stock ecu vs. s-squared tune
05-03-2005, 07:30 PM
#122
Troll
Join Date: Sep 2004
Location: Pimpin' tards
Posts: 895
Originally Posted by BAN SUVS
jimbotko, your arguments would carry much more merit if you weren't posting them anonymously.
Should I have posted that under my original screen name that I rarely use (intentioanlly at least) even though everyone that is worth a crap knows who I am?
05-03-2005, 07:44 PM
#124
Registered User
Join Date: Nov 2004
Location: Middle Earth
Posts: 551
Car Info: 2005 SRR WRX
Originally Posted by topnotchwrx
*takes bowl of popcorn away from Killian and then grabs a coke*
05-03-2005, 07:53 PM
#125
Originally Posted by BAN SUVS
Keegan, I'd like you to step back and examine events as they unfolded in hindsight and consider a few questions I have.
First, prior to going to the Road Dyno gala event sponsored by EQ Tuning, had you ever, even once, suspected there was anything wrong with your car? Did you ever feel as if it was pulling timing? If so, why did you not contact Nate about it prior to discussing it with Ed? Looking at the logs Ed is so kindly hosting for you, I see no indication that you experienced any knock during your road dyno datalogging runs. Now, there IS a drop in absolute ignition advance on both, right at 3700 RPM. I presume this is what Ed interpreted as knock, except that the IAM was at 16 and knock signal zero on both runs. You did NOT have any detonation with Ed in the car, unless he for some reason chose not to host those files. Doubtful.
So Ed, being the good samaritan he is, went and looked at the fine learning, and lo and behold, did not like what he saw. However, for those of us that know how to draw and read 2D graphs, you'll note that the cells Ed took issue with, the ones with the negative numbers, are not active on either of road dyno graphs. They are at moderate load levels, which the car was out of well before the supposed knock incident took place.
Keegan, I want you to understand something here- if your IAM is at 16, you have not experienced knock. Period. No matter what you think you hear. While Ed can debate tuning philosophy with Nate all day long and neither one prove themselves absolutely right or the other absolutely wrong, there was no basis in those 2 datalogs for Ed to tell you that your car was tuned unsafely.
As for claiming that the car was low on power, using road dyno graphs as evidence... well, that's just asinine. For one thing, Nate's Mustang dyno reads lower than any other dyno within 2000 miles of here. Period. Gruppe-S reads higher, Cobb reads higher, Vishnu reads higher, everyone. So saying that Nate left the car 10 whp short shows more about Ed than it does Nate. But here's the best part- even if it WAS short on power, a DeltaDash road dyno is not even close to conclusive, concrete evidence of such. Please allow me to demonstrate, using my own copy of DeltaDash and my humble 2.5RS.
First, prior to going to the Road Dyno gala event sponsored by EQ Tuning, had you ever, even once, suspected there was anything wrong with your car? Did you ever feel as if it was pulling timing? If so, why did you not contact Nate about it prior to discussing it with Ed? Looking at the logs Ed is so kindly hosting for you, I see no indication that you experienced any knock during your road dyno datalogging runs. Now, there IS a drop in absolute ignition advance on both, right at 3700 RPM. I presume this is what Ed interpreted as knock, except that the IAM was at 16 and knock signal zero on both runs. You did NOT have any detonation with Ed in the car, unless he for some reason chose not to host those files. Doubtful.
So Ed, being the good samaritan he is, went and looked at the fine learning, and lo and behold, did not like what he saw. However, for those of us that know how to draw and read 2D graphs, you'll note that the cells Ed took issue with, the ones with the negative numbers, are not active on either of road dyno graphs. They are at moderate load levels, which the car was out of well before the supposed knock incident took place.
Keegan, I want you to understand something here- if your IAM is at 16, you have not experienced knock. Period. No matter what you think you hear. While Ed can debate tuning philosophy with Nate all day long and neither one prove themselves absolutely right or the other absolutely wrong, there was no basis in those 2 datalogs for Ed to tell you that your car was tuned unsafely.
As for claiming that the car was low on power, using road dyno graphs as evidence... well, that's just asinine. For one thing, Nate's Mustang dyno reads lower than any other dyno within 2000 miles of here. Period. Gruppe-S reads higher, Cobb reads higher, Vishnu reads higher, everyone. So saying that Nate left the car 10 whp short shows more about Ed than it does Nate. But here's the best part- even if it WAS short on power, a DeltaDash road dyno is not even close to conclusive, concrete evidence of such. Please allow me to demonstrate, using my own copy of DeltaDash and my humble 2.5RS.
