Frustrated with knock

[phantom6294]

Originally Posted by bova80 while cruising the car trys to find minimum best torque which allows the car to run with best emissions, yes emissions suck but it seems that mazda has the car trying to find the point where it has the least, so in doing so the ecu will sometimes hit knock points but i was reassured from christian that knock while cruising is not dangerous since there is no cylinder pressure.

I'd love to hear how the car in any possible way determines the Minimum Spark for Best Torque while driving down the highway. Unless the car has a built in dynamometer????? The only fathomable way for an ECU to determine MBT is to monitor road speed as it hunts down MBT. I guess if the car slows down slightly its because we veered too far from MBT, the car produced less power and thus we lost speed. I cannot fathom any other way for a car to sense it has reached MBT. Of course, did the car slow down because the road's grade changed or did it slow down because it veered too far from MBT?

The whole point to R&D is to determine the MBT in 'the lab' and once determined, program the ECU accordingly. It isn't really that hard to do either. MBT isn't fundamentally going to change. The ECU may ''tweak' the timing in real time to account for things such as IAT/BAT, knock retard, ECT, etc. For exmaple, higher IAT/BAT would increase the chance of detonation so the timing may be adjusted down accordingly. MBT didn't fundamentally change though. If the engine operating conditions change (high load, lower load, RPM increases, decrease), the ECU looks to a new cell in the timing map to be given the proper timing (most likely, MBT). An engine operating at a certain RPM, under certain load has a defined MBT. The ECU is going to have a timing map against which it pulls what the ECU programmer has empirically determined to be the MBT for certain operating conditions. The idea that the ECU can 1) determine when MBT has been reached and 2) is programmed to do so just doesn't pass the basic BS test.

From an emissions stand point... the car can't detect emissions beyond sampling the readings from pre and post cat O2 sensors to ensure the cat appears to be working. Once again, there is no way for the ECU to determine when it is producing more or less emissions. That would have to be determined 'in the lab' apriori by Mazda. Of course, once MBT is determined, through testing, it's a mere matter of ensuring the car is running at stoich AFRs. That is the perfect place for the engine to be for the best trade off between emissions and safety. So, if MBT is known a priori, then there is no need for the ECU to be on a scavenger hunt for a best emissions. All it needs to do is read from a timing map, set the timing for the current operating conditions (adjusted as according to other parameters as previously mentioned), and simultaneously inject the fuel according to the fuel map, trimming it as necessary as determined by the O2 sensor.

Two arguments have put forth: 1) the ECU continues to advance the timing until the onset of knock in an attempt to maximize fuel economy and/or 2) the ECU continually attempts to find MBT. I've provided evidence against argument 1 in the form of some logs and have yet to see evidence supporting it. If our ECU truely worked that way, it should be a piece of cake for someone to log this happening. Argument 2 makes little sense because I fail to see how an ECU can, in real time, determine when it has or hasn't reached MBT and lower emissions. Once the proper timing has been determined empircally, the only way for the ECU to really change emissions and/or economy is by the amount of fuel injected into the engine. Too little becomes dangerous as you run lean and increase NOx emissions. Too rich and you produce CO and HC. As such, the ECU trims the fuel according to readings from the O2 sensor.

Fundamentally, MBT isn't a phantom that the ECU needs to continually hunt down. It can easily be determined with the proper testing. MBT for given engine operating conditions is NOT going to change.

Of course, who knows... maybe Mazda was just being really lazy and decided writing complex algoritms to allow the ECU to somehow magically determine MBT on its own was better than just strapping the engine to a dynamometer and quickly figuring it out for themselves. Who knows....

[bova]

you don't need a dyno to find torque there bud, that kind of thing can be calculated. and MBT is basically the same as saying knock thresh hold so maybe if i said the car was looking for its knock thresh hold you would understand that.

with all the different sensors on this car i'm sure its quite easy to calculate the readings it needs to determine that stuff. its already been said this ecu is quite complex.

