Tune Fix
Oct 4, 2007, 09:11 AM
#1
Tune Fix
I have not tuned an Evo 8 for a while and I missed doing that. This Evo 8 belongs to a friend of mine from SoCal. It had the following mods:
TBE
O2 housing
264 cams
Walbro fuel pump
720 injectors
Greddy EBC
The car was tuned with Ecuflash and a piggyback that controls fuel only. The AFR was set using the narrowband O2 sensor on the car.
When I first logged the car I forgot to do the 2 byte load and did not log loadcalc. So the charts below use injector pulse width (IPW) for one axis and RPM for the other. IPW moves somewhat closely to load. So I will use it as a substitute for load in the chart. So an IPW of 15-17 is equivalent to load 240-260, an IPW of 11-13 is equivalent to load 200-220.
Here is what the data showed when I first logged the car. These are two back-to-back logs. I did not want to do more since the car was knocking like crazy.
The boost was allegedly set @ 21 psi using the eye-ball-the-gauge method. But when the boost was logged the average boost @ peak was 23.37 psi. That is the AVERAGE boost on 91 octane gas!!!! The maximum boost was 24.57 psi. If you think that is too much on CA 91 octane gas, then you are correct.
As I mentioned before, the AFR was set using the narrowband and not a wideband. The fuel map was left stock and it was left to the piggyback to adjust fuel. Here is what the AFR looked like:
The car is generally running too lean for an Evo 8 on 91 octane gas. The AFR is rich at 4000 rpm and get leaner as the rpm increases. This is the opposite of what you have taught me in this forum. The AFR is also very erratic. The above chart does not show the erratic nature of the AFR, but the one below sure does.
The chart also shows you the knock that this car was experiencing. It had a max of 12 counts of knock in this log and a maximum of 11 counts of knock in the other log. Before the knock hit in this log, the timing was at 7* BTDC and it got pulled all the way to 3* BTDC. The 4* pull is in line with our assumption on this board that every three counts of knock pull 1* of timing. The timing hit 13* by 7500. That is exactly where I set the timing to eliminate the knock. The knock sensor was telling the tuner what timing to run, but he was not listening. The timing that the car ended up with after I tuned it was 2-3* @ peak boost and 13* by redline/cutoff. That is exactly what the knock sensor was telling me to do. This is another instance where listening/reading the knock system helps the tuner set the best timing that an Evo likes rather than set the timing to what he likes. Any timing increase beyond 13* on this car triggered knock. Below is the entire knock profile for this Evo:
There are 4 counts of knock at 4000 rpm and then the knock ratchets up to 10-12 counts by redline.
The first thing I did when I tuned the car was to turn down the boost. The boost was turned down by an average of 1.2 psi from 3500 to 5000 rpm and an average of 0.6 psi from 5500 rpm to redline. This is what the new boost profile looked like:
The second objective was to give the Evo the timing that it wanted. Reading the knock sensor showed me what the car wanted: 2-3* at peak boost and slowly increasing to 13* by redline/cutoff.
Finally, I set out to tune the AFR. I set the AFR around 11.5 at peak and slowly tapered it to 10.8:1 by redline. I generally set Evo 8s a tad rich than Evo 9s since they are not blessed with the better cooled cylinder head.
What about the knock? Let the logs speak for themselves. No more than 1-2 knock counts at peak boost torque. This is a far cry from the 11-12 counts of knock that the car was registering at the top end and the 4 counts of knock at the low end.
So did the car make more power than before? No, it did not. But when the boost is turned down and the timing is retarded, you have to expect some power loss. The upside is that the car is not knocking anymore and the owner was willing to sacrifice some power for safety, smoothness and consistency.
TBE
O2 housing
264 cams
Walbro fuel pump
720 injectors
Greddy EBC
The car was tuned with Ecuflash and a piggyback that controls fuel only. The AFR was set using the narrowband O2 sensor on the car.
When I first logged the car I forgot to do the 2 byte load and did not log loadcalc. So the charts below use injector pulse width (IPW) for one axis and RPM for the other. IPW moves somewhat closely to load. So I will use it as a substitute for load in the chart. So an IPW of 15-17 is equivalent to load 240-260, an IPW of 11-13 is equivalent to load 200-220.
