Part throttle tuning on the street
05-17-2010, 04:47 PM
#1
Former Sponsor
Thread Starter
Part throttle tuning on the street
Since I'm an amatuer tuner with no access to a dyno, I've been trying to figure out a good way to do part throttle tuning without having to pay for dyno time, and I think I finally got it. Here is what I did.
1) Start with the fuel maps from the "Stock Eqiuvalent" calibration. This is very important, since the FlashPro Manager expects the fuel map to be uniform with parallel load lines. If you start with a map that you've tried tuning with before with an old school aproach, the interpolated suggested fuel changes will probably be wrong.
2) Alter the Target Throttle Plate table so that there are 2 big "zones" of throttle level so that you can easily achieve a smooth / steady target manifold pressure across the whole RPM range. Like this:
The goal with this step is to create a datalog in 2nd or 3rd gear with a very smooth pass through each part of the fuel map at each load level (300 mBar, 400 mBar, 500 mBar, etc.). The slope in the target throttle map is needed because you have to reduce restriction as RPMs rise to maintain the same manifold pressure. The reason for the 2 "zones" is to help you maintain that pressure without having to control the throttle pedal perfectly.
Note: The car will be driveable with this throttle map, but it will be very slow. If you don't have a big vacant parking lot near your home, it might be best to drive the car to the testing area using a canned calibration, then flash this tuning calibration once you are at the testing area.
3) Once the new calibration is loaded, drive the car around for a few minutes to get rid of any heat soak. Then when you are ready to begin the datalog, accellerate to about 1500 RPMs in 2nd gear, hit the datalog button, then push the throttle somewhere in the 1st zone (about 30% throttle). Once you hit the rev limitter (or max speed in 2nd gear), slow back down to about 1500 RPMs then put the throttle all the way to the floor (100% throttle). Once you hit the rev limitter again, slow down and hit the datalog button again to end the datalog.
4) Go back home to analyse the datalog and apply changes to the fuel maps. (Note: I've never used Live tuning, but this might be a good time to try it instead of going back home)
5) Open the datalog.
As you can see, the manifold pressure will be pretty constant through the 2 zones (400 mBar and 600 mBar). Even though the throttle pedal was at about 37% in the first zone, the DBW system automatically forces the throttle plate to follow the ramp to maintain about 400 mBar.
6) Go to the fuel tables to see what FlashPro Manager shows for suggested changes. If the suggestions make sense (obvious trend), and seem to match the L trim and S trim values that are in the datalog that you viewed in step 5.
If there is a clear trend in fuel suggestions, then apply them to the map. In the example above, I have applied what I believe would be good changes to the rest of the map as well.
7) Alter the Target Throttle Plate table again to increase the target manifold pressure for both zones by about 100 more mBar as shown below. As you can see, this does overlap with the previous map, but that is on purpose. The overlap will give you an opportunity to fune tune and validate the changes you just made in step 6.
8) Repeat steps 2 through 7 until you've climbed up to your full throttle manifold pressure. I believe it is better to work from the low load levels up to the higher ones. If you want to go slower for a more precise tune, make smaller increments in the throttle maps. BTW, if you are going to interpolate the suggested fuel changes for cells that you didn't hit in the datalogs, you should only do that for areas that you haven't already hit. Basically, don't interpolate down to lower load levels that you won't be hitting again.
1) Start with the fuel maps from the "Stock Eqiuvalent" calibration. This is very important, since the FlashPro Manager expects the fuel map to be uniform with parallel load lines. If you start with a map that you've tried tuning with before with an old school aproach, the interpolated suggested fuel changes will probably be wrong.
2) Alter the Target Throttle Plate table so that there are 2 big "zones" of throttle level so that you can easily achieve a smooth / steady target manifold pressure across the whole RPM range. Like this:
The goal with this step is to create a datalog in 2nd or 3rd gear with a very smooth pass through each part of the fuel map at each load level (300 mBar, 400 mBar, 500 mBar, etc.). The slope in the target throttle map is needed because you have to reduce restriction as RPMs rise to maintain the same manifold pressure. The reason for the 2 "zones" is to help you maintain that pressure without having to control the throttle pedal perfectly.
