The EZK computer- safeguarding or self tuning?
09-28-2015, 03:01 PM
#1
Rennlist Member
Thread Starter
The EZK computer- safeguarding or self tuning?
Reading the current thread regarding 32V supercharging options and wisdom of doing such prompted me to sit down and write this post to see what opinions it solicits.
In the oil industry I work in modern design philosophy specifically tries to differentiate between systems that control and systems that safeguard and "never the two shall meet" is the guiding mantra. In earlier times safeguarding was implemented by totally different systems but nowadays the systems used for control can be used for safeguarding and vice versa with some limitations. The safeguarding system is based on integrity of operation and the consequence of failure is taken into consideration in determining how much money to spend on making such systems as safe as reasonably practical. Similarly a control system is predicated more on accuracy and ease of operation given that failure is backed up by the safeguarding system as the feature of last resort.
Superficially the EZK is portrayed as a safeguarding system in that according to Porsche literature it can pull back timing by up to 9 degrees of advance if one has dodgy fuel, and you automatically lose 6 degrees of timing if one of the knock sensors or the Hall trigger fails. I do not know why the Hall trigger features in this equation I just know that is what happens.
I got into the Sharktuning business after John introduced the mark 2 version with the PEMS. This is a great system in my opinion but the dilemma is this potentially creates as many problems as it solves because anyone with a hint of intelligence overnight becomes much more "aware" of what is going on inside his motor. Knowledge is very much a two sided weapon in that awareness confirms what is right but also highlights what is less than perfect and thus leaves doubt as to what to do about what one sees.
I live in an environment where the gasoline is not the best in the world. Our gasoline is clean and well regulated but it is limited to 95 RON.
"Super" gasolines typically are rated at 98 RON [that is the case in the UK & Dubai for instance]. When we go to Dubai and fill up with Super 98 we feel a difference in performance on a stock car- why?
My take on this is that the stock tuning is not as limp as some seem to think it is. Thus with our local gasoline and typical summer time temps of 44C- sometimes higher] the engine is being pulled back just that normally you would never know what is happening- no pinging or funny noises.
My sharktuning experience over a number of years was very much a mixed affair. On the one hand I was expecting to simply dial in more advance and feel more power but it did not work like that. In fact on full load and high end rpms I found myself pulling timing. Discussed the matter extensively with the likes of John, Jim, Louie, Colin and more recently with Ken [Porken] and Tuomo, and eventually came to some interesting conclusions.
First of all Sharktuner told me that the stock maps are designed for 98 octane. It also told me that for models with cats fitted they seemed to run slightly more advance. Initially I found this a little strange but when I investigated this I learnt [rightly or wrongly] that a slower exhaust needs more advance as the mix takes longer to burn. Perhaps this in part explains why I had to pull some timing but then of course I have a faster than stock exhaust so maybe that explained why max advance was limited to about 24 degrees. Then a short while ago John Gill in upside down land stated that utilising the sharktuner system would typically limit advance to about 24 degrees- exactly what I had experienced.
The other "problem I found was that the inner cylinders were the ones controlling the advance I could apply, specifically 6,2, 3 & 7 in that sequence. Initially I could not resolve this but when Ken came up with his excellent and very generous posts covering his work on the S4 chips and the the control system in general, he offered an interesting patch wherein I could differentially designate the timing on each cylinder. What I found was that by pulling about 3 degrees on the inner cylinders I started to see/find limits wherein I was starting to see knock events on all cylinders.
Reading the thread on supercharging today I then saw an interesting post from Dr Greg Brown commenting on how knock events do in fact occur on a regular basis but presumably, just as I commented earlier, most folks are simply not aware of what has been going on in the computers brain because there are no external symptoms. Sharktuner catches all these events and tells you which cylinder is being impacted and how much retard has been applied to that cylinder to calm things down.
The other interesting thing I found was that whereas AFR does impact the tendency to knock especially if lean, I found that once you get down to about AFR 13, any richer has little impact in reducing knock tendency which at this point is effectively controlled by removing advance.
Another observation I made was that the EZK system is very sensitive and although consistent on 6 & 2, I saw signs that suggested there are also quite a lot of "noise" related events that are probably false signals- this I noted by setting the step retard from the stock 3 degrees to as little as 0.4 degrees- the clue being that upon onset the retard applied was the same as the step value and often no more.- suspicious?
Finally, once I had achieved a balance across the board [i.e. 3 degrees differential retard applied to the inner cylinders] I then applied 5 degrees of advance across the board on the higher load cells and removed the differential correction. Sure enough with sharktuner engaged I picked up knock events on the inner cylinders and then did a run with ST2 disengaged. I could not hear or detect anything going on at all but the motor was noticeably more lively.
This begs the question as to whether the EZK should be set to zero knocks for safety or whether it should be allowed to control knock events by over tuning to some threshold such as say something in the range of 3 degrees up to perhaps even 5 degrees? Note I do not mean compared to stock maps I am talking specifically relative timing compared to what superficially appears to be a stable tune with no knocks at all that considers the gasoline being used and the local running environment [air temps].
I also noted that fitting plenum spacers changed the tendency to first ping from No2 to No6 cylinder. Running with a permanent timing differential between inner and outer cylinders is not realistic because it improves some operating cells but probably has a deleterious effect or more cells. For this reason I think Ken was very much onto something when he created the ability for the system to self learn individual load cells and apply such to different groups of cylinders.
Another point that was noted in another thread was that the EZK system although well out of date computing technology wise was still a very quick system in its day and even at 6k rpm it is fast enough to apply a retard correction by the time the next firing stroke occurs.Another point that was noted in another thread was that the EZK system although well out of date computing technology wise was still a very quick system in its day and even at 6k rpm it is fast enough to apply a retard correction by the time the next firing stroke occurs.