Here is what the data showed when I first logged the car. These are two back-to-back logs. I did not want to do more since the car was knocking like crazy.
The boost was allegedly set @ 21 psi using the eye-ball-the-gauge method. But when the boost was logged the average boost @ peak was 23.37 psi. That is the AVERAGE boost on 91 octane gas!!!! The maximum boost was 24.57 psi. If you think that is too much on CA 91 octane gas, then you are correct.
As I mentioned before, the AFR was set using the narrowband and not a wideband. The fuel map was left stock and it was left to the piggyback to adjust fuel. Here is what the AFR looked like:
The car is generally running too lean for an Evo 8 on 91 octane gas. The AFR is rich at 4000 rpm and get leaner as the rpm increases. This is the opposite of what you have taught me in this forum. The AFR is also very erratic. The above chart does not show the erratic nature of the AFR, but the one below sure does.
The chart also shows you the knock that this car was experiencing. It had a max of 12 counts of knock in this log and a maximum of 11 counts of knock in the other log. Before the knock hit in this log, the timing was at 7* BTDC and it got pulled all the way to 3* BTDC. The 4* pull is in line with our assumption on this board that every three counts of knock pull 1* of timing. The timing hit 13* by 7500. That is exactly where I set the timing to eliminate the knock. The knock sensor was telling the tuner what timing to run, but he was not listening. The timing that the car ended up with after I tuned it was 2-3* @ peak boost and 13* by redline/cutoff. That is exactly what the knock sensor was telling me to do. This is another instance where listening/reading the knock system helps the tuner set the best timing that an Evo likes rather than set the timing to what he likes. Any timing increase beyond 13* on this car triggered knock. Below is the entire knock profile for this Evo:
There are 4 counts of knock at 4000 rpm and then the knock ratchets up to 10-12 counts by redline.
The first thing I did when I tuned the car was to turn down the boost. The boost was turned down by an average of 1.2 psi from 3500 to 5000 rpm and an average of 0.6 psi from 5500 rpm to redline. This is what the new boost profile looked like:
The second objective was to give the Evo the timing that it wanted. Reading the knock sensor showed me what the car wanted: 2-3* at peak boost and slowly increasing to 13* by redline/cutoff.
Finally, I set out to tune the AFR. I set the AFR around 11.5 at peak and slowly tapered it to 10.8:1 by redline. I generally set Evo 8s a tad rich than Evo 9s since they are not blessed with the better cooled cylinder head.
What about the knock? Let the logs speak for themselves. No more than 1-2 knock counts at peak boost torque. This is a far cry from the 11-12 counts of knock that the car was registering at the top end and the 4 counts of knock at the low end.
So did the car make more power than before? No, it did not. But when the boost is turned down and the timing is retarded, you have to expect some power loss. The upside is that the car is not knocking anymore and the owner was willing to sacrifice some power for safety, smoothness and consistency.
Oct 4, 2007, 09:53 AM
#2
Evolving Member
iTrader: (1)
Join Date: Mar 2007
Location: Coastal Maine
Posts: 315
Likes: 0
Received 0 Likes
on
0 Posts
Good job. I love it when people post there tuning experiences with hard numbers. As a person that is very new to tuning, it helps me back up my own numbers. As mine a FAR from what my two mail-in flashes look like. (one was on the car when I bought it).
Oct 4, 2007, 10:10 AM
#3
Evolving Member
iTrader: (4)
Join Date: Dec 2003
Location: Ventura County, CA
Posts: 357
Likes: 0
Received 0 Likes
on
0 Posts
You're a pro at this stuff now, naj.
I was wondering why does peak torque always seems to come around 4000-4500 rpm area for tuned cars? If I remember correctly, don't stock evos have their peak at 3500? Or is this is because of the 10.5 turbine?
I was wondering why does peak torque always seems to come around 4000-4500 rpm area for tuned cars? If I remember correctly, don't stock evos have their peak at 3500? Or is this is because of the 10.5 turbine?
Oct 4, 2007, 10:29 AM
#4
Evolving Member
iTrader: (8)
Join Date: Jul 2006
Location: Silicon Valley
Posts: 179
Likes: 0
Received 0 Likes
on
0 Posts
nj - nice retune....i would recommend that your proceed with setting 1) boost 2) afr then 3) timing. you could possibly gain more power by tuning timing last after setting the correct afr beforehand.
