When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.
Yeah, I wasn't going to pull the trigger but you guys got the wheels turning hahaha. I need something to tinker with in the meantime and the small power bump, more compact packaging, and induction noise pushed me over the edge.
I ended up stealing your design with the two 45* elbows, MAF sensor housing, and a long 90* bend, but bought it all off Amazon. I think it came out to a bit more than your total cost but I had it all by Sunday so it was worth it haha. Thanks for the insight on that.
More tuning stuff, yay! I've got a WOT lean condition at high altitude and ambient temp that reared its head just after things started getting hot around here. The Fab9 manifold is slated to land Monday, and given I'll be rescaling for the larger MAF/intake and (hopefully) higher VE, I figured I should get rid of this current issue beforehand.
Preface first: This ECU does its primary fueling calculations based off of the MAF, but also allows for injector pulse width tweaks to be made on a load scaling table which plots engine load (from the MAF) and RPM. I only ever modified the two highest load rows on my load compensation map (1.0g/rev and 1.16g/rev). This is all I ever needed to mess with so far, as at WOT I've historically only ever bounced between those cells regardless of altitude and temp. Since I never really hit the lower load cells at high RPM, I just kinda interpolated those rows with the low load cell values. With the VE-modifying changes I've made to the engine (cams mainly, I suspect), I've had to increase the multipliers as load increases. Long story short, I just interpolated the mid-high load values for fueling (since I'm almost never there and have few datalogs in this zone), and so far this has worked. The O2 readings showed that we were hitting our target AFR of 12.7-13:1 at just below WOT.
Screengrab of the old load compensation table. Note how the cells in the lower right corner taper down as load decreases.
So I figured everything was well in the fueling world, but as this Summer heated up, the car started to run at 13.5-13.7:1 as it neared WOT at redline in the canyons. I should've jumped on this quickly. However, with me and my buddies' program of meeting up at the crack of dawn, driving a few runs, socializing, driving again, and then racing home before our significant others are out of bed, I somehow went four or five canyon runs without attempting to address the issue lol. Depending on temps and altitude, the 13:5-7:1 condition would start to occur above ~6,300rpm, however the rest of the rev range would go on the lean side too, closer to 13.2:1.
A couple cool cars I guess...
So this weekend I finally busted out my laptop at altitude and ran a bunch of datalogs and tune revisions to clear up the lean condition at high ambient temp/altitude. This is where having an OEM ECU that needs to be reflashed every revision is a bit of a PITA. With a standalone, I could've ran autotune or at least made changes to the fuel map on the fly. My cycle looked more like: run a log, pull over, review log, make changes to fuel map, flash ECU (3-4 minutes of chill time), run another log. Not extremely efficient but I got the changes I wanted to make dialed in over the course of a couple hours, with the engine holding a steady 12.5-12.7:1 at WOT all the way from 3k-7.2k rpm.
Updated load compensation map. Took some more datalogs when I got back down to sea level and the engine was asking for less compensation at max load near redline. I don't know why this is but those 1.14 values in the 1.16g/rev row kept it holding 12.5:1 till redline where I live. For reference, Load usually peaks at 1.12g/rev at 4k rpm, then tapers off to ~1.07 at 5k, then back up to ~1.1 at 6k, then down to ~1.05 at redline (sea level and ~70*f IAT's).
Damn, I guess I could've condensed that all to: I finally tuned three more rows of my fuel map.
Oh well, I'm excited to have it dialed in so it was worth a post.
Also, the evap canister and shield on these cars is worth 7 pounds. I deleted my evap purge valve and lines at the beginning of this year but left the rest of the system unmodified and have been dragging around this damn dead weight the whole time. Pulled the canister out last week and looped the vent lines with some random tubing I had laying around, 1/2" I think...
