my first real cotritbution here. this is derived from about 400hrs of maf car tunning on several 1227302 ecms. so this most likely will apply to the 1227165 but maybe not.


the first issue is the maf tables themselves.all is not what it appears to be. the maf tables are a represention of the maf input FQ in MGHZ.i could get into the various calculations to explain how they arrive at counts but every ecm will do it differntly.

but in short heres how they do it

x = FQ
y = ?? whatever it might be in the scaler

x / y = counts

ok so now we have a table that looks like this


counts gm/sec
1583 86.0

etc etc etc

whats critical here is not to gte thinking that the airflow value here is absolute becuase its not. the airlfow value is what you tell it to be. so if the ecm see

900mghz it come to the counts and look for the value associated with 900mghz. in this case we will call it 20.
so the ecm will know when the value for counts is 20 it will look up what the airlfow should be. in this case 20gm/sec
so now it has 20 counts 20gm/sec. this is great but how do we arrive at airflow. i tolied over this for a while and i got tired of geussing and borrowed a CFM meter from a friend. i used a flow bench and a scope ad created a paper table of the various CFM vs FQ

so now that i knew how much air i had flowing thorugh the maf i could simply trunciate the results to gram/sec. well in my case this was a stock maf with the screen removed. so i had some very minor devitations from the stock values by just a bit. so i simply moved the appropriate values for flow vs FQ(which is really counts) and that solved the issue with the maf being out of calibration. now with that striaghtned out i moved to the.

injector pulse width VS load value

which oddly enough is dependent on the maf tables. here is where the real tunning is. becuase the ecm really only uses the MAF flow table to come to here for the injecotor calcs.

so this is where i spent my time. there is really no need to adjust the maf tables unless you have the following circumstance met

larger maf
removed screens ported housings.
or a LT1 maf with an interpreter.

other wise leave them alone unless you feel for sure that they are inacurate.

so in short

ecm is doing this for fueling

FQ =to AIRFLOW

AIRFLOW to INJECTOR PULEWIDTH. then on to fuel trim etc etc etc.


the big adder here is the line of slope. when there is no value the ecm is simply drawing a line from value to value and creating a mid point or an approximation. the ecm does this for all the fuel tables.maf flow, injector pulsewidth. i would be impossable to have all the various values represented so they just put in some fairly crical amounts and leave it at that. then allow the ecm to calcualte the remander.

i hope this helps somebody out there shaking there head about maf. if anyone has anything to add please feel free. this is by no means the end all be all of maf tunning. also it would be good to finally have a realy good discusion on scalar tables and what they do.

oh and on a foot not there is really no need to buy a LT1 maf sensor and hardware either. you could simply build a bigger housing for the MAF card to flow more air as the bigger housing will simply reduce the amount of veloctity around the wire and still max out at 255gs/sec. even though itll flow more that still does not matter. one its up to you then rebuild the injector pulse wdith vs LV8 table and be fine. im going to be doing this soon and ill let you all know whats going on little later on in the project.

 

Good work! One comment though, on a 165 MAF using Tunercat 6E definition file, there is no "injector pulse width vs. LV8" table. That would be nice though.

 

Nice work. You did a great job on that post. Thanks

 

Funstick

You up to sharing your findings on MAF flow vs FQ to the rest of us that have similar MAFs minus screens...

great work...:hail:


I have started a working/playing with calculating cfm needs vs rpm and your finding may help fine tune my calculations to adjust MAF counts based on engine size and efficencies.

??? did flow max out on the MAF (in other words - max cfm of stock MAF without screen?)

 

You know why I signed on to this board today? To make a post ASKING questins about what you just put! Thanks so much.

*Bookmarked!*

 

Good job. You're right on target with this thinking I believe. If anyone has a motor that actually will flow more than 255g/s and they want to run MAF, then THIS is the way to do it!

