I'm pretty sure the reason it happens is the solenoid and vacuum line that goes to the EGR. It sounds like it introduces a "vacuum leak" when it thinks its not opening enough or at all.

I wonder if disconnecting the vacuum source to the solenoid will keep the car from stalling (mine had the vacuum line disconnected from the EGR system for some time while I was working on another little mod)

 

Yeah, I was thinking "air leak" also. Stopping the EGR system from operating sounds safe enough to me, so I'll try disconnecting the EGR vacuum line (and plug the vacuum line) and see what happens.

 

Boost can be logged in Zeitronix software currently as stated in that post.

It is currently being added to Mitsulogger and Evoscan at this time.

 

There are two air lines going from the manifold to the EGR solenoid. Did you disconnect the line before the throttle butterfly or the one after (or both)?

 

Ah. I misunderstood. Cant wait for Evoscan to log boost!

 

There are two air lines on the intake manifold that feed into the EGR system. One before the throttle butterfly and one after. I tried pinching off each line individually, and then pinching off both together. Either way, the engine still dies within two seconds of starting it.

Did you disconnect these lines or different lines?

 

bump...

 

I never addressed this post..

There are several connections behind the intake manifold related to EGR, its a weird setup back there so just disconnecting the vacuum source from the manifold will foul things up.

What I had done was disconnect and plug the two vacuum lines to the EGR valve itself and the car would run fine.. It would occasionally throw an EGR code which hopefully will be found in the ECU eventually (as there are the OBD test and CEL code issues to deal with) but after doing that, I was able to play with a GM MAP sensor in place of the stock sensor and read boost. The big issue was the errors that would get thrown from the aftermarket sensor not reading properly. I'm hoping eventually we isolate this code so the test always passes, and there is a way to disable the EGR error code if the valve, or solenoid, or MDP sensor mis-reads or is disconnected.

After I have a chance to get my car inspected, and I can renew its registration, I will resume testing on this stuff.

 

you rock! subscribed.

 

Has anyone tested to see if the rear o2 sensor port can read up to 5V?

 

I'm pretty sure its scaled to 1v at the moment..

 

So what would happen if an input greater than 1V is sent to the ECU through the rear o2 port? Would Mitsulogger be able to log the RAW value?

 

the current ECU programming would read 255 over 1.0v as far as I can tell..

 

Okay, that's good to know. I'll use the stock MAP sensor port then.

 

Well, with a some help from bez_bashni to find the MDP/MAP translation table in the Evo 9 ROM, I got the JDM MAP sensor working with no CELs in my USDM Evo 9, and I am now logging boost in EvoScan. After a few simple mods to the ROM file using ECUFlash, installing and using the JDM Evo 9 MAP sensor is plug-and-play.

After bez_bashni found the table for me, I wrote up the xml to display the tables in ECUFlash. For the USDM Evo 9 ROMs (8859XXXX), the following text should be added to the ECUFlash xml ROM definition. (I added the text at the end of the xml file just before the </rom> tag.) The first item is the translation table. The second item is the MDP fault check value.

<table name="MDP translation" category="Misc" address="670c" type="2D" level="1" scaling="uint8">
<table name="MDP voltage" address="800e" type="Y Axis" elements="9" scaling="uint16"/>
</table>

<table name="MDP fault check" category="Misc" address="29d6" type="1D" level="1" scaling="uint16"/>

After modifying the ROM xml file, launch ECUFlash, and find the new entries. (I set the User Level to Developer, so you'll have to switch to Developer mode in the ECUFlash options.) For the USDM Evo 9 ROM, the translation table should read:

43 11
70 20
97 31
124 40
151 48
178 57
205 67
232 80
255 255

And the fault check value should read:

246

I changed the USDM values to values that jcsbanks had recently posted on aktivematrix for his Evo 9 FQ360. The Evo 9 FQ360 values are:

26 20
46 40
66 57
86 78
106 97
126 117
146 138
166 158
186 255

And the fault value should be changed to:

232

(Using the MAP/MDP sensor calibrations I had measured several months ago along with some algebra, I was able to verify (within experimental accuracy) that the JDM sensor calibration combined with the JDM translation would give the same translated values in the ECU as the USDM sensor calibration and translation, so I was very hopeful that the combination would work.)

Save the ROM file, and flash your ECU with it.

After I made those changes, I installed the JDM MAP sensor, the car fired up, idled perfectly, and had no CELs.

To log boost in psi in EvoScan, the MAP logging definition in the EvoScan Data.xml file should be changed to:

<DataListItem DataLog="Y" Display="MAP" LogReference="MAP" RequestID="38" Eval="1.334*x*14.5/100-14.5" Unit="psi" MetricEval="" MetricUnit="" ResponseBytes="1" GaugeMin="-15" GaugeMax="30" ChartMin="-15" ChartMax="30"/>

After updating the EvoScan Data.xml file, I took the car for a drive with EvoScan running the entire time. Result is shown below. The car ran perfectly the entire time. No CELs, no odd behavior of any sort. During the drive, I did a WOT run in 3rd gear. That result is shown below also. The logged MAP values correspond exactly with the OEM MR/RS boost gauge values I observed during the run.

So, if anyone is interested, I'd be happy to import some JDM Evo 9 map sensors. I'm actually importing a few boxes of other JDM items in the next few weeks, so this would be a great time to place an order. Here is the thread I set up for the group buy:

https://www.evolutionm.net/forums/sh...82#post4074482