MR2 Owners Club Forum banner

MAF pipe discussions and fuel trims

25K views 138 replies 15 participants last post by  Gouky 
#1 ·
So there's a discussion in the Hacked ECU that started about MAF pipe sizing and how it affects the ECU's fueling strategy.

A couple of the talking points so far:
-Excessively low MAF flow readings and the need to reduce the diameter of the pipe
-stock airbox isn't necesairly equivalent to a straight pipe
-do air flow straighteners fix anything?
-does a velocity stack really help?

I think we could just sit and argue the point back and forth and get nowhere. i think we need some way to measure real data as well as someone to do it.

I was thinking of a setup to do the tests but i'm having a hard time coming up with something that sucks as hard as a 2GR when running at idle all the way to redline. a shopvac will move a similar amount of air at the top end but when we try to restrict it it fails to pull nearly as much vacuum. I think this may be OK for this test anyways but if someone has a better idea please say so.

here's my thought for a quick test rig:
 
See less See more
1
#4 ·
Thanks Marc. First autocross of the season today (for me), and I didn't have time to start a thread before I left.

First of all, I will briefly recap my experience on the subject. More detail can be found in my swap thread http://www.mr2oc.com/188-v6-mr2-forum/650833-my-2gr-fe-swap-thread.html

1. Started with a 3" pipe, with no flow straightener. I don't recall paying attention to fuel trims at any point other than idle, but it would throw system lean and a MAF Sensor error codes after about two drive cycles.
2. Added a honeycomb flow straightener, 3/4" thick, 1/8" cells. Eventually added a 1/16" sleeve to the inside of the intake, reducing the diameter ID by 1/8". This fixed the error codes, but fuel trims cruising at 75mph (3500rpm ish) were still around +18-20%. Data logging a 2nd or 3rd gear pull also showed air fuel ratios under load where quite a bit leaner than the ECU was requesting.
3. Built a new intake with 2.75" pipe, still using the same flow straightener. I haven't data logged any pulls with it yet, but cruise trims are 0 on one bank and +5 on the other. At this point my idle fuel trims are about -10, so there seems to be an error in the "slope" of the MAF output, as well as the scale. At least now I seem to have it zeroed out in the middle of the range instead of right at the bottom.

What I am thinking might be a useful starting point here is get some data from anyone who is will/able to provide it. Basically I am wondering how many people have high fuel trims and don't realize it. It seems like a few people who have really tried to get the fuel trims close to zero have ended up with smaller intake piping. Of course, it's also possible that I simply have a bad MAF... but I hate to throw parts at things and would rather not buy a new one without more evidence that this one is bad. *Maybe* it is, but there are a lot of other factors that could be effecting the readings as well.

Here's my data:

Intake ID: 2.75"
Flow straightener?: Y
Fuel pressure: 50psi
LTFT @ steady highway cruise in 5th gear (3000-3500rpm ish): +18

Intake ID: 2.6"
Flow straightener?: Y
Fuel pressure: 50psi
LTFT @ steady highway cruise in 5th gear (3000-3500rpm ish): +2


A test rig would be pretty neat, and throwing an OEM airbox / intake on it would be a nice data point as well, but I'm not sure how necessary it is. In the end, all we are trying to do is match the MAF reading to the OEM MAF tuning. To that end, the ECU itself will tell us how far off we are, as effectively the fuel trims will approach zero as the measured flow rate approaches the actual flow rate.
 
  • Like
Reactions: Myxalplyx
#5 ·
Alex,

I did not realize you were also increasing fuel pressure at the same time. in theory your changes have increased the airflow where the MAF is by about 33% and you've increased the fuel flow in the injectors by 7%.

This explains why you were getting lean fault codes. the ECU will only apply up to 35% correction before is gives up. While many people get fuel trims in the 10-20% added fuel range I've only seen one other build that had lean code issues and in that case it was something wrong with the intake (several years ago, i do not remember the details)

i'd go get numbers from my car but with the supercharger on there the numbers aren't relevant. best i remember my setup was around +10% with the NA setup. I'll be switching back to that setup in the near future and i'll be able to get numbers at that point.

