PDA

View Full Version : Ways to remove the AFM



Trance4c
12-22-2004, 10:16 PM
I would like to start a discussion about ways we can take the AFM out of the loop. This one piece is a HUGE restriction for our engines, and if we could remove it we could gain a very good amount of power from a simple operation. The question though really is.. how simple is it?

The thought has come up between myself and CelicaLaxer (Mike) about hooking up an ST205 ECU and disabling the I/C pump/water setup.

Does anyone have an ECU pin out diagram for the ST205 ECU? This would be a mod, that if possible.. I would be willing to do. The cost of a standalone and dyno time is not worth it for the power range I want.

ChrisD
12-22-2004, 10:54 PM
Hehe just want nuke and i were tossing around last night: http://www.celicatech.com/forums/showthread.php?p=10534#post10534

OK here's my take on it. I am almost certain I'm going to ditch the AFM. I am actually shocked no one in the community has done it before, because of how simple it is all seeming to me.

Of course there are lots of ways of doing it. Nuke is looking at some hotwire setups that take a much different approach than I plan on doing.

I plan on doing an AFM to MAP conversion with my SMT-6. Here is what I need to do.

1. Measure and record the voltages of the AFM at different load levels. Based on what I've seen so far, the idle reading is roughly 3.0 V and WOT is roughly 0 V. I'll find out more when I do the actual measurements. I need to measure the voltage at each load point on my SMT-6. The SMT6 gives me 8 points.

2. With the AFM still hooked up, connect the MAP sensor. Drive and test the engine at the same 8 load points, and take note of the MAP voltages at each point.

3. Determine what correction factor is needed to make the MAP voltage equal the AFM voltage at each load point. On the SMT6, enter in these correction factors at each load point for all RPMs in the table used for fuel correction (AFM modification table).

4. Cut the AFM output wire at the AFM (THA wire), and connect this wire to the MAP sensor.

5. Start the car to make sure it works.

6. Remove AFM, and then remove the air intake sensor from the AFM (or find a donor intake temp sensor from another car).

7. Wire in the intake sensor so that it is mounted just inside the air filter.

8. Start the car. Drive and test to make sure all is well.

9. Redefine the load points for my specific tune, such that I have maps up to desired boost level. This is where it is a little complicated for me to explain. I want to have maps up to 17psi, however I do not want 17psi to be at 0volts because that corresponds with a near 100% duty cycle. So I will have the voltage a little higher than that. Maybe .5 volts or something. Also determine what load % is best for obtaining partial throttle, wide open throttle, cruise, and idle points. This might represent something like 20in/hg, 10in/hg, 0psi, 5psi, 10psi, 13psi, 15psi, 17psi. It will depend on the tune.

OK hopefully this makes sense. Someone please let me know if Im on the right track here, and if there's anything that should be added. Of course this isn't the be-all end-all method of doing this conversion, but given the tools I have, I think it's the best way.

ChrisD
12-22-2004, 10:58 PM
Here is a web page on a ford truck doing this type of conversion using the same method I plan on using. http://www.idaautomotive.com/changing_over_a_maf_to_a_map_sen.htm

Also, I am not going to use the Toyota turbo pressure sensor. I am either going to use the pressure sensor for my greddy boost gauge (if it is of the right voltage range), or I will have to buy some aftermarket MAP sensor.

Hooligan
12-23-2004, 12:47 AM
HKS used to make a VPC that would get rid of the AFM in favor of a MAP setup, but I think it's been discontinued.

Also, hotwire setups still have a machined housing that all of the air must pass through. Basically like a MAF withut the flapper door. There is still some restriction there.

nuclearhappines
12-23-2004, 01:06 AM
dave i don't quite see the restriction with a hot wire, check this out

http://www.blown4x4.com/Engine_Mods/MAF_Porting/maf_porting.html

it's almost an open pipe... there's a few wires running through the center on most of them and that's it... and you can use a 3" or a 3.8" (i think those are the common two sizes) and you can just yank it from the yard from an ls6, lt1 ...etc and they have the ability to meter in the 500-600 hp neighborhood

-nuke

Hooligan
12-23-2004, 01:16 AM
Nuke, you're right, there's hardly anything in there, but the pipe itself might not be the correct size. If you can find a 3"+ one that you can decipher the wiring scheme from, you're in good shape. I have seen a few in the 3" range, but wasn't sure how large you wanted to go.

Also, some of the hotwires that I've seen have intentional ribs/vanes in them to get a cleaner flow of air across the hotwire itself in order to get a more accurate reading. Many of the newer ones are accurate to 99+% if properly calibrated and modification of the internal geometry could make it less accurate. I'm sure this inaccuracy would be more noticeable at lower RPMS as the air is slower moving and more turbulent.

Trance4c
12-23-2004, 01:21 AM
hrmm... I imagine a 3" piece like that though wouldn't be much of a restriction in such a small pipe like that.. the question though is.. do I want to put this in front of my turbo without anything? Then.. with the extension of this off the turbo if I don't want it bolted right on it.. how much of a restriction does that pipe become.

