Mitsubishi 3000GT & Dodge Stealth Forum banner

1 - 20 of 30 Posts

·
BAWC: In the Mix
Joined
·
5,168 Posts
Discussion Starter #1 (Edited)
Code changes for injector swaps are fairly straight forward, but swapping in larger turbos is going to change the air flow characteristics the stock ECU is programmed for. More air flow means higher engine load.

Think of engine load like this:



Most cars spend 90% of their lives idling and cruising so car manufacturers need to tune emissions for that period of a cars life. That is why there are so many columns setup under vacuum.

Been getting a few PMs from members about tuning the fuel and timing tables. First off, in order to really get a good fine tune, you need to know which cells to adjust in the tables. I haven't verified the load logging functions in EVOScan are correct and I suspect they're just calculated guesses. If that's the case they're not super accurate. This is something I'm working on and is my first priority. There's a limitation to how much engine load we can log as well. A write-up and will be created once I figure it out 2-byte logging!

I suspect even with a turbo as small as a 13G, we're going to be hitting load cells which go past the safety's in the ECU and limitations of the MAF.

The 3/S primary load sensor is the Mass Air Flow Meter (MAF). It is the ECUs biggest asset, but it can be a limiting factor to the engines ultimate performance when attempting to tune beyond the stock parameters. The MITSUBISHI air-flow meter utilizes the Karman-Vortex principal to measure the air-flow using ultra-sonic sound, which after processing outputs an alternating voltage (AC) proportional to the air-flow. The frequency of the stock 3/S MAF signal ranges from about 30Hz to 2600Hz, with very good resolution at low air-flow rates. This allows very precise AFR trimming at light engine loads for emissions purposes.

A larger turbo (or even running more boost with the stockers) is going to suck more air thru the MAF so we're going to hit load cells much higher than what you see in the tables. This isn't a big deal as anything larger than the last column of cells, the ECU just uses whatever value is in the last cell. I suspect there is a max load value programmed into the ECU somewhere, but none has yet been found.

Yes, you can change the scaling value in the axis for load. I don't recommend changing the scaling for the columns under 100 engine load though. The reason is, there are other tables that use this axis scale that we haven't identified yet. If you change the axis, whatever value is in those tables, is going to get utilized in whatever column you just changed it to. This is going to make the car do all kinds of strange things, especially during warm-up. It's also going to screw up all the emissions tests!

We're working on setting up with a separate and much larger fuel and timing table to use for tuning high engine loads. This is in the distant future right now though.

FUEL CUT

Another limitation is the safety restrictions programmed into the ECU. You know that sensation where it feels like your call hits a wall during hard acceleration? Yeah, that annoying fuel cut! Guys running a piggyback don't have to worry about fuel cut because they fool the ECU into seeing lower load values so that is why they don't ever hit it.

We do have the ability to change this using the Boost Limit table though!



The column on the right is the load value the ECU will activate fuel cut. I'm not sure (yet) how long the ECU will allow that load before it shuts off the fuel injectors.

Stock MAF

The next limitation is the physical amount of air the stock MAF can count. On a 3/S this is about 375 g/s (ECU limits it to 288 g/s) at 2700 HZ. A 15G class turbo or larger, can EASILY flow more than that since 375 g/s = 410HP!

Fortunately, mitsu makes a MAF sensor that can count a lot more air than that and I'll do a write up about how to swap it in once I get a chance to test it. So for now guys, I suggest holding off on the large turbo swaps with the stock ECU until all of these limitation issues have been solved.
 

·
BAWC: In the Mix
Joined
·
5,168 Posts
Discussion Starter #2 (Edited)
Reserved for MAF swap directions
 

·
BAWC: In the Mix
Joined
·
5,168 Posts
Discussion Starter #3 (Edited)
Reserved for tuning calculations with a wideband!
 

·
SDSU Alumnus
Joined
·
2,982 Posts
Re: How to: Tuning fuel and timing tables with ECUFlash

EvoM sure had good insight on the software. Thanks again for replying to my PM's.

In regards to fuel cut and limitations, I think there are many ways we could go about it. To name a few: partially/fully gutting the factory MAS, EVO8 MAS conversion as "TT Eric" has been successfully using, or even a GM MAF retaining stock MAS for IAT/BARO and factory drivability. What are your thoughts?
 

