90550001 Speed density alpha test - includes 10 bit MAP sensor accuracy update [Archive] - Mitsubishi Lancer Register Forum

90550001 Speed density alpha test - includes 10 bit MAP sensor accuracy update

johnbanks
28-03-2009, 18:20
Change to the earlier version that read MAP sensor with only the standard 8 bit accuracy, this version reads 10 bits like the Evo 9 (hence scaling changes from *1.334 to /3). Untested. Please report any bugs. You are responsible for your engine, not me. As usual, always check AFR and knock.

ROM:
banks.myzen.co.uk/90550001-v5.10fix-mods-SDv3.zip
31/3/9 UPDATE: v3 contains the same default baro code that I've used in other versions. Might resolve some issues with fuel trims not updating.

xml additions (or changes from earlier version):

<scaling name="MAP 16bit" units="kPa" toexpr="x/3" frexpr="x*3" format="%.1f" inc="0.3333"
storagetype="uint16" endian="big"/>
<scaling name="MAP VE load" units="Load (%)" toexpr="x*(5/32*0.596)" frexpr="x/(5/32*0.596)"
format="%.1f" inc="0.093125" storagetype="uint16" endian="big"/>
<scaling name="Percent (128)" units="%" toexpr="x/1.28" frexpr="x*1.28" format="%.1f"
storagetype="uint8" endian="big"/>
<table name="SD MAP sensor VE and calibration" category="SD" address="3428" type="2D"
level="1" scaling="MAP VE load">
<table name="MAP" address="340a" type="X Axis" elements="8" scaling="MAP 16bit"/>
</table>
<table name="SD RPM VE" category="SD" address="357c" type="2D" level="1" scaling="Percent (128)">
<table name="RPM" address="4d22" type="X Axis" elements="17" scaling="RPM"/>
</table>

cossie1
28-03-2009, 18:22
Cheers John, I will test it at some point ;)

cossie1
28-03-2009, 18:38
Just opened the rom, after checking the xml twice.

Still only getting 8bit instead of 10 ??

johnbanks
28-03-2009, 18:43
It should look the same, but I've rewritten the analog to digital converter routine for the MAP sensor and the internal calculation is four times more accurate. You'll find that the MAP sensor table has a top row that is now able to be set to the nearest one third of a kilopascal rather than one and a third.

cossie1
28-03-2009, 18:46
Ahh I see.

Does that explain why the SD RPM VE is all XXX's then ?

johnbanks
28-03-2009, 18:54
Damn, it is always a nightmare taking chunks out of an xml file and then hoping the tables come out right, especially axes etc. The addresses are correct above, but I need to rush out before I have chance to fix it.

Here is how it looks in mine, but the xml I've posted will need tidying up to be portable, looks like a difference between Ecuflash versions and how they handle xml.

cossie1
28-03-2009, 18:56
Don't worry found the issue.

Because of the space in percent (128), when you copy and paste it into the xml for some reason it add's an extra space :lol:

NINJA
28-03-2009, 19:41
good job man!!!
very interesting !!!:smthumbup

sub7
28-03-2009, 23:56
I wish I could be part of the test until my GM air temp arrive! :(

cossie, did you install air temp near the map sensor or throttle body?

johnbanks
29-03-2009, 10:28
Cossie1, hopefully the way I've shown percent (128) in the first post will now work (keeping the percent and (128) on the same line).

cossie1
29-03-2009, 13:10
I wish I could be part of the test until my GM air temp arrive! :(

cossie, did you install air temp near the map sensor or throttle body?

I fitted an IX inlet manifold, with the map and MAT/IAT sensor located in the normal places ;)

Cossie1, hopefully the way I've shown percent (128) in the first post will now work (keeping the percent and (128) on the same line).

Yep that will copy over fine now :smthumbup

NINJA
30-03-2009, 21:34
I fitted an IX inlet manifold, with the map and MAT/IAT sensor located in the normal places ;)


cossie,where on your 7 ecu do you have wiring your jdm map?
and your MAT?

thanks

cossie1
30-03-2009, 21:42
cossie,where on your 7 ecu do you have wiring your jdm map?
and your MAT?

thanks

MAT goes to pin 72 and ground.

I just pulled pin 72 out of the connector (as when the MAF is disconnected you don't need it going to the ecu anyway), and connected it there.

For the 3bar see below.