Interesting first thoughts.. curious, chris.. how much was the SMT6?

Hooligan
12-23-2004, 01:25 AM
Clayton, I definitely don't think you'd want to bolt this right to the turbo. I don't think it would like the extra heat, especially since heat is how it measueres airflow.

How big is the inlet on the turbo? If you could find a larger diameter hotwire and then transition down to the inlet size with the largest radius as possible on the transition, there would not be a large efficiency reduction.

nuclearhappines
12-23-2004, 01:29 AM
the vr4 owners run this either in draw through or blow through setup (ie after the airfilter before the turbos or after the turbos right before the throttle body)

there's alot of debate on draw through vs blow through... the main reason for draw through (pre turbo) is that by going this route you don't get oil (from possibly bad turbo seals) or really high temps on your sensor element so you're not likely to damage the hotwire because it is sensitive (i've heard of people using k*n's having to clean hotwires because of the k&n filter oil coating the sensor and altering the readings).... but then again i also know of alot of GM owners spraying 100-150 dry shots through this type of MAF.

The main reason for blow through (right before the throttle body) is that the sensor measures the air entering the engine...which gives you a much more accurate reading, and a car who'se performance is much less affected by boost leaks (because the sensor measures air after most of leaks... ) so you don't end up running seriously rich because of a leak... the counter arguement is the high temps +oil possibly killing the sensor.

I think a hotwire is much less restrictive compared to your airfilter that is on the end of hte pipe... however i think the flapper afm is more restrictive than both...

jmho

-nuke

Weasy2k
12-23-2004, 01:42 AM
I am planning to use the SMT7 to remove the AFM from my near stock mr2 turbo. I will tune without a AFM and dyno...i will then take my trucks MAF sensor (96 tacoma V6 3.4L) and wire it in to take over from the MAP and tune to that and see the dyno results.

Reason why i am doing this is because with a MAF sensor if you ever decide to change a cam you wont have to retune the whole car at all, this is becuase the SMT6/7 will be reading the air ENTERING the engine and not just the amount of PSI. As the engine ages the VE also goes downhill so it wont be sucking in as much air (even with pressure)and the effiency goes down hill as well (nothing DRASTIC). The MAF will be able to correct this by continuing to read the Airflow vs the MAP will just read pressure and not adjust the loss of effenciacy over time.

With that all said and done ill post all the results on my website and in the forums.

The SMT6 is currently selling for 442usd or 575cdn.
For 505usd you can include a 2.5bar MAP sensor.

The SMT7 (once released) will cost 580usd/755cdn (at current time)

(sorry about spelling im typing to fast...we need speel check :P)

Trance4c
12-23-2004, 10:40 PM
This is what I want... NOOOO restrictions!

http://www.floridadrift.com/forum/album_pic.php?pic_id=940

Lagos
12-24-2004, 02:10 AM
happy wiring...let me know how it goes..

EDIT: save the pic and zoom into it to read it better. i also have ths full st205 bgb in pdf format if you really need it.

http://www.gunstig.org/6gc/images/ST205ECUdiagram2.jpg

JGS3SGE
12-24-2004, 06:10 AM
The GM MAF is VERY different from the common hotwire ford, I believe they are less sensitive to oil/dirt. But the way they work are suppose to be completely different. :confused: You guys who think a 3.14" ID MAF is restrictly to your non-standalone using ass should have your heads checked LOL, J/K but really, thats a 450+HP draw thru MAF...wheres the restriction?

As for the temp thing... its a HOT-ASS wire, WTF is some hot air going to do? :lol:

Flapper door to hotwire is the easiest conversion, it'll relate to the flapper ECU farely close. A little tweaking and it'll work great.
MAP sensor ECUs require the air temp input to work correctly all the time. And while the ECU may have air temp input, its function is likely only used for timing correction as the AFM corrects for air temp automatically.
Does anybody know if the VPC had ACT input?

Weasy2k
12-24-2004, 06:40 AM
Well then you can easily fix that :) connect the Intake temp sensor to the smt6/7 and tune it according to its temp to...just like you would for water temp multiplier :P done!

Conrad_Turbo
12-24-2004, 09:15 AM
Nuke, since you are better versed at building solid state circuits do you know or have any that will convert an rpm signal into a 0-5V output? If so that could make things a lot more interesting in the grand scheme of things since I could write up some code that would allow for a correction factor at an indicated rpm range.

See the other thread for some of the basic things I have done that I would use in the programming.

I don't think the air temp sensor is used solely for timing control, I would say it is there to calculate the mass flow rate of air. See the other thread for my reasoning behind it.

Lagos
12-24-2004, 03:17 PM
i think the best way to ditch the afm would be to upgrade to an st205 ecu. all the calculations are done for you, and you have fuel and timing maps up to 17psi of boost. so it would be a nice over all upgrade and well worth the effort. tho you would probably have to upgrade to the st205 injectors and ditch your t-vis for things to work correctly.