At 50psi people normally report a tiny bit of black smoke on hard tip-in. this is from the part of the map that does not use the trim compensation.

Have you considered sending in your injectors to be cleaned/flow tested?
 
#6 ·
The increased fuel pressure was the first thing I tried, before getting the flow straightener even. To be honest it never really made that much difference. I doubt I saw a 7% change from it (maybe half that), although I was mostly looking at idle trims at that point (for what that's worth). At this point I could probably put that back closer to stock, I just ended up where I am by making incremental changes. I have considered sending the injectors out for cleaning, but kinda thought it might not be worth it since the motor was pretty low mileage. But, maybe there is something going on there.

When you say you had +10% fuel trims, what sort of RPM/Load was that at?
 
#7 ·
the 10% i recall being pretty stable across the entire range but this was a while ago. I'll have to get new numbers for you when i return to stock.

as for cleaning the injectors I've seen some pretty nasty stuff depending on how the junkyards treat the engine while the fuel line is open. there's just something about your situation that isn't adding up and i'm not sure what it is.
 
#10 ·
One other thing I forgot to mention was that part of the reason I haven't looked too hard at the injectors is because my airflow numbers didn't seem right either (and neither do the numbers that a lot of other people have reported). Assuming 18% drivetrain loss, a swapped motor in an MR2 (280whp) is making about 340 crank hp, or about 25% more than stock (assuming Toyota's 269 crank HP number is accurate).

Reportedly a stock Highlander flows about 205-210g/s peak (http://www.mr2oc.com/188-v6-mr2-forum/459195-2gr-p0171-p0174-system-too-lean.html). I would expect one of our swapped motors with intake, headers, etc, to flow quite a bit more than that, given the HP numbers we are seeing. Gavin's comment on facebook: https://www.facebook.com/groups/mr2oc/permalink/10154223835075728/ about using a stock Exige airbox and seeing 225g/s peak flow got me to re-examine my setup. That number seems more reasonable, although maybe still a little low since it's only about 10% more than stock. Assuming 25% power gain = 25% airflow increase, we should be seeing more like 250g/s. I realize there is a lot of assumptions going into that, but I think it's safe to say that our flow numbers should be at least 225g/s at sea level.

At that point, I was seeing peak flow numbers of only about 175g/s. At my altitude I would expect mass airflow to be about 13% lower than at sea level, but when you apply that correction to my 175g/s reading you only get about 200g/s. Less even than the stock Highlander.
 
#11 ·
It's important to distinguish actual intake flow from measured intake flow. I'm not seeing this distinction being made clearly enough. To make the MAF read higher (i.e. to get a higher "fictional" MAF reading), you make the MAF tube smaller. This does not mean the flow has increased - with a smaller more restrictive tube you get less flow not more, even though the ECU is detecting more as a result of the MAF seeing higher velocity! You can make the ECU detect any g/s number you want from the MAF by varying the intake geometry. But (and this is a big but) this has almost no bearing on the actual g/s flowing into the chambers. The g/s number is based on a MAF calibration that (presumably!) was obtained for a stock intake. Ok so what happens when you make the MAF read higher than actual by making the tube smaller? The base maps and tables will call for increased fueling as would be called for at that actual flow level. Presumably the region of maps we care most about is out of closed loop so there is no trim correction. The stock ECU is known to run rich hindering power production, and a smaller MAF tube makes it run even richer. Marc showed the calculations for how much bias can be induced in the MAF readings by reducing the tube diameter - 33% between 2.75 and 3 inch! Does that translate into running 33% richer at WOT? This would be a good question to address.

My preliminary data collection of AFR and timing with my 3-inch MAF shows AFR's in the mid-12's at WOT (below, published previously in my 2GR test and tune thread), which is exactly what I would shoot for if I were tuning, as this is a "magic number." Admittedly there are some artifacts in the data collection because of OBD-2 limitations. I am now working to resolve this by using non-OBD2 data collection (stay tuned! LOLOL). And with that done I'll be ready to do some more meaningful testing.



