

After doing research on here for a while I noticed it was a normal problem. People using the 97 sunfire knock sensor for testing but stopped. But the main issue I was having it seemed like it was the dreaded bad harness, id like to say a "sheilded" does give the pcm a better signal/more pure this is not needed for testing it seems for the quick drives I was doing.

I ordered through my parts house a new knock sensor....4 actually when all said and done....anyone know what list is...well I paid only a little under that for the knock sensor. Just like the other fourm says its like a. "Poor mans fuse". Well sadly when the car would hit this fuel cut due to a bad knock sensor harness it was actually also taking out the knock sensor I found out during my testing. I used a wire externaly ran so I would not waste anymore time with the damn harness. Found the sheild was actualy touching the (ks) wire right at the knock sensor due to bad crimp and the oil that has broken down the knock sensor. I bought the last sensor id ever buy from toyota and sat down with my trusted father to figure out what type of frequency the sensor was putting out and what range it worked at when mounted to the engine. The sensor generates an a/c signal and the computer looks at the certin ranges in which the changes of pre-ignition. Finding the sunfire one works very well within the same range, the engine seemed to respond better and still react damn near a little better than the stock sensor and not blow when a episode of code 52 from too much boost/I did force the code while on the rack with both sensors in. I was pulling the current motor I was doing all the testing on so it really didn't matter to me, I just wanted to fix my most expensive vital sensor the car needs in my eyes. From my parts house the(sunfire97}} sensor was about 20 or so and the harness from GM was roughly I think lie 12 maybe. I've ran this sensor for almost 3 years on the engine before actually pulling the engine to do a freshen up on here. I did not need to run sheilded wire for my testing, I did moniter the entire time when running between the two sensors, and from the other research I beliveve this is the way I will go from here on. Since putting the sunfire ks on I've had zero problems, and I'm never buying a 150$ sensor that weak again.

There's nothing wrong going to stock, but when the sensor:
1) same thread pitch/goes in to the block
2) is that cheap.
3) won't break when you screw something up....unless you are dumb and don't inspect your car when getting the code 52 for safe mode
4) well I have too many reasons why. I like using factory parts as offeten as possible, but one that breaks that easy....

I do have every sensor moniter and data logged on that car and the data shows good, and the car has benn a dream, I belive the sensor responds faster as well.

I will further say on my other 100% stock, well, I do have the factory snesor......
 
C

ccsouls

New member
tubasteve":2bbzi97l said:
I would like to majorly chime in on this!!!

I hah this dreded code after getting some gas here about a year ago.(nothing to do with the gas station, but I did go back and ask...)

After doing research on here for a while I noticed it was a normal problem. People using the 97 sunfire knock sensor for testing but stopped. But the main issue I was having it seemed like it was the dreaded bad harness, id like to say a "sheilded" does give the pcm a better signal/more pure this is not needed for testing it seems for the quick drives I was doing.

I ordered through my parts house a new knock sensor....4 actually when all said and done....anyone know what list is...well I paid only a little under that for the knock sensor. Just like the other fourm says its like a. "Poor mans fuse". Well sadly when the car would hit this fuel cut due to a bad knock sensor harness it was actually also taking out the knock sensor I found out during my testing. I used a wire externaly ran so I would not waste anymore time with the damn harness. Found the sheild was actualy touching the (ks) wire right at the knock sensor due to bad crimp and the oil that has broken down the knock sensor. I bought the last sensor id ever buy from toyota and sat down with my trusted father to figure out what type of frequency the sensor was putting out and what range it worked at when mounted to the engine. The sensor generates an a/c signal and the computer looks at the certin ranges in which the changes of pre-ignition. Finding the sunfire one works very well within the same range, the engine seemed to respond better and still react damn near a little better than the stock sensor and not blow when a episode of code 52 from too much boost/I did force the code while on the rack with both sensors in. I was pulling the current motor I was doing all the testing on so it really didn't matter to me, I just wanted to fix my most expensive vital sensor the car needs in my eyes. From my parts house the(sunfire97}} sensor was about 20 or so and the harness from GM was roughly I think lie 12 maybe. I've ran this sensor for almost 3 years on the engine before actually pulling the engine to do a freshen up on here. I did not need to run sheilded wire for my testing, I did moniter the entire time when running between the two sensors, and from the other research I beliveve this is the way I will go from here on. Since putting the sunfire ks on I've had zero problems, and I'm never buying a 150$ sensor that weak again.