Trust the above of interest and it will be interesting to learn about what others have found utilising sharktuner or systems that permit individual cylinder tuning. Whether or not access to a dyno would have eliminated some of these issues is a point of some interest to me.
Rgds
Fred
In the oil industry I work in modern design philosophy specifically tries to differentiate between systems that control and systems that safeguard and "never the two shall meet" is the guiding mantra. In earlier times safeguarding was implemented by totally different systems but nowadays the systems used for control can be used for safeguarding and vice versa with some limitations. The safeguarding system is based on integrity of operation and the consequence of failure is taken into consideration in determining how much money to spend on making such systems as safe as reasonably practical. Similarly a control system is predicated more on accuracy and ease of operation given that failure is backed up by the safeguarding system as the feature of last resort.
Superficially the EZK is portrayed as a safeguarding system in that according to Porsche literature it can pull back timing by up to 9 degrees of advance if one has dodgy fuel, and you automatically lose 6 degrees of timing if one of the knock sensors or the Hall trigger fails. I do not know why the Hall trigger features in this equation I just know that is what happens.
I got into the Sharktuning business after John introduced the mark 2 version with the PEMS. This is a great system in my opinion but the dilemma is this potentially creates as many problems as it solves because anyone with a hint of intelligence overnight becomes much more "aware" of what is going on inside his motor. Knowledge is very much a two sided weapon in that awareness confirms what is right but also highlights what is less than perfect and thus leaves doubt as to what to do about what one sees.
I live in an environment where the gasoline is not the best in the world. Our gasoline is clean and well regulated but it is limited to 95 RON.
"Super" gasolines typically are rated at 98 RON [that is the case in the UK & Dubai for instance]. When we go to Dubai and fill up with Super 98 we feel a difference in performance on a stock car- why?
My take on this is that the stock tuning is not as limp as some seem to think it is. Thus with our local gasoline and typical summer time temps of 44C- sometimes higher] the engine is being pulled back just that normally you would never know what is happening- no pinging or funny noises.
My sharktuning experience over a number of years was very much a mixed affair. On the one hand I was expecting to simply dial in more advance and feel more power but it did not work like that. In fact on full load and high end rpms I found myself pulling timing. Discussed the matter extensively with the likes of John, Jim, Louie, Colin and more recently with Ken [Porken] and Tuomo, and eventually came to some interesting conclusions.
First of all Sharktuner told me that the stock maps are designed for 98 octane. It also told me that for models with cats fitted they seemed to run slightly more advance. Initially I found this a little strange but when I investigated this I learnt [rightly or wrongly] that a slower exhaust needs more advance as the mix takes longer to burn. Perhaps this in part explains why I had to pull some timing but then of course I have a faster than stock exhaust so maybe that explained why max advance was limited to about 24 degrees. Then a short while ago John Gill in upside down land stated that utilising the sharktuner system would typically limit advance to about 24 degrees- exactly what I had experienced.
The other "problem I found was that the inner cylinders were the ones controlling the advance I could apply, specifically 6,2, 3 & 7 in that sequence. Initially I could not resolve this but when Ken came up with his excellent and very generous posts covering his work on the S4 chips and the the control system in general, he offered an interesting patch wherein I could differentially designate the timing on each cylinder. What I found was that by pulling about 3 degrees on the inner cylinders I started to see/find limits wherein I was starting to see knock events on all cylinders.
Reading the thread on supercharging today I then saw an interesting post from Dr Greg Brown commenting on how knock events do in fact occur on a regular basis but presumably, just as I commented earlier, most folks are simply not aware of what has been going on in the computers brain because there are no external symptoms. Sharktuner catches all these events and tells you which cylinder is being impacted and how much retard has been applied to that cylinder to calm things down.
The other interesting thing I found was that whereas AFR does impact the tendency to knock especially if lean, I found that once you get down to about AFR 13, any richer has little impact in reducing knock tendency which at this point is effectively controlled by removing advance.
Another observation I made was that the EZK system is very sensitive and although consistent on 6 & 2, I saw signs that suggested there are also quite a lot of "noise" related events that are probably false signals- this I noted by setting the step retard from the stock 3 degrees to as little as 0.4 degrees- the clue being that upon onset the retard applied was the same as the step value and often no more.- suspicious?
Finally, once I had achieved a balance across the board [i.e. 3 degrees differential retard applied to the inner cylinders] I then applied 5 degrees of advance across the board on the higher load cells and removed the differential correction. Sure enough with sharktuner engaged I picked up knock events on the inner cylinders and then did a run with ST2 disengaged. I could not hear or detect anything going on at all but the motor was noticeably more lively.
This begs the question as to whether the EZK should be set to zero knocks for safety or whether it should be allowed to control knock events by over tuning to some threshold such as say something in the range of 3 degrees up to perhaps even 5 degrees? Note I do not mean compared to stock maps I am talking specifically relative timing compared to what superficially appears to be a stable tune with no knocks at all that considers the gasoline being used and the local running environment [air temps].
I also noted that fitting plenum spacers changed the tendency to first ping from No2 to No6 cylinder. Running with a permanent timing differential between inner and outer cylinders is not realistic because it improves some operating cells but probably has a deleterious effect or more cells. For this reason I think Ken was very much onto something when he created the ability for the system to self learn individual load cells and apply such to different groups of cylinders.