The ultimate arbiter of this question would be in my opinion back-to-back testing on the dyno with various intake geometries - not just WOT runs but constant load runs to explore the entire load vs. rpm map. While the calibration of MAF as suggested in this thread should be informative, I don't know how it translates into practical results. I believe the questions raised about the MAF are more directly addressed on the dyno. As I've said before I wish I were in a position to do this kind of testing, but I'm not. :(

One thing I found illuminating in the Lotustalk discussion (850 posts very little of it with any real relevance) of the intake is the effect of the tube length. Marc has pointed previously to the effect of the harmonics and these I believe are determined primarily by length (along with other factors). It never occurred to me before that harmonics could induce higher than 100% VE (this was pointed out by Marc in analysis of the torque curve). I see that BOE has settled on a short intake tube design. Presumably this is based on their dyno testing. I'm also looking closely at the pic and it looks like the 45-degree silicone bend may be a reducer from the 3-inch throttle body to a smaller tube diameter - hard to tell on this one. I wonder if anyone has access to one of these units to make measurements.

Another factor in play here is the ACIS. When this thing comes on it gives a real kick in the butt and the MAF readings should just jump right there. So just thinking about this now it occurred to me here that I should be looking more closely at my MAF vs RPM data logs. Why this didn't occur to me sooner (and why none of you learned gents suggested it).... let's not go there LOLOL. Anyway I find this plot of MAF versus RPM collected on successive 1/4 mile runs at the track EXTREMELY illuminating despite some inherent limitations. I don't see anything there directly relevant to the ACIS but I do see something else. Take a look:



So.... there is visibly a restriction on my airflow that manifests at the very top end of the range - see how it flattens out. My working hypothesis is it's due to a. tube length, or b. air filter, or c. exhaust? I've changed air filter since then and immediately felt a change on the butt dyno. No measurements yet to confirm. Intake tube length should be easy to change and I intend to. Of course this could be due to other factors. We shall find out with more data.

One more thing: thanks Marc! This community can count itself very fortunate thanks to you.
 
#14 ·
It's important to distinguish actual intake flow from measured intake flow. I'm not seeing this distinction being made clearly enough. To make the MAF read higher (i.e. to get a higher "fictional" MAF reading), you make the MAF tube smaller. This does not mean the flow has increased - with a smaller more restrictive tube you get less flow not more, even though the ECU is detecting more as a result of the MAF seeing higher velocity! You can make the ECU detect any g/s number you want from the MAF by varying the intake geometry. But (and this is a big but) this has almost no bearing on the actual g/s flowing into the chambers. The g/s number is based on a MAF calibration that (presumably!) was obtained for a stock intake. Ok so what happens when you make the MAF read higher than actual by making the tube smaller? The base maps and tables will call for increased fueling as would be called for at that actual flow level. Presumably the region of maps we care most about is out of closed loop so there is no trim correction. The stock ECU is known to run rich hindering power production, and a smaller MAF tube makes it run even richer. Marc showed the calculations for how much bias can be induced in the MAF readings by reducing the tube diameter - 33% between 2.75 and 3 inch! Does that translate into running 33% richer at WOT? This would be a good question to address.

My preliminary data collection of AFR and timing with my 3-inch MAF shows AFR's in the mid-12's at WOT (below, published previously in my 2GR test and tune thread), which is exactly what I would shoot for if I were tuning, as this is a "magic number."
Agreed (at least mostly). However, my hypothesis is that actual flow is considerably greater than what we are measuring. Roughly speaking, airflow = hp. So if our swapped motors are making 25% more power than the motor does in it's stock application, then they ought to be flowing 25% more air. Or at least, that is the general trend I would expect to see. Yet your logging (with cams no less!), shows only 208g/s peak flow, the same as the logs floating around from a stock Highlander. To me, that just doesn't add up. What I am suggesting is that by going to a smaller pipe we would not be making the reading higher than actual, but in fact making it match up with the actual flow.