Typically your evap canister has the vent solenoid integrated into it. Not the case with this car. The vent solenoid is mounted to the body below the trunk and above the rear subframe, along with another valve of some sort downstream of it. I'm pretty sure I can pull those off and shave a whole additional 1.5 pounds, but I only had time for the canister last week.
More interesting stuff next week. I already have my Bosch 68mm TB and everything needed to make a 3" intake, and the Fab9 IM and throttle body adapter harness are supposed to show up on Monday. Should be able to pop that sucker in next week and see it really is good for another 10whp as advertised.
Last edited by Z_WAAAAAZ; Sep 5, 2025 at 11:53 AM.
Reason: Corrections
I hadn't even thought about your car being DBW until just now. Makes sense give the year, I think my 07 Subaru was the first year they converted over to ethrottle as well.
Pretty plug and play then aside from wiring and a recalibration (or do you even have to calibrate those)? Any advantages over the stock throttle body aside from more airflow?
I didn't either until I first drove it My first time driving it hard, I was thrown way off by the throttle control. Ended up getting used to it but now that I have control of the DBW tables, I have the throttle set up fairly linear to try and emulate the cable actuated throttle on my last car.
I thought for sure there was going to have to be some recalibration for the throttle body but by some miracle apparently both TP sensors on the Bosch TB read within .05 volts of the OEM TB through the full throttle sweep. No recalibration required aside from playing with the DBW tables if desired. Seriously, I'm kinda blown away that it's PNP aside from the wiring.
Main advantage should just be more airflow. My maximum WOT MAP at sea level is around 95kpa up until 5k rpm when it drops down to around 93kpa, so there's an intake restriction somewhere. Apparently the upsized throttle body and a 3" intake (stock is 2.375") is good for another 1-2kpa and 3-5whp on an N/A car. The intake manifold is supposed to be good for another 1-2kpa and even more top end at 5k+ rpm.
Really grasping at straws for more power without building another motor or going FI haha.
Last edited by Z_WAAAAAZ; Sep 17, 2025 at 01:11 AM.
More tweaking tonight because I induced a new hiccup with my tuning session last week. I got the lean spot at high altitude and ambient temps sorted out but this morning I did a pull on the way to work and AFRs went rich, about 12:1, near redline. Did a bunch of messing around after work, including adding some resolution to the load cells at WOT, checking my MAF and MAP sensors for discrepancies, and confirming via the logs that the ECU was simply adding more pulse width than usual near WOT.
Long story short, I found out the software interpolates between roughly 3 or 4 vertical rows in the load table, not only two as the visual aid shows while the car is running. If you look at my post above, you'll see that map with the two leaned out cells in the bottom right corner was nailing my AFR's in warm temps. This morning and evening (cooler temps out), it resulted in a rich condition, with load at ~1.1 g/rev across the board. I added a load cell at 1.08g/rev for better resolution in the area and kept leaning out that area bit by bit. No substantial change occurred even though I was running between the 1.08 and 1.16g/rev cells at WOT. Example below. You'd think that leaning out the circled cells would fix the target area but it was only making a marginal difference. I was hitting 12:1 in said cells.
So after a couple tries, I decided to lean out the 0.875 and 1.0 cells for sh*ts and giggles, even though I wasn't theoretically touching them during the pull. Boom, leaning out those cells resulted in AFRs in the higher load area dropping back down to 12.5:1.
So I guess I have some interpolating and blending to do. Gonna richen up that 0.875 row again tomorrow and see if it has any effect on AFRs in the rev range there. If it does have an effect, that means the ECU is using four of the vertical rows at one, seemingly equally, for its calculations. That's really weird to me, but good to know. Hope this helps somebody eventually because I have nothing else to contribute to the site this week
More tweaking tonight because I induced a new hiccup with my tuning session last week. I got the lean spot at high altitude and ambient temps sorted out but this morning I did a pull on the way to work and AFRs went rich, about 12:1, near redline. Did a bunch of messing around after work, including adding some resolution to the load cells at WOT, checking my MAF and MAP sensors for discrepancies, and confirming via the logs that the ECU was simply adding more pulse width than usual near WOT.
