GM knock sensor may be bad. R&D here
by DatSRBoi » Wed Feb 06, 2008 8:27 pm
Dawgz says hi. Aim him up and tell him you love him.
Anywho. Ever wonder why the GM knock sensor "PERFORMS" So well that it never blows?
well how come everyone can blow a stock knock sensor, throw a GM one in and their problems go away? obviously the knock didnt just magicly leave when you stuck the adapter+GM sensor in.."
So Grumpy SolaraGuy did his homework and search because Dawgz told him to.
So what kind of test can we design to determine whether the GM sensor(s) are compatible within original Toyota specifications.Well Replicate the the same frequency to compare the two
The GM sensor is Pstamped as P.N. 10456287
The Toyota sensor is from a Gen2 JDM 3SGE (not a 3SGTE)
The phase inflexion of the GM part was around 5.6kHz
The phase inflexion of the Toyota part was around 6.3kHz
The AC impedance of the GM part at 5.6kHz was around 410ohms
The AC impedance of the Toyota part at 6.3kHz was around 350ohms
The DC resisance of the Toyota part was infinite (AC-coupled piezo)
The DC resistance of my GM part was ...100k?
S%*T!!!
With the sensors hooked up to a fast sampling 'scope, the Toyota unit would give a nice resonant sine wave around 6.9kHz, starting at about 1.4Vpp, then tapering out over about 100 cycles.
On the other hand, the GM sensor only gave a spiky nonresonant pulse of very high amplitude... makes me think its busticated now...
Which also means the GM sensor can not pick up a lot of stuff your motor is giving off.
Toyota Sensor
GM Sensor
These response plots are courtesy of 510rob who ran them on the spectrum analyzer.
After digging around a little bit more, it turns out that the detonation frequency of an engine is a function of the cylinder bore. The equation for determining the frequency of detonation in KHz is approximately 573/cylinder bore in mm. This works out to a frequency of 6.7 KHz for a stock 3SGTE with an 86mm bore, which agrees very closely with the response plot 510rob produced for the Toyota sensor. The GM sensor that 510rob tested appears to be tuned for an engine with a bore of 100mm. This leads me to believe that the GM sensor tested is NOT going to hear knock on a 3SGTE given the sharp frequency filter applied to tune out extraneous engine noise.
What does this all leads too? BOOM
For those still struggling to understand what the deal is, here is another way I can explain it: A detonation event is similar to hitting the block with a hammer. The block is a cast piece of iron that acts much like a bell. The bigger the bell, the lower the frequency of the sound that it makes (compare the note of a big church bell with that of a tiny one mounted above a shop door to announce that someone has entered). The GM sensor that folks have been using is tuned to hear only the note of a larger bell while the Toyota sensor is tuned to hear only the note of the smaller bell.
A say from ATS
Repeating one line from the qoute We have NEVER sold it as a cure for knock, only as an alternative to expensive OEM sensors
A few years ago, a GM knock sensor guy sent me a list of the most commonly used sensors. At the time, they were only making them in three frequencies to cover all their engines. The frequencies were 5.2 kHz, 6.0 kHz, and 7.0 kHz.
We aren't using the 7.0 kHz. I know of two 7 kHz GM sensors. One is from the Saturn, part number 21021879, with 10 x 1.5 mm threads. The other is from the Fiero, part number 10497163, with 1/4 tapered pipe threads. Our sensor kit are not any of those.
The difference between the GM sensor and the OEM sensor is that the OEM sensor forces you to discontinue whatever driving conditions resulted in the knock condition, because it breaks, and sends the car into limp mode.
The GM sensor will NOT break, so if you keep causing the knock, you will break your motor.
The strategy for the OEM sensor assumes that a child is driving the car.
Don't run the GM sensor unless you are a responsible adult, that is capable of making a rational decision to stop doing whatever is causing the knock response.
Dat----- So to sum it up. The GM sensor we have on our solara does its job yes. BUT it does not pick up signal pass a certain "knock". Because of that we are not seeing the almighty GM sensor breaking on the solara under boost. That is why they last so long yet on our SAFC they are going nuts. The solara ECU DOES NOT SEE THIS KNOCK because the GM sensor are not giving off that signal to the ECU. So you think your car is doing good, when its actually knocking behind your back.
So who ever said long ago the GM sensor is more sensitive.. Wrong. I was wrong. I search and this is our results. GM sensor is only meant as a cheap replacement, not a fix it.
