only evo X runs SD/MAF combo... earlier ones are MAF only. EUDM evos dont even have a map sensor.
The reason I switched to SD is because big cams and fast spooling turbo wanted to mess with my MAF and it would not idle with MAF and S2 cams/hks 7460 kai turbo.
Sudden stall on decel I also experienced with SD but solved by adjusting the table that turns fuel on on decel. Problem was compounded by a light flywheel.

Yeah, my SD setup doesnt like short cold starts either.

 

The rom you are using is the one responsible for creating and presenting to you the log, so how much time it takes to do so for the specific sensor input and how accurately is presented into the numeric form for you to read is very relative, the same goes for any other input it gets from any type of sensor. So how well the rom patch is written and how bug free it is reflects on the result. The pin you are using on your evo 9 usdm Ecu is different than the one I was using on my 2003 jdm evo 8 ecu. The matter has very much to do with the ECU vs the specific sensor controller you are using, as it is the controller that calibrates and sends the voltage frequency signal to the Ecu chipset responsible for receiving and converting that into the signal from which the Eprom of the ecu reads, it is the same chipset that controls and receives the input frequency signal and then forwards it to the Eprom in the final form. This specific chipset works in a certain frequency spectrum, one that is based on the oem o2 sensor voltage/frequency signal output. This frequency spectrum is of a much more narrow bandwidth than the one the bosch 4.9 LSU wideband sensor puts out. If the external controller/translator of the specific sensor package you are using translates the ECU signal into a set of frequencies that are on the limits of what the oem chipset translator can read then the ECU will take more time to read the signal and it may not be as accurate as it should. At 0-1v analog output from your controller there should have been no problem for the set of frequencies presented to the ECU translator/analyzer, and there should not be any need to dictate the signal into a single specific frequency of .45v in order for the oem Ecu analyzer to be able to rectify a signal, filter it, calculate its mean value, display the histogram of its amplitude, and so forth and present that to the Eprom of the ECU, as the chipset can receive and identify the whole spectrum of frequencies that belong to a 0-1 voltage margin. The stock ecu chipset signal analyzer can read from 0-5v as it is not o2 sensor specific but it receives and analyzes all the signals from all the sensors connected to the ECU. I have not used the specific plx controller nor the innovate one but there is no need for a wideband sensor controller to have a voltage adjuster, as I said the ecu chipset can identify frequencies that belong to 0-5v output. Also I see no reason,( I've used a few different brands of sensor packages that included a external display device, such as gauge, both analog and digital, or a flat screen), for an external standalone controller to need 30sec to calibrate to atmosphere vs CO2 condition as that is the only process it takes an extended amount of seconds, usually between 30-50 sec, and that is done in the initial installation and first use of the controller, after that initial calibration the controller stores the data, and it only takes it 8-10 seconds max to present a reading to the external display. The actual o2 frequency signal is readily available from the sensor itself, as the sensor itself translates the co2 reading and the percentage of oxygen in that reading into voltage frequency signal by itself. The controller/translator is only there to power and heat the sensor, and to translate the receiving frequency into a numeric form that you can read, the same process and function the oem chipset does for the oem narrowband sensor. On the Power fc I have the sensor directly connected to the ecu, before its controller that is, and I have a digital gauge connected to it controller. This way I can verify that the controller is doing its job, but I also have it present no interference and manipulation of the actual sensor signal. As I have live mapping and logging on the power fc, as with all standalones, I can compare the timing and accuracy translation of the external controller's through the gauge, to my laptop AFR numeric display. The external controller process of connecting the sensor signal to the store calibration data and displaying a numeric afr reading on the digital gauge takes 4-5 seconds, and 3 seconds for the power fc and there is no difference value wise. The reason you had a problem with the a specific sensor controller/translator vs the oem ecu is because a set of frequencies of a certain spectrum that belong to a 0-1voltage , may differ slightly in certain frequency characteristics, such as the ones I mentioned above. than another spectrum of frequencies that also belong to another 0-1voltage output. This slight difference in characteristics can delay translation and analysis on the receiving end. You would not have the same problem if you used the plx controller with any aftermaket standalone ecu, as the chipset analyzer these contain on their board is designed and constructed to both read narrow and wideband sensors of any type.



Above 900whp on a Dynojet the oem honeycomb becomes a restriction, and it might actually de-attach itself, unless one bonds it thoroughly to the plastic frame, and or become deformed. I have never seen the mesh collapse or become deformed at 600whp DJ. I'm sure there are plenty of examples over there as well that run between 550-650whp on a Dynojet on oem MAF sensor as is. Removing the honeycomb does not require to rescale nor affects the maf related tables in any way. The sensor relies on pressure, mass and temperature of air to calculate its density, and it only needs an area of 1cm2 of air to do so, and these parameters do not change by removing the mesh, so nor will the sensor frequency output spectrum change. Only if one changes one of these parameters, i.e boost, might need to rescale the maf table so that it matches the higher intake mass of air that goes through it. In regards to MAF vs hp, you can check out Chris(Melon racing) results.