I disagree that the actual flow will necessarily decrease with the change from 3" pipe to 2.75" pipe. By that I mean, I doubt we are maxing out the flow capability of a 3" pipe, so reducing the diameter doesn't necessarily restrict the flow. Speed it up, yes, but not necessarily restrict it.

Now, the subject of the ideal AFR for power is another topic entirely. Looking at the requested AFR log from Torque, the ECU want's to run 11:1 at the top end, which does seem excessively rich. But what about the lower end of the RPM range? In the facebook post I linked to Gavin said that he reduced the presence of the torque lumps at the 2500-3500 range by changing his intake. Looking at my logs with a 2.75" ID intake, the actual AFR in that RPM range is 14.5:1, and the ECU is requesting 13.7:1. Seems plausible to me that running that lean at WOT could be costing some torque. See post #150 of my swap thread: http://www.mr2oc.com/188-v6-mr2-forum/650833-my-2gr-fe-swap-thread-3.html

The ultimate arbiter of this question would be in my opinion back-to-back testing on the dyno with various intake geometries - not just WOT runs but constant load runs to explore the entire load vs. rpm map. While the calibration of MAF as suggested in this thread should be informative, I don't know how it translates into practical results. I believe the questions raised about the MAF are more directly addressed on the dyno. As I've said before I wish I were in a position to do this kind of testing, but I'm not. :(

One thing I found illuminating in the Lotustalk discussion (850 posts very little of it with any real relevance) of the intake is the effect of the tube length. Marc has pointed previously to the effect of the harmonics and these I believe are determined primarily by length (along with other factors). It never occurred to me before that harmonics could induce higher than 100% VE (this was pointed out by Marc in analysis of the torque curve). I see that BOE has settled on a short intake tube design. Presumably this is based on their dyno testing. I'm also looking closely at the pic and it looks like the 45-degree silicone bend may be a reducer from the 3-inch throttle body to a smaller tube diameter - hard to tell on this one. I wonder if anyone has access to one of these units to make measurements.
I am thinking about taking my car back to the dyno with the smaller intake pipe to see if there is any difference. I wish I had logged AFR the first time, but at least I can do that on the street.

I still need to go read the Lotustalk threads...

Another factor in play here is the ACIS. When this thing comes on it gives a real kick in the butt and the MAF readings should just jump right there. So just thinking about this now it occurred to me here that I should be looking more closely at my MAF vs RPM data logs. Why this didn't occur to me sooner (and why none of you learned gents suggested it).... let's not go there LOLOL.
In theory if something like the ACIS is tuned right, there shouldn't be a jump when it switches. It should switch at the point where it becomes better to have it open. If there is a jump in power / flow when it switches, that probably means it is switching over too late. Personally I have not noticed when it switches on my motor, but your cams could very well change the optimal switch point.

So.... there is visibly a restriction on my airflow that manifests at the very top end of the range - see how it flattens out. My working hypothesis is it's due to a. tube length, or b. air filter, or c. exhaust? I've changed air filter since then and immediately felt a change on the butt dyno. No measurements yet to confirm. Intake tube length should be easy to change and I intend to. Of course this could be due to other factors. We shall find out with more data.
This could also be cams, valve size, etc. Granted you have the larger cams, but you are also revving an extra 1000 RPM higher than the rest of us and the limit of those larger cams has to show up at some point.
 
#12 ·
back to back tests would be great to a point but this ECU is pretty fantastic at compensating for many MAF deficiencies. it takes about 60-100 miles of varied driving for the ECU to tune itself so you can't really just do it on the dyno between tests.

you can actually see it in action here. the blue line was after an ECU reset, the two red lines were captured on the ECU before the reset and had been driven plenty of miles to get it to tune itself: (note this is a supercharged dyno run but the effect will be similar)


good call on mentioning the difference between measured reading and actual. I failed to make a point of this earlier and this could certainly throw some people off.
 
#15 ·
Here's the 2017 Camry 2GR stock intake.

There's a longer length, smaller diameter intake pipe that bypasses the main intake opening (which is closed by a VSV at low revs). Above a certain rev range (maybe 3500rpm?), the VSV opens the intake control valve, which opens a butterfly and allows full airflow via a shorter and much larger intake opening (combined with the smaller one, I think too).