Another point that was noted in another thread was that the EZK system although well out of date computing technology wise was still a very quick system in its day and even at 6k rpm it is fast enough to apply a retard correction by the time the next firing stroke occurs.Another point that was noted in another thread was that the EZK system although well out of date computing technology wise was still a very quick system in its day and even at 6k rpm it is fast enough to apply a retard correction by the time the next firing stroke occurs.
Trust the above of interest and it will be interesting to learn about what others have found utilising sharktuner or systems that permit individual cylinder tuning. Whether or not access to a dyno would have eliminated some of these issues is a point of some interest to me.
Rgds
Fred
Last edited by FredR; 09-28-2015 at 03:27 PM.
09-28-2015, 04:12 PM
#2
Inventor
Rennlist Member
Rennlist Member
Nice read, Fred!
IIRC, the hall sensor is needed to determine whether to use the front or rear knock sensor.
IIRC, the hall sensor is needed to determine whether to use the front or rear knock sensor.
09-28-2015, 04:19 PM
#3
Rennlist Member
............ enabling the EZK to identify which cylinder is firing... it that cylinder has a knock event, then the EZK retards the ignition on that individual cylinder only. Clever stuff !
09-28-2015, 07:29 PM
#4
Rennlist Member
The stock knock system does its job very well , but the issue is which is a knock event and what is not. Engines are particular noisy units , listening to them in Dyno that I do professionally , you will be amazed , my booth Dyno knock detection system, uses cylinder positioning to attempt to help the likelihood of encroaching detonation. THe stock system is over sensitive , that is why companies like Motec etc , dont use it , there is a lot of variables, the ability to discern between noise is only just being explored now . In the dyno you will see torque plateau from advancing the ignition, to match the volumetric efficiency , of the engine , continuing to add ignition advance will only see torque starting to decrease, and due to the fact that the 928 crankshaft bearings are rubbish, engine damage ensures quickly in the blink of eye.
As good example my GT engine runs 29 degrees engines advance at 6000 rpm, 28 at the soft cut at 7000, full cut at 7200 . It has been tracked in the local sprints for 3 years , with slicks and is very fast, (to fast as it has no roll cage)
The best I have been able to get with the stock system is 23 degrees before the system starts to retard.
Interesting that the delco system used in the holden (GM) also have the same issues with over sensitive retard, they make their power with cubic capacity, ie 6 and 6.2 litre. Once of the trick in tuning these units is too make the system less sensitive, ie reduce the voltage reacting point, they respond instantly, which you see in the increase in the amount of torque produced.
As good example my GT engine runs 29 degrees engines advance at 6000 rpm, 28 at the soft cut at 7000, full cut at 7200 . It has been tracked in the local sprints for 3 years , with slicks and is very fast, (to fast as it has no roll cage)
The best I have been able to get with the stock system is 23 degrees before the system starts to retard.
Interesting that the delco system used in the holden (GM) also have the same issues with over sensitive retard, they make their power with cubic capacity, ie 6 and 6.2 litre. Once of the trick in tuning these units is too make the system less sensitive, ie reduce the voltage reacting point, they respond instantly, which you see in the increase in the amount of torque produced.
09-28-2015, 09:22 PM
#5
Nordschleife Master
Catalytic converters do not flow as well, because of this the cylinders are not evacuated as well as with a free flow system (x-pipe), because of this you can run more advance.
More charge (fresh air and fuel) = less ignition advance.
There are other factors on the ignition system. I used to think it was over sensitive too. More work and understanding has shown otherwise.
More charge (fresh air and fuel) = less ignition advance.
There are other factors on the ignition system. I used to think it was over sensitive too. More work and understanding has shown otherwise.
09-29-2015, 11:08 AM
#6
Rennlist Member
Thread Starter
Some interesting thoughts already. My specific problem lies in trying to tune the motor without the benefit of a dyno. Presumably on the dyno you dial in more [or less] until the readings max out- not a simple task but systematic at least.
The whole point of my efforts was to test whether I could in fact tune the motor to some extent such that I might be able to add something like a supercharger, turbo or stroker in due course and do so knowing that I could control the changes.
Initially I could not get my head around what was going on and figured there had to be something wrong with my motor. Oil consumption was excessive but eventually put that down to the mistake of modifying my breather system to the GTS version. When I isolated the oily recycle I got some improvement and now I have a closed system that as yet I have not been able to fully evaluate [I have been off the road for a few weeks with other problems] but initial evaluation was positive.
During my initial attempts at sharktuning both john and Jim C advised that some degree of retard in the region of two to three degrees would be "acceptable". Knowing the debate about the stock bearing metal being relatively soft and the known weakness of the 2/6 journal I put two and two together and wondered if the 2/6 failure may be linked to the fact that I was seeing knock events on No6 and No2 leading the charge and this lead me to be somewhat cautious with advance.
However I got somewhat perplexed when reading about Ken's work on the early 32V models. They have no anti knock protection but presumably have the same metallurgy in the big end shells. If the system could not take tuning then logic might suggest that his chips could possibly hammer the crap out of the big ends in these models-but that seemingly did not happen.
I have never heard any "knocking" in my motor although years ago I remember GTS owners complaining about such to the local agents. Now i suspect that was a mix of the local gasoline, the breather system ingesting oil/building deposits and high ambient temperatures along the lines Greg Brown was writing about the other day.
My motor always gave good compression readings and when I eventually carried out a leak down test a few months ago the results were nigh on perfect. I suppose I could still have a problem with weak oil control rings but the evidence to date does not warrant a motor strip down. The mileage covered to date is currently running around the 100k miles so nothing mega as it were.