·
BAWC: In the Mix
Joined
·
5,168 Posts
Discussion Starter #5 (Edited)
The EVO 8, 9 and 10 MAF's were a few I was refering to. ;)

The EVO guys had shitty luck with the GM MAFs. Could never get consistant load readings and all kinds of ideling and closed loop load issues... I will probably not look into this option.

Gutting the stock MAF removes the key feature about it. That is to count the air the turbos suck in. When we have the ability to modify the calibration for different sensors, I see no need to gut it. We'd have better luck opening holes behind it than tearing it appart.
 

·
SDSU Alumnus
Joined
·
2,982 Posts
Re: How to: Tuning fuel and timing tables with ECUFlash

Just in case you missed it, Eric did a writeup here: My Evo 8 Maf conversion! <<
Would it be as simple as entering the new "MAF Size" into the rom though? There is good discussion on capacities and power.

I`ve seen over 3100hz on the stock MAF with td04-16g`s, all the screens in..
FWIW, my friend's E01/EMB read over 3800hz on a fully gutted factory MAS. This was with 15-20% in correction on 16G's, so it is definitely capable of handling over 3k. I saw the logs in-person too.

My idea behind the partial/full "gutting" was to allow more headroom by presenting less load/airflow to the ECU. Could this not be an alternative for those who do not want to swap sensors and have a moderate setup? Reference of mentioned idea: Getting the most out of your Apexi S-AFC= Stock MAF honeycomb removal <<
It seems like an effective way to "increase" its capacity although less accurate at metering airflow.

My point of thought is that this "enhancement" should only place us a little lower in the load table than usual right? by kind of stretching it out physically anyway. It would prevent us from hitting any tuning walls. IF this works, would there be a way to enter a baseline adjustment for honeycomb removals? i.e. entering injector latencies for injectors that flow 10% more than what you actually have as compensation [720 numbers for 660]

Sorry if my logic is flawed or totally incorrect. I am willing to test this out though if you think it may work... possibly trying both gutted factory MAS and Evolution conversion.
 

·
BAWC: In the Mix
Joined
·
5,168 Posts
Discussion Starter #7 (Edited)
Just in case you missed it, Eric did a writeup here: My Evo 8 Maf conversion! <<
Would it be as simple as entering the new "MAF Size" into the rom though? There is good discussion on capacities and power.
Yes, I am familiar with the swap. That's just the first step on how to install the MAF. What still needs to be done are the MAF calibrations on the ROM (this is what the reserved post is for).

The MAF size is only one table the ECU uses. Like the injectors, there are more calibration tables that need to be adjusted. We have all 5 tables found in the 99 ROM (MAF size, scaling, smoothing, adder and sensor filtering). In the EVO rom, they named them different but I have a hunch, they're still the same. What I'm still trying to figure out is the correlation between those tables. I'm afraid, I might have to manually tune these tables in the 99 rom on a real car...

Hopefully this kid will buy my EVO soon so I can buy this 98 VR4 I've been eying out...

What's really going to boggle your mind is I fully expect a MAF recalibration will be needed on cars with aftermarket intake pipes to get the fuel trims with-in spec.

FWIW, my friend's E01/EMB read over 3800hz on a fully gutted factory MAS. This was with 15-20% in correction on 16G's, so it is definitely capable of handling over 3k. I saw the logs in-person too.
With no honeycombs to resonate, I question the accuracy of that number... The MAF uses an air stream-centered conical vortex generator to precisely-shape all the air disturbances in the measurement chamber. It does this through a honeycomb grid to straighten airflow and reduce turbulence.

Located on either side of the chamber are a transmitter and receiver that send and receive a signal using ultrasonic waves. The amount of distortion in signal caused by the vortices, whose frequency increases with airflow, is measured and compared against a known set data points, from which an inference about air flow is made. Without the honey combs, the airflow is chaotic which I could see as beeing read much higher than what's actually there.

Unmodified, the MAF can count higher 1604hz BUT the ECU doesn't export anything higher than 1 byte information (2*6.29 = 1604) to a logger. This was the limitation I was talking about. I tried to research what the MAX hz readings the MAF is physically limited to but didn't have any luck.

I was wondering how people were seeing values greater than 1604hz though. I completely forgot the E-manage had that logging ability...