[Silver Ecstasy]

Did we ever determine if the KR was pre or post Fuel Pump/ECU update involved? Meaning, could the fuel pump replacement/ecu flash solve this issue? Contribute? Or has nothing to do with it at all?

Just an innocent thought (since I know nothing about KR, and don't even monitor it otherwise i'll just be paranoid all the time).

[08_ms3_gt]

i haz knock...

running the AP in monitor mode today, switched from boost to KR for shits.

normal driving + one quick 1-2-some of 3, look at the AP and suddenly i'm reading 2.5. WTF?!??!?!

so i got on it a few more times in a few gears and didn't find any readings above 0.3 at WOT.

damn phantom knock... i almost shit my pants.

[phantom6294]

Originally Posted by bova80 you don't need a dyno to find torque there bud, that kind of thing can be calculated.

Then please, explain to me how a car calculates torque in real time while the car is cruising down the interstate at a sustained load and sustained speed. Explain to me what sensor gives input to the ECU to indicate the engine is producing less torque once the ECU advances the timing beyond MBT.

Originally Posted by bova80 and MBT is basically the same as saying knock thresh hold so maybe if i said the car was looking for its knock thresh hold you would understand that.

WRONG. MBT does NOT equal the knock threshold. 1 degree beyond MBT does not magically mean you begin to knock. Yes, the further you advance beyond MBT the greater the likelihood of knock, but MBT and knock threshold are not directly related. Hell, it is possible that an engine is unable to operate at MBT because of conditions such as high IAT/BAT or even high ECT. Yet, MBT doesn't change for given engine operating conditions (Load, RPM). If X amount of air (and cooresponding fuel) enters a cylinder, it takes a a relatively fixed amount of time for that mixture to burn. The more dense that mixture is the less time, the less dense the more time. Further, there is a critical time during the piston's movement that you want peak pressure to occur so that said pressure is utilized most efficiently to push the piston down. This is normally around 14 ATDC. Peak pressure before or after this critical point results in wasted energy. Things like IAT/BAT, ECT, etc do not change when MBT is for a given load and RPM, it only changes the threshold of knock. Since detonation/knock is a function of too much heat/pressure for the fuel to withstand, high IAT/BAT directly lowers the threshold of knock. An abnormal ECT (an engine running hot) will also lowers the knock threshold because the intake charge would be heated more than if the engine were running at normal temps. Again, these things don't change how long it takes X amount of air/fuel to burn (and hence MBT), it only changes the knock threshold.

Further, as I've already said, advancing the timing beyond MBT yields NO advantages to engine output, and more than likely has negative effects on power output. With the spark advance set to MBT, the engine is most decidedly producing peak cylinder pressure at the correct time (normally around 14 degrees ATDC) which yields the most torque. If an engine is at MBT, it is making the most efficient use of the fuel being injected.

However, for the sake of argument with regards to emissions... It could be argued that advancing the timing beyond MBT may give the mixture more time to burn resulting in lower emissions. We all know the ECU uses the O2 sensor to indicate the AFR of the mixture. Well, if the engine was already operating at a stoich AFR, then that AFR would increase. The ECU would see a lean AFR and then add fuel. Why would we add fuel if we're attempting to maximize economy?? If however, the O2 was reporting a rich AFR condition, why not subtract fuel???? Wouldn't subtracting fuel be better for fuel economy and emissions???? You don't use timing to change the AFR, you change the amount of fuel going into the engine.

This is basic engine tuning techniques. The whole point to the concept of MBT and proper timing of peak pressure is that this is where the engine is MOST EFFICIENT and makes the most use of the energy stored in the fuel.

So, fundamentally:
- MBT can easily be determined during testing
- MBT is not going to change for an engine under specific operating conditions -- once it is determine you program the ECU as such
- There are no advantages when advancing the timing beyond MBT
- SO WHY DESIGN AN ECU TO ADVANCE THE TIMING BEYOND MBT????
- WHAT NEED IS THERE TO DESIGN AN ECU TO ATTEMPT THE IMPOSSIBLE TO DETERMINE MBT IN REAL TIME?