My problem now is trying to understand what is signal and what is noise and how much I can afford to dial in without doing any damage given the motor seems to want more advance. Clearly I cannot dial in more advance on some cylinders so I am very much tempted to take Ken up on his offer and try to run with his modified EZK programme but even that relies on signals from the EZK that may be noise.
The alternative is to dial in more advance and simply let the EZK take out what it perceives is not desirable hence the title of this thread. Given that I am probably within 3 degrees of an optimal setting and the system has 9 degrees of correction available to it, my logic tells me if i dial in about 5 degrees of advance over and above what I [conservatively] have at the moment then I should be able to let the system "self optimise".
What I have also concluded is that tuning is a very interactive thing. Start playing with the exhaust, x-pipes, plenum spacers and inlet system as I have, throw in a set of GTS inlet cams etc etc and there are many interactions in play that need custom optimisation. I am also convinced that Ken was onto something when he was trying to smooth out the MAF signal. I have a feeling that system is also too reactive and gives transient outputs that that are just not real world accurate. I suspect that alone may explain why stock fuelling on top end is as rich as it is. I found it quite difficult to lean out the AFR at high rpms and get accurate results- it was as if the beast had a mind of its own. Louie once advised me that the 928 makes max power at AFR of about 13.3 due to having an efficient cylinder head design whereas stock fuelling generates an AFR of about 12.4.
I have enjoyed trying to fathom this lot out and concluded that it is not something you should take on ***** nilly. Thank goodness we have some pretty smart resources available on this list.
Regards
Fred
The whole point of my efforts was to test whether I could in fact tune the motor to some extent such that I might be able to add something like a supercharger, turbo or stroker in due course and do so knowing that I could control the changes.
Initially I could not get my head around what was going on and figured there had to be something wrong with my motor. Oil consumption was excessive but eventually put that down to the mistake of modifying my breather system to the GTS version. When I isolated the oily recycle I got some improvement and now I have a closed system that as yet I have not been able to fully evaluate [I have been off the road for a few weeks with other problems] but initial evaluation was positive.
During my initial attempts at sharktuning both john and Jim C advised that some degree of retard in the region of two to three degrees would be "acceptable". Knowing the debate about the stock bearing metal being relatively soft and the known weakness of the 2/6 journal I put two and two together and wondered if the 2/6 failure may be linked to the fact that I was seeing knock events on No6 and No2 leading the charge and this lead me to be somewhat cautious with advance.
However I got somewhat perplexed when reading about Ken's work on the early 32V models. They have no anti knock protection but presumably have the same metallurgy in the big end shells. If the system could not take tuning then logic might suggest that his chips could possibly hammer the crap out of the big ends in these models-but that seemingly did not happen.
I have never heard any "knocking" in my motor although years ago I remember GTS owners complaining about such to the local agents. Now i suspect that was a mix of the local gasoline, the breather system ingesting oil/building deposits and high ambient temperatures along the lines Greg Brown was writing about the other day.
My motor always gave good compression readings and when I eventually carried out a leak down test a few months ago the results were nigh on perfect. I suppose I could still have a problem with weak oil control rings but the evidence to date does not warrant a motor strip down. The mileage covered to date is currently running around the 100k miles so nothing mega as it were.
My problem now is trying to understand what is signal and what is noise and how much I can afford to dial in without doing any damage given the motor seems to want more advance. Clearly I cannot dial in more advance on some cylinders so I am very much tempted to take Ken up on his offer and try to run with his modified EZK programme but even that relies on signals from the EZK that may be noise.
The alternative is to dial in more advance and simply let the EZK take out what it perceives is not desirable hence the title of this thread. Given that I am probably within 3 degrees of an optimal setting and the system has 9 degrees of correction available to it, my logic tells me if i dial in about 5 degrees of advance over and above what I [conservatively] have at the moment then I should be able to let the system "self optimise".
What I have also concluded is that tuning is a very interactive thing. Start playing with the exhaust, x-pipes, plenum spacers and inlet system as I have, throw in a set of GTS inlet cams etc etc and there are many interactions in play that need custom optimisation. I am also convinced that Ken was onto something when he was trying to smooth out the MAF signal. I have a feeling that system is also too reactive and gives transient outputs that that are just not real world accurate. I suspect that alone may explain why stock fuelling on top end is as rich as it is. I found it quite difficult to lean out the AFR at high rpms and get accurate results- it was as if the beast had a mind of its own. Louie once advised me that the 928 makes max power at AFR of about 13.3 due to having an efficient cylinder head design whereas stock fuelling generates an AFR of about 12.4.
I have enjoyed trying to fathom this lot out and concluded that it is not something you should take on ***** nilly. Thank goodness we have some pretty smart resources available on this list.
Regards
Fred
10-18-2015, 11:11 AM
#7
Rennlist Member
Thread Starter
Just recycling this thread to see if I can generate any interest from the shark tuning community at least.
What prompted this post was the recent thread on supercharging kits available and their relative merits. Something I spotted a couple of days ago caught my eye and that was how Carl's basic kits seemed to be somewhat dismissed because they did not have custom tuned eeproms. Initially I might also have agreed with this criticism but given that John [Speake] reckons that when boosting a car a good safe starting point is to pull timing to the tune of 1 degree per psi of boost I had to rethink this one a bit. Carl responded to the critiques posed by stating that in his experience his basic kits did not need such. Given that this kit runs about 5 or 6 psi of boost might suggest pulling timing by about 5 degrees and this is well within the capability of the EZK system that can retard individual cylinders by up to 9 degrees. Carl has no shortage of experience supercharging cars so maybe he is relying on the EZK system as a form of self regulating control and maybe this is quite acceptable- thoughts? Of course if the working margin is being used up it leaves less room if/when poor gasoline is in the tank but realistically is that likely in a place like the USA?