I guess there was what you could call discussion on the topic some time back. Here's the link:
https://www.thirdgen.org/techbb2/sho...hreadid=114335

(not trying to chum the waters here...)

 

i think i failed to propery communicate some thing abot MAF sensro flow.


ok the maf senor actually measures air based on speed.by enlrging the sensor housing you will reuce the velocity.


so heres how it works in theory at the moment.

2inch pipe flows 500 cfm V= velocity

v= 345ft per sec

now you get a load of 255grm/sec

now if the pipe were say 3inchs you would see a V of 250 ft per sec so now you have this sort of reading

3 inch pipe flows 500cfm at 250 ft per sec.

now what we have done is reduce the oveall output FQ of the MAF by reducing the incomig air velocity across the maf wire. this basically allows you to rescale yr inputs at the MAF vs COunts tables.

now all i have to do i build a bigger housing and calibrate it. will also try ot get a couple of FQ vs FLOW charts up soon.

hey guys i am trying to demistfy the MAF ok. if you guys could dig out the various calibration form teh code so we can how each maf so we know that x FQ equals x flow in grms per sec then we have got the complete handle on the ecm.

im gona keep it comming.

 

ooh and for the guy who asked about the screens. i only saw a real flow differcne in the 3-6 gram region and the 130 gram area. as the V6 i am working with really doe not pull more air than this just yet. simply using tunercat and the scale feautre i multipliehe values by 1.05 ot correct them. i hope that helps. i have zero idea if the V8 maf cars need more calibrtion up top.maf table 1 and maf table 6 but yet agian this is for a V6.

also there apears to be a table in the 165 ecm $6e code that is mising that is present in other maf systems. the INJECTOR PULSE WIDTH vs LV8. this table needs to be found.

 

Alright, gotta pick up the chalk on this one...

vel2 = vel1 * (d1/d2)^2
where d1=2 inches and d2=3 inches. That's the ratios of the cross-sectional areas (pi*r^2 or pi*d^2/4) not the diameters.

So, for 500scfm (that's standard cubic feet per minute at 60 degrees F and zero inches of vacuum, or 1 atmosphere absolute pressure) the AVERAGE velocity through a 2" pipe will be dependent on pressure and temperature of the gas in the pipe. Generally, I guess you can assume that the pressure will be near 1atm (no vacuum) and temperature is near ambient (60 degrees). For a 2" pipe, your cross-section will be 0.0218 sqft. That means that at 500scfm at standard conditions, your gas velocity is around 382 ft/sec (or 4.3 mph).

For a 3" pipe, you can just correct the velocities by the ratio of diameters, so the 3" pipe velocity for 500 scfh will be 382 * (2/3)^2 or 170 ft/sec.

So by increasing your diameter by 50%, you can decrease your velocity (and hence your mass flux across the wire) by 55%. You would think at first you'd have to double your pipe size to do that, but it's just the cross-section you have to double. In fact, the cross section of a 3" pipe is 2.25 times that of a 2" pipe.

Okay, well you may ask what about the velocity at the actual wire? That, coupled with gas density, which is determined by temperature and pressure, is what gives me the mass flux reading at the MAF, right? Well, is it different than the average gas velocity? Let's see.

If the gas flow is turbulent, then you will have a relatively uniform profile of gas velocity across the radial distribution of the duct. To determine if it is turbulent we have to look at something called the Reynolds number (Re). If it is greater than, say, 5000 (2100 is considered the 'transition' point), then the flow will be turbulent largely. It is equal to:
D*u*rho/mu , where D=pipe diameter, u=average gas velocity, rho=gas density, and mu=gas viscosity. This is air at/near ambient conditions.
D=2"=0.0508m
u=382ft/sec=116m/s
rho=MP/RT=28*1/(.08206*293)=1.1gm/L=1.1kg/m3
mu=1.7e-5N-s/m2=kg/m-s
Units cancel, and Re=381,298.8, way turbulent.