Lets first address the issue of the knock that is very obviously present in the logs. First take a look at the first tuned log and pay close attention to the knock correction values just past 3700RPM. You see the sharp drop from 7 to 3 degrees and then a gradual climb back 6+ degrees? To a trained eye, this very obviously represents knock and the ECU pulling timing to get around it. Now look further where knock correction drops again around 5000RPM and finally goes NEGATIVE by 6200RPM. The ONLY way knock correction will go negative is either if there is knock present at the moment the log is being recorded, or there has been enough knock previously to cause the ECU to learn enough negative correction to pull timing bellow the base timing map. Show me one car that's running well without knock that has negative knock correction numbers! Now lets look at the second tuned log where you can see the exact same pattern except now its more severe. Knock correction drops sharply from 7 to 2 degrees now at 3800RPM and again goes negative by 6300RPM. On this run we could actually hear the knock just bellow 4000RPM and could feel the flat spot caused by the ECU pulling even more timing than it had already learned in that spot.
Now gaining this new information, maybe you'd like to change your statement about the IAM. Saying that just because the IAM is at 16, there is no way that the car had experienced knock is complete nonesense and just about any tuner will back me up on this. Like I said before, there are parameters that define under what conditions the IAM is learned and seperate parameters that define under what conditions fine learning is active. These are all controllable by the tuner and can make the IAM act vastly different than it does in a stock ECU.
On to the issue of the fine learning table. If you had ever used the Ecutek tuning software, you would know that the load and RPM labels on the axis are not taken from the ECU but rather from the specific rom that is loaded into the software at the time. In this case, a stock rom was loaded, so those labels are as they would be in a stock ECU, not at all how they are in a propperly tuned ECU, so the labels in the screen capture are irrelevent. You did, however stumble on a good observation that those extremely negative values on the top end are not in the higher load cells, but rather in the more moderate ones. Again, if you had any experience tuning or even interest in tuning, you would know that at higher RPM's, calculated load decreases significantly and causes the ECU to trace to the left within the map as RPM's increase. This puts you right around those -6 and -9.5 values as you hit the higher RPM's.
As for claiming that the car was making poor power, even though that is not at all the main concern of this thread, I'll address that as well. I agree dynos read differently and every car is different. The really important thing is the comparisson of numbers from the same car, same dyno, and in this case the same road. As you can see, that comparisson is clearly presented here by Keegan. Same road, same car, same dyno software, same settings, two different ECU's that both made extremely similar power. And for the record, this is not 10whp lower than similar cars. The difference is actually closer to 20whp and almost 30 ft-lbs when compared to other uppipe/turboback/tuned cars I've run on this road dyno. This is of course no surprise as that's generally the gain most cars with these mods see when going from a stock ECU to a custom tune. And as per your example DD road dyno graph, care to show us the parameters you used to acquire that graph? I have no problem showing everyone that the parameters I used remaine constant from run to run... but maybe if you had read my previous replies, you would have seen that.
I think its fairly clear that you need to learn a bit more about the subject before making such claims.
Thanks
05-03-2005, 07:59 PM
#126
I'd just like to say that this kind of attitude is ridiculous. Would you guys rather I had not said anything to Keegan letting him think that his car was running fine when it was knocking all over the place? I suppose I should have also posted his road dyno chart with all the others pretending that there was nothing wrong with the car and showing that his car was making less power than a Cobb stage 2 car with a stock uppipe and of course crediting SS with the tuning? I guess I should have expected this type of response for being honest and up front with someone who's car I was concerned about.
05-03-2005, 08:00 PM
#127
Registered User
iTrader: (9)
Join Date: Mar 2003
Location: Sacramento, CA
Posts: 2,451
Car Info: 1996 Mustang GT/2013 Outback Limited
Originally Posted by EQ Tuning
Lets first address the issue of the knock that is very obviously present in the logs. First take a look at the first tuned log and pay close attention to the knock correction values just past 3700RPM. You see the sharp drop from 7 to 3 degrees and then a gradual climb back 6+ degrees? To a trained eye, this very obviously represents knock and the ECU pulling timing to get around it. Now look further where knock correction drops again around 5000RPM and finally goes NEGATIVE by 6200RPM. The ONLY way knock correction will go negative is either if there is knock present at the moment the log is being recorded, or there has been enough knock previously to cause the ECU to learn enough negative correction to pull timing bellow the base timing map.