Long story short, I found out the software interpolates between roughly 3 or 4 vertical rows in the load table, not only two as the visual aid shows while the car is running. If you look at my post above, you'll see that map with the two leaned out cells in the bottom right corner was nailing my AFR's in warm temps. This morning and evening (cooler temps out), it resulted in a rich condition, with load at ~1.1 g/rev across the board. I added a load cell at 1.08g/rev for better resolution in the area and kept leaning out that area bit by bit. No substantial change occurred even though I was running between the 1.08 and 1.16g/rev cells at WOT. Example below. You'd think that leaning out the circled cells would fix the target area but it was only making a marginal difference. I was hitting 12:1 in said cells.
So after a couple tries, I decided to lean out the 0.875 and 1.0 cells for sh*ts and giggles, even though I wasn't theoretically touching them during the pull. Boom, leaning out those cells resulted in AFRs in the higher load area dropping back down to 12.5:1.
So I guess I have some interpolating and blending to do. Gonna richen up that 0.875 row again tomorrow and see if it has any effect on AFRs in the rev range there. If it does have an effect, that means the ECU is using four of the vertical rows at one, seemingly equally, for its calculations. That's really weird to me, but good to know. Hope this helps somebody eventually because I have nothing else to contribute to the site this week
Mazda has a table called High load enleanment with these exact cells leaned, I don't know why did that.
Maybe because they have the AFR super rich to begin with. If you have that table anywhere in romdrop, zero it out.
Yo, thanks for the heads up man. I'm aware of the enleanment table, although it's called the "fuel compensation table" in ECUtek. This table had a bunch of values populated in the high load/rpm cells when I first started tweaking the tune but I zeroed it out. Don't remember what the initial values were.
Took the car up to even higher altitude today and fueling was more dialed in, lean spot at near redline mostly gone. I added a little more load compensation to the 0.6875 g/rev row even though I was hitting 0.90g/rev at redline. After that the lean spot cleared up. It seems like the engine just wants vertically linear compensation values for each RPM cell at high load, which I guess makes sense because we're shooting for 12.5-12.7:1 in this entire part of the map, instead of 14.7:1 like everywhere else. I was treating it like a VE table before, tapering upward with load. I'm learning, slowly...
The Fab9 intake manifold is showing up tomorrow so I pulled the OEM manifold out of the car today. Pulling the manifold on one of these cars is normally an ordeal, but so many things have been removed from this car's engine bay at this point that it took all of ten minutes to get it out. Sweet, mission accomplished on my end lol.
New 68mm Bosch TB vs the OEM unit.
The most time consuming part of the install was actually pulling the front bumper to route the new 3" intake elbow and air filter. Kinda a PITA for serviceability but the consensus is that longer intake = more power on these engines when naturally-aspirated. My previous TDR intake went straight into the area behind the bumper but didn't make a 90* and snake further along the inside of the bumper.
Yes, longer intake adds torque, robs responsiveness. That column of air has to be accelerated and decelerated and the more air involved the more inertia in it. My first tuner on the original GM (****, 20+ years ago ) did some experimentation. This on an atmo NB8A with a cowl intake, and a Haltech E8 from memory. best power was on the long intake, dropped off as we removed segments.
Yes, longer intake adds torque, robs responsiveness. That column of air has to be accelerated and decelerated and the more air involved the more inertia in it. My first tuner on the original GM (****, 20+ years ago ) did some experimentation. This on an atmo NB8A with a cowl intake, and a Haltech E8 from memory. best power was on the long intake, dropped off as we removed segments.
I'm surprised a longer intake makes more power, intuitively it seems like it would just add more pressure drop at the throttle body. Must have something to do with a resonance.
That column of air has to be accelerated and decelerated and the more air involved the more inertia in it.