Look, you put forth two theories. One theory could be easily proven if you'd log the ECU continuously advancing the timing until the onset of knock and then the log showing a back and forth of spark advance as it is advanced and retarded. I've already posted portions of a log while cruising where no such thing happened. While I won't go so far as to say I've proved that theory of yours wrong, you've yet to post logs refuting mine. In response to your second theory, I discuss concepts like MBT and give what seems to me to be a logical refute to your theory... but your only response is...

Originally Posted by bova80 with all the different sensors on this car i'm sure its quite easy to calculate the readings it needs to determine that stuff. its already been said this ecu is quite complex.

[danesti]

i get audible knock in the 6 range if im in 4th gear or above. all the time... i wont even get close to using the torque of this engine around 2500.. ill step on it from 3500 and above because im paranoid.
im not one to play chicken with this shit

[bova]

I got the following info from this article: http://www.freepatentsonline.com/7289899.html

Electronic control systems associated with motor vehicle powertrains utilize certain data for performing control functions. Some data that is processed by control algorithms to develop control data is obtained directly from sensors while other data used by control algorithms is developed by processing certain data according to various other algorithms. Because driving conditions frequently change during typical trips made by a motor vehicle, data that is used by control algorithms must be updated with sufficient regularity to accurately reflect changes, and processing performed by an electronic control system must be fast enough to keep pace with changing conditions in order to assure optimal vehicle performance. Transients in engine/powertrain operation are especially challenging for control systems.

Because of particular dynamics of particular engine/powertrain components and/or the manner in which data for, or related to, them is developed, certain data that is important for optimal vehicle performance may not track operational changes, especially transients, with sufficient timeliness for optimal vehicle performance as those changes are occurring.

For example, shifting of an automatic transmission may be controlled by the amount of engine brake torque being produced by an internal combustion engine. When control of transmission shifting is electronic in nature, a processing system must develop accurate engine brake torque data for use by an automatic transmission controller to assure shifting at the proper time. Because shifting occurs as a consequence of the vehicle being accelerated or decelerated, the powertrain is in a dynamic state, and so is data related to it.

A known processing strategy for calculating engine brake torque data during transient operation, such as during engine acceleration, uses engine rotational inertia data and engine speed change (acceleration) data to compensate basic torque data that is being calculated according to what is essentially a steady-state algorithm. While the iteration rate of the steady-state algorithm, in conjunction with compensation for acceleration transients, might seem sufficiently fast to yield accurate brake torque data as the engine speed changes, various factors, such as component dynamics, affect accuracy and lead to the need for even better transient compensation.

One of the present inventor's prior patents, U.S. Pat. No. 6,584,391, relates to a system and method for engine torque calculation. The algorithm that is embodied in that system and method calculates gross engine torque by processing engine speed and fueling data. Net torque, i.e. brake torque, is calculated by calculating torque losses and subtracting those losses from calculated gross torque.

One component of torque loss is engine pumping loss, which is a function of different factors that include intake manifold pressure and exhaust backpressure. When the engine accelerates, an engine speed derivative term calculation contributes to transient compensation for engine torque.

In certain motor vehicles having turbocharged engines and automatic transmissions whose shift points are electronically controlled, it has been observed that shift quality may be affected by the nature of certain turbochargers, such as certain two-stage turbochargers. Through recognition that the quality of transmission shifts is attributable to the effect of such turbochargers on the prior method for brake torque calculation, the inventors have created a novel and improved strategy for calculating engine brake torque.

Here there is even a patent on it: http://www.freepatentsonline.com/6584391.html

and if you ask me this log i did which i posted on the first page showed spark advance increase and then the on sight of knock then shows it decreasing. rpms are bascially the same as well as throttle percentage. all i have seen you do is shoot down everyone's theories and never come up with one of you own.