Nevertheless, going down the route you chose in regards to the usage of the honeycomb, presents no issues whatsoever, and since you feel better keeping it, why not.





Marios

 

You are correct regarding factory tune not being a true SD because it doesn't use pressure to define load. However, when you convert to "SD" it only uses pressure to correlated back to Load using the "KPA to Load" table in order to modify the Load based HIgh Octane Fuel map when in open loop, just like a MAF tune. The VE table is simply a multiplier table for the Load based High Octane Fuel map when in open loop. BTW, a "SD" tune uses Load for MIVEC and Ignition timing just like a MAF tune.

 

You typed a lot of words so it was more than I was able to read. As mentioned, the wideband controller has two (2) analog outputs, 0-1V simulated narrowband and 0-5V wideband (10-20AFR). I had the 0-1V simulated narrowband output going to pin 71 (Front O2 Signal) of the stock ECU, which I had been running it this way for 5+ years. With the old PLX wideband controller it put out a good 0-1V simulated narrowband signal immediately so there was no issue. However, when that unit failed I purchased the new PLX widband controller and it goes through a calibration process before it spits out a good 0-1V simulated narrowband signal. The setup works just fine once the wideband controller finishes its calibration process upon startup, but it takes about 30 seconds to get a clean signal and until that happens the car sputters. Innovate also goes through a calibration process, but you can program the until to output a predefined value (usually 0.45v for 14.7:1) during the calibration phase so the ECU doesn't trim. However, I ditched the Innovate because it was killing the LSU 4.9 sensors in less than 2 months. I'm sure there is a mod to the ROM that could have been done to delay going into closed loop, but I decided to just add the NTK sensor on the downpipe and problem solved.


Based on your input I will try the OEM MAF screen and see what happens. I can also pull the screen while the motor is running and see what effects, if any, it has on the MAF signal. Glad to hear someone say it will work up to 900whp with the MAF screen. I remember Melon Racing claiming ~680whp, but he had the MAF screen removed.

 

You've got some patience Sir. There is no way I could chase my tail on the stock ecu speedbumps after being on a standalone. Tweaks, hours of logging and hopes that some change may sometimes maybe work somewhat. The lack of fuel correction and fail safe protection is a no brainer for me.

Wouldn't life be easier by swapping the ecu once a year.

 

With zero exception, there has not been a single software (OEM or aftermarket ECU) I've used that didn't have its gremlins. A few months back I plugged in the Haltech and actually setup the VE table to where the car would idle well, but just haven't had time to mess with it since. At the time I was running the Haltech, it was a painful process to get the OEM ECU running each year because I had several things controlled by the Haltech that had to be reversed. I suppose if I went back to the haltech I could make it a painless process to convert back to the stock ECU for the yearly inspection. I'm going to try the factory MAF and if I still have issues I guess I will just bite the bullet and stick the Platinum Pro back in the car. It was setup for the 10:1 2L on 1000cc injectors so not many changes needed to optimize for the 2.15L on 1600cc injectors with the larger SX362.

 

As I said I have not used neither the PLX nor the innovate sensor controller, and based on what you say here neither will I. I have used quite a few 4.9 LSU controller packages and none has had to calibrate after initial installation for 30 sec every time I start the car, and since you might have not read my comment on it, I will repeat that the stock ECU sensor contoller/analyzer chipset although it is designed around a specific set of frequencies that have to do with the specific sensors the 4g63 engine bears, this enhances accuracy and translation speed, its spectrum capability is up to 5v it has a range of 0-5V that is and it is responsible not just for the o2 narrowband, but for the receiving signals of other engine sensors. It can receive, control and analyze a signal from a five wire wideband sensor such as the 4.9 LSU through the use of the Eprom, but the appropriate rom is needed so that the numeric translation is presented in the log. Most of the roms out there do not have such coding in them, but the coding they use as far as logging is concerned is for a signal that goes through the external standalone controller first. And enough people have reported problems in logging afr and or using it for closed loop through a wideband sensor vs the stock ecu due to the bugs the roms they used had. There is also no rom that delays closed loop, but there was a rom if I am not mistaken that gave the option of bypassing closed loop on idle but still keeping it active during cruising i.e. That was a few years back and I do not remember if it actually worked as it should. I have used the 4.9 LSU sensor with external controller on both the 2003 JDM evo 8 ecu and the Power FC L and D-jetro, with no issues whatsoever. Also know that on a standalone you can setup and map idle on open loop alone.




I will once more repeat myself just to be clear on this and not get misunderstood. The mesh will not be any restriction up to 900whp on a dynojet, but it will eventually detach itself through time, so you need to bond it further more on the plastic casing or remove it.