Oct 4, 2007, 10:54 AM
#5
I also noticed that on cars that have MBC/EBC the peak boost is slightly delayed than on cars running ECUboost control.
Last edited by nj1266; Oct 4, 2007 at 10:58 AM.
Oct 4, 2007, 10:57 AM
#6
After I was done setting the timing, boost and AFR on this car, I went back and increased the timing by 1* from 5500 rpm to 7500 rpm. The car knocked and did not like it. So I set back the timing to the way it was.
Oct 4, 2007, 04:03 PM
#7
First of all nice job on the tune. Was the piggyback an AFC?
The way i prefer to do it is in steps with the boost and AFR increasing in steps. I start at 18psi and leave the fuel tables stock and see if it has knock. Decrease timing to cure knock and then adjust the AFR to the desired level. Then do another pull and up the boost to 20psi. After the AFR and boost are set by doing it in steps I then start increasing the timing til it knocks and then back off 2 degrees. I will do at least 3 pulls on the final MAP to ensure it is good to go.
I am not saying your method is wrong or anyone elses method is wrong I just prefer to do it in steps.
The way i prefer to do it is in steps with the boost and AFR increasing in steps. I start at 18psi and leave the fuel tables stock and see if it has knock. Decrease timing to cure knock and then adjust the AFR to the desired level. Then do another pull and up the boost to 20psi. After the AFR and boost are set by doing it in steps I then start increasing the timing til it knocks and then back off 2 degrees. I will do at least 3 pulls on the final MAP to ensure it is good to go.
I am not saying your method is wrong or anyone elses method is wrong I just prefer to do it in steps.
Trending Topics
Oct 4, 2007, 05:10 PM
#8
The way i prefer to do it is in steps with the boost and AFR increasing in steps. I start at 18psi and leave the fuel tables stock and see if it has knock. Decrease timing to cure knock and then adjust the AFR to the desired level. Then do another pull and up the boost to 20psi. After the AFR and boost are set by doing it in steps I then start increasing the timing til it knocks and then back off 2 degrees. I will do at least 3 pulls on the final MAP to ensure it is good to go.
Oct 7, 2007, 05:24 PM
#9
This is another 05 Evo 8 that I tuned yesterday. It had the following mods before I re-tuned it:
2.75” Down pipe
Stock Cat
RRE Stealth Exhaust
Drop-in filter
Walbro FP
Dejon Tool MBC set @ 23 psi (eyeball-the-gauge method)
Tune (same tuner as the car in first post)
We changed the car to:
RRE High Flow Cat
ECU boost control with #65 pill
Looney Tune
Due to time constraints I did not get a chance to do a baseline log. I had to install the HFC and the boost hose assembly with the pill just before the tuning session. We also found a pin hole leak in the HFC and had to go to a muffler shop to fix it.
What I do have are the logs that were given to the owner by the famous tuner. The logs do not include boost and AFR numbers since the famous tuner does not log boost and uses a narrowband sensor to set the AFR. It baffles me that a famous tuner does not have a wideband in his tuning kit!!!!
So here is what a typical log looked like:
In this chart the car knocks from around 5K rpm to redline. There are two knock spikes, one at 5000 rpm and one at 6900 rpm. Timing is pulled 4* and 3* respectively. As with the prior car, the tuner refused to listen what the knock sensor was telling him. By 7000+ rpm that car wanted no more than 14* of timing, but the tuner set the timing in the map at 17* in the 200 load cell and 20* in the 180 load cell. The knock profile of all 4 logs looked like this:
NB: The ECU_Load in this chart is actually LoadCalc.
Most of the serious knock is concentrated in the top of the rpm range from 5000 rpm all the way to redline. When the owner of the car asked about the high knock count, the tuner told him that this level of knock is acceptable.
This is the way the tune looked after I finished it. These are 4 back-to-back 3rd gear WOT logs:
The average boost was set @ 21 psi and tapered all 18.xx psi by redline. I wish I had the before boost charts with the Dejon Tool to compare, but judging from prior logs of a Dejon Tool MBC set @ 21 psi it should taper to 17-18 psi by redline.