[Deadman]

this car is dumb LOL

I am cruising home in 6th @ about 85-90mph for probably 20mins straight. Every so often i would see the DH go BLUE and alert me Kr 3.9 4.1 in that area... for no reason at all. Seems to do it if im around 3200-3500rpm. It would do it for no more than a split second.. all i can really see is the screen go blue thats all.

[pidass]

Originally Posted by 08_ms3_gt i haz knock... running the AP in monitor mode today, switched from boost to KR for shits. normal driving + one quick 1-2-some of 3, look at the AP and suddenly i'm reading 2.5. WTF?!??!?! so i got on it a few more times in a few gears and didn't find any readings above 0.3 at WOT. damn phantom knock... i almost shit my pants.

I've seen mine spike to 5.9 while just cruising on the interstate

[Phantom]

Originally Posted by Deadman this car is dumb LOL I am cruising home in 6th @ about 85-90mph for probably 20mins straight. Every so often i would see the DH go BLUE and alert me Kr 3.9 4.1 in that area... for no reason at all. Seems to do it if im around 3200-3500rpm. It would do it for no more than a split second.. all i can really see is the screen go blue thats all.

Always seems to be between 3k and 3.5k RPM.

[bova]

Originally Posted by phantom6294 Then please, explain to me how a car calculates torque in real time while the car is cruising down the interstate at a sustained load and sustained speed. Explain to me what sensor gives input to the ECU to indicate the engine is producing less torque once the ECU advances the timing beyond MBT. WRONG. MBT does NOT equal the knock threshold. 1 degree beyond MBT does not magically mean you begin to knock. Yes, the further you advance beyond MBT the greater the likelihood of knock, but MBT and knock threshold are not directly related. Hell, it is possible that an engine is unable to operate at MBT because of conditions such as high IAT/BAT or even high ECT. Yet, MBT doesn't change for given engine operating conditions (Load, RPM). If X amount of air (and cooresponding fuel) enters a cylinder, it takes a a relatively fixed amount of time for that mixture to burn. The more dense that mixture is the less time, the less dense the more time. Further, there is a critical time during the piston's movement that you want peak pressure to occur so that said pressure is utilized most efficiently to push the piston down. This is normally around 14 ATDC. Peak pressure before or after this critical point results in wasted energy. Things like IAT/BAT, ECT, etc do not change when MBT is for a given load and RPM, it only changes the threshold of knock. Since detonation/knock is a function of too much heat/pressure for the fuel to withstand, high IAT/BAT directly lowers the threshold of knock. An abnormal ECT (an engine running hot) will also lowers the knock threshold because the intake charge would be heated more than if the engine were running at normal temps. Again, these things don't change how long it takes X amount of air/fuel to burn (and hence MBT), it only changes the knock threshold. Further, as I've already said, advancing the timing beyond MBT yields NO advantages to engine output, and more than likely has negative effects on power output. With the spark advance set to MBT, the engine is most decidedly producing peak cylinder pressure at the correct time (normally around 14 degrees ATDC) which yields the most torque. If an engine is at MBT, it is making the most efficient use of the fuel being injected. However, for the sake of argument with regards to emissions... It could be argued that advancing the timing beyond MBT may give the mixture more time to burn resulting in lower emissions. We all know the ECU uses the O2 sensor to indicate the AFR of the mixture. Well, if the engine was already operating at a stoich AFR, then that AFR would increase. The ECU would see a lean AFR and then add fuel. Why would we add fuel if we're attempting to maximize economy?? If however, the O2 was reporting a rich AFR condition, why not subtract fuel???? Wouldn't subtracting fuel be better for fuel economy and emissions???? You don't use timing to change the AFR, you change the amount of fuel going into the engine. This is basic engine tuning techniques. The whole point to the concept of MBT and proper timing of peak pressure is that this is where the engine is MOST EFFICIENT and makes the most use of the energy stored in the fuel. So, fundamentally: - MBT can easily be determined during testing - MBT is not going to change for an engine under specific operating conditions -- once it is determine you program the ECU as such - There are no advantages when advancing the timing beyond MBT - SO WHY DESIGN AN ECU TO ADVANCE THE TIMING BEYOND MBT???? - WHAT NEED IS THERE TO DESIGN AN ECU TO ATTEMPT THE IMPOSSIBLE TO DETERMINE MBT IN REAL TIME? Look, you put forth two theories. One theory could be easily proven if you'd log the ECU continuously advancing the timing until the onset of knock and then the log showing a back and forth of spark advance as it is advanced and retarded. I've already posted portions of a log while cruising where no such thing happened. While I won't go so far as to say I've proved that theory of yours wrong, you've yet to post logs refuting mine. In response to your second theory, I discuss concepts like MBT and give what seems to me to be a logical refute to your theory... but your only response is...

