Knock Sensors
04-11-2016, 07:51 PM
#1
Knock Sensors So I'm having issues with my abs inop light and codes for the AIR system on my car. When I had my motor freshened up a couple of years ago the knock sensors were not changed and were giving the Tuner a fit so he tuned them out. I'm going to remove the AIR system so it will need to go to the tuner and if I'm going there I will replace the knock sensors and have him turn them back on and maybe he can read the abs codes and tell me what needs replacing in the system.
Advance sell Delphi sensors and I'm thinking they should be a good fit for this car. Should I just go ahead and replace the intake manifold bolts while I'm doing the job? I have a large selection of bolts where I work, can anyone tell me the exact size of the bolts for the intake manifold? Oh yea I will replace the harness for the knock sensors while I'm at it.
Advance sell Delphi sensors and I'm thinking they should be a good fit for this car. Should I just go ahead and replace the intake manifold bolts while I'm doing the job? I have a large selection of bolts where I work, can anyone tell me the exact size of the bolts for the intake manifold? Oh yea I will replace the harness for the knock sensors while I'm at it.
04-11-2016, 08:08 PM
#2
So I'm having issues with my abs inop light and codes for the AIR system on my car. When I had my motor freshened up a couple of years ago the knock sensors were not changed and were giving the Tuner a fit so he tuned them out. I'm going to remove the AIR system so it will need to go to the tuner and if I'm going there I will replace the knock sensors and have him turn them back on and maybe he can read the abs codes and tell me what needs replacing in the system.
Advance sell Delphi sensors and I'm thinking they should be a good fit for this car. Should I just go ahead and replace the intake manifold bolts while I'm doing the job? I have a large selection of bolts where I work, can anyone tell me the exact size of the bolts for the intake manifold? Oh yea I will replace the harness for the knock sensors while I'm at it.
Advance sell Delphi sensors and I'm thinking they should be a good fit for this car. Should I just go ahead and replace the intake manifold bolts while I'm doing the job? I have a large selection of bolts where I work, can anyone tell me the exact size of the bolts for the intake manifold? Oh yea I will replace the harness for the knock sensors while I'm at it.
Heres a thread example
https://ls1tech.com/forums/lt1-lt4-m...ng-timing.html
Notice there is a constant "chasing" of the problem, driving and diagnostics, then finally after several trial/error attempts, it gets "finished".
04-12-2016, 02:43 PM
#4
The factory goes for maximum safety. The OEM tune is very sensitive and tends to pull a lot of timing when it detects knock, and this can impact performance greatly. The factory tune is also very "safe" (there is some power locked away because of that) so between the safe tune, and the safe knock parameters, the engine should last a loooooong time because it will consistently run in a reduced timing state at the slightest hint of knock noise. Great for grandmom to drive a daily, but not what we want as performance enthusiasts.
Next, the OEM knock sensors are "tuned" to specific frequencies which the original engineers found to be associated with the onset of engine knock (pre audible). Those frequencies are strongly tied to the internals of the engine i.e. the crankshaft, connecting rod, piston, flywheel, anything that spins/rubs/bounces/jostles and creates a tone that the sensor could "see". Some tones are more "visible" than others, (thus the reason we say it is "tuned" to those frequencies) So anything you have done to the engine, lets say you installed a new set of lifters/camshaft/springs and a new flywheel, now you have to go drive the car and see what the knock sensor thinks of your new additions. You are now a candidate for "false knock". You also may have moved the pre-audible frequency of engine knock around, i.e. you might have just made real knock invisible to your OEM knock sensors with the current settings. They may need to be re-configured, or it might even be outside of the capability of the sensor to detect engine knock and you need to find a new sensor.
So how can you tell what is real knock, and what is false knock, and when you are knocking without it showing up at the sensor? That is the hard part everybody struggles with, that is what takes time, logging, patience, tuning, experience.
Here is how I would do it.
Based on experience, I already know more or less what sort of timing to use with a given engine. So that is 51% of the battle. I still start low, lets say an N/A engine takes 25* btdc at WOT. Lets say no knock shows up. While still on the street, I would also try 27* and 23* and look for changes. Sometimes 23* will start showing knock, where 27* will not. These kinds of strange behaviors are what cost alot of time to sort out. Knock sensor is a tool, but not a guaranteed thing, just because it shows or does not show noise doesn't really mean anything unless we can put together a picture and convince ourselves of what is really happening. So back to our example, lets say the knock sensor seems inconsistent, I can't rely on it, what else can I do? Well first, I would stop and pull the plugs. Always check your plugs for abnormal signs while/after "street tuning" especially when something like this happens (inconsistent behavior from a sensor you think is trolling you). Next, I take the car to a chassis dyno. Having an EGT gauge at this point would help but isn't necessary.