What if you were running at 20scfh and a 3" tract? Re would be 381,298 * (20/500) * (2/3)^2 = 6778. This is kinda in the 'transitional flow' regime, and that's why (at least partly) at the low end of your MAF curves you see some nonlinearity. When flow is not 'turbulent' you have some parabolic profiling where the velocity near the middle of the tube is higher than near the wall. None of this of course takes into account the actual center 'tube' of most MAFs, where D=1/2" or so.

What the screens can do is to 'reset' the parabolic velocity profile that develops at lower flows, as well as some of the ill-characterized eddies and swirls that can develop locally in the tract. Then the flow will not develop to a profile again by the time the gas crosses the wire. What the fins can do is something similar. They introduct an obstruction in terms of both cross-section occupied and drag associated with the fin surface area. This will prevent swirling from occuring around the sample port, and it will also cause some flow to be preferentially forced through the port. Taking out screens and fins will not only allow swirl to occur, but it will also let more flow bypass the sensor. Don't know what the balance of the effect will be, but it's likely to change the 'shape' of your volts-vs-gms curve down low compared to at higher flows.

If I had to guess, I'd say that keeping the story straight at the lower flows is the real challenge, because at higher flows you're fully turbulent and it doesn't matter. Perhaps a combination solution, where the screens and fins are kept in tact, would be good.

Different line of thought. How about using a stock MAF, and instead of gutting it, descreening it, and adding a big honkin pipe, you could just add a little 'occlusion' to the sensing port, like a baffle, that would redirect flow away from the wire. Assuming the pressure drop across the whole piece didn't kill you in terms of engine VE, you could flow 4-500scfh through the sensor and it would read like 230 or so. Just a thought, but then again, why take the drop? Bigger tube would help. Or just a bypass pipe...

I'll shut up now, I don't know why this MAF subject gets everyone worked up, me included. I guess it is kinda interesting. Same principles apply throughout the engine flow stuff.

 

craig i simply posted a thoery the math i had not done. but i did undestand what i was saying i simply hadnt done the math yet. it just looks like the simplest way to fix the maf flow issues associated with it being a restriction. i am also planning on implementing the maf screens on the larger housing to help organize the airlfow.

 

That table, the injector pw vs. lv8 is usable in the ecu file for the $32 165 files, but not in the later $6e tables. ive always wondered why. maybe someone can find them, and get tc and winbin to add them :hail: . Great thread. i just hope it doesnt get "ugly" in here, as some of the other threads have.
Bob

 

That's pretty sweet...

Can you give some more details around your procedures for the measurement? This is really interesting. What MAF style? Freq and not volts? Intake/outlet ducting? Any baffling? Compare 'raw' readings to those interpreted at the ECM in terms of gms/sec? Any plans to extend to different 'conditions' of MAF? A 'bypass/gutted' version? Etc?

I'd like to hear more, this sounds like a pretty solid piece of work. Thanks for the contribution!

BTW, what other naming would that table go by, and what does it look like? I can dig around some...

 

well think of the lv8 vs injector pulse width as a VE table of sorts. as for the maf testing its a little ways away. i had to return the CFM meter and i cant find the notes i had scribbled down as to fq vs cfm. arrrgghh. this was a standard isseu gm maf sensor found on mpfi 2.8 v6 cars. doing some crosee refing it tunr up on sevral other cars as well. it look like a generic part. im glad to see this topic generate so much interest. hopefully if well al coloberate we can come up with a viable and real way to tune maf. with other than stock maf sizez.

craig as for other names for the table ???? i got no clues. just look for someting that pulls values from the maf tables then calculates injector pulse width. i would imagine it goes somewhere betwen the maf scalar to the actually injector command. i know thats alot of code. but the good news is that it is just a 2d table.

if you tell me how to do screen shots i will be glad to put up the tables.