Note: The car will be driveable with this throttle map, but it will be very slow. If you don't have a big vacant parking lot near your home, it might be best to drive the car to the testing area using a canned calibration, then flash this tuning calibration once you are at the testing area.
3) Once the new calibration is loaded, drive the car around for a few minutes to get rid of any heat soak. Then when you are ready to begin the datalog, accellerate to about 1500 RPMs in 2nd gear, hit the datalog button, then push the throttle somewhere in the 1st zone (about 30% throttle). Once you hit the rev limitter (or max speed in 2nd gear), slow back down to about 1500 RPMs then put the throttle all the way to the floor (100% throttle). Once you hit the rev limitter again, slow down and hit the datalog button again to end the datalog.
4) Go back home to analyse the datalog and apply changes to the fuel maps. (Note: I've never used Live tuning, but this might be a good time to try it instead of going back home)
5) Open the datalog.
As you can see, the manifold pressure will be pretty constant through the 2 zones (400 mBar and 600 mBar). Even though the throttle pedal was at about 37% in the first zone, the DBW system automatically forces the throttle plate to follow the ramp to maintain about 400 mBar.
6) Go to the fuel tables to see what FlashPro Manager shows for suggested changes. If the suggestions make sense (obvious trend), and seem to match the L trim and S trim values that are in the datalog that you viewed in step 5.
If there is a clear trend in fuel suggestions, then apply them to the map. In the example above, I have applied what I believe would be good changes to the rest of the map as well.
7) Alter the Target Throttle Plate table again to increase the target manifold pressure for both zones by about 100 more mBar as shown below. As you can see, this does overlap with the previous map, but that is on purpose. The overlap will give you an opportunity to fune tune and validate the changes you just made in step 6.
8) Repeat steps 2 through 7 until you've climbed up to your full throttle manifold pressure. I believe it is better to work from the low load levels up to the higher ones. If you want to go slower for a more precise tune, make smaller increments in the throttle maps. BTW, if you are going to interpolate the suggested fuel changes for cells that you didn't hit in the datalogs, you should only do that for areas that you haven't already hit. Basically, don't interpolate down to lower load levels that you won't be hitting again.
05-17-2010, 05:14 PM
#2
Former Sponsor
Thread Starter
BTW, I made some changes to the lambda overlay settings that made this work better or worse. I changed the target lambda for all 3 load levels (Low, Medium, High) to 14.7 so that it wouldn't suggest adding tons of fuel while it was still in closed loop. I would change these back to appropriate values for open loop before doing full throttle tuning. I also changed the transition delay and maximum interpolation to different levels to see what would give the best trend.
05-17-2010, 10:32 PM
#4
Registered User
Join Date: Aug 2009
Location: Los Angeles, CA
Posts: 370
Likes: 0
Received 0 Likes
on
0 Posts
Nice write up! however, there are other ways (i.e. controlling the throttle via right foot at x ratio x rpm x time) to do it that way you don't waste too much time. It's best to datalog part throttle upto your set vtec point. I highly doubt anyone want's to stay in vtec at part throttle for a long period. However, your write-up is very though out and I respect that!
05-18-2010, 05:01 AM
#5
Former Sponsor
Thread Starter
Thanks. I have tried controlling the throttle manually for part throttle tuning, but I just couldn't hold it steady enough to get a good sample. Basically, at low RPMs, very small changes in throttle plate cause large changes in manifold pressure, which cause large changes in short term fuel trims. I've never been able to get good fuel change suggestions from FlashPro Manager for anything other than full throttle without using this technique.
I agree about most people not needing to worry about part throttle on the high cam, but since I autocross and track my car, I spend LOTS of t ime at all throttle levels on the high cam.
I agree about most people not needing to worry about part throttle on the high cam, but since I autocross and track my car, I spend LOTS of t ime at all throttle levels on the high cam.