Show me one car that's running well without knock that has negative knock correction numbers! Now lets look at the second tuned log where you can see the exact same pattern except now its more severe. Knock correction drops sharply from 7 to 2 degrees now at 3800RPM and again goes negative by 6300RPM. On this run we could actually hear the knock just bellow 4000RPM and could feel the flat spot caused by the ECU pulling even more timing than it had already learned in that spot.
Show me one car that's running well without knock that has negative knock correction numbers! Now lets look at the second tuned log where you can see the exact same pattern except now its more severe. Knock correction drops sharply from 7 to 2 degrees now at 3800RPM and again goes negative by 6300RPM. On this run we could actually hear the knock just bellow 4000RPM and could feel the flat spot caused by the ECU pulling even more timing than it had already learned in that spot.
Are you also aware that the KC will fluctuate throughout the RPM range depending on load and situations that the car sees?
Now gaining this new information, maybe you'd like to change your statement about the IAM. Saying that just because the IAM is at 16, there is no way that the car had experienced knock is complete nonesense and just about any tuner will back me up on this. Like I said before, there are parameters that define under what conditions the IAM is learned and seperate parameters that define under what conditions fine learning is active. These are all controllable by the tuner and can make the IAM act vastly different than it does in a stock ECU.
I think its fairly clear that you need to learn a bit more about the subject before making such claims.
05-03-2005, 08:02 PM
#128
Registered User
iTrader: (9)
Join Date: Mar 2003
Location: Sacramento, CA
Posts: 2,451
Car Info: 1996 Mustang GT/2013 Outback Limited
Originally Posted by EQ Tuning
I'd just like to say that this kind of attitude is ridiculous. Would you guys rather I had not said anything to Keegan letting him think that his car was running fine when it was knocking all over the place? I suppose I should have also posted his road dyno chart with all the others pretending that there was nothing wrong with the car and showing that his car was making less power than a Cobb stage 2 car with a stock uppipe and of course crediting SS with the tuning? I guess I should have expected this type of response for being honest and up front with someone who's car I was concerned about.
05-03-2005, 08:10 PM
#130
Originally Posted by dz
You are aware that the knock sensor is disabled by the ECU when the RPM hits 6200 and then does not come back on until the RPM drops below 6100 aren't you?
Are you also aware that the KC will fluctuate throughout the RPM range depending on load and situations that the car sees?
Nate keeps all stock safeguards in place on all of his tunes. Any tuner that does not should not have access to the tuning software. Speaking of which, what version of the tuning software are you currently running on your tuning rig?
Damn, pass some of that popcorn this way please...
Are you also aware that the KC will fluctuate throughout the RPM range depending on load and situations that the car sees?
Nate keeps all stock safeguards in place on all of his tunes. Any tuner that does not should not have access to the tuning software. Speaking of which, what version of the tuning software are you currently running on your tuning rig?
Damn, pass some of that popcorn this way please...
And yes, KC fluctuates throughout the RPM range and load range dependant mainly on the ignition correction table and and IAM. Your final timing is roughly calculated as follows: base timing map(RPM, load) + (IAM/16)*ignition correction map(RPM, load) + fine learning table(RPM, load). There are also corrections based on IAT, coolant temp, etc. that are applied, although these are fairly constant throughout a third gear pull. Knock correction as reported in DD is the second part of that equation... (IAM/16)*ignition correction map(RPM, load) + fine learning table(RPM, load).
Keeping stock safeguards in place is not the issue here. Even a stock ECU can knock in places and leave the IAM unaffected. The IAM is an overall adjustment of the timing and mainly gets effected when there is knock present throughout the load and RPM range. It also depends on how long the car has been running after a reset. The IAM is very finicky immediately after a reset and will react to just about any ping, but after driving for a while, it becomes much more stable and the fine ignition learning is put into effect to absorb any small problem spots.
Edit: I'm running the latest Ecutek Flash2002 software v1.37. Not sure what this has to do with, but I have nothing to hide.
Last edited by MethodBuilt; 05-03-2005 at 08:18 PM.
05-03-2005, 08:19 PM
#132
Just out of curiousity, and please forgive my ignorance, but the "Knock Signal 1" column on both .csv files you posted are filled with zeroes. Does that typically mean that no knock was detected or is it only more serious knock issues that are incremented in that field.
