Engine Mechanical Topics Discussion for motor builds, cams, head work, stripped bolts and other engine related issues. The good and the bad. If it goes round and around or up and down, post it here.

TC -ION Sensing Detection Explained

Thread Tools
 
Search this Thread
 
  #1  
Old 08-22-2005, 10:57 AM
HDF Tech's Avatar
HDF Tech
HDF Tech is offline
Banned
Thread Starter
Join Date: May 2005
Location:
Posts: 14,498
Likes: 0
Received 147 Likes on 41 Posts
Default TC -ION Sensing Detection Explained

First ION knock detection is the direction most applications are going to use. The old PEZIO sensor relied on a vibration ( caused by some level of knock ) to register at the sensor, It is very hard to detect low levels of knock and seperate the singal ( knock from detonation ) from normal engine vibrations, such as valve closings on a Harley engine. On most vehicle applications the knock sensor can sense true knock from noise, but may not READ every cylinder in a multi-cylinder engine at every cylinder at every RPM, But most system use a GLOBAL retard scheme where any knock will retard spark to all cylinders at the same amount. New systems are going to indidvidaul cylinder knock detection where only the cylinder producing knock is retarded and only until knock ( detonation stops) so in a V8 or V6 you might only have one or two cylinders with any amount of retard.

The advantage of the ION system is it is not dependent on a sensor location and it can be calibrated to detect each cyclinder at all engine speeds and loads.


ION sensing knock detection does not use a Sensor that is externally mounted on the engine. The name ION sensing is derived from the fact that as detonation happens IONIZATION at the plug tip changes the resistence to fire the plug. The actual detection to this change in resistance is internal to the coil and the signal is sent back to the ECM. Software tables are then calibrated using model based software. The way it is calibrated is important to understand WHY some folks are having TROUBLE. A very sophisticated Combustion measurement system is placed on a development engine and actual cylinder pressures are recorded vs. crank angle degrees and coil operating ranges, as the engine is running at set speed and loads. By reading "NORMAL" outputs from the ignition module electronics to the ECM input and noting the change when detonation casues Peak cylinder pressures to rise, a model can be developed that as long as the signal from the ignition module is outputing the "expected NORMAL" signal back to the ECM no retard will occur. A change in resistence value and amount can be correlated to a given rise in pressure in the chamber during combustion.Detonation causes very large spikes in pressure even at very low levels of detonation, so the model can and does detect even before "audilble detonation ( that that can be heard ) happens. It is a very good system and very reliable with one HUGE exception. Detection is dependant on the values calibrated in the software that represent knock, and are based on a stock engine ( or a known modified engine such as the STAGE I,II etc.) with a GIVEN set of Production intent parts.ie , Spark plugs, wires, and even compression ratio can and does change the resistance at the coil. So once folks start changing wires, Plugs ( to non-factory specs) and modifing the engine cams, compression ratio etc,( and even expected A/F ratios, and therefore combustion temperature) the values derived form a production engine are no longer valid and may cause the software to "determine" that knock is present when it's not or fail to dectect knock when it is.

Just as production vehicles such as cars and trucks use similar schemes for many functions but most folks don't change parts or "hop-up" current cars with fuel injection systems... As an aside most people really don't know that if you change the aircleaner system or duct work ,even with a MASS AIRFLOW sensor you change the output of the mass airflow sensor. Those sensors are calibrated to read correctly only with a known set of induction parts. There is some amount of adaptablity in the fuel system but it is possible to get even a closed-loop system, to where the softeware cannot adapt enough for some changes that one might make. EXHAUST, intake etc.


So Harley knock detection is really not that different than other vehicle manufactures, it's US the users that may not understand the impact on a calibrated system
 
The following users liked this post:
abhibary (12-26-2018)
  #2  
Old 07-05-2010, 11:57 AM
Johncr's Avatar
Johncr
Johncr is offline
Advanced
Join Date: Aug 2007
Location:
Posts: 89
Received 0 Likes on 0 Posts
Default Interesting Point

"Detection is dependant on the values calibrated in the software that represent knock, and are based on a stock engine ( or a known modified engine such as the STAGE I,II etc.) with a GIVEN set of Production intent parts.ie , Spark plugs, wires, and even compression ratio can and does change the resistance at the coil. So once folks start changing wires, Plugs ( to non-factory specs) and modifing the engine cams, compression ratio etc,( and even expected A/F ratios, and therefore combustion temperature) the values derived form a production engine are no longer valid and may cause the software to "determine" that knock is present when it's not or fail to dectect knock when it is."


Now, if we change our cams, expected A/F, compression ratio ETC. Is there a way to set the expected values so the knock detect system works as expected on a modified engine?
 
