Seems to be some amount of mystery around this. so... Knock - Detonation, ping, pre-ignition. Basically boils down to your piston shuddering or slapping the sides of the cylinder wall. Usually because of the charge being ignited too early (while the piston is still going up) or very uneven combustion. Visualize hitting the top of your piston with a sledge hammer while it's still on it's way up. There are a number of things that cause knock. Timing that's too far advanced, fouled or hot plugs that can touch off combustion without a spark, etc. Knock is bad. Knock can lead to busted rings, burn holes in the tops of pistons, "oil burners", etc. Bad knock is very bad and very bad in a very short amount of time. We have knock sensors and an automatic mechanism to reduce/minimize knock. Up to a point? There are 2 vibration sensors, one bolted to each side of the block just below the exhaust headers, that are tuned to detect the vibration signature of engine knock. These knock sensors are hooked to the PCM which is programmed to pull (retard) timing in response to knock detection.This is the part that I think could benefit from some discussion to see if anyone knows and if we can get some idea of how that works. Twice now, in the SRT Engineer Chat sessions, a couple of questions on knock retard have been answered with something to the effect of "It's not that simple and unless you are working with equipment in the 100's of thousands of dollars range it wouldn't do any good to tell you anything". While that may very well be true I think some surface knowledge could be useful if not interesting. These are some of the things I wonder about...: The PCM will report the amount of knock retard it is applying to a scanner like the Predator or DashHawk. We're all becoming familiar with stuff like "It's expected to see maybe up to 4* or 5* of knock retard around shift points" or "You really don't want to see more than about 2* of knock retard at WOT", etc. The PCM determines timing from a number of inputs and sensors. Static tables, dynamic tables, and various sensors. This would be the desired timing without taking knock into consideration. If knock is detected the PCM will reduce the applied timing by the amount necessary to reduce/eliminate knock. If you look at Timing Advance, Knock Retard, and actual timing (usually for cylinder #1) it should roughly work out to Advance - KnR = Actual. It's just roughly because there are other things factored in there as well. So If you're seeing 5* of knock retard does it mean that your engine is actually knocking more than when you're seeing 2* of knock retard. I don’t think so. It's just that the PCM had to pull more timing to get the knock down under the same threshold. (assuming it's not a variable threshold)? So is running at 5* of knock retard worse than running at 2* of knock retard? Once the timing has been pulled to correct it, I don't think so but getting to that corrected state could be different. - Certainly there are many conditions and calculations that figure into this threshold where the PCM pulls timing in response to knock. Just how “big” is that threshold? If we ran at that threshold all the time. would the engine life would be reduced by 20%? Something like that? - When we cross the threshold where the PCM pulls timing, is the response to pull a lot of timing (relatively speaking) initially and then gradually add it back? Or does it gradually pull just a little bit of timing until the knock drops below the threshold? Does it matter how bad the knock is initially? From graphs it looks like the onset of intervention is usually faster than the removal. For example the knock retard graph will jump almost immediately to 4* and then trail off gradually back to 0. Occasionally it looks like it comes on gradually. - How fast are the knock sensors sampled (or whatever else) to detect knock? Does it only take 1 "knock" before intervention or does it knock for a while before intervention? Once a correction is made, and the condition is no longer present, how long does it take to get back to no correction? Maybe it is better to run at 2* of knock instead of 5* of knock because of the lead time to correct it or the taper off time after it's gone? - It seems to me that if you're running some degrees of knock retard you would retard your timing that many degrees to get rid of it. Lets say the timing advance is 21* and you're seeing 6* of knock retard. If you wanted to eliminate the knock retard, would you have to set your timing at 15*? If that’s what it takes would there be any difference between setting your timing to 21* and having the PCM correct it or setting your timing at 15* without the PCM needing to correct it? - What is the range in which the PCM can handle knock? Is there a limit? Some people have blown up their engines with detonation. Did they all have the knock sensors disabled? Or is it something like the PCM will only pull up to 10* and if it's worse than that you get (harmful) knock? Or were they just outside the range that the PCM can handle? Haha, there’s more I’m wondering about but is there any interest in this? and is anyone who knows, interested in sharing.