The AFR on this Evo 8 came out nearly perfect. 12.xx:1 during spool-up, 11.4:1 at peak boost, and slowly tapering to 11:1 by redline. I have Logworks to thank for this. Reading Evoscan logs in Logworks virtually takes out the guessing out of tuning. The charts that you see are exact replicas of the tables in the rom.
The timing on the car was set up on the conservative side. I believe that I can still add 1-2* of timing from 5000 rpm to redline on this Evo. Unfortunately, we ran out of time to test this theory. So the timing was set as you can see below:
The reason why I believe that there is more timing advance to be had with this Evo is because the knock was almost non-existent, especially at the top end. Four back-to-back WOT logs and not even 1 count of knock 6000 rpm and above.
Since I did not do a base run, I did not have an accurate before and after power comparison. But I did have the logs that were done on the car when it was first tuned. I was able to collect data that allowed me to account for weight and weather variations between the two logs you see below. Two variables I could not account for were: road conditions and the stock cat vs. HFC. So the comparison is not really a good one.
2.75” Down pipe
Stock Cat
RRE Stealth Exhaust
Drop-in filter
Walbro FP
Dejon Tool MBC set @ 23 psi (eyeball-the-gauge method)
Tune (same tuner as the car in first post)
We changed the car to:
RRE High Flow Cat
ECU boost control with #65 pill
Looney Tune
Due to time constraints I did not get a chance to do a baseline log. I had to install the HFC and the boost hose assembly with the pill just before the tuning session. We also found a pin hole leak in the HFC and had to go to a muffler shop to fix it.
What I do have are the logs that were given to the owner by the famous tuner. The logs do not include boost and AFR numbers since the famous tuner does not log boost and uses a narrowband sensor to set the AFR. It baffles me that a famous tuner does not have a wideband in his tuning kit!!!!
So here is what a typical log looked like:
In this chart the car knocks from around 5K rpm to redline. There are two knock spikes, one at 5000 rpm and one at 6900 rpm. Timing is pulled 4* and 3* respectively. As with the prior car, the tuner refused to listen what the knock sensor was telling him. By 7000+ rpm that car wanted no more than 14* of timing, but the tuner set the timing in the map at 17* in the 200 load cell and 20* in the 180 load cell. The knock profile of all 4 logs looked like this:
NB: The ECU_Load in this chart is actually LoadCalc.
Most of the serious knock is concentrated in the top of the rpm range from 5000 rpm all the way to redline. When the owner of the car asked about the high knock count, the tuner told him that this level of knock is acceptable.
This is the way the tune looked after I finished it. These are 4 back-to-back 3rd gear WOT logs:
The average boost was set @ 21 psi and tapered all 18.xx psi by redline. I wish I had the before boost charts with the Dejon Tool to compare, but judging from prior logs of a Dejon Tool MBC set @ 21 psi it should taper to 17-18 psi by redline.
The AFR on this Evo 8 came out nearly perfect. 12.xx:1 during spool-up, 11.4:1 at peak boost, and slowly tapering to 11:1 by redline. I have Logworks to thank for this. Reading Evoscan logs in Logworks virtually takes out the guessing out of tuning. The charts that you see are exact replicas of the tables in the rom.
The timing on the car was set up on the conservative side. I believe that I can still add 1-2* of timing from 5000 rpm to redline on this Evo. Unfortunately, we ran out of time to test this theory. So the timing was set as you can see below:
The reason why I believe that there is more timing advance to be had with this Evo is because the knock was almost non-existent, especially at the top end. Four back-to-back WOT logs and not even 1 count of knock 6000 rpm and above.
Since I did not do a base run, I did not have an accurate before and after power comparison. But I did have the logs that were done on the car when it was first tuned. I was able to collect data that allowed me to account for weight and weather variations between the two logs you see below. Two variables I could not account for were: road conditions and the stock cat vs. HFC. So the comparison is not really a good one.
Last edited by nj1266; Oct 7, 2007 at 07:26 PM.
Oct 7, 2007, 05:38 PM
#11
I had to use 9 since the base run was very choppy. I used 9 on both run. I usually use 8 which is 1 above the default of 7.
Oct 7, 2007, 06:31 PM
#14