Weasy2k
12-24-2004, 05:40 PM
I don't think the air temp sensor is used solely for timing control, I would say it is there to calculate the mass flow rate of air. See the other thread for my reasoning behind it.
You are right
It shoudlnt be[just for timing] as most vehicles use the air temp as a modifier for the air density/flow.

nuclearhappines
12-24-2004, 11:09 PM
do you know or have any that will convert an rpm signal into a 0-5V output

yes i believe they have chips that do this for you ... let me look around a bit....

-nuke

nuclearhappines
12-24-2004, 11:16 PM
http://www.analog.com/en/prod/0%2C2877%2CAD650%2C00.html

$10-20 though...

much better

http://www.national.com/pf/LM/LM2917.html

there's that too...digital ckt

http://www.national.com/ms/LB/LB-45.pdf


If we are using a micro controller (depending on how much we are doing with it) ... you can always sample the tach (frequency) you're interested in ... sample at a high enough rate... increment a counter.... after a long enough period has gone by .... normalize the counted number .... push that out to one of your DAC's... reset the counter to zero... and go over...

nuclearhappines
12-24-2004, 11:55 PM
conrad... complex... remember my other thread about using the right thing for hte job....

what's your microchip clocked at ? Do you have a spare input pin on it ?

if you have spare time on the chip....

8000 rpm * 4 sparks per round = 36000 pulses per minute
600 puleses per second

sample the ignitor signal at least 6000 timers per second
if the ignitor is high/ low increment a counter...

every so many samples :
normalize the counted number by the maximum possible count for that period of time ... use this number as your rpm signal
output to a DAC (if you want someone else to see this voltage)
reset your counter and start sampling again

-nuke

Conrad_Turbo
12-25-2004, 12:51 AM
conrad... complex... remember my other thread about using the right thing for hte job....

what's your microchip clocked at ? Do you have a spare input pin on it ?

if you have spare time on the chip....

8000 rpm * 4 sparks per round = 36000 pulses per minute
600 puleses per second

sample the ignitor signal at least 6000 timers per second
if the ignitor is high/ low increment a counter...

every so many samples :
normalize the counted number by the maximum possible count for that period of time ... use this number as your rpm signal
output to a DAC (if you want someone else to see this voltage)
reset your counter and start sampling again

-nuke That'd burn quite a bit of processing time, which is why I'd want to go with a chip that would eliminate the processor from having to do any counting, it can just check the pin, see the voltage, and calculate the rpms. I do have a few LM2917M-8 (surface mount), but I should try and get the ones that I could actually solder myself. Also shouldn't it be 300 pulses per second since it's a 4 stroke engine?

The microchip is very similar to the BasicX 34 or 35 (can see it at www.basicx.com (http://www.basicx.com)) except mine is a modified version with more ram, eeprom space and is configured a bit differently.

Nuke so basically for this project I will use a fresh processor that's sole purpose is to convert a MAP sensor into having an AFM ouput but while doing this still the input from the temp. sensor, then in the end giving the user the ability to change the modified AFM output by plus or minus X% at 1000 or even 500 rpm increments. Although with my processor I am using on the Starlet I have it running about 10-15 different tasks and controlling them all according to the software that I had written.

So really you are right, the chip is overkill as the one I am using has more input/output pins and has much more memory and ram than is needed...but it'd make it so easy to build once the code is written, it is very versatile and can be easily tweaked in the future.

Merry Christmas Nuke. :D

nuclearhappines
12-25-2004, 01:53 AM
2907 it is : )

i just want to see this done before i get another celica ;)

Conrad_Turbo
12-25-2004, 08:02 AM
You can have 300cfm of 100 degree celcius air and it will open the flapper door the same as if there was 300cfm of -30 degree celcius air. The mass flow rates are much different but it will open the flapper door the same. The air/fuel mixture ratio is based on the mass flow rate system, not volume, so the air temp sensor is what allows Xcfm to be converted into a mass flow rate unit, ie: lb/hr or whatnot.

If the air temp sensor was only for timing control, it would make more sense to install it at the throttle body, not before the turbo.

KoreanJoey
12-29-2004, 07:48 PM
I have a MAF sensor on my 3rd Gen :D

Trance4c
12-29-2004, 08:14 PM
i think the best way to ditch the afm would be to upgrade to an st205 ecu. all the calculations are done for you, and you have fuel and timing maps up to 17psi of boost. so it would be a nice over all upgrade and well worth the effort. tho you would probably have to upgrade to the st205 injectors and ditch your t-vis for things to work correctly.

This is the route I'm looking to go. I have an extra TVIS plate that I'm going to have gutted and setup for it. In the mean time, I'll be getting ahold of some MKIV Supra 550's and then I'll go for an ST205 ECU and see what I can do.