I love these types of posts. I can't answer most of your questions but I can say that most modern PCM are "Real Time" but can't give you the definition of what they mean is "Real Time". For my truck (05 Ram Srt10) the knock sensor voltage will reach 1.2 volts before the PCM retards the timing back 2 degrees at WOT. When my tuner is tuning the truck there are many variables to engine knock like fuel air mixture, fuel pressure, timing, air intake temperatures, altitude, fuel octane ratings, air flow via CAI or air filters, ETC.... ETC.... The general rule of thumb for my truck is for every degree of timing that I can advance it translates to 10RWHP. In addition the leaner the air fuel mixture the more power you can make but the hotter the engine runs resulting in knocking which translates to piston damage. This is a fine line. Be very carefull and only let an experienced tuner do the job. Some engines like the 04 Ram Srt10's don't have knock sensors. Do all the Srt engines have knock sensors other than the one I mentioned above?
When it comes to knock there is also ghost knock...and I am an semi expert in that department. My motor shows at times 8*-13.5* of ghost knock. I call it ghost knock cause it is caused by noise from HUGE springs I run and other valve train noise. We have run 100 octane and 120 octane additive to my fueling system with the same knock results...meaning it is not REAL knock. I was going to run a race tune without knock sensors this weekend with my 120 octane additive (known as water-Methanol injection) and had tuned for it but a little thing called breaking my new rear end's spider gear knocked that plan out..no pun intended. My car still runs in the 11s with 10* of KR....I know it sucks and there is a possibility I could lessen it with different length push-rods/other valve train mods and we are exploring those once I need new springs again.
I guess your problem would be the ghost knock masking real knock? I guess(?) if you went from 100 octane to 120 octane and it didn't change that you can assume it's all ghost knock? at least right now?
To answer this; knock sensors are what is called a pezio-electric sensor (not sure if i spelled that right). basically it creates its own voltage signal from 0-5 volts based on vibrations in the motor. if the vibrations cross a certain threshold, which the pcm would see as knock, timing is retarded. now you can also get into this in a whole other way, and redfox touched on this a bit. larger displacement engines, different valvetrain (springs, pushrods etc.) and such can cause the harmonic vibrations of the engine to differ from stock, sometimes causing a signal that the knock sensors could see as knock, but in reality, it is normal vibrations for the motor.
X2 exactly x2 I resemble this. New diff not available until May 14th. X2 exactly what I am doing. Exploring different bore and stroke too. C.
So that amount of ghost knock is preventing you guys from getting your timing anywhere near where it should be right?
Don't want to answer for Nathan, but my tune overrides the knock retard. I did the same things as Nathan. I bumped up the octane and still saw the knock. I knew it wasn't real at that point. It is a delicate line as someone else mentioned. I am still not 100% in the a/f and kr zone, but pretty close. C.
My 11.66 run was on 13* of timing...most SRTs run in the mid to high 20s...I think I am going to make a dampener and longer bolts and see if I can isolate the knock sensor more and get it to read less.
Excellent thread! I spend a great deal of time dealing with "knock" at work every day...albeit on methane burning monsters driving huge gas compressors. One quick point before I shut up and listen: Detonation is actually two flame fronts colliding in your combustion chamber...usually the true spark-plug driven flame front and some other undesired flame front. Many things can start the undesired flame front. Pre-ignition is just that...something like a glowing point that starts ignition before the intended time. You can have both on one combustion stroke. Both cause the same result as far as our knock sensors are concerned. Video of normal combustion: http://paultan.org/archives/2008/04/09/video-inside-a-combustion-chamber/ We have a video somewhere at work of detonation inside a combustion chamber...truly amazing to watch. I'll see if I can get it electronically and post it up.