Thanks, have a nice day.
Anywho. Ever wonder why the GM knock sensor "PERFORMS" So well that it never blows?
From a MR2 member word made a point. "
when you blow a stock knock sensor.. somthing is wrong with your car.. your knockin... right..
well how come everyone can blow a stock knock sensor, throw a GM one in and their problems go away? obviously the knock didnt just magicly leave when you stuck the adapter+GM sensor in.."
So Grumpy SolaraGuy did his homework and search because Dawgz told him to.
So what kind of test can we design to determine whether the GM sensor(s) are compatible within original Toyota specifications.Well Replicate the the same frequency to compare the two
I have the GM knock sensor in my car currently and I'm monitoring it with the TEC3 and in my datalogs I can see knock way before I can hear it. Looking through the datalogs it is showing up at what I would consider to be the common knock areas. High load mid to low RPM. But with my tune as I get to about 5000 rpm the knock will start to fall off to almost nothing. This whole time I've never actually heard anything from the engine.
The GM sensor is Pstamped as P.N. 10456287
The Toyota sensor is from a Gen2 JDM 3SGE (not a 3SGTE)
The phase inflexion of the GM part was around 5.6kHz
The phase inflexion of the Toyota part was around 6.3kHz
The AC impedance of the GM part at 5.6kHz was around 410ohms
The AC impedance of the Toyota part at 6.3kHz was around 350ohms
The DC resisance of the Toyota part was infinite (AC-coupled piezo)
The DC resistance of my GM part was ...100k?
S%*T!!!
With the sensors hooked up to a fast sampling 'scope, the Toyota unit would give a nice resonant sine wave around 6.9kHz, starting at about 1.4Vpp, then tapering out over about 100 cycles.
On the other hand, the GM sensor only gave a spiky nonresonant pulse of very high amplitude... makes me think its busticated now...
Which also means the GM sensor can not pick up a lot of stuff your motor is giving off.
Toyota Sensor
GM Sensor
These response plots are courtesy of 510rob who ran them on the spectrum analyzer.
After digging around a little bit more, it turns out that the detonation frequency of an engine is a function of the cylinder bore. The equation for determining the frequency of detonation in KHz is approximately 573/cylinder bore in mm. This works out to a frequency of 6.7 KHz for a stock 3SGTE with an 86mm bore, which agrees very closely with the response plot 510rob produced for the Toyota sensor. The GM sensor that 510rob tested appears to be tuned for an engine with a bore of 100mm. This leads me to believe that the GM sensor tested is NOT going to hear knock on a 3SGTE given the sharp frequency filter applied to tune out extraneous engine noise.
What does this all leads too? BOOM
Back in '99 and '00 when I was venturing out into the unknown with making good power on the CT26, I did all the wrong stuff; replacing my toyota sensor with a GM one, and using a AFC to control 550s. I went through 3 motors from detonation problems. I returned to a toyota sensor in 2001, and my motor survived until I broke a Ferrea inner valve spring from improper machining.
I dug around the 'net trying to figure out the differences in between the 2 sensors, and the biggest difference was in bore size for the 2 engines.
For those that do not live and breathe tech speak, think of it this way: We all know that there is a "speed of sound". Sound travel is not instantaneous. We have all seen firework shows, and noticed that the "burst" of the fireworks in the sky can be seen visually before the "boom" happens, just as we have seen the flash of lightning before we hear the thunder.
A knock sensor is made in such a way for each engine so that the sound from a detonation event is picked up by the sensor based on a frequency. Consider that the knock sensor is "tuned" just like a guitar string. Most musical people can recognize when a guitar is out of tune, but an inappropriate knock sensor is looking for a narrow range of frequencies and cannot tell if it is out of tune. The wrong knock sensor is much like that annoying tone deaf drunk at karaoke that sings a song so poorly that you want to retch. If the frequency occurs outside of the listening range of the sensor, then the "sound" falls on deaf ears (as Ricky said), and the ECU does not retard timing to try and keep the motor intact. We lose ring lands and pop head gaskets!
Bottom line, please address your knock issues before throwing in a GM sensor and adapter. Your motor will live a longer life.
For those still struggling to understand what the deal is, here is another way I can explain it: A detonation event is similar to hitting the block with a hammer. The block is a cast piece of iron that acts much like a bell. The bigger the bell, the lower the frequency of the sound that it makes (compare the note of a big church bell with that of a tiny one mounted above a shop door to announce that someone has entered). The GM sensor that folks have been using is tuned to hear only the note of a larger bell while the Toyota sensor is tuned to hear only the note of the smaller bell.
