Knock Sensor
09-11-2006, 08:09 PM
#12
Registered User
Join Date: Aug 2006
Location: Albuquerque
Posts: 221
Likes: 0
Received 0 Likes
on
0 Posts
Hmmm. Interesting that everyone seems to have slightly different experiences with thier knock sensors.
djpeak10: Its possible that I may not be pulling as hard at top under WOT. but its not a drastic difference. I saw 22 degrees on the OBD tool, like I said, 22 was an ideal number for the 2.0L turbo Mitsu engine. When I get the new sensor in, I'll check the timing again and report back.
ek9: I don't know much about Honda timing and how it is affected by knock, thats why I'm here to talk and learn, but... On the Mitsu engine, only under hard acceleration and high RPMs is the knock sum used to pull the timing. High knock counts at part throttle and low RPMs do not pull timing, at least on the Mitsu. I've never seen mechanical noise registering as knock in the upper RPMS under load on a mechanically sound mitsu engine, true knock could always be reduced or removed by increasing the octane. Is the s2k knock sensor really this ineffective at high RPM, granted the s2k revs higher than 7500, that it must interpolate?
With the sensor malfunctioning on the s2k I would expect timing to be pulled down to 0 degrees under hard acceleration. Since I still see 22 degrees, I must ask is 22 the normal/target number for this engine at WOT? Or is it higher and my malfunctioning sensor is causing it to be pulled down to 22?
djpeak10: Its possible that I may not be pulling as hard at top under WOT. but its not a drastic difference. I saw 22 degrees on the OBD tool, like I said, 22 was an ideal number for the 2.0L turbo Mitsu engine. When I get the new sensor in, I'll check the timing again and report back.
ek9: I don't know much about Honda timing and how it is affected by knock, thats why I'm here to talk and learn, but... On the Mitsu engine, only under hard acceleration and high RPMs is the knock sum used to pull the timing. High knock counts at part throttle and low RPMs do not pull timing, at least on the Mitsu. I've never seen mechanical noise registering as knock in the upper RPMS under load on a mechanically sound mitsu engine, true knock could always be reduced or removed by increasing the octane. Is the s2k knock sensor really this ineffective at high RPM, granted the s2k revs higher than 7500, that it must interpolate?
With the sensor malfunctioning on the s2k I would expect timing to be pulled down to 0 degrees under hard acceleration. Since I still see 22 degrees, I must ask is 22 the normal/target number for this engine at WOT? Or is it higher and my malfunctioning sensor is causing it to be pulled down to 22?
09-12-2006, 05:46 AM
#13
On the mitsu, are the part throttle knock counts used to create the knock sum which is then used at WOT, high loads?
On the EMS, I was looking at plots of knock sensor voltage versus RPM and beyond a certain RPM, the engine noise caused the K-Sensor voltage to increase to the point it was drowning out an signal. From that I gathered they determine the "knock proneness" from part-throttle running while perhaps not applying it till high load conditions.
On the EMS, I was looking at plots of knock sensor voltage versus RPM and beyond a certain RPM, the engine noise caused the K-Sensor voltage to increase to the point it was drowning out an signal. From that I gathered they determine the "knock proneness" from part-throttle running while perhaps not applying it till high load conditions.
09-12-2006, 05:57 AM
#14
I'm not sure if the knock sensor on the S2000 is different than on the Civic, but here's how the Civic knock sensor works:
As EK9 stated, the knock sensor is essentially a microphone tuned to a certain frequency that sends an electric signal to the ECU, telling it the engine is knocking. To test a knock sensor, you're supposed to tap it with a metal object and measure the slight voltage spike.
When this signal is received by the ECU, it retards the timing to prevent predetonation (knocking). A retarded timing reduces performance and fuel efficiency. But since engine knock is harmful to your engine, it is better to retard timing than to harm your engine. A knock sensor that fails to emit this signal does not hurt engine performance much: Predetonation means that combustion takes place before TDC, so there's a slight amount of work that the engine has to do to get to TDC after combustion.
If your knock sensor isn't working properly, it will just not send a signal when it is knocking. This will not trigger a CEL unless the knock sensor is showing an open circuit. In other words, a defective knock sensor may or may not give you a CEL, depending on whether your ECU perceives that it is not connected.
As EK9 stated, the knock sensor is essentially a microphone tuned to a certain frequency that sends an electric signal to the ECU, telling it the engine is knocking. To test a knock sensor, you're supposed to tap it with a metal object and measure the slight voltage spike.
When this signal is received by the ECU, it retards the timing to prevent predetonation (knocking). A retarded timing reduces performance and fuel efficiency. But since engine knock is harmful to your engine, it is better to retard timing than to harm your engine. A knock sensor that fails to emit this signal does not hurt engine performance much: Predetonation means that combustion takes place before TDC, so there's a slight amount of work that the engine has to do to get to TDC after combustion.
If your knock sensor isn't working properly, it will just not send a signal when it is knocking. This will not trigger a CEL unless the knock sensor is showing an open circuit. In other words, a defective knock sensor may or may not give you a CEL, depending on whether your ECU perceives that it is not connected.