Thus, all things being equal, if the EZK system is good enough to use as a control system it will pull some timing as and when needed and as I have stated previously, the only time I have ever been aware of the EZK actually doing something is when it is on sharktuner and I can see from the datalogs what has been going on [or it pings when on line]. During normal running one would not know it was in fact playing along. For sure no audible knocking or rattling going on and if it was I am sure someone like Carl would recognise something was not right. Start playing along with 10 psig of boost and a different matter altogether.
Porsche themselves designed the system so that if you run into a location that has poor gasoline when travelling transcontinental it can still run safely on a lesser octane fuel if that is all that is available. My 928 when running stock chip sets clearly ran better when fuelled with super 98 up in Dubai. This suggested to me that when using our native 95 RON fuel it might kick in with some retard-and it does. My motor seems to like more advance on top end than I have currently programmed and thus why I am keen to get a consensus as to what we can and cannot/should not do with the EZK.
I have concluded that without a dyno to support me, the best thing I can do is tune the system until I see no pings and then add 4 or 5 degrees of advance and leave the system to "do its thing" as needed.
Further thoughts?
Rgds
Fred
What prompted this post was the recent thread on supercharging kits available and their relative merits. Something I spotted a couple of days ago caught my eye and that was how Carl's basic kits seemed to be somewhat dismissed because they did not have custom tuned eeproms. Initially I might also have agreed with this criticism but given that John [Speake] reckons that when boosting a car a good safe starting point is to pull timing to the tune of 1 degree per psi of boost I had to rethink this one a bit. Carl responded to the critiques posed by stating that in his experience his basic kits did not need such. Given that this kit runs about 5 or 6 psi of boost might suggest pulling timing by about 5 degrees and this is well within the capability of the EZK system that can retard individual cylinders by up to 9 degrees. Carl has no shortage of experience supercharging cars so maybe he is relying on the EZK system as a form of self regulating control and maybe this is quite acceptable- thoughts? Of course if the working margin is being used up it leaves less room if/when poor gasoline is in the tank but realistically is that likely in a place like the USA?
Thus, all things being equal, if the EZK system is good enough to use as a control system it will pull some timing as and when needed and as I have stated previously, the only time I have ever been aware of the EZK actually doing something is when it is on sharktuner and I can see from the datalogs what has been going on [or it pings when on line]. During normal running one would not know it was in fact playing along. For sure no audible knocking or rattling going on and if it was I am sure someone like Carl would recognise something was not right. Start playing along with 10 psig of boost and a different matter altogether.
Porsche themselves designed the system so that if you run into a location that has poor gasoline when travelling transcontinental it can still run safely on a lesser octane fuel if that is all that is available. My 928 when running stock chip sets clearly ran better when fuelled with super 98 up in Dubai. This suggested to me that when using our native 95 RON fuel it might kick in with some retard-and it does. My motor seems to like more advance on top end than I have currently programmed and thus why I am keen to get a consensus as to what we can and cannot/should not do with the EZK.
I have concluded that without a dyno to support me, the best thing I can do is tune the system until I see no pings and then add 4 or 5 degrees of advance and leave the system to "do its thing" as needed.
Further thoughts?
Rgds
Fred
Trending Topics
10-18-2015, 12:09 PM
#8
Nordschleife Master
Here are some thoughts. These are informed based on dyno experiments, simulations, and logic.
There are two separate knock regimes in the 928 engine: Mid range rpm and high rpm. Of course, everything is continuous but I think it's helpful to think about it in those discrete terms.
At mid-range rpm, the main knock drivers are 180-degree exhaust blowdown interference, relative cooling of middle versus corner cylinders and cylinder heads, and long intake runners. I believe the exhaust interference is the more significant driver of knock and the intake runner length is less significant. This is because the exhaust gas is hot.
At mid range and low rpms, cylinder #2 blows an exhaust pulse towards cylinder #3 180-degrees later which coincides with the overlap of cylinder #3 and hinders the evacuation of the combustion chamber. The same happens 1->4, 8->5, and 6->7. This is a big problem until the rpm is high enough that the exhaust pulse traveling at the sound of speed no longer arrives in the victim cylinder before the victim cylinder's exhaust valve has closed. So either long tube headers or high rpms will diminish the effect of the 180-degree interference problem.
Note that to the extent the long intake runners increase the pressure at the intake valve at IVO, they may in fact reduce knock! That's because the higher pressure will help with evacuation. To the extent the long intake runners increase the pressure at the intake valve at IVC, it increases the knock tendency by increasing the cylinder pressure (in net, a good problem to have). In net, the long runner cylinders seem to knock more at the mid range rpms than short runner cylinders because of either the higher temperatures of the center cylinders or long intake runners filling them better.
At higher rpms, the 180-degree exhaust blowdown interference is no long a problem because the crankshaft is moving so fast that the exhaust blowdown pulse arrives at the other cylinder after the EVC. Now, however, the 90 degree exhaust blowdown interference is the big problem. Cylinder #3 victimizes #1 and cylinder #5 victimizes cylinder #6. At lower rpms, the two blow about on top of each other which just means a little more pumping losses which is a small problem in this context. At high rpms, however, the 90-degree exhaust blowdown pulse arrives late enough to screw up the victim cylinder's overlap and again leave a bunch of hot exhaust gas in the combustion chamber. Since the physical distance between cylinders #1 and #3 is longer than that between #5 and #6, #1 starts suffering from this 90-degree blowdown interference first.