oh craig all mafs ouput the older ones anyways have a stable output of 2.5 volts. with a fq range of 34mghz to 150mghz.check it with a multimeter

opps craig i made a bobo on the fq's i fixedit.also on the procdeure for maf flow. i mounted the stock TB and the intact pipping ona flow bench. we calibrated teh bench then ran teh airflow up. i started grabbing FQ readings. then i compared those reading to what i had on the stock engine at various loads and airflows. then i pulled the maf and noted the change in FQ at a given airflow and simply amended the tables.at some point soon i will actaully get of my *** and make a complete table set including oversized housings. but im busy with work so it could take a few day or even weeks:lala: :lala: :lala: :lala: :lala: :lala:

 

You sure on those freq's? Just looking at the 165 stuff, I thought the freq-based stuff went from 34-150Hz and the volt-based stuff went from (roughly) 0-5V, and that we mostly have the 0-5v dealy.

Didn't find much, how about "Default air flow offset per pct TPS vs RPM", sounds interesting. I must be missing something. I guess what you're looking for is a 'fudge factor' table of MAF table bias correction/adjustment as a function of LV8? That's pretty nifty...

 

Thats because it doesnt exist.

 

oh no the table i am talking about actually is a fuel table. it uses the maf flow in grams and turn it into injetcor pulewidth. maybe so far is ha simply been missed or overlooked by other tunners. i spent 2 hrs today working on a new calibrtion today for a fresh 2.8 we just built and let me tell you. with the knowledge i learned formt eh other 400hrs of tweaking ive got this new motor just about dead on. ive got 128's every where. ive got awesome acceleration and i can comand the injector pulsewidth. this table is missing form tunercat on the $6e. maybe they used diferent method to calculate injector pulsewitdh on the other ecms ?? i dont know but if i take say the value

lv8 vs pulsewidth

33g/s = 2380 msec

and i change 2380 to say 4000 i goes seriuosly rich.
its not a correction its the actuall msec time for the injector at a given airflow. its a great table to have becuase i cn tune the a/f dead on. wihtout having to fudge any numbers

 

Note where the table is at, and the lines just above it. Its used as a verification of flow rate based upon TPS and RPM for code 33 and 34 check.

 

I know which table you're talking about, the earlier $32 MAF codes have it. Its not in the 6E anywhere. I'm not reading from Tunercat or the like, but the code itself instead. The table (or similar) isnt there. Its calculated directly off the load and airflow but I am not sure what constants and such it is using to do so. I have been dealing with timing calculations and not fuel stuff since my fuel curve pretty close anyway after some adjustments.

 

well maybe cash $6e isnt the hot ticket after all.

 

I like it better. Better driveability than with the 32. Highway mode too...

 

well if thats the case maybe its time to rewrite teh $6e code for a lv8 vs injetcor pulsewidth table. ??? or maybe add the highway spark mode to teh $32. at first i thought the 302 $3a was gonna be really hard to get good driveability but after gleming the lv8 vs injetcor pulsewidth table i find i can get the ecm to do exactly what i want it to. maybe the drivebility is poor becuase its poorly tuned.i dont have any firm answers on which code is truly better $32 or $6e but if you ask me $6e is easie to tune.

 

ok heres the first of the 2d tables

 

now here is where you can see the fuel curve in the 2d graph.i figured maybe it would be easier to explain if there were some decent screen shots of what i was talking about.this is in fact the fuel table. it takes the values asigned at the maf tablea for airflow and turns it into injector opening times. the trick is in realizing this early in the game. study these table and it will all become clear as to what i am talking about.the maf tables are only used to convert input maf FQ into a fixed airfllow value. this is where the ecm gets fueling info from. stop playing with injetcor constants and start tweaking the fuel the right way.

 

You know when you do 'ctrl-printscreen' it will do a screen capture right out of Windows, and then you can 'paste as new image' in Photo Editor or whatever...