This is how they've improved low-end torque, which is combined with ACIS, as ACIS uses the plenum as extra runner length with intake piping as plenum/air chamber (long runner), and extra air chamber in the plenum and intake piping in the open position (short runner).

Also ensure your open element filter isn't skewing MAF readings too. Note the MAF readings with just bare pipe, no filter and with a filter. You might some differences if the filter itself is disturbing airflow within the pipe.
 

Attachments

#20 ·
Also ensure your open element filter isn't skewing MAF readings too. Note the MAF readings with just bare pipe, no filter and with a filter. You might some differences if the filter itself is disturbing airflow within the pipe.
I experimented with this briefly early on, before I got the flow straightener. Removing the filter seemed to make a significant difference (increased the MAF reading), but I really only checked at idle. Adding the flow straightener seemed to have a similar effect, but I haven't re-visited the effect of the filter since adding the straightener.

Do you use something like a pre-made housing to build your intake off of?
For the 3" intake I used one of the many 2GR Camry intake pipes available on ebay, with some modification. For my new 2.75" version I built my own with a 2.75" MAF flange I bought off of ebay and a 2.75" 45° aluminum bend. If I had it to do over I would have gotten a 90 actually, as the required bend is actually slightly over 45°, at least in my application. If you don't have the stock coolant neck location to clear (I know you don't on your swap), less bend would be required.
 
#17 ·
if you wanted to do that it would be good to look at the subaru stuff. i know the 2008 WRX uses the same MAF sensor in the same size pipe and a similar shaped airbox. i'm sure the years stretch more than 2008 but i don't know what the range is.
 
#18 ·
Funny you should mention the 08 WRX as this is my DD. While it used to do a fair amount of track duty, it's since been retired to navigate the snowy roads and shuttle kids and groceries. Anywho, I've both the stock and aftermarket intakes so let me know if I can provide any pics or answer questions. BTW, the WRX has been a fantastic car. While it's limits are lower than the MR2, it's a fun and easy car to drive in any condition.
 
#19 ·
I completely agree. the reason i know is because it's what my wife uses as her DD. She's had it since new for about 9.5 years now and it's been a fantastic and relatively low maintenance car except for the clutch. it's now on its 5th clutch and I've had to reweld the clutch pedal bracket to the firewall. Subaru finally admitted that 3 spot welds to hold the clutch bracket on wasn't enough last year... about 4 years after we experienced the issue and fixed it. That clutch is just way too small for the power going through it. Seems like hers is probably one of the few out there that hasn't been modified at all.

but back to the intake, yeah it uses the same MAF and the pipe size at the airbox/MAF is the same size. it then reduces diameter to go into the turbo but that's after where we care about.

Interestingly, going to rockauto to bring up a cross reference it seems the part numbers aren't showing the same applications for both of those despite the fact that they look identical and fit exactly the same so maybe there's a calibration difference between the units but they are physically the same size.

here are the applications shown:
LEXUS ES350 2007-2012
LEXUS GS350 2007-2011
LEXUS IS F 2008-2014
LEXUS IS250 2006-2013
LEXUS IS350 2006-2013
LEXUS RX350 2007-2009
SCION TC 2007-2010
SCION XB 2008-2014
SCION XD 2008-2012
TOYOTA AVALON 2005-2012
TOYOTA CAMRY 2007-2011
TOYOTA COROLLA 2009-2011
TOYOTA HIGHLANDER 2008-2013
TOYOTA MATRIX 2009-2013
TOYOTA RAV4 2006-2012
TOYOTA SIENNA 2007-2010
TOYOTA VENZA 2009-2015
SUBARU B9 TRIBECA 2006-2007
SUBARU BAJA 2003-2005
SUBARU FORESTER 2009-2014
SUBARU IMPREZA 2008-2014
SUBARU IMPREZA SPORT 2008-2009
SUBARU LEGACY 2003-2015
SUBARU OUTBACK 2003-2015
SUBARU TRIBECA 2008-2014
SUBARU WRX 2015
SUBARU XV CROSSTREK 2013-2014

it may be worth digging through that list and finding out which ones use the same size MAF pipe and from there seeing what is available aftermarket that may work well for this swap.
 
