Dealing with the "crappy tune zone"
02-19-2012, 05:37 PM
#1
Former Sponsor
Thread Starter
Dealing with the "crappy tune zone"
Have you ever wondered why totally stock Hondas have such "spikey" AFR's at WOT? Do you think Honda just did a crappy job on the factory tune? After all, everybody knows that a good custom tune has a nice, flat AFR on the dyno (I'm kidding)! Here's an old AFR plot from my car when I first started tuning it. The blue line was from the stock tune. Note that the AFR bounces all around between ~12.5 and ~14.5!
Well, I finally understand why! It's because Honda didn't tune the car at WOT! They tuned the entire map at partial throttle, since that's the only way to be sure that the car will perform well at any altitude and also have decent fuel economy! Since our ECU is based on manifold pressure (not throttle position), the ECU gives the same amount of fuel at 900 mBar no matter what the position of the throttle plate is. However, the "ideal" amount of fuel (and timing) changes a LOT based on the position of the throttle plate! This is due to the huge impact that intake resonance has on ideal fuel and ignition without having much impact on manifold pressure.
Since a stock '06+ AP2 achieves full MAP at about 30% throttle at 3K RPMs, which is also the resonant RPM for the stock intake, Honda had to choose between a good AFR at 30% throttle OR 100% throttle. If they tuned for 30% throttle, they could achieve good part throttle response, driveability, and fuel economy, but if they chose 100% throttle, they would have to sacrifice that for max power AND have specific tunes for each city! Now it's clear that they chose part throttle, and just allow the AFR to bounce all over the place at WOT.
Since I don't drive my S2000 at significantly different altitudes, and am much more interested in peak performance than I am fuel economy, I went the other way with it. I tuned for optimal AFR and ignition timing at 100% throttle. That choice costs me quite a bit when MAP is near peak with a partially closed throttle (the "crap zone"). Below is a screen shot of the target throttle map that I used to see how part throttle / full load differed from full throttle / full load. I made it so that the throttle plate would slowly open as the RPMs increased even though my foot was all the way to the floor. The resulting manifold pressure was almost the same as if the throttle plate was fully open. Below that is the resulting torque plot and AFR from that throttle map compared to the "stock" throttle map. I should also mention that the torque plot shows dashed line that represent the fuel curve. I wish I would have fixed that, since it is very misleading. I had already tried "tuning out" the difference unsuccessfully using the fuel table. Let me assure you that the data I'm showing is NOT caused by that. These are just the best screen shots I have to deliver the message. I don't have any more time to invest ...
Note how much the curves differ between 3K - 4.5K, which is where most driving occurs! Also note that it would be MUCH, MUCH better if the AFR for the part throttle condition was LEANER than the full throttle condition, but it's actually much RICHER! Now I know why my back bumper is black with SOOT and my fuel economy sucks!
After thinking about it for a year or two, I realized that the issue might be correctable using the target throttle map. Below is the stock throttle map with some illustrations. Note that the values in the table don't represent actual throttle plate position. The values are actually multiplied by ~1.5, so if you have 66% throttle plate at 66% throttle pedal, that results in an actual throttle plate of 100% at 66% throttle pedal.
I spent a few days geeking out on it, and found that if I optimized my target throttle map so that the throttle plate would transition very, very quickly from mostly closed to fully open once I was near full load, I could avoid the "crap zone"! Now I'm able to make a small change in throttle pedal position, and the throttle plate will go from ~30% to 100% rapidly to avoid the area where the AFR goes to crap. It seems to also give a better throttle response, since that part of the throttle used to do virtually nothing.
Well, I finally understand why! It's because Honda didn't tune the car at WOT! They tuned the entire map at partial throttle, since that's the only way to be sure that the car will perform well at any altitude and also have decent fuel economy! Since our ECU is based on manifold pressure (not throttle position), the ECU gives the same amount of fuel at 900 mBar no matter what the position of the throttle plate is. However, the "ideal" amount of fuel (and timing) changes a LOT based on the position of the throttle plate! This is due to the huge impact that intake resonance has on ideal fuel and ignition without having much impact on manifold pressure.
Since a stock '06+ AP2 achieves full MAP at about 30% throttle at 3K RPMs, which is also the resonant RPM for the stock intake, Honda had to choose between a good AFR at 30% throttle OR 100% throttle. If they tuned for 30% throttle, they could achieve good part throttle response, driveability, and fuel economy, but if they chose 100% throttle, they would have to sacrifice that for max power AND have specific tunes for each city! Now it's clear that they chose part throttle, and just allow the AFR to bounce all over the place at WOT.