When tuning cars, I almost never log the MAF. Engine load is the cruicial info. The reason I want the ability to log HZ on a 3/S is because I know the stock MAF can only count so high and want to know where that limitation is.

Sorry if the way I stated the origonal post was confusing.

My idea behind the partial/full "gutting" was to allow more headroom by presenting less load/airflow to the ECU. Could this not be an alternative for those who do not want to swap sensors and have a moderate setup? Reference of mentioned idea: Getting the most out of your Apexi S-AFC= Stock MAF honeycomb removal <<
It seems like an effective way to "increase" its capacity although less accurate at metering airflow.

My point of thought is that this "enhancement" should only place us a little lower in the load table than usual right? by kind of stretching it out physically anyway. It would prevent us from hitting any tuning walls. IF this works, would there be a way to enter a baseline adjustment for honeycomb removals? i.e. entering injector latencies for injectors that flow 10% more than what you actually have as compensation [720 numbers for 660]

Sorry if my logic is flawed or totally incorrect. I am willing to test this out though if you think it may work... possibly trying both gutted factory MAS and Evolution conversion.
If that's the case, then tuning the calibration tables for a hacked MAF would probably be the best approach. The load numbers you'd be seeing would still be false though... I also believe it'd be MUCH more difficult to get a good tune that way though.

I would NOT alter the tune for the injectors to compensate for false MAF readings. That's just another variable. I'd first tune the injectors till the trims were happy, then gut the MAF and tune the MAF tables.
 

·
SDSU Alumnus
Joined
·
2,982 Posts
Re: How to: Tuning fuel and timing tables with ECUFlash

Ah, I see your point clean and clear. BTW, is there an updated rom out? because I only have eight tunable tables currently [not that it really matters anyway]. The 2-byte logging sounds exceptionally useful for our application. This would be vital for the "Map Tracer" feature, correct?

What's really going to boggle your mind is I fully expect a MAF recalibration will be needed on cars with aftermarket intake pipes to get the fuel trims with-in spec.
That IS interesting! It reminds me of the Imprezas and Evolutions requiring tunes for their staged mods. This is probably going to be the same scenario on the 98/99's with at least BPU/breather enhancements, yet alone larger turbos with greater CFM ratings.

I would NOT alter the tune for the injectors to compensate for false MAF readings. That's just another variable. I'd first tune the injectors till the trims were happy, then gut the MAF and tune the MAF tables.
That was my other thought. I guess I will see what can be accomplished with it until you are able to test+finalize the Evo8 calibrations.

On another note, I stumbled upon this while browsing the DSM guys' methods for calibrating their GM MAFs:
maffrequencylimits [ECMtuning - wiki] <<
How accurate is that though? 11% increase while maintaining reliable readings.

Heh, they even referenced JL's site: 3000GT MAS Karman Hz to Air Flow Conversion <<

Apparently, 2700hz seems to be the magical number agreed upon, no?

//edit//
I also found this useful piece of information for those interested in the conversion:
DSM Wiseman said:
The Evo MAS uses a different plug, although the wiring is arranged the same fashion so making a harness is simple if you have a harness from an Evo or a host of other cars:

01-04 Chrysler Sebring Coupe
02-04 Dodge Stratus Coupe
00-02 Eclipse
99-03 Montero Sport 3.0
03 Outlander
99-03 Galant
 

·
BAWC: In the Mix
Joined
·
5,168 Posts
Discussion Starter #9 (Edited)
Apparently, 2700hz seems to be the magical number agreed upon, no?
It's 2*11.3x so between 2600 and 2700hz.

Good ole Jeff! Thanks for posting that link. That chart is a good indicator if you even need a larger MAF depending on your turbo setup.



From what I understand, it takes about 150 cfm per 100hp... So for 300 horsepower you need a turbo that puts out about 450cfm.

I'd need to take a MAF reading from a completely stock 1G but 450cfm according to stealth316 is about 1800hz. That means even on a stock car, you're maxing the logger... That can't be right, I'm missing something there...

According to this link you posted Alan, the stock 3/S MAF is good for about 680cfm or 450hp. The EVO8 MAF is good for a bit under 600hp.
 