Once on the dyno, I can see what my timing adjustments do to the torque curve, This is part of tuning 101. With my initial 25* compared to the 27*, I only gained 2-3 torque units, which tells me it was probably unnecessary, but wait on that for now. When reducing to 23*, I lost 8 units of torque. Still within range of "proper" but now I know for sure that 23-25 is where I want to be, most likely. But here is the kick. What is the temperature of the inlet charge? are any components on the engine changing air density (like a turbocharger?) I can't go into extreme detail, but you would want to factor the possibility of conditions varying widely, especially on a street driven turbo car, for example it can get very hot idling in traffic with the A/C on, and this will impact full throttle performance AND full throttle timing numbers. If I set the engine to 25* and "walk away" then the next time the car gets stuck in traffic and really heat soaks good, the engine will likely knock. How can we avoid that situation? Can you rely on a knock sensor to pull timing under those conditions? should you?
The answer is no. Experience tuning vehicles has taught me never to rely on the knock sensor to "do the work for you". The knock sensor is a last resort, a final line of defense. It should never be used to pull timing if at all possible. The tune you put in the computer should be free from knock to begin with, and the engine should run in all conditions without knock, even without a knock sensor. So whats our plan now?
The solution is to never be caught in those situations to begin with. In the above example, the real culprit is the heat soak, not the tune. A well thought out turbocharger system has all of the necessities in place to prevent that exact situation: coatings, wraps, methanol injection, dividers, intercooling, reflective surfaces/shields, and it was tuned on the rich/safe side just like the factory would have done for daily drivers. (this all pertains to daily drivers). Once we have obliterated the chances of rising underhood temperatures from ruining our engine, a set of sensors designed to keep things in check (properly placed IAT / methanol level sensors for example) help us maintain our setup and warns us when something gets out of control.
Now back to the car/dyno
So we found our 25* (this engine was N/A forget about the turbo for a minute) was a reasonable number, if we had an EGT gauge it would have told us 900 or 1100*F depends where we put it, but I like to see a nice cold non piston melting number, backed by plenty of fuel. Lets flip our scenario now and say that there was knock showing up, all the time. It knocked at 23, 25, and 27. I even tried 18* and it still knocked. Well that just seems like false knock. How can we be sure? What should we do? The first thing I do: Put race fuel in the Tank. If we put race fuel in the tank, and see the same exact knock pattern, the next thing I do is yank, clean, and wrap with teflon tape the knock sensors, then try again with the race fuel. You see where I am going with this? You try to stop the sensor from reacting with the race fuel while using pump gas timing numbers (or nearly), then, go back to the pump gas and see if the "knock" comes back. The reason this is so important isn't because you need to depend on the knock sensor... it is because you want to verify whether or not the knock sensor can even be trusted to begin with. If I keep seeing false knock with or without race fuel, then I start stethoscope-ing the engine, looking for potential internal problems, otherwise the knock sensors get replaced, and we try again.
This situation goes round and round, all day and night, I have hundreds maybe thousands of logs from cars and trucks just trying to dial in whether or not the knock sensor is "lying" or "sometimes lies" or "cant be trusted" etc... It really is a hassle, a big ordeal, but the end result is worth it if you can get it to work, because once you KNOW the thing can be reliable (I put 87 in the tank once just to see, and guess what lit up like a christmas tree?) you can tack it on at the end of the final tune as an additional safety measure.
Next, the OEM knock sensors are "tuned" to specific frequencies which the original engineers found to be associated with the onset of engine knock (pre audible). Those frequencies are strongly tied to the internals of the engine i.e. the crankshaft, connecting rod, piston, flywheel, anything that spins/rubs/bounces/jostles and creates a tone that the sensor could "see". Some tones are more "visible" than others, (thus the reason we say it is "tuned" to those frequencies) So anything you have done to the engine, lets say you installed a new set of lifters/camshaft/springs and a new flywheel, now you have to go drive the car and see what the knock sensor thinks of your new additions. You are now a candidate for "false knock". You also may have moved the pre-audible frequency of engine knock around, i.e. you might have just made real knock invisible to your OEM knock sensors with the current settings. They may need to be re-configured, or it might even be outside of the capability of the sensor to detect engine knock and you need to find a new sensor.
So how can you tell what is real knock, and what is false knock, and when you are knocking without it showing up at the sensor? That is the hard part everybody struggles with, that is what takes time, logging, patience, tuning, experience.
Here is how I would do it.
Based on experience, I already know more or less what sort of timing to use with a given engine. So that is 51% of the battle. I still start low, lets say an N/A engine takes 25* btdc at WOT. Lets say no knock shows up. While still on the street, I would also try 27* and 23* and look for changes. Sometimes 23* will start showing knock, where 27* will not. These kinds of strange behaviors are what cost alot of time to sort out. Knock sensor is a tool, but not a guaranteed thing, just because it shows or does not show noise doesn't really mean anything unless we can put together a picture and convince ourselves of what is really happening. So back to our example, lets say the knock sensor seems inconsistent, I can't rely on it, what else can I do? Well first, I would stop and pull the plugs. Always check your plugs for abnormal signs while/after "street tuning" especially when something like this happens (inconsistent behavior from a sensor you think is trolling you). Next, I take the car to a chassis dyno. Having an EGT gauge at this point would help but isn't necessary.