Not trying to flame here, I'm just trying to understand what I'm reading here.
Not trying to flame here, I'm just trying to understand what I'm reading here.
05-03-2005, 08:21 PM
#133
Registered User
Join Date: Nov 2004
Location: Middle Earth
Posts: 551
Car Info: 2005 SRR WRX
*moves on to a pint of ben and jerry's*
05-03-2005, 08:22 PM
#134
Registered User
iTrader: (12)
Join Date: Nov 2002
Location: Reno, NV
Posts: 18,369
Car Info: 1993/2000/2001 GF4 mostly red
Originally Posted by EQ Tuning
How much experience do you have tuning cars or subaru's for that matter? Its fairly obvious by your post here that you don't have a clue about how the ECU works.
Originally Posted by EQ Tuning
Lets first address the issue of the knock that is very obviously present in the logs. First take a look at the first tuned log and pay close attention to the knock correction values just past 3700RPM. You see the sharp drop from 7 to 3 degrees and then a gradual climb back 6+ degrees? To a trained eye, this very obviously represents knock and the ECU pulling timing to get around it. Now look further where knock correction drops again around 5000RPM and finally goes NEGATIVE by 6200RPM. The ONLY way knock correction will go negative is either if there is knock present at the moment the log is being recorded, or there has been enough knock previously to cause the ECU to learn enough negative correction to pull timing bellow the base timing map. Show me one car that's running well without knock that has negative knock correction numbers!
Originally Posted by EQ Tuning
Now lets look at the second tuned log where you can see the exact same pattern except now its more severe. Knock correction drops sharply from 7 to 2 degrees now at 3800RPM and again goes negative by 6300RPM. On this run we could actually hear the knock just bellow 4000RPM and could feel the flat spot caused by the ECU pulling even more timing than it had already learned in that spot.
Originally Posted by EQ Tuning
Now gaining this new information, maybe you'd like to change your statement about the IAM. Saying that just because the IAM is at 16, there is no way that the car had experienced knock is complete nonesense and just about any tuner will back me up on this. Like I said before, there are parameters that define under what conditions the IAM is learned and seperate parameters that define under what conditions fine learning is active. These are all controllable by the tuner and can make the IAM act vastly different than it does in a stock ECU.
Originally Posted by EQ Tuning
On to the issue of the fine learning table. If you had ever used the Ecutek tuning software, you would know that the load and RPM labels on the axis are not taken from the ECU but rather from the specific rom that is loaded into the software at the time. In this case, a stock rom was loaded, so those labels are as they would be in a stock ECU, not at all how they are in a propperly tuned ECU, so the labels in the screen capture are irrelevent. You did, however stumble on a good observation that those extremely negative values on the top end are not in the higher load cells, but rather in the more moderate ones. Again, if you had any experience tuning or even interest in tuning, you would know that at higher RPM's, calculated load decreases significantly and causes the ECU to trace to the left within the map as RPM's increase. This puts you right around those -6 and -9.5 values as you hit the higher RPM's.
Originally Posted by EQ Tuning
As for claiming that the car was making poor power, even though that is not at all the main concern of this thread, I'll address that as well. I agree dynos read differently and every car is different. The really important thing is the comparisson of numbers from the same car, same dyno, and in this case the same road. As you can see, that comparisson is clearly presented here by Keegan. Same road, same car, same dyno software, same settings, two different ECU's that both made extremely similar power. And for the record, this is not 10whp lower than similar cars. The difference is actually closer to 20whp and almost 30 ft-lbs when compared to other uppipe/turboback/tuned cars I've run on this road dyno. This is of course no surprise as that's generally the gain most cars with these mods see when going from a stock ECU to a custom tune. And as per your example DD road dyno graph, care to show us the parameters you used to acquire that graph? I have no problem showing everyone that the parameters I used remaine constant from run to run... but maybe if you had read my previous replies, you would have seen that.
Originally Posted by EQ Tuning
I think its fairly clear that you need to learn a bit more about the subject before making such claims.
Thanks
Thanks
05-03-2005, 08:26 PM
#135
Originally Posted by nKoan
Just out of curiousity, and please forgive my ignorance, but the "Knock Signal 1" column on both .csv files you posted are filled with zeroes. Does that typically mean that no knock was detected or is it only more serious knock issues that are incremented in that field.
Not trying to flame here, I'm just trying to understand what I'm reading here.
Not trying to flame here, I'm just trying to understand what I'm reading here.
Thanks for the good question.