05-18-2010, 08:07 AM
#6
Registered User
Great idea! I have been logging and trying to get my part throttle fueling back in shape and so far the fuel suggestions have been less than ideal. I will have to give this a shot.
05-18-2010, 03:22 PM
#7
Former Sponsor
Thread Starter
I started over on this today so that I could see how well Live Tuning would improve the process. I was BLOWN AWAY by how easy it is now. I created a modified version of the target throttle map with 3 zones so that I could cover much more of the load levels at once.
Then I created a display that would include 3 guages for throttle level so that I could see when I was within each of the zones. The first guage shows when the throttle is between 1.1% and 12.5%, the second shows when it is between 16.7% and 50%, and the third shows when it is between 66.7% and 100%. Note that you need to avoid the gaps between the zones (12.5-16.7 and 50-66.7), since the DBW system will interpolate and screw up the manifold pressure. When datalogging, it can be difficult to stay in the zones, but try to keep the zone guages in the middle. There seems to be some strange behaviors in the target throttle mapping that I've tried reporting to Hondata months ago, but since it doesn't impact "regular" driving, they haven't responded. One thing to note in particular is that the right 4 columns (75%-100%) seem to be broken, since you can't achieve zone 3 unless it starts on column 66.7.
I drove around an undeveloped subdivision for about 10 minutes hitting all 3 zones several times from about 1500 RPMs to 7000 RPMs, then looked at the suggested fuel changes. They looked VERY GOOD, so I applied them via Live Tuning (Very Nice!) where the obvious trends were, and extended those trends to the right-most columns (high load).
Then I drove around again for another 10 minutes, and checked out the new suggested fuel changes. They looked even better, so I applied the trends again.
Then I drove around again for another 10 minutes, and checked them again. I couldn't believe my friggin' eyes!
I plan to shift the target throttle map up so that the 3 zones fall between and above the previous zones so that I can verify that the interpolation went well.
Then I created a display that would include 3 guages for throttle level so that I could see when I was within each of the zones. The first guage shows when the throttle is between 1.1% and 12.5%, the second shows when it is between 16.7% and 50%, and the third shows when it is between 66.7% and 100%. Note that you need to avoid the gaps between the zones (12.5-16.7 and 50-66.7), since the DBW system will interpolate and screw up the manifold pressure. When datalogging, it can be difficult to stay in the zones, but try to keep the zone guages in the middle. There seems to be some strange behaviors in the target throttle mapping that I've tried reporting to Hondata months ago, but since it doesn't impact "regular" driving, they haven't responded. One thing to note in particular is that the right 4 columns (75%-100%) seem to be broken, since you can't achieve zone 3 unless it starts on column 66.7.
I drove around an undeveloped subdivision for about 10 minutes hitting all 3 zones several times from about 1500 RPMs to 7000 RPMs, then looked at the suggested fuel changes. They looked VERY GOOD, so I applied them via Live Tuning (Very Nice!) where the obvious trends were, and extended those trends to the right-most columns (high load).
Then I drove around again for another 10 minutes, and checked out the new suggested fuel changes. They looked even better, so I applied the trends again.
Then I drove around again for another 10 minutes, and checked them again. I couldn't believe my friggin' eyes!
I plan to shift the target throttle map up so that the 3 zones fall between and above the previous zones so that I can verify that the interpolation went well.
Trending Topics
05-21-2010, 05:59 AM
#9
Former Sponsor
Thread Starter
I'm glad it's working out for you. Did you start over with a "virgin" fuel map, or did you try to work with the map you already had? My theory is that it wouldn't help fix a map that is already screwed up, since it wouldn't be able to interpolate the suggestions correctly.
05-21-2010, 08:46 AM
#10
Banned
the typical rule of thumb is that when viewed as a 2d graph, there should be 3 "groups" of lines and they should all be relatively parallel with the lower load lines being "flatter" than the upper load lines.
They should be smooth, if you have huge differences from one point to the next it could probably be better.
They should be smooth, if you have huge differences from one point to the next it could probably be better.