  #3  
Old 05-15-2015, 09:11 AM
joenash's Avatar
joenash
joenash is offline
Cruiser
Join Date: Nov 2010
Location: Phoenix AZ
Posts: 210
Likes: 0
Received 0 Likes on 0 Posts
Default

Originally Posted by Johncr
"

Now, if we change our cams, expected A/F, compression ratio ETC. Is there a way to set the expected values so the knock detect system works as expected on a modified engine?

Good Question. I put I guess you call it a stage one screaming eagle air filter and mufflers on my bike. bought the SESPT tuner from harley, to tune it cause they said when you changed all that the bike will be too lean and not run correctly, and to see how the tune was, how the bike was running I used the VCI to record the data. Upon reviewing the data I noticed knock in the front cylinder and some in the rear cyl, so I thought the plugs were bad changed them , still the same mostly knock in the front cyl, so I have been changing my VE tables to see if that helps but not really , my last run both the front and rear cyls had a lot of knock wich was weird cause it was allwas in the front.
Anway you would think that since you are getting the software from the manufacturer that they would know or change the values for ion sensing knock retard, especially if you are buying their brand parts, air filter exhaust and such. I know the bike is still running a bit lean because I went a little old school and pulled the plugs to inspect them and they are white and clean lean looking. I know their is a setting in the program where you can turn knock control off, but that kind of defeats the purpose of it and protecting your motor.
???? I think harley needs to figure a different way to detect knock cause a lot of their customers modify their bikes and probably have knock problems.
 
  #4  
Old 05-15-2015, 11:21 PM
Ed Ramberger's Avatar
Ed Ramberger
Ed Ramberger is offline
Stellar HDF Member
Join Date: May 2015
Location: Then Wisconsin, now North Carolina
Posts: 3,172
Received 887 Likes on 621 Posts
Default

Interestingly, for decades, no knock sense was incorporated and properly tuned motors ran just fine.

The biggest offender that causes false knock readings are people using after-market plug wires or plugs as these components become part of the ion sense system. Some crazy exhaust systems also affect it.

Most vehicles will show a few degrees of knock - especially with poor quality fuels, and hot engine temps. That's nothing to worry about.

Most of the time, when picking a map to start tuning with (selecting one with the closest cam profile is key - then compression ratio) ion sense is fine. If false knock is seen, ion sense may have to be turned off. The bike should be checked for false knock after it is tuned. Ion sense should also be turned off during the tuning process.
 
  #5  
Old 05-17-2015, 05:02 PM
Sierra977's Avatar
Sierra977
Sierra977 is offline
Ultimate HDF Member
Join Date: Jul 2014
Location: Windwardside, Saba N.A.
Posts: 8,250
Received 3,988 Likes on 2,065 Posts
Default

So what happens when spark plug gap is increased? Say to 0.040 or more? Used to do this to get a "hotter" spark ie. higher voltage required to initiate spark over wider gap. Does this affect ion-sensing?
 
  #6  
Old 05-18-2015, 01:18 AM
joenash's Avatar
joenash
joenash is offline
Cruiser
Join Date: Nov 2010
Location: Phoenix AZ
Posts: 210
Likes: 0
Received 0 Likes on 0 Posts
Default

Originally Posted by Ed Ramberger
Interestingly, for decades, no knock sense was incorporated and properly tuned motors ran just fine.

The biggest offender that causes false knock readings are people using after-market plug wires or plugs as these components become part of the ion sense system. Some crazy exhaust systems also affect it.

Most vehicles will show a few degrees of knock - especially with poor quality fuels, and hot engine temps. That's nothing to worry about.

Most of the time, when picking a map to start tuning with (selecting one with the closest cam profile is key - then compression ratio) ion sense is fine. If false knock is seen, ion sense may have to be turned off. The bike should be checked for false knock after it is tuned. Ion sense should also be turned off during the tuning process.
I was at the harley dealer the other day cause I want to get some cams for my bike and was talking to one of the techs who specialzes in doing the motor builds and tuning, So I asked him cause if you have the sespt you can turn off knock control, if they do. and he said they do not, I guess that is the best idea cause if you have a hot motor and maybe it is in tune but you might get a bad tank of gas or something you would not want detonation in the combustion camber!

Originally Posted by Sierra977
So what happens when spark plug gap is increased? Say to 0.040 or more? Used to do this to get a "hotter" spark ie. higher voltage required to initiate spark over wider gap. Does this affect ion-sensing?
Hmm good question I don't know, I am trying to figure all this out espicially since I put an air cleaner on my bike now. Mufflers were first. The exhaust was fine but now everyone says I have to change the map, tune the bike so I figured how hard could it be. Well I got the sepst cause I figured it was from harley they would surely make a product that would work well with their bikes. Anyway with the tuner you can do a data recording then look at later on the computer, omg their is knock retard all over the place I tried changing ve timing afr but it still shows knock retard. I don't know if any of the guys using the flash tuners are having any knock or not cause they load a map in and just ride it.
And I would think their would be some sort of leeway in the software for variances , after all every spark plug is not exactly the same resistance and the wires the front one is longer than the rear.
As for a wider gap and a hotter plug, the bikes do not have a carburetor and are not dumping fuel into the motor the computer precisely monitor the fuel going into the motor, and the spark is hotter than on the older bikes.
 