A say from ATS
Sometime between OEM sensor #1 and OEM sensor #3 is when I installed the MSD knock alert. It comes with a GM sensor and (crappy) adaptor. I originally ran it off the GM sensor, but swithced to the OEM sensor to see if it read any differently (it didn't). The night I popped GM sensor #3 I thought, hmm if the MSD knock alert reads the same on the OEM sensor as it does on the GM sensor I wonder if they will interchange? So I crawled under the car and swapped the wires around. Reset the ECU and fired it up, no more code 52.
Anyway, ATS will continue to sell the GM sensor as an alternative to the OEM sensor. We have NEVER sold it as a cure for knock, only as an alternative to expensive OEM sensors. I have had many, many conversations with customers where I try to get them to grasp the concept that the cause of the detonation has to be found.
Aaron
Repeating one line from the qoute We have NEVER sold it as a cure for knock, only as an alternative to expensive OEM sensors
A few years ago, a GM knock sensor guy sent me a list of the most commonly used sensors. At the time, they were only making them in three frequencies to cover all their engines. The frequencies were 5.2 kHz, 6.0 kHz, and 7.0 kHz.
We aren't using the 7.0 kHz. I know of two 7 kHz GM sensors. One is from the Saturn, part number 21021879, with 10 x 1.5 mm threads. The other is from the Fiero, part number 10497163, with 1/4 tapered pipe threads. Our sensor kit are not any of those.
The difference between the GM sensor and the OEM sensor is that the OEM sensor forces you to discontinue whatever driving conditions resulted in the knock condition, because it breaks, and sends the car into limp mode.
The GM sensor will NOT break, so if you keep causing the knock, you will break your motor.
The strategy for the OEM sensor assumes that a child is driving the car.
Don't run the GM sensor unless you are a responsible adult, that is capable of making a rational decision to stop doing whatever is causing the knock response.
I am running a stock longblock with an AEM + TD06 + 870cc with 93 oct fuel. I tuned without the aid of a knock sensor (it was bolted on but not wired due to other circumstances) and have been running 18 psi @ 11.5:1 AFR or richer, with minimal (5-8*) timing. I could not hear any audible knock so I went ahead and added in a little more timing. I would say I am still running a little less than "moderate" timing, definitely not aggressive at all (10-13* ).
After all that the car was running great but I am going to be going to the dyno tomorrow so I wired up the knock sensor, forseeing running a little more boost and perhaps some more timing on the dyno to see what happened. I went out and logged 2nd thru 4th gears on highway pulls...and the knock sensor was going crazy!! I had assumed (I dont' know why) that anything over 1.5V from teh knock sensor meant "knock". Well I'm seeing ~3.9V at ANY time under boost. So I say oh crap, and lower my boost to 14 psi. Do another pull in 2nd gear, same knock. So I reduce back to my "minimal" timing map, and boost to 12-13psi, same knock. So I take out another TWENTY PERCENT timing (now 6-9*) in the boost range and its STILL knocking at just under 4V. At this point I bet I'm making 220 whp (educated guess) and the car is a dog. I know that unless my motor or gas has a problem it should NOT be knocking here, so I'm starting to wonder if the knock sensor is working. It shows 0.1-0.5V when in the vacuum range of the map.
Dat----- So to sum it up. The GM sensor we have on our solara does its job yes. BUT it does not pick up signal pass a certain "knock". Because of that we are not seeing the almighty GM sensor breaking on the solara under boost. That is why they last so long yet on our SAFC they are going nuts. The solara ECU DOES NOT SEE THIS KNOCK because the GM sensor are not giving off that signal to the ECU. So you think your car is doing good, when its actually knocking behind your back.
So who ever said long ago the GM sensor is more sensitive.. Wrong. I was wrong. I search and this is our results. GM sensor is only meant as a cheap replacement, not a fix it.
Thanks, have a nice day.
Last edited by DatSRBoi on Wed Feb 06, 2008 9:37 pm, edited 1 time in total.
[b][color=red]WARNING: MY SIGNATURE MAY CAUSE SEIZURES![/color][/b]
-
DatSRBoi - Grumpy SolaraGuy
- Posts: 5411
- Joined: Sat Jul 10, 2004 5:33 pm
- Location: DFW Texas