Once on the dyno, I can see what my timing adjustments do to the torque curve, This is part of tuning 101. With my initial 25* compared to the 27*, I only gained 2-3 torque units, which tells me it was probably unnecessary, but wait on that for now. When reducing to 23*, I lost 8 units of torque. Still within range of "proper" but now I know for sure that 23-25 is where I want to be, most likely. But here is the kick. What is the temperature of the inlet charge? are any components on the engine changing air density (like a turbocharger?) I can't go into extreme detail, but you would want to factor the possibility of conditions varying widely, especially on a street driven turbo car, for example it can get very hot idling in traffic with the A/C on, and this will impact full throttle performance AND full throttle timing numbers. If I set the engine to 25* and "walk away" then the next time the car gets stuck in traffic and really heat soaks good, the engine will likely knock. How can we avoid that situation? Can you rely on a knock sensor to pull timing under those conditions? should you?
The answer is no. Experience tuning vehicles has taught me never to rely on the knock sensor to "do the work for you". The knock sensor is a last resort, a final line of defense. It should never be used to pull timing if at all possible. The tune you put in the computer should be free from knock to begin with, and the engine should run in all conditions without knock, even without a knock sensor. So whats our plan now?
The solution is to never be caught in those situations to begin with. In the above example, the real culprit is the heat soak, not the tune. A well thought out turbocharger system has all of the necessities in place to prevent that exact situation: coatings, wraps, methanol injection, dividers, intercooling, reflective surfaces/shields, and it was tuned on the rich/safe side just like the factory would have done for daily drivers. (this all pertains to daily drivers). Once we have obliterated the chances of rising underhood temperatures from ruining our engine, a set of sensors designed to keep things in check (properly placed IAT / methanol level sensors for example) help us maintain our setup and warns us when something gets out of control.
Now back to the car/dyno
So we found our 25* (this engine was N/A forget about the turbo for a minute) was a reasonable number, if we had an EGT gauge it would have told us 900 or 1100*F depends where we put it, but I like to see a nice cold non piston melting number, backed by plenty of fuel. Lets flip our scenario now and say that there was knock showing up, all the time. It knocked at 23, 25, and 27. I even tried 18* and it still knocked. Well that just seems like false knock. How can we be sure? What should we do? The first thing I do: Put race fuel in the Tank. If we put race fuel in the tank, and see the same exact knock pattern, the next thing I do is yank, clean, and wrap with teflon tape the knock sensors, then try again with the race fuel. You see where I am going with this? You try to stop the sensor from reacting with the race fuel while using pump gas timing numbers (or nearly), then, go back to the pump gas and see if the "knock" comes back. The reason this is so important isn't because you need to depend on the knock sensor... it is because you want to verify whether or not the knock sensor can even be trusted to begin with. If I keep seeing false knock with or without race fuel, then I start stethoscope-ing the engine, looking for potential internal problems, otherwise the knock sensors get replaced, and we try again.
This situation goes round and round, all day and night, I have hundreds maybe thousands of logs from cars and trucks just trying to dial in whether or not the knock sensor is "lying" or "sometimes lies" or "cant be trusted" etc... It really is a hassle, a big ordeal, but the end result is worth it if you can get it to work, because once you KNOW the thing can be reliable (I put 87 in the tank once just to see, and guess what lit up like a christmas tree?) you can tack it on at the end of the final tune as an additional safety measure.
04-12-2016, 04:35 PM
#5
Well I read all of what you wrote and still...........
When I had the parts in my sig installed the mechanic did not install new knock sensors, after all the money invested the tuner told me he could not get the old sensors to function because when they are old and they get disturbed they will not work so he turned them "off". I have been driving around like this for 60,000 miles or so and from time to time I hear a ping. I am bringing my car back to a tuner to have the AIR tuned out and I would feel more comfortable with knock sensors than without. I have no desire to burn holes in the tops of my pistons. So what I am deriving from you is to not bother installing new knock sensors because nobody can make them work??? This Camaro with the updated Z06 motor is my weekend toy for car shows and such, I drive a corvette during the week.
When I had the parts in my sig installed the mechanic did not install new knock sensors, after all the money invested the tuner told me he could not get the old sensors to function because when they are old and they get disturbed they will not work so he turned them "off". I have been driving around like this for 60,000 miles or so and from time to time I hear a ping. I am bringing my car back to a tuner to have the AIR tuned out and I would feel more comfortable with knock sensors than without. I have no desire to burn holes in the tops of my pistons. So what I am deriving from you is to not bother installing new knock sensors because nobody can make them work??? This Camaro with the updated Z06 motor is my weekend toy for car shows and such, I drive a corvette during the week.