  #7  
Old 06-08-2015, 10:23 AM
joenash's Avatar
joenash
joenash is offline
Cruiser
Join Date: Nov 2010
Location: Phoenix AZ
Posts: 210
Likes: 0
Received 0 Likes on 0 Posts
Default

Originally Posted by HDF Tech

Detection is dependant on the values calibrated in the software that represent knock, and are based on a stock engine ( or a known modified engine such as the STAGE I,II etc.) the values derived form a production engine are no longer valid and may cause the software to "determine" that knock is present when it's not or fail to dectect knock when it is.

Originally Posted by Johncr
"Detection is dependant on the values calibrated in the software that represent knock, and are based on a stock engine ( or a known modified engine such as the STAGE I,II etc.) with a GIVEN set of Production intent parts.ie , Spark plugs, wires, and even compression ratio can and does change the resistance at the coil. So once folks start changing wires, Plugs ( to non-factory specs) and modifing the engine cams, compression ratio etc,( and even expected A/F ratios, and therefore combustion temperature) the values derived form a production engine are no longer valid and may cause the software to "determine" that knock is present when it's not or fail to dectect knock when it is."

?
So do you think Harley changes the knock detection values if you use their download or do they keep it the same since they figure you are using their BOLT ON PARTS?
Their may be Values in the software that we can not see or change?
 
  #8  
Old 06-08-2015, 10:36 AM
joenash's Avatar
joenash
joenash is offline
Cruiser
Join Date: Nov 2010
Location: Phoenix AZ
Posts: 210
Likes: 0
Received 0 Likes on 0 Posts
Default

Originally Posted by Ed Ramberger
Interestingly, for decades, no knock sense was incorporated and properly tuned motors ran just fine.



Most vehicles will show a few degrees of knock - especially with poor quality fuels, and hot engine temps. That's nothing to worry about.

Most of the time, when picking a map to start tuning with (selecting one with the closest cam profile is key - then compression ratio) ion sense is fine. If false knock is seen, ion sense may have to be turned off. The bike should be checked for false knock after it is tuned. Ion sense should also be turned off during the tuning process.
Well like you said their was no knock for years what would happen if you left knock detection off? especially if you had a lot of mods?

How would you tell if you had false Knock?

If you were the average joe and did not have a dyno to put a load on the bike and an AFR meter to check your tune , And you had to drive your bike around the block and look at your data you recorded on your VCI, how would you know if you were to lean? or too rich according to the way I understand the ion sensing if their was too much fuel also it would see incorrect resistance values also?
????????????????
 
  #9  
Old 06-08-2015, 01:32 PM
98hotrodfatboy's Avatar
98hotrodfatboy
98hotrodfatboy is offline
Seasoned HDF Member
Join Date: Mar 2010
Location: Poolville
Posts: 17,594
Received 4,793 Likes on 3,304 Posts
Default

Originally Posted by joenash
Well like you said their was no knock for years what would happen if you left knock detection off? especially if you had a lot of mods?

How would you tell if you had false Knock?
I would think the same way we always did. Pay close attention to your build and always keep a keen ear open. Its worked successfully like that for decades...
 
  #10  
Old 06-08-2015, 11:53 PM
sloufoot's Avatar
sloufoot
sloufoot is offline
Moderator
Join Date: Jan 2012
Location: Southeast Idaho
Posts: 26,410
Received 4,081 Likes on 3,483 Posts
Default

When you are looking at the sespt tuner all it will let you do is stay within the EPA standards.
On an engine that is already lean from factory almost anything you change will make it run leaner. The only tuner from Harley that will let you tune past that limit is the SEPST Pro tuner.
But that tuner isn't compatible with your warranty.
The knock sensor system is a good thing to not mess with. If the spark settings & fuel settings are set correctly there won't be any knock with the standard knock sensor settings.
One other thing to check is the plug wires to see that they are the correct resistance. Many aren't correct new.
If your plug is white white you most likely need more fuel in those knock areas. Not much of a change can be enough to take care of it. Sometimes it does take a large adjustment in both spark & fuel .
But remember what tuner you have........ Also get a lot of reading in the tuning forums & ask questions. Don't just shoot in the dark, it will cost ya !
 


Quick Reply: TC -ION Sensing Detection Explained



All times are GMT -5. The time now is 12:44 AM.