#21 ·
Here's another $0.02 on the intake geometry with more to come later as I collect more data:



This video shows the fuel trims at idle, in the cheap eBay 3-inch intake with a salvage sourced $25-dollar used MAF. Can't complain too much about 0% LTFT on both banks.

Inspired by BOE, I've shortened up the tube, removing the portion that curved to the side vent. Haven't yet tested under load. The thought here is that BOE did document testing of different length intakes on the dyno and they settled finally on a short intake.
 
#24 ·
Very minimal then.


Question for you all, the diameter of your intakes, whether it's 2.75" or 3", is this the ID or OD? Reason I ask is because I got my induction pipe from someone on twobrutal who made a batch up years back. I want to extend it so the filter is down nearer the side vent, yet my pipe has a 2.75" ID, everything I can find in the UK has this as an OD, so mot sure if this is a UK weird thing or if I just have a strange sized pipe?

Sent from my SM-G930F using Tapatalk
 
#25 · (Edited)
Tubing is *usually* speced by OD. Perhaps yours is metric? 70mm tube would be 2.756" OD. Maybe not enough oversize to explain the 2.75" ID measurement though...

Regarding IAT's, that was with my "short" intake. The pipe is a little over a foot long with a 45° bend at the end, but the filter isn't really all the way down in the side vent. I'm not sure if it's really worth extending it. But, I do have pretty good airflow through the engine bay, what with no rain guard and side vent ad-on's to scoop a little more air. Perhaps those are helping more than I know.
 
  • Like
Reactions: monkeyra
#27 ·
With tubing we're always talking about OD. Water hoses on the other hand....

My IAT is a few degrees F over ambient as long as the car is moving more than 40mph. If I come to a stoplight the temp quickly rockets up 10, then 15, then 20F above ambient. There is a heat soak/hysteresis meaning if I get moving again the temp does go down but never seems to go back below ambient+10F, until I shut off the engine and let the car cool.
 
#28 ·
Have you wrapped your intake stop heat from contacting it?
Have you done a throttle body coolant bypass?

I don't want to make any assumptions that you have already. This same situation in my car was from the throttle body coolant. Bypassed that, the temp rise was lower. Until I wrapped the intake. Then temps dropped down to either ambient temp or up to +5F from ambient. Sitting idle the temp rise would be much much slower but once moving, quickly drop back down to close to ambient. Same on my Honda Fit.
 
#30 ·
The throttle body coolant is after the MAF sensor, I wouldn't expect it to have any significant effect on measured temps. It's as simple as removing the hoses and capping the ports on the water pipe / coolant neck. Or loop the hose back on itself. I have always heard that the throttle body heat is intended to prevent it freezing up in cold weather, and perhaps to improve fuel atomization (also in cold weather). For a summer car, I don't see any down side to removing it. Probably no huge benefit either... but every little bit helps right? Especially on these motors with their plastic intake manifold, not heating the throttle body should keep it significantly cooler.
 
#31 ·
I bypass the throttle body heat on the race car because it probably gives 0.02hp and with a race car every last bit matters. but on any other car i would rather not have to worry about the throttle body freezing. throttle bodies can freeze even above freezing temperatures, it matters what the dew point is and what the freezing point is and keep in mind when the throttle body is not wide open there's a pressure drop causing a temperature drop. it can be a 10-20F drop depending on the factors.

the DBW will throw a fault code if it takes too much power to close and will limit power anyways so it's probably not a big risk but again you're probably talking about 10-20watts of difference in power output.
 
#32 ·
As indicated, every little bit counts. When my Rav was N/A, I didn't remove it due to the possibility of unexpected freezing in the winter. Boosted though! Bypass that throttle body. Noway it should freeze on a boosted car, ever. The heat from the air pressure after the supercharger/turbo should be enough to keep any ice from building up on it.
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top