i also want to add that after having a discussion withchristian from cobb on you said, you are in fact wrong, and this car does find data based on MBT and that JDM ECU's have been able to do this since the 90s with complex algorithms (Subaru, Nissan, Mitsubishi,.....) so maybe you should go find out facts more before you start spoutin off.

[phantom6294]

Originally Posted by bova80 i also want to add that after having a discussion with christian from cobb on you said, you are in fact wrong, and this car does find data based on MBT and that JDM ECU's have been able to do this since the 90s with complex algorithms (Subaru, Nissan, Mitsubishi,.....) so maybe you should go find out facts more before you start spoutin off.

First, I've never conclusively said how exactly our ECU works. Go re-read my posts, the fundamental principles are sound, or perhaps you need to work on your reading comprehension? I've never once stated our ECU doesn't use these techniques. I was wrong on the point that engine could calculate engine torque in real time, but it doesn't change the fundamental principles with which I argue. What I have done, is taken direct issue with your reasoning as to why you think our ECU might use these techniques and how it relates to knock.

Let's go back and review some of your statements:

Originally Posted by bova80 i believe part throttle knock is how the car maximizes its fuel efficiency... the car keeps advancing timing for the best gas mileage and then it gets some knock and backs it down a bit and the cycle continues always trying to reach that optimal fuel efficiency ... MBT is basically the same as saying knock thresh hold

I can only infer, and I think its fair conclusion, that you believe that the car continues to advance the spark timing until the onset of knock in order to locate MBT. The fundamental flaw in that logic is that MBT and knock threshold are directly related; THEY AREN'T. An engine, under certain operating conditions could operate MANY degrees beyond MBT without knock. Also, an engine,under certain operating conditions, might be unable to operate at MBT. For example, if you increase the temperature of the intake charge (and hold air mass constant) you decrease the knock threshold. Why, because knock is a direct effect of too much heat/pressure in during combustion. If the intake charge starts at a higher temperature, it reaches a higher temeprature during the combustion event. You do not, however, change the MBT. Furthermore, operating the engine beyond MBT (i.e. advancing the timing beyond MBT), in and of itself, will not change the fuel economy. IT CAN'T. If you do not change the amount of fuel entering the engine, you cannot change your fuel economy. UNLESS, the engine was not already operating at MBT. If the engine is not operating at MBT, you advance the timing to MBT, then adjust the fuel trim to the desired AFR. That's how you maximize fuel economy -- you operate the engine at MBT and trim fuel to obtain desired AFRs. It really is that simple. For a sustained load and RPM (cruising on flat surface) the MBT will not change.

Furthermore, you have yet to provide ANY evidence that the ECU does in fact continue to advance the timing until the onset of knock. I post portions of a log cruising at a sustained speed at sustained loads. Is there any continued spark advance??????? NO. Is there any onset of knock???? NO. So, please, explain why my ECU isn't continuing to advance the timing while I cruise along???

You have yet to provide any log evidence or any logical reasoning behind any of your statements. Do me a favor and actually read, then comprehend my arguments and refute my reasoning and evidence in the form of logs and spend less time responding with "Christian said so..."

Hey, I just got off the phone with Mazda in Japan. Turns out, I was right, you're wrong, end of conversation.