I do believe Chris@ Mellon made more than that on the stock ecu/maf, around 740whp and MAP made over a 1000whp on the stock ecu setup but I am not sure if it was SD based or MAF.


It is really nice to see fellow gearheads share the same desire and interest at a great degree.




Marios

 

Understood and thank you for the clarification. I would be a little apprehensive to permanently bond the screen to the MAF housing, but it makes sense this approach would keep it from failing.

 

I corrected my previous comment, and added to it, as Chris made 927whp but that was on the AEM V2, he did make though 740whp on the stock ecu and maf with the mesh attached.





Marios

 

MoTec doesn't have Gremlins, it just works. Its alot of setup but it will do exactly what you program it to do .

But even your old pro box will work and has fuel correction incase something odd takes place you aren't swapping an engine. Or having a painful experience

 

I sure would like to get the MoTec, but all of my free cash has been going to build the 4.15L stroker motor for the GTR. The bare block was $5600/per and I purchased 2 of them and I just dropped more cash on new heads, intake valves, upper oil pan, oil pump and valve springs. For now its MAF or the Haltech.

I created a MAF tune last night and got the car warmed up on the SD tune and then downloaded the MAF tune and swapped in the MAF. Car fired right up and Low trims were -3% after idling for 20 minutes so I took her out for a drive. The car drove smooth, but I could tell something was slightly off and the car then threw a P0134 code despite my having a new O2 sensor and despite the car was trimming. Ding Ding, car is on MAF and I'm running the VTA BOV! Ugh, I do have the Tial BOV that is designed for recirculation, but I'm not sure how I feel about welding a tube to the intake and then running a stupid looking hose from the BOV to the turbo inlet. I'm going to plug the BOV for now and see how it does tomorrow morning on cold start since temps are in the 30's.

 

I plugged the BOV yesterday and the car no longer throws the P0134 code so that is good news. Keep in mind the MAF tune I am using is a bone stock 88590715 bin file that I simply transferred over my MIVEC, Ignition timing, Injector parameters, and coil charge parameters from my SD tune so I have yet to do any injector tuning. This morning the temps were below 32F and the car fired right up and I immediately drove the car and it had ZERO hesitation. The car idles well at about 950rpm, but I will likely need to do some slight tweaking. At this point I think I'll spend the money to redo the intake so that I can integrate the BOV recirculation and start tuning the setup on MAF.

 

I had time to tune the injectors on MAF in preparation for boost tuning as soon as I get the BOV re circulation setup. The ECU doesn't have much resolution for the injector scalar so it was either 1392cc or the next step up, which was 5% higher at 1462 so you just have to pick your poison. The latency had much better resolution, but you can only go so far getting the Low & Medium trims dialed in using these parameters. Once you find the best possible set of latency and Injector scalar values you need to clear the ECU and start a fresh log so you can log STFT without the influence of trims.

Below is a log of MAF Frequency vs STFT when cruising and I made the yellow line to show the general STFT trend across the frequency range logged.




Once you know the STFT at a given frequency you just go into your MAF table and raise or lower the "g/s" values accordingly since they are just a fuel multiplier. Below is the stock graph vs the current graph after a few iterations and as of now I have only dialed in values between 35Hz and ~300Hz. This is really no different than how I tune MAF on the GTR, but the GTR has 64 breakpoints as opposed to 21 on the EVO as seen below.




Once you get the MAF table dialed in your trims should get pretty tight. Below is STFT vs MAF Frequency after tuning the MAF Table and I added the yellow line to show the general trend, but definitely better than before tuning the MAF table.





These were my trims after sitting at idle in a parking lot for 10 minutes and then about 15-20 minutes of driving, which was mostly steady state highway driving so load mostly below 300Hz.

 

The parts arrived to fabricate the dump from the BOV back to the turbo inlet, which was a 1.375" aluminum tube and a 1.375" ID flexible silicone hose. I got the tube welded to the turbo inlet so the only thing remaining is to recoat it with black wrinkle paint. Unfortunately, the $50 flexible silicone heater hose wasn't so flexible so when I tried to make the bend it just collapsed. I ended up going to Advanced Auto and finding a flexible radiator hose that was a perfect fit, but not the best aesthetics.

Anyway, loaded up the MAF tune and all is good so far and I will hopefully get this tuned out this week. The main focus will be to get the AFR's dialed in , but will also dial in cold start AFR and idle stepper lookup table. Idle is just fine when warm, but it idles too high on cold start and has some dips that need to be addressed.

 

Awesome thread, loving the info. Want to see more detail on your PCV system, I did see some posts in your S366 thread. What pump are you using? Can it flow oil also if you should decide to have it suck from the bottom of the can?

Also i noticed that your MAF Hz is reading to 1610, have a look around I'm sure there is a way to increase that.

Thanks for the interesting reading content, loving it.

PS why did you decide to go with the Garret 35r when you had such good success on with the Airwerkes?