Since I don't drive my S2000 at significantly different altitudes, and am much more interested in peak performance than I am fuel economy, I went the other way with it. I tuned for optimal AFR and ignition timing at 100% throttle. That choice costs me quite a bit when MAP is near peak with a partially closed throttle (the "crap zone"). Below is a screen shot of the target throttle map that I used to see how part throttle / full load differed from full throttle / full load. I made it so that the throttle plate would slowly open as the RPMs increased even though my foot was all the way to the floor. The resulting manifold pressure was almost the same as if the throttle plate was fully open. Below that is the resulting torque plot and AFR from that throttle map compared to the "stock" throttle map. I should also mention that the torque plot shows dashed line that represent the fuel curve. I wish I would have fixed that, since it is very misleading. I had already tried "tuning out" the difference unsuccessfully using the fuel table. Let me assure you that the data I'm showing is NOT caused by that. These are just the best screen shots I have to deliver the message. I don't have any more time to invest ...
Note how much the curves differ between 3K - 4.5K, which is where most driving occurs! Also note that it would be MUCH, MUCH better if the AFR for the part throttle condition was LEANER than the full throttle condition, but it's actually much RICHER! Now I know why my back bumper is black with SOOT and my fuel economy sucks!
After thinking about it for a year or two, I realized that the issue might be correctable using the target throttle map. Below is the stock throttle map with some illustrations. Note that the values in the table don't represent actual throttle plate position. The values are actually multiplied by ~1.5, so if you have 66% throttle plate at 66% throttle pedal, that results in an actual throttle plate of 100% at 66% throttle pedal.
I spent a few days geeking out on it, and found that if I optimized my target throttle map so that the throttle plate would transition very, very quickly from mostly closed to fully open once I was near full load, I could avoid the "crap zone"! Now I'm able to make a small change in throttle pedal position, and the throttle plate will go from ~30% to 100% rapidly to avoid the area where the AFR goes to crap. It seems to also give a better throttle response, since that part of the throttle used to do virtually nothing.
02-19-2012, 06:00 PM
#2
Former Sponsor
Thread Starter
I forgot to include the shots of my updated fuel maps. I added a column for 950 mBar, since that is just below the atmospheric pressure that I've ever seen where I live, so I feel confident that any time the MAP is below 950 mBar, it's definitely partial throttle. I tuned that new column for partial throttle plate.
02-19-2012, 06:02 PM
#3
Former Sponsor
Thread Starter
Sheesh ... now I realized that I forgot to include the screen shot of my new AFR plots showing the difference between the true full throttle and "crap zone". It isn't perfect, but it's on the right side of perfect (not uber-rich)!
02-22-2012, 03:24 PM
#5
Great posts! Your posts always seem to highlight the benefits of data logging and analysis.
Here's a post from the AEM forums a few years ago on a similar topic. Lots of experienced tuners chimed in.
http://forum.aempower.com/forum/index.php/topic,13630
Explanation:
With the AEM, if you tune boost comp, a throttle modifier is often used to reduce AFR at part throttle. The fuel table is flat for any given RPM and fueling is delivered based primarily on throttle modifier in vacuum and the boost comp table in boost (double pressure double fuel assumption). I tried this sort of tuning at first using the throttle mod to achieve leaner part throttle AFR. At least with the AEM, I found drivability to be lacking and AFRs to be all over the place when using the throttle modifier alone. I found better control using boost compensation in vacuum and boost, eliminating the throttle modifier correction, and tuning cell by cell. I got the cell by cell tuning and drivability very close on the road and polished it off on a load bearing dyno where I was able to hit and sustain loading at the vast majority of the cells. I can't say enough how awesome a good load bearing dyno with real time tuning capability is for tuning part throttle. Said tuning achieved ~30 mpg on the highway with great drivability.