There's nothing wrong going to stock, but when the sensor:
1) same thread pitch/goes in to the block
2) is that cheap.
3) won't break when you screw something up....unless you are dumb and don't inspect your car when getting the code 52 for safe mode
4) well I have too many reasons why. I like using factory parts as offeten as possible, but one that breaks that easy....

I do have every sensor moniter and data logged on that car and the data shows good, and the car has benn a dream, I belive the sensor responds faster as well.

I will further say on my other 100% stock, well, I do have the factory snesor......
Click to expand...

Thank you for this post.

I too have had the same issue Autocrossing my car. If I get into high boost sometimes it will blow my knock sensor as well.

After MONTHS of diagnostic testing & frustration...and after people telling me knock sensors don't "blow" and I had to find a problem....I finally realized the Toyota knock sensors must blow somehow.

I had not heard of the sunfire replacement....although many have written the GM knock sensor replacement was a bad idea because the GM sensor checked for a range dissimilar from the Toyota. I assume your testing of the sunfire knock sensor uses a similar frequency as the Toyota and that is helpful. (Especially if the cost is as low as you describe.)

I appreciate the upate.
 
T

tubasteve

New member
It is the same....if it did not produce the same range using the 2.2/2.4 knock sensor from a cavi same vintage it would work for two pulls before throwing the 52....and was close to the same...again %close% to the same frequency.
-jr
 
G

gearhead313

New member
Did the OP ever get the sensor changed? Just curious...




staying on the offtopic... I think overboosting causing the knock sensor to blow out is peculiar. The only thing that makes sense is if the knock (ping) is so bad that the KS generates a signal that the sensor itself cannot handle for such a period of time and rate. If this is true, then it makes sense. As far as the "frequency" of it, it does depend on the engine characteristics.

Also, remember that a CEL light doesnt mean the engine is knocking, its just the circuit malfunctioning! In all my alltrac experience, replacing the wire with a new, shielded, direct shot to the ecu fixes the problem. I've had good luck with the GM sensor not failing on me, but who knows if its ACTUALLY detecting any knock. I dont have a way to monitor it right now. Ultimately, the vehicle should very rarely if ever knock in stock to mildly upgraded form.



This is a good read:

http://www.deviantmethods.com/bigmoose/pages/knock.htm
 
T

tubasteve

New member
The car I was testing was on my dd engine.also it does detect knock.the sensor generated a signal to keep computer happy.

Mine was blowing cause of bad wiring. However, when code 52 during heavy accel would cause a boost spike and not all the time, but would blow the sensor. I wouldn't run past 12 lbs, or more timing. I don't use cheap or crappy fuels. Its been a while since I did the testing...I'll see what data sheetsI still have.
, jr
 
T

tubasteve

New member
If like to point out...not same knock sensor for one, and also its using hondas knock....hondas mount the sensor 1" lower than the toyota.

I would test a • New sensor as well. The crystals inside the knock sensors ware out...all my test were brand new factory. And I do agree. Not all knock sensors (like napa to gm) are the same.
 
C

ccsouls

New member
tubasteve":1t51idwz said:
The car I was testing was on my dd engine.also it does detect knock.the sensor generated a signal to keep computer happy.

Mine was blowing cause of bad wiring. However, when code 52 during heavy accel would cause a boost spike and not all the time, but would blow the sensor. I wouldn't run past 12 lbs, or more timing. I don't use cheap or crappy fuels. Its been a while since I did the testing...I'll see what data sheetsI still have.
, jr
Click to expand...

Keep in mind if boost goes up above...17-19lbs(maybe less)....if you don't have race fuel....that will blow the sensor too. This I know for sure.

At least in tests run on MR2 3SGTEs.

You have to have an ECU tune for boost levels above 15lbs. to adjust for that.
 
You must log in or register to reply here.
Top