[bova]

did you not see my log?????? timing was advanced and then knock occurred, what more do you want to see? depending on which map the ecu looks at on the current conditions of timing in closed loop will be how it advances the timing, yes there are more than 1 timing map in closed loop and it learns and modifies them, if you don't believe just wait until cobb releases their documentation on this because i've already seen it.

but whatever you believe what you want and i'll believe what i want. the car knocks while cruising and that is all there is to it. its not dangerous because its spark knock and not detonation knock. spark knock that generally produces the light pinging sound at light throttle settings and lean mixtures and detonation knock under heavy load and cylinder pressure that is severely destructive in that it can collapse rods and burn through a piston before you can lift off the throttle

[GongShow]

Not to be a knob, but is this accomplishing anything other than winning the "I'm right" contest? I appreciate the depth of analysis and commitment to knowledge, but this is turning into a typical car forum (which I hope it doesn't become).

You can disagree without taking "there".

[phantom6294]

Originally Posted by bova80 did you not see my log?????? timing was advanced and then knock occurred, what more do you want to see? depending on which map the ecu looks at on the current conditions of timing in closed loop will be how it advances the timing, yes there are more than 1 timing map in closed loop and it learns and modifies them, if you don't believe just wait until cobb releases their documentation on this because i've already seen it. but whatever you believe what you want and i'll believe what i want. the car knocks while cruising and that is all there is to it. its not dangerous because its spark knock and not detonation knock. spark knock that generally produces the light pinging sound at light throttle settings and lean mixtures and detonation knock under heavy load and cylinder pressure that is severely destructive in that it can collapse rods and burn through a piston before you can lift off the throttle

I didn't see your log because you edited your post and I didn't think to go back and re-read that particular post. However, I've analyzed said log and here are my thoughts. This is mere 17 seconds with very little sustained load. Not very convincing. Please tell me this isn't the best you can do????



You seem to have a fundamental lack of understanding of the correlation between spark advance and engine load. It ain't that hard of a concept to grasp. At a sustained RPM, engine load increases, spark advance needs to decrease. Engine load decreases, spark advance needs to increase. Your log fundamentally shows EXACTLY WHAT I'VE BEEN SAYING. The only significant period of sustained load has sustain spark advance. IMAGINE THAT.

NO WHERE IN YOUR LOG, IS THERE SUSTAINED LOAD WITH INCREASED SPARK ADVANCE TO THE ONSET OF KNOCK.

The first instance of KR in your log seems random but spark advance was not changing when it occurred. Knock was detected, occurred, the ECU pulled timing accordingly, and then you took your foot off the throttle. The second instance occurs as you are opening the throttle -- the ECU initially sets the timing per the operating conditions (as you leave open loop and enter closed loop) and then begins to decrease spark advance because engine load increases. The third instance as spark advance is being decreased. (Didn't you say spark advanced is increased until the onset of knock). Spark advance is dropping because, engine load is increasing!!!

Yes, the car knocks while cruising. Thank you Captain Obvious. I've never disputed that in the least!!! I've never said it was dangerous and repeatedly told people it can ignored. I'm just tired of your half baked theories that do not follow basic spark timing principles.

SUSTAINED LOAD, SUSTAINED RPM, SUSTAINED SPARK.

I'm still waiting for a log showing sustainED load and the spark advance being advanced to the onset of knock. Until then.... I won't hold my breath.

[bova]

so why dont' you go do it?

i'm done with these knock threads, quite frankly knock while cruising doesn't really matter cause it happens to everyone. you have good logic but yet aren't proving anything either.

[tagkat]

Here's something interesting. And to throw a wrench into it ....Noticed today that 60 degrees + no knock. 59 degrees below instant knock. The highest knock I had at 60+ was 1.3 after 60 miles. Below 60 degrees highest on same path and similar speeds 7kr.
Same day only diff was mainly temp.

[MS3 Ventura]

For those people that bent a rod, and had the dreaded load ticking sound before popping their engine, what did they see when monitoring KR? Were the KR readings constantly above zero? Or, did KR spike excessively?