Typical boost comp tuning:
http://forum.aempower.com/forum/index.php?topic=11494.0
A later post I made with screenshots of the final tune:
https://www.s2ki.com/s2000/topic/804...ase-map-setup/
Tim
Here's a post from the AEM forums a few years ago on a similar topic. Lots of experienced tuners chimed in.
http://forum.aempower.com/forum/index.php/topic,13630
Explanation:
With the AEM, if you tune boost comp, a throttle modifier is often used to reduce AFR at part throttle. The fuel table is flat for any given RPM and fueling is delivered based primarily on throttle modifier in vacuum and the boost comp table in boost (double pressure double fuel assumption). I tried this sort of tuning at first using the throttle mod to achieve leaner part throttle AFR. At least with the AEM, I found drivability to be lacking and AFRs to be all over the place when using the throttle modifier alone. I found better control using boost compensation in vacuum and boost, eliminating the throttle modifier correction, and tuning cell by cell. I got the cell by cell tuning and drivability very close on the road and polished it off on a load bearing dyno where I was able to hit and sustain loading at the vast majority of the cells. I can't say enough how awesome a good load bearing dyno with real time tuning capability is for tuning part throttle. Said tuning achieved ~30 mpg on the highway with great drivability.
Typical boost comp tuning:
http://forum.aempower.com/forum/index.php?topic=11494.0
A later post I made with screenshots of the final tune:
https://www.s2ki.com/s2000/topic/804...ase-map-setup/
Tim
03-01-2012, 06:24 AM
#6
Former Sponsor
Thread Starter
Thanks for the feedback. I finally had an opportunity to read it. 
I don't have boost, so my issue is entirely about the huge change in the amount of intake resonance that occurs over a very small load range. I did my part throttle tuning by simulating a load bearing dyno on the street by using a restricted throttle plate. This allows me to use my software dyno to tune part throttle just like I do the full throttle. Here is a thread about it.
https://www.s2ki.com/...-on-the-street/
Over the past week, I've continued to tweak my strategy. I realized that I could reallocate the top half of my load indexes (unused boost columns) back down to the load range I actually use. I looked at the historical weather data to see what the min / max barometric pressure range is in my area, then concentrated those extra load columns across that range. Then I analized my MAP curve to see how much the MAP value changes across the RPM range, and overlaid that on my maps to determine which cells would always be WOT and which cells would always be PT. There several columns (spaced at about 10 mBar each) that cross from WOT to PT depending on RPM. By incorporating this with the throttle mapping above, my fuel economy and throttle response has improved significantly. I normally don't make comments about throttle response, since it's so subjective, but this truly provided a large change. I've never felt much change in power when I went from ~40% throttle to 60% at 4K RPMs, but I sure do now!
Below are my new fuel and ignition maps.
I don't have boost, so my issue is entirely about the huge change in the amount of intake resonance that occurs over a very small load range. I did my part throttle tuning by simulating a load bearing dyno on the street by using a restricted throttle plate. This allows me to use my software dyno to tune part throttle just like I do the full throttle. Here is a thread about it.
https://www.s2ki.com/...-on-the-street/
Over the past week, I've continued to tweak my strategy. I realized that I could reallocate the top half of my load indexes (unused boost columns) back down to the load range I actually use. I looked at the historical weather data to see what the min / max barometric pressure range is in my area, then concentrated those extra load columns across that range. Then I analized my MAP curve to see how much the MAP value changes across the RPM range, and overlaid that on my maps to determine which cells would always be WOT and which cells would always be PT. There several columns (spaced at about 10 mBar each) that cross from WOT to PT depending on RPM. By incorporating this with the throttle mapping above, my fuel economy and throttle response has improved significantly. I normally don't make comments about throttle response, since it's so subjective, but this truly provided a large change. I've never felt much change in power when I went from ~40% throttle to 60% at 4K RPMs, but I sure do now!
Below are my new fuel and ignition maps.
Trending Topics
03-02-2012, 07:29 AM
#8
Registered User
Join Date: Dec 2011
Location: Decatur, AL
Posts: 57
Likes: 0
Received 0 Likes
on
0 Posts
I know this is a bit off topic but I cant resist. You have obviously put a lot of effort into optimizing your tune. What surprises me so much is the difference in ign advance between the low and high cams. When you switch cams at 3600 your timing jumps from roughly 16 degrees to roughly 27 degrees. Is that where you found the best power or is this something you did just for this test?
03-02-2012, 09:19 AM
#9
Former Sponsor
Thread Starter
Dynamic compression is higher on the low cam, so that demands less ignition advance. If I give it any more advance on the low cam, I get tons of knock, but I could give the high cam several more degrees without any knock.
03-02-2012, 09:21 AM
#10
Former Sponsor
Thread Starter