At high rpms, the 8" short intake runners and 12" long intake runners work about equally well. They happen to tune a different order to IVO at the same rpm (slightly above 6000 rpm) with the long runners having a small edge and are about equally effective at the IVC with the short runners having a small edge. At 6000-6500 rpm, my guesstimate is that intake runner length variation for the stock S4 manifold is not a big factor in determining the knock.
Based on all these considerations, the theory would predict that #3 would knock the most at low rpms (center cylinder, long intake runner, 180-degree interference victim), #6 would knock the most at mid range rpms (center cylinder, long intake runner, 90-degree interference victim, and #1 would knock the most at high rpms (the worst 90-degree interference victim when that problem is at its worst).
On the turbo car, this theory is to some extent validated with #6 and #1 knocking the most easily at those respective rpms. #2 isn't particularly knock prone based on my experience. However, there are many other variables at play in my system and I don't know how much that experience can be applied to normally aspirated motors.
One interesting take home message about this is that for a normally aspirated car, the long tube headers have the potential of completely changing the relative tendency of different cylinders to knock at different rpms. So I'd be careful to not extrapolate the findings from long-tube header equipped cars to short-tube exhaust manifold cars.
There are two separate knock regimes in the 928 engine: Mid range rpm and high rpm. Of course, everything is continuous but I think it's helpful to think about it in those discrete terms.
At mid-range rpm, the main knock drivers are 180-degree exhaust blowdown interference, relative cooling of middle versus corner cylinders and cylinder heads, and long intake runners. I believe the exhaust interference is the more significant driver of knock and the intake runner length is less significant. This is because the exhaust gas is hot.
At mid range and low rpms, cylinder #2 blows an exhaust pulse towards cylinder #3 180-degrees later which coincides with the overlap of cylinder #3 and hinders the evacuation of the combustion chamber. The same happens 1->4, 8->5, and 6->7. This is a big problem until the rpm is high enough that the exhaust pulse traveling at the sound of speed no longer arrives in the victim cylinder before the victim cylinder's exhaust valve has closed. So either long tube headers or high rpms will diminish the effect of the 180-degree interference problem.
Note that to the extent the long intake runners increase the pressure at the intake valve at IVO, they may in fact reduce knock! That's because the higher pressure will help with evacuation. To the extent the long intake runners increase the pressure at the intake valve at IVC, it increases the knock tendency by increasing the cylinder pressure (in net, a good problem to have). In net, the long runner cylinders seem to knock more at the mid range rpms than short runner cylinders because of either the higher temperatures of the center cylinders or long intake runners filling them better.
At higher rpms, the 180-degree exhaust blowdown interference is no long a problem because the crankshaft is moving so fast that the exhaust blowdown pulse arrives at the other cylinder after the EVC. Now, however, the 90 degree exhaust blowdown interference is the big problem. Cylinder #3 victimizes #1 and cylinder #5 victimizes cylinder #6. At lower rpms, the two blow about on top of each other which just means a little more pumping losses which is a small problem in this context. At high rpms, however, the 90-degree exhaust blowdown pulse arrives late enough to screw up the victim cylinder's overlap and again leave a bunch of hot exhaust gas in the combustion chamber. Since the physical distance between cylinders #1 and #3 is longer than that between #5 and #6, #1 starts suffering from this 90-degree blowdown interference first.
At high rpms, the 8" short intake runners and 12" long intake runners work about equally well. They happen to tune a different order to IVO at the same rpm (slightly above 6000 rpm) with the long runners having a small edge and are about equally effective at the IVC with the short runners having a small edge. At 6000-6500 rpm, my guesstimate is that intake runner length variation for the stock S4 manifold is not a big factor in determining the knock.
Based on all these considerations, the theory would predict that #3 would knock the most at low rpms (center cylinder, long intake runner, 180-degree interference victim), #6 would knock the most at mid range rpms (center cylinder, long intake runner, 90-degree interference victim, and #1 would knock the most at high rpms (the worst 90-degree interference victim when that problem is at its worst).
On the turbo car, this theory is to some extent validated with #6 and #1 knocking the most easily at those respective rpms. #2 isn't particularly knock prone based on my experience. However, there are many other variables at play in my system and I don't know how much that experience can be applied to normally aspirated motors.
One interesting take home message about this is that for a normally aspirated car, the long tube headers have the potential of completely changing the relative tendency of different cylinders to knock at different rpms. So I'd be careful to not extrapolate the findings from long-tube header equipped cars to short-tube exhaust manifold cars.
10-18-2015, 02:29 PM
#9
Rennlist Member
I understand that Carl's kits typically run at 6psi, so let's take that as a starting point.
So at maximum boost the timing should typically be retarded 6degs or so. As the operating range of the EZK stock system is 9deg that only leaves a margin of 3deg to cope with other variable factors that are mentioned above.
Personally I wouldn't be comfortable with that.
As an interesting side note, there are quite a few moderately boosted 85/86 32v engines, mostly with no knock control whatsoever. How do they survive ? Well the answer is that there is a huge margin of "safety" timing retard on the standard EPROMs - something like 6deg.
With modern fuels it is possible to advance the ignition by 6deg or more safely for non boosted engines. That is why Ken's 85/86 chips have delivered such good improvements. Also the same for my 84-86 16v chips.
So at maximum boost the timing should typically be retarded 6degs or so. As the operating range of the EZK stock system is 9deg that only leaves a margin of 3deg to cope with other variable factors that are mentioned above.
Personally I wouldn't be comfortable with that.