Ah, makes sense! I was always aware of intake length having an effect on torque and response, but never stopped to think about why. The air mass explanation for response is almost too simple haha.
Originally Posted by crxguy52
I'm surprised a longer intake makes more power, intuitively it seems like it would just add more pressure drop at the throttle body. Must have something to do with a resonance.
I'm not gonna pretend to be an expert on the topic but I've seen a couple back to back dyno plots (stock airbox vs long AEM CAI) for these cars, and the CAI produces a couple resonance windows throughout the rev range.
Just searched for longer than I'm willing to admit and can't find that dyno chart, though. IIRC, the AEM CAI was good for a 4-5whp bump around 4,500rpm due to resonance as well as a couple more HP/TQ above 5.5k.
Last edited by Z_WAAAAAZ; Sep 9, 2025 at 02:03 AM.
Ah, makes sense! I was always aware of intake length having an effect on torque and response, but never stopped to think about why. The air mass explanation for response is almost too simple haha.
I'm not gonna pretend to be an expert on the topic but I've seen a couple back to back dyno plots (stock airbox vs long AEM CAI) for these cars, and the CAI produces a couple resonance windows throughout the rev range, torque
Just searched for longer than I'm willing to admit and can't find that dyno chart, though. IIRC, the AEM CAI was good for a 4-5whp bump around 4,500rpm due to resonance as well as a couple more HP/TQ above 5.5k.
Yup, just recently tuned a car with an aem intake and cams. You can clearly see a hefty bump in load midrange. And intake temps were bang on ambient temps, on a hot summer day.
Quite an impressive intake. Not getting enough credit.
No joke, I would've grabbed one when I got my car if I'd been able to find any buzz on it.
Manifold is in! And we're making power baby. Or at least a little more power lol.
A couple things need tidying up, but she's running again!
Install went pretty smoothly. Luckily I only ran into a couple hiccups. Fab9 didn't include the throttle body harness adapter and my power steering delete pulley bracket also didn't clear the #1 intake runner. Luckily only a few millimeters needed to be ground off of the bracket and I was able to overnight a connector for the TB.
Clearance after taking the PS delete bracket to a grinder. Tight! Hadn't heard a peep about from anyone about this which is kinda weird.
Had the install done last night but didn't have the TB connector yet. Also forgot my rotary tool and had to do a temporary cut on the intake elbows with a razor blade. Don't look at those I had a bunch of electrical pins laying around though, and used those to "adapt" the factory TB wiring and get the car running so I could check my work.
I updated the MAF scaling table but decided to hold off on modifying the DBW tables until driving the car. Multiplied my previous MAF tables by 1.59 to account for the change from the factory 2.375" intake to the new 3" intake. The car started right up and ran, which was rad! I didn't try to drive it, though. Some of the pins were on the loose side and I didn't trust them to stay put while driving.
Got the TB connector today and wired it in. I started the car up and thought, "Wow, this manifold amplifies the chop at idle so much!" It wasn't the manifold. The engine was misfiring.
I drove up the street and immediately knew something was wrong. Lack of power, rough running, and the AFR was going way rich. Of course first I had to go through thinking it was one of many worst case scenarios. Bro I probably left a stack of quarters on the fender and they all somehow wound up going through one of the intake ports and now my valves are bent there's craters in my combustion chambers. Once I got past that stage, I started pulling coil connectors with the engine running. Found the misfire was localized to cylinder 2, and then found it would disappear intermittently if I wiggled the coil wiring. Ended up finding the wiring broken at the base of the coil connector's ground pin. This was annoying because I replaced all of the coil connectors last year due to the factory coil wiring being a common failure point. I probably just manhandled the wiring harness too hard too many times over the last year. Ended up jerry rigging a temporary fix (don't even think about asking) and ordered some new connectors in case another one has an issue.