Or turn off the long term function and only use the short term like I do. I have used 40 degrees BTDC timing and let the knock sensors pull the timing back automatically for you. By shutting off long term there is no memory to start from. Just my 2 cents......... J
I was doing some data loggin yesterday and I was gonna start a post about knock for the SRT4 but I believe the it holds true for most modern PCM's. . Obviously I am running a different motor but you guys have also over looked one aspect of modern PCM's. Most of them also use a predictive algorithim for knock, meaning that based on a few of the motors parameters it calculates where the most likely spot where detonation could occur and pulls out a few degrees. For example I have 0 knock till about 4500 rpms then it progressively pulls out about 5* of timing till it reaches 5200 then is puts it right back in.
Ahh, that sounds sort of kind of related to this (well, at least it's in the same area). The Predator allows fuel and timing adjustment in 3 rpm ranges. 1k-3.8K, 4K-4.8K, and 5K-7K (actually I think it's 5.2K or5.3K). I was adjusting my AFR using the canned tune and was seeing a bump in the AFR in the 3.8K-4K area. Seemed like it was about a 0.5 AFR bump. Through that ~200 rpm range the AFR would gradually climb about 0.5 and gradually go back down. Did that on their canned tune without any changes also. I asked Diablo about that. Mike mumbled something about it being a transition area where they "smoothed" the transition in settings from one rpm range to the next in order to guard against a "spike" or something like that. I was saying that I was trying to get my AFR flat all the way through (~12.2), and that on the canned tune I was running about 12.6 and it would go to 13.1 or 13.2 in this area and I didn't understand the bump and why it would bump up instead of down, yada, yada. So instead of trying to explain it to someone completely clueless, Johan told me to get things where I wanted them and send him the tune. He sent it back saying that there was a bump in my VE table that he had smoothed out. ?? It helped. So could this be something like that? and where can I find the Cliff notes on a VE table? and is it dynamically populated?
Some people were talking about a nylon gasket/bushing against where the knock sensor bolts to the block. I am not sure what became of that. C.
That sounds like shooting in the dark. What are you trying to do? You don't want to disable the knock sensors but you want to drop their sensitivity. You're hoping to reduce what you're fairly sure is ghost knock while still detecting severe real knock? As far as the sensor is concerned it can't distinguish between ghost knock and real knock. So if you're going to leave your knock sensors enabled, how do you determine how much knock is acceptable to you if it was real knock? If you were able to say "this much knock is ok for me", then how do you calibrate your knock sensors for that using a nylon gasket/bushing? Just doesn't sound like a solution to me. Way closer to just turning off your knock sensors than reducing the ghost knock a little bit.
I can go on and on about KR, KS and more. Your knock sensors are 0v-5v. False knock or "ghost knock" happens and this can be low voltage. A breakdown in voltage. This is a typical scenario that I use and it may not be suitable in your application. Outside temp and under hood temps, to incorrect heat range spark plugs are a factor. False Knock: 0v-1.5v typical Burst Knock: 1.5v-2.3v Light Knock: 2.3v-3.0v Heavy Knock: 3v-5v Knowing what voltage your KR activity is happening at is a must. You may be seeing KR at a knock voltage of only 1.5v. Now the PCM is going to see KR regardless of the voltage and it's going to look at the tables in the PCM and start pulling out your valuable timing. Some tuners will see this and just add to the WOT or Part throttle table the amount of retard in as +timing. So in other words at 4,000 rpm's your seeing -5 degrees of retard, and in the table it's 25 they will put it at 30 total. Bottom line is if your seeing overe 3v of KS activity your running pretty hot or you have one really noisy motor. Some of you may want to look into this little gadget: http://www.turbomagazine.com/tech/0703_turp_knock_sensor_monitor/index.html
Josh, The DashHawk will monitor/log the knock sensor voltage as well as the amount of timing retard the PCM is applying in response to that. I don't know if the Predator will monitor/log this as well. I know it can report knock retard. You pointed to a knock sensor. I can't see much use for that unless you're going to use it as a sort of stethoscope and attach it where you think the false knock is coming from and then compare by sticking it on other spots on the engine. Or are you saying that the stock sensors are not that good?