09-12-2006, 06:40 PM
#15
Registered User
Join Date: Aug 2006
Location: Albuquerque
Posts: 221
Likes: 0
Received 0 Likes
on
0 Posts
Something just hit me, do we even have a way to see the knock count on the s2k with the stock ecu? I didn't look through the obd tool menus to see what else it could do. I'll try to check next weekend. The EMS, depending on the sensor and its programming could behave entirely different than what I'm used to.
ek9: On the mitsu, are the part throttle knock counts used to create the knock sum which is then used at WOT, high loads?
Hmmmm. I don't really think this is true. I'll check the few datalogs that I have to see the highest RPM at which knock starts and how the timing was corrected. I do know the ECU remembers where it sees knock and can pull the timing on a run even if there is no knock. A simple reset of the ECU fixes this but then the fuel trims must be re-learned also. From what I've seen, on a properly running mitsu engine, the knock that pulls timing that is determintal to acceleration usually comes on at full boost and the mid to top of the RPMs, WOT. Its possible to see very high knock counts when you first hit on your gas while cruising or stoped but this is probably more likely due to too much loading, being in too high a gear at too low an RPM, but this knock will dissappear once the car winds up.
Usually while cruising with part throttle the timing is maxed. When you first hit on the gas the knock can increase and the timing my dip but as you start to accelerate because you are only at part throttle the timming slowly climbs back to max advance regardless of the knock count. I found a few good datalog examples, one being a quarter mile run. Showing that as I let off the pedal from WOT the timing instantly maxes out, and that at part throttle on my launch with knock the timing kept on climbing. I need to figure how to put this on here.
e9k, were those s2k plots you were looking at? Can I see them. How can you tell the difference between engine noise and real knock on the datalogger? Do you have any datalogs showing knock, timing, rpm, and throttle position on the s2k?
slalom44:
I agree with most of what your saying. The knock sensor is basically a microphone that is being filtered by the ecu. Since it is a basic microphone, there is an AC signal being generated from the sensor not just a simple DC voltage. This AC signal contains the specific frequency signature of the sound that it is picking up. Hook up the sensor to an O-scope and tap the sensor with diffenent objects and see how the signature is different. The ecu should only be looking for pinging sounds, not normal engine noises. I do realize that other mechanical components may generate similar sounds to knock but this is usually when there is something wrong with them. I have seen and heard of many contributors to knock but all of them could be fixed, once they were identified.
slalom44:
I agree if your knock sensor has an attenuated signal then the ECU may never see a large enough knock value to pull the timing. But... with an open knock sensor circuit the ECU should pull the timing to a safer value, say 0, to protect the engine under load. It doesn't make any sense to me that any make of engine with a knock sensor would allow full timing advance with an open knock sensor circuit at WOT.
So is 22 degrees normal for the s2k?
ek9: On the mitsu, are the part throttle knock counts used to create the knock sum which is then used at WOT, high loads?
Hmmmm. I don't really think this is true. I'll check the few datalogs that I have to see the highest RPM at which knock starts and how the timing was corrected. I do know the ECU remembers where it sees knock and can pull the timing on a run even if there is no knock. A simple reset of the ECU fixes this but then the fuel trims must be re-learned also. From what I've seen, on a properly running mitsu engine, the knock that pulls timing that is determintal to acceleration usually comes on at full boost and the mid to top of the RPMs, WOT. Its possible to see very high knock counts when you first hit on your gas while cruising or stoped but this is probably more likely due to too much loading, being in too high a gear at too low an RPM, but this knock will dissappear once the car winds up.
Usually while cruising with part throttle the timing is maxed. When you first hit on the gas the knock can increase and the timing my dip but as you start to accelerate because you are only at part throttle the timming slowly climbs back to max advance regardless of the knock count. I found a few good datalog examples, one being a quarter mile run. Showing that as I let off the pedal from WOT the timing instantly maxes out, and that at part throttle on my launch with knock the timing kept on climbing. I need to figure how to put this on here.
e9k, were those s2k plots you were looking at? Can I see them. How can you tell the difference between engine noise and real knock on the datalogger? Do you have any datalogs showing knock, timing, rpm, and throttle position on the s2k?
slalom44:
I agree with most of what your saying. The knock sensor is basically a microphone that is being filtered by the ecu. Since it is a basic microphone, there is an AC signal being generated from the sensor not just a simple DC voltage. This AC signal contains the specific frequency signature of the sound that it is picking up. Hook up the sensor to an O-scope and tap the sensor with diffenent objects and see how the signature is different. The ecu should only be looking for pinging sounds, not normal engine noises. I do realize that other mechanical components may generate similar sounds to knock but this is usually when there is something wrong with them. I have seen and heard of many contributors to knock but all of them could be fixed, once they were identified.
slalom44:
I agree if your knock sensor has an attenuated signal then the ECU may never see a large enough knock value to pull the timing. But... with an open knock sensor circuit the ECU should pull the timing to a safer value, say 0, to protect the engine under load. It doesn't make any sense to me that any make of engine with a knock sensor would allow full timing advance with an open knock sensor circuit at WOT.
So is 22 degrees normal for the s2k?
Thread
Thread Starter
Forum
Replies
Last Post
unowned
S2000 Under The Hood
1
04-11-2021 08:19 AM