Oh right, the performance. The logs show measurable gains! Got it out on the road tonight and took a few datalogs at WOT. At 72* ambient temps, I'm seeing MAF readings about 10% higher than before, along with higher load values that I've never seen the engine hit. My MAF scaling settings should produce readings that are close to or spot on where they were before upsizing the sensor housing to go with the new 3" intake. MAP is also holding at 96-97kpa across the board whereas before it would hit 95kpa and taper down to 93kpa by 5,500rpm. I've gotta imagine my MAF scaling is accurate because fuel trims are all pretty close to zero like they were previously, and I'm seeing the ~5g/sec reading at idle like before as well.
There should be more gains I can make up top as well. I'm currently tapering down from 25* to 10* of VVT advance at 6k+ rpm, but have been told that with my setup I'll probably make more power with 15-20* at redline. I was running zero advance at redline with the stock manifold, so figured I'd just start with 10* and work my way up with this new manifold to figure out what it likes. Still waiting on my revision from Fab9 so I'll just keep playing with the tune myself in the meantime.
Quick log. Biggest thing I noted was the additional load and pulse width around the 5k rpm range. Historically, I've seen a pretty pronounced dip in the torque curve on my 2.5's at 5k rpm. With the new manifold, however, load ratings actually peak at 5k (in addition to being higher everywhere else also). Pulse widths in the 6.1-6.2ms range would result in way rich AFRs at that rpm but now they're fairly close to where I want them. I'll continue tweaking fueling this week. This log was with no tune revisions aside from updating the MAF scaling and VVT table.
Oh, and finally, the sound. Holy heck, the induction noise with this thing is awesome. Way more grunt than with the stock intake manifold. I'll get a sound clip this week and post it up here.
Phone microphone doesn't quite do it justice but you get the gist. Intake noise resonates really nicely through the cabin when you get on it. I'm sure the engine's making more power but half my perception might just be from the increased drama when you get on the throttle now lol.
More tuning today. I bumped the VVT up to 15* at redline and dialed in the fuel map a little more. The engine seems to like 15* of advance at redline. MAF readings went up a touch more, hitting 139g/s where they'd previously hit 132-133g/s in the same ambient temps. What's really noteworthy, though, is the broadness of the powerband. Previously, the engine wouldn't generally hit 120g/s until about 6,500rpm, and it would just touch 130g/s at 7k. Now we're hitting 120g/s around 6k and holding above 130g/s from ~6,500rpm on. Really excited to get this thing dyno'd at the next race at WSIR at the end of this month.
Next up will be trying 20* of advance at redline and making a second tune with limited throttle opening in case I go over my power limit at WSIR.
Phone microphone doesn't quite do it justice but you get the gist. Intake noise resonates really nicely through the cabin when you get on it. I'm sure the engine's making more power but half my perception might just be from the increased drama when you get on the throttle now lol.
More tuning today. I bumped the VVT up to 15* at redline and dialed in the fuel map a little more. The engine seems to like 15* of advance at redline. MAF readings went up a touch more, hitting 139g/s where they'd previously hit 132-133g/s in the same ambient temps. What's really noteworthy, though, is the broadness of the powerband. Previously, the engine wouldn't generally hit 120g/s until about 6,500rpm, and it would just touch 130g/s at 7k. Now we're hitting 120g/s around 6k and holding above 130g/s from ~6,500rpm on. Really excited to get this thing dyno'd at the next race at WSIR at the end of this month.
Next up will be trying 20* of advance at redline and making a second tune with limited throttle opening in case I go over my power limit at WSIR.
0
My first time driving it hard, I was thrown way off by the throttle control. Ended up getting used to it but now that I have control of the DBW tables, I have the throttle set up fairly linear to try and emulate the cable actuated throttle on my last car.
) did some experimentation. This on an atmo NB8A with a cowl intake, and a Haltech E8 from memory. best power was on the long intake, dropped off as we removed segments. 
I had a bunch of electrical pins laying around though, and used those to "adapt" the factory TB wiring and get the car running so I could check my work.