By the way, Exactly what is LV8? Yeah, I know, it's a load variable. Product of MAF, TPS, RPM, etc. But exactly how is it calculated and what does it really mean?

 

www.wellsmfgcorp.com/counterpoints/Counterpoint3_2.pdf

Here is a tidbit from Mark_ZZ3 from the TPI forum....

 

craig the LV8 in this table is quite simply the incoming airflow. teh LV8 #'s apppearing in the tables i have shown are quiet simply the gram/sec of incoming air. its the same for the timing table.


but in general LV8 is the actuall load of the engine exspressed in terms of airflow. obviously if this is the case we can often times just equte LV8 = grams/sec. there are some cases of lv8 or more correctly COUNTS in the maf tables where this relation ship is not true. the maf tables are what trunciate MAF FQ to actuall airflow. then the LV8 vs Inj pulse width table translates the airflow into injetcor openging time. just think of it as a 2d map vs rpm VE table.

the tps sensor in MAF systems is largely ignored. its used mainly for AE, Tcc, WOT and a couple of other functions. the critical eleent in MAF is that if yoreally wan to get yr tune on you have to adjust the correct tables. i sort of just stumbled onto this one in tunercat. and i started palyig with it. i then began to realize its importance. so in the case of the 3d sprak table and the 2d lv inj table we have what is a great tool. its no harder than tunning a s/d system.

as for rpm its only used in one table in the maf system. it used to build a line of slope for injector pulsewdith. its MAX airflow vs RPM. now what you have to do is grab the maf FQ at each rpm and trunciate it to gram/sec and plug those value in to the MAX airflow vs rpm table. this is important if you pull more air the maf systm will ignore it and stop using the injecotr slope if you get outside the aiflow boundries here a example table if at a given RPM you exceed the boundries the egine will go lean.

 

ok here is the MAX AIRFOW vs RPM TABLE. is serves as secondary line of slope. if you find yr running a hair rich at a given rpm reduce the allowable amount of airflow by a couple of grams. this will retard the line of slope and help lean it up a bit.

 

anyone who care to add to this list can. ecms so far without LV8 vs INJECTOR PULSE WIDTH tables are


t31 1227148 gn ecm i cant imagine they missed this one ??
$6e 1227165 tpi 89

 

ok maf systems to date that tune as i have been discussing so far are

1227870 $1f v6 v8
1227165 $32 $32e v8
1227302 $3a

so far this is what i have confirmed. i will go ahead and post more as i have a chance to play with more ecm's. i also plan on doing a huge right up on S/D systems sometime in the next month or so.

 

Funstick

I may have missed this but am hoping to have it clarified.

On the ARAP BIN the representation in the MAF Tables is as follows.
A certain voltage=A certain gm/sec.

In your ???? BIN the representation in the MAF Tables is as follows.
A certain mghz=A certain gm/sec.

Is this correct? If it isnt I have no idea where you are getting mghz from then.

Thanks!

Mark

 

You're correct. On most all of the MAFs you'll see in thirdgen systems it will essentially be 0-5 volts scaled to 0-255g/s. If it is frequency, it will be low-hertz as was mentioned above. The MHz are in newer systems I guess.

I'm not familiar with the 1227302 referred to with MHz, but I guess it exists.

 

put a FQ counter on the maf ouptu wire. it will put out a FQ.

 

Craig

Thanks a bunch.

This is kind of a far off question but anyway.

What do you know about Algebra and Calculus, specifically Linear Modeling?

Mark

 

PDE's baby, Newton Raphson all the way... Go numerical...

 

Craig

Have you modeled the data in the MAF Tables? Looks like GM was fudging around with values for the lower voltage, MAF Table 1 values, maybe? Dont understand why a point is graphed outside the line in MAF Table 2. The rest of the values look as if they were just plugged in right from a formula though.

What do you think?

Mark

Well any I have graphs of what I am talking about but cant get them to post, they are.jpg. HELP!