NINJA
30-03-2009, 22:00
thanks a lot man!!!:adore:



:smthumbup

cossie1
30-03-2009, 22:01
np's

BarryC
31-03-2009, 16:49
well I gave this a good shot today tried to change the ve tables and the map cal table the car still trimmed to -11% low LTFT no matter what I did :cry:

attached logs for JB

johnbanks
31-03-2009, 19:50
You are idling at a much higher MAP than most Evos. Mine idles at 33-40kPa at warm idle depending on ignition timing, yours is over 60kPa. The problem is that your trims are contradictory - the short term (O2 FB) is positive, yet your low and high long term fuel trims are negative. Because you're idling with quite a lot of air, it may be that at idle you're not actually using the low trim, but could be on the medium.

There is another potential issue: 90550001 was one of the first SD patches I did other than my home ROM. Whilst I have updated the MAP sensor accuracy for this post, I am killing the baro fault codes rather than substituting a standard value in the ADC routine. Tephra and I both noticed that fuel trims were lazy to update when we tested this early method, but we weren't sure as fuel trims can be lazy anyway.

I will update the baro code method to match the other ROMs that have worked well, and then ask you to retest with an ECU unplug so that fuel trims are zero when you try again.

It is a bit of a nightmare to do this without the car in front of you as I don't know when people have issues whether it is the code or them getting to know this particular SD tuning method. However, you and Cossie1 seem to have problems compared with others.

I had no driveability or trim issues even on the original 8 bit baro killed method, but I did collect some careful VE logs before hand. Perhaps you are leaning too heavily on the trims which are not working as well as they should. Let's see what the update shows. Perhaps you could collect some 2 byte load, RPM, MAP logs using the MAF sensor for a variety of conditions so we can calculate your VE?

johnbanks
31-03-2009, 22:00
First post updated with v3 which contains the same baro code I've used on other ECUs. Hopefully the fuel trims will update better.

cossie1
31-03-2009, 22:03
Nice work John, I am gutted I can't do any testing at the moment :(

johnbanks
31-03-2009, 22:05
Broken?

cossie1
31-03-2009, 22:05
Broken?

Spec update ;)

BarryC
01-04-2009, 21:24
First post updated with v3 which contains the same baro code I've used on other ECUs. Hopefully the fuel trims will update better.

cheers John will test it tomorrow ;)
have a log done with lots of driving conditions on the MAF but its a big file if ya want me to email it to ya I can :smthumbup

johnbanks
02-04-2009, 07:11
Use logworks to do a VE plot - MAP vs RPM and the item to plot is load/MAP. Cossie1 has done it, threads on evom about it.

NINJA
25-05-2009, 17:51
hi everybody

again me!!!:lol:

my gm map is installed,and scalled in evoscan with this formule:

For GM Sensor:
Boost PSI – 0.17573*x-14.5 (This is relative, leave the -14.5 for absolute) In evoscan scale to 10
Boost Bar – (0.012116157*x)-1 (Also relative, leave the -1 off for absolute) In evoscan scale to 100


but i have a question,where must i "write" (remplace to jdmmap scalling) this scalling in the xml??

thanks

rob2006
15-10-2009, 23:57
Change to the earlier version that read MAP sensor with only the standard 8 bit accuracy, this version reads 10 bits like the Evo 9 (hence scaling changes from *1.334 to /3). Untested. Please report any bugs. You are responsible for your engine, not me. As usual, always check AFR and knock.

ROM:
banks.myzen.co.uk/90550001-v5.10fix-mods-SDv3.zip
31/3/9 UPDATE: v3 contains the same default baro code that I've used in other versions. Might resolve some issues with fuel trims not updating.

xml additions (or changes from earlier version):

<scaling name="MAP 16bit" units="kPa" toexpr="x/3" frexpr="x*3" format="%.1f" inc="0.3333"
storagetype="uint16" endian="big"/>
<scaling name="MAP VE load" units="Load (%)" toexpr="x*(5/32*0.596)" frexpr="x/(5/32*0.596)"
format="%.1f" inc="0.093125" storagetype="uint16" endian="big"/>
<scaling name="Percent (128)" units="%" toexpr="x/1.28" frexpr="x*1.28" format="%.1f"
storagetype="uint8" endian="big"/>
<table name="SD MAP sensor VE and calibration" category="SD" address="3428" type="2D"
level="1" scaling="MAP VE load">
<table name="MAP" address="340a" type="X Axis" elements="8" scaling="MAP 16bit"/>
</table>
<table name="SD RPM VE" category="SD" address="357c" type="2D" level="1" scaling="Percent (128)">
<table name="RPM" address="4d22" type="X Axis" elements="17" scaling="RPM"/>
</table>

does anyone have an xml with this info already in it. I cant modify my xml i just get a stupid error in word pad.

Cheers

sponners
16-10-2009, 12:35
have you tried editing with notepad? should work ok.

Barny644
16-10-2009, 13:17
this is all very very clever and way out my league. impressive chaps!