As an interesting side note, there are quite a few moderately boosted 85/86 32v engines, mostly with no knock control whatsoever. How do they survive ? Well the answer is that there is a huge margin of "safety" timing retard on the standard EPROMs - something like 6deg.
With modern fuels it is possible to advance the ignition by 6deg or more safely for non boosted engines. That is why Ken's 85/86 chips have delivered such good improvements. Also the same for my 84-86 16v chips.
10-18-2015, 05:01 PM
#10
Rennlist Member
Thread Starter
...As the operating range of the EZK stock system is 9deg that only leaves a margin of 3deg to cope with other variable factors that are mentioned above.
Personally I wouldn't be comfortable with that.
...With modern fuels it is possible to advance the ignition by 6deg or more safely for non boosted engines.
Generally I find that 3 degrees can makes quite a noticeable difference. What I have been contending with is inconsistency- i.e. sometimes retard will be initiated at a given operating point and then at other times it does not. With the spread of individual cylinders having different knock characteristics [as discussed by Tuomo] it seems to me that when I tune to eliminate knock events the motor is stymied somewhat but when I throw in an extra 4 degrees of advance, whereas I get some retard on some cylinders here and there, the motor feels much livelier and I do not detect anything going on unless ST2 is connected.
For this reason I suspect I am being to conservative and if I had access to a dyno it would prove this [but I do not]. Thus having a few more degrees of advance and relying on the EZK to pull it back on the offending cylinders if it gets out of hand seems a good idea to me.
The notion that different cylinders are impacted by different resonances makes every sense. By using Ken's patch and handicapping different cylinders gave quite an insight into what is going on and my findings were very similar to what Ken reported in his threads.
My dilemma now is knowing where to stop and where I can push the envelope a bit more. I am hoping to give Ken's self modulating EZK chip set a workout shortly when I have got my current raft of irksome age related maintenance issues sorted out. Seems I am at or close to the next round of replacements given I have had this 928 for 10 years now!
One of the many good things about ST2 is that it shows the cells that are high load centric [and there are not too many of them]- so they are the focus of my attention. I will also be doing some cooler air temperature calibrations in Dec/Jan if I manage to pick up some more work.
Your comment on gasoline is interesting- I am now somewhat out of touch on gasoline quality- why today's brews should take more advance I am not sure but your comment makes me wonder if what we brew here is "out of date". For sure it is not brilliant. I find 1.6 litre repmobiles quite lively when I visit the UK- maybe the gasoline is a factor.
Rgds
Fred
Regards
10-18-2015, 06:24 PM
#11
Hi Fred. You're in Oman, right? I think the 95 fuel there is even lower octane. My prado would fly from Dubai to Muscat / Salalah but would struggle and ping all the way back. A few years ago now, but may be relevant to your safety margin thinking. Dave
10-18-2015, 10:20 PM
#12
Rennlist Member
Thread Starter
Many friends over here, particularly in Porsche circles have commented same as you but not from driving in a Prado. Many report that the motor just does not feel alive and that is understandable if the std tune [optimised for 98 RON I believe] leaves the motor down on power due to retardation as may well happen- especially in summer conditions.
Dubai markets a Super 98 blend and for sure it is superior- I have tried to do some lobbying here to get some imported to a garage or two but no joy to date. The Sharktuner is a bit of a double edged thing in that it is truly brilliant but it also shows you what is going on "warts and all".
The fuel is tested at 95 RON in the refinery [I used to be the chief engineer there] but I have a theory that they use too much light ends in the gasoline pool to get the octane test passed but by the time some of it has vaporised the real world octane in cars like ours that recycle fuel is probably less than what it is rated to be.
That being said I also feel that batch firing is impeding matters in that 6 and 2 clearly lead the charge followed by the two inner cylinders thus a need for differential advance cylinder to cylinder at different rpm's. Thus my challenge is whether we can rely on the EZK to carry out such a role. For sure I have never heard a whimper from the motor never mind the ghastly noise/rattle you actually get when this phenomena happens. Some GTS motors actually reported this happening at low to mid range rpms many years ago.
Rgds
Fred
10-18-2015, 11:03 PM
#13
Nordschleife Master
The fuel is tested at 95 RON in the refinery [I used to be the chief engineer there] but I have a theory that they use too much light ends in the gasoline pool to get the octane test passed but by the time some of it has vaporised the real world octane in cars like ours that recycle fuel is probably less than what it is rated to be.
If the test point was moved to the bowser, then the refinery would be forced to modify the blend so that even after storage, transport etc. it was still 95 RON, which would fix the fuel quality issue. The challenge is getting legislation to enforce the interests of the consumers, versus the power of $$$ spent by the oil company avoiding the cost of that change.
Interesting thread, thanks Fred
10-19-2015, 04:56 AM
#14
Rennlist Member
Thread Starter
That's a process issue I've encountered in a number of industries - they've designed the process around the test point rather than the product specification. Great for minimising short-term production costs, terrible for minimising the cost of change.
If the test point was moved to the bowser, then the refinery would be forced to modify the blend so that even after storage, transport etc. it was still 95 RON, which would fix the fuel quality issue. The challenge is getting legislation to enforce the interests of the consumers, versus the power of $$$ spent by the oil company avoiding the cost of that change.
Interesting thread, thanks Fred
If the test point was moved to the bowser, then the refinery would be forced to modify the blend so that even after storage, transport etc. it was still 95 RON, which would fix the fuel quality issue. The challenge is getting legislation to enforce the interests of the consumers, versus the power of $$$ spent by the oil company avoiding the cost of that change.