·
BAWC: In the Mix
Joined
·
5,168 Posts
Discussion Starter #10 (Edited)
The 2-byte logging sounds exceptionally useful for our application. This would be vital for the "Map Tracer" feature, correct?
Yes but Maptracer isn't going to work when you have EVOScan configured for a 3/S. :p
 

·
SDSU Alumnus
Joined
·
2,982 Posts
Re: How to: Tuning fuel and timing tables with ECUFlash

I'd need to take a MAF reading from a completely stock 1G but 450cfm according to stealth316 is about 1800hz. That means even on a stock car, you're maxing the logger...
Did you really want those test results? A buddy of mine has a completely stock '93 VR-4 with 80k. The only "mod" was removing the BCS restriction for 2G-spec. I might even be able to temporarily install an EMU to log higher than the 1600 1-byte limit if the data will be of any value.

Yes but Maptracer isn't going to work when you have EVOScan configured for a 3/S. :p
HAHA! Where is 3Scan when you need it...

Would my statement via PM still stand though?
...would it be safe to assume that we can TRY to tune higher boost but should avoid any huge spikes that taper?
This is in reference to using the last load column onward in the maps for everything greater/unseen.
 

·
BAWC: In the Mix
Joined
·
5,168 Posts
Discussion Starter #12 (Edited)
Did you really want those test results? A buddy of mine has a completely stock '93 VR-4 with 80k. The only "mod" was removing the BCS restriction for 2G-spec.


HAHA! Where is 3Scan when you need it...

Would my statement via PM still stand though?

This is in reference to using the last load column onward in the maps for everything greater/unseen.
I'd just tune the last column for your boost spikes. Take note, I updated the picture in the first post to show the load values I was seeing from the couple cars I was tuning.

You can also rescale the axis from 120 to 160 (remember to not touch the columns under 100 load) to be greater than it but since you can't log more than 160 engine load anyway (yet) so it's kind of a crap shoot.

I haven't posted up my formula yet on how to tune the fuel table using a wideband. This makes tuning a lot more accurate than simply "guess and checking". Give me a bit of time. I even made an excell spread sheet to do it for you! Just got to adjust it to match the 3/S fuel table.

OH and yes please send me the logs from his car.

3Scan :lol:
 

·
SDSU Alumnus
Joined
·
2,982 Posts

·
Registered
Joined
·
4,353 Posts
I`ve taken all the screens out of my MAS bar the centre one and mildly ported it (Snakeskinners idea).

Granted it needs a re-tune but afterwards it`s still rock solid and reliable.
 

·
BAWC: In the Mix
Joined
·
5,168 Posts
Discussion Starter #15 (Edited)
I`ve taken all the screens out of my MAS bar the centre one and mildly ported it (Snakeskinners idea).

Granted it needs a re-tune but afterwards it`s still rock solid and reliable.
People have done it and yes it can work. That's not the point. You can tune for the air, it's just not the best way to set it up. The idea isn't to get a MAF that flows more air, it's to get a MAF that will COUNT more air.

Something tells me, many guys would rather gut their stock MAF than buy another. It's their preference.
 

·
Registered
Joined
·
4,353 Posts
Fair do`s, the EVO MAF looked like a restriction over the stock MAF though.
 

·
BAWC: In the Mix
Joined
·
5,168 Posts
Discussion Starter #17 (Edited)
Fair do`s, the EVO MAF looked like a restriction over the stock MAF though.
No doubt but If guys like Mellon can make 700whp with the EVO MAF (on his silver 8), I'm confident it'll do fine on a 3/s.

Actually I think his MAF is gutted too.... :p
 

·
Breaks everything
Joined
·
9,937 Posts
Greg, wanna come street tune my 99 this fall/winter whenever i finish being lazy and install the rest of the mods. Would be a simple one, 450 injectors, 19T's, want to run 15psi and no more. Just reliable/simple/fun.

Jason
 

·
BAWC: In the Mix
Joined
·
5,168 Posts
Discussion Starter #19
Greg, wanna come street tune my 99 this fall/winter whenever i finish being lazy and install the rest of the mods. Would be a simple one, 450 injectors, 19T's, want to run 15psi and no more. Just reliable/simple/fun.

Jason
I'd be happy to.

19Ts and only 15psi though?! :p
 

·
Breaks everything
Joined
·
9,937 Posts
Or whatever it'll make on pump fuel...I want to build the driveline before adding real power so i can bang on it and not think its just gonna break every time i fancy a drive(wishful thinking i know..haha)

Jason
 
1 - 20 of 30 Posts
Top