Interesting thread, thanks Fred
A point that many folks do not understand is that octane rating has nothing to do with power generation potential chemistry wise- rather it facilitates extraction of the power potential within the fuel. Thus is something one can feel [or measure] and more developed power invariably means that something is being varied and in our engines it will be the amount of advance being applied thus my perception that the EZK is just as much a control mechanism as it is a safeguarding [protective] tool.
Dialing in too much advance does nothing for the big ends and dialing in too little is just flushing power down the can. Getting the right amount of advance dialed in is very critical to developed power and I feel as though I am struggling with this. I get the impression that 3 degrees too conservative is enough to lose a bunch of power and we have a system that can by its own brain take out up to 9 degrees.
This is why I am thinking that one should adjust the advance to remove the need to apply knock control and then add a fixed margin of say 3 degrees [maybe 5 degrees?] across the board and then let the thing adjust itself as it needs to self optimise. Remember I am seeing [perceiving?] a need for a 3 or 4 degree differential between the inner [more advance] and outer cylinders.
Rgds
Fred
10-19-2015, 10:42 AM
#15
Rennlist Member
Fred,
Are you running the same plugs (heat-wise) across all cylinders? WR7DC is stock, in your climate you probably want WR6DC. Then try WR5DC in the hotter cylinders.
There's nothing wrong with letting the EZK do its job and take out a few degrees on a few cylinders. When it starts retarding all (or nearly all) of the cylinders, then you are not accomplishing anything. And if it is pulling out more than 3-4 deg then you are eating into margin that you might otherwise need.
A couple of interesting data points: Fuel does vary of course, and altitude and temperature are big variables. Heading to Wichita, we filled in Denver and then headed downhill across Colorado and into Nebraska, car running fine in the increasing heat (an decreasing altitude). At one point, when Sharktuner happened to be logging, I nailed it to get around some pesky traffic ... Bwahhhh, lovely sound, but some smoke out the back and when I looked at ST's log it showed shows 9 deg retard on all 8 cylinders. Yikes! Higher-altitude fuel mix (lower octane) and high temps used up 9 deg and then some.
Our GT has been tuned as you suggest for at least the last 100K miles. Louie's theory was to let the EZK do its job and I have followed that also, setting the top-end advance so that the EZK would pull out a few degrees up to 5 deg) from 2-3 cylinders. But as our Kansas trip pointed out, we need margin for varied terrain, temperature and fuel. So I always had this nagging worry about what we were doing to rod bearings.
Another data point: Doing some tuning on a local dyno shop, I am watching Sharktuner and the tech is watching the dyno screen and rearview mirror. "Try taking out a couple of degrees above 5000" he says, sure enough Sharkplotter shows some retard there-- but just the usual 3-4 deg. He is seeing smoke-rings in the mirror, a sign of detonation. I take out 3 deg, we run again, more power-- 4 or 5 more hp, not less. "Try a couple more" he says, another few hp. Interesting, very interesting.
So what about the rod bearings? At 250K miles we disassemble the motor to see what's in there, and the rod bearings are perfect. Never been changed as far as we know, all marked identically with a Porsche logo, can't tell 2/6 from any others and no signs of detonation on bearings or pistons.
So I've got no evidence that letting the EZK do its job on a few cylinders does any harm, but it is also clear that you want a pretty fat margin for the unexpected. And adding timing doesn't necessarily lead to more horsepower. And neither can logging, but it certainly can tell interesting stories.
Cheers,
Are you running the same plugs (heat-wise) across all cylinders? WR7DC is stock, in your climate you probably want WR6DC. Then try WR5DC in the hotter cylinders.
There's nothing wrong with letting the EZK do its job and take out a few degrees on a few cylinders. When it starts retarding all (or nearly all) of the cylinders, then you are not accomplishing anything. And if it is pulling out more than 3-4 deg then you are eating into margin that you might otherwise need.
A couple of interesting data points: Fuel does vary of course, and altitude and temperature are big variables. Heading to Wichita, we filled in Denver and then headed downhill across Colorado and into Nebraska, car running fine in the increasing heat (an decreasing altitude). At one point, when Sharktuner happened to be logging, I nailed it to get around some pesky traffic ... Bwahhhh, lovely sound, but some smoke out the back and when I looked at ST's log it showed shows 9 deg retard on all 8 cylinders. Yikes! Higher-altitude fuel mix (lower octane) and high temps used up 9 deg and then some.
Our GT has been tuned as you suggest for at least the last 100K miles. Louie's theory was to let the EZK do its job and I have followed that also, setting the top-end advance so that the EZK would pull out a few degrees up to 5 deg) from 2-3 cylinders. But as our Kansas trip pointed out, we need margin for varied terrain, temperature and fuel. So I always had this nagging worry about what we were doing to rod bearings.
Another data point: Doing some tuning on a local dyno shop, I am watching Sharktuner and the tech is watching the dyno screen and rearview mirror. "Try taking out a couple of degrees above 5000" he says, sure enough Sharkplotter shows some retard there-- but just the usual 3-4 deg. He is seeing smoke-rings in the mirror, a sign of detonation. I take out 3 deg, we run again, more power-- 4 or 5 more hp, not less. "Try a couple more" he says, another few hp. Interesting, very interesting.
So what about the rod bearings? At 250K miles we disassemble the motor to see what's in there, and the rod bearings are perfect. Never been changed as far as we know, all marked identically with a Porsche logo, can't tell 2/6 from any others and no signs of detonation on bearings or pistons.
So I've got no evidence that letting the EZK do its job on a few cylinders does any harm, but it is also clear that you want a pretty fat margin for the unexpected. And adding timing doesn't necessarily lead to more horsepower. And neither can logging, but it certainly can tell interesting stories.
Cheers,
