Injector Scaling at low pulse width / non linear scale

td-d · **Posted:** Thu Oct 11, 2012 7:09 am

Following from this discussion: viewtopic.php?f=15&t=8290&start=15 following Cobb's discovery of these, I've dug them out in a couple of roms.

This table and thresholds provide additional pulse width compensation at very low pulse width, where the flow characteristics no longer match the linear section of the curve, as described in this Injector Dynamics article: http://www.injectordynamics.com/LowPulseTech.html - the so-called 'knee'.

This will be a tremendous help in scaling very large injectors, where you may be dipping in and out of the non-linear area at idle constantly.

For AZ1G400W, here are the definitions:

Code:

   <table name="Low Pulse Width Fuel Injector Compensation" address="c886c" type="2D" level="1" scaling="InjectorPulseWidthCompensation">
      <table name="IPW compensation %" address="c884c" type="X Axis" elements="8" scaling="BasePulseWidth(ms)"/>
   </table>
   
      <table name="Low pulse width fuel injector compensation maximum RPM" address="c7c6c" type="1D" level="1" scaling="RPM"/>

   <table name="Low pulse width fuel injector compensation maximum IPW" address="c7c70" type="1D" level="1" scaling="BasePulseWidth(ms)"/>

To find them:

The easiest way is to go to the Table_Tipin_Enrichment_Compensation_RPM location. 1 to 4 Tables up, this is a table with 8 cells that looks like this:

Code:

ROM:000861EC Table_Fuel_Injector_Trim_Small_IPW:.data.w 8 ; DATA XREF: sub_33720+1Ao
ROM:000861EC                                         ; ROM:off_3377Co
ROM:000861EE                 .data.w h'400
ROM:000861F0                 .data.l flt_C884C
ROM:000861F4                 .data.l byte_C886C      ; = 1.266
ROM:000861F8                 .float 0.0078125
ROM:000861FC                 .float 0.0

That's the comp table. If you X-ref back to the data location, you will find the other 2 parameters. You can verify that you at at the right place with the 1.5 and 0.5 float values (which are the limits to the compensations):

Code:

ROM:0003376A                 .align 4
ROM:0003376C off_3376C:      .data.l Ram_Engine_Speed ; DATA XREF: sub_33720+2r
ROM:00033770 off_33770:      .data.l unk_FFFF6F5C    ; DATA XREF: sub_33720+6r
ROM:00033774 off_33774:      .data.l Fuel_Injector_Trim_Small_IPW_max_RPM
ROM:00033774                                         ; DATA XREF: sub_33720+Ar
ROM:00033778 off_33778:      .data.l Fuel_Injector_Trim_Small_IPW_max_IPW
ROM:00033778                                         ; DATA XREF: sub_33720+12r
ROM:0003377C off_3377C:      .data.l Table_Fuel_Injector_Trim_Small_IPW
ROM:0003377C                                         ; DATA XREF: sub_33720+1Ar
ROM:00033780 off_33780:      .data.l Fun_GetMapValue2D_8B_16B_Float
ROM:00033780                                         ; DATA XREF: sub_33720+1Cr
ROM:00033784 off_33784:      .data.l unk_FFFF76B4    ; DATA XREF: sub_33720+22r
ROM:00033784                                         ; sub_33720+34r ...
ROM:00033788 off_33788:      .data.l Fun_check_high_low_thresholds
ROM:00033788                                         ; DATA XREF: sub_33720+26r
ROM:0003378C                 .float 1.5
ROM:00033790                 .float 0.5

A short cut is to look for the 1.5 and 0.5 values - 3F C0 00 00 3F 00 00 00 - it's normally the second last one found.

td-d · **Posted:** Thu Oct 11, 2012 7:10 am

Reserved for other ROM addresses

AZ1G101N

Code:

M:000C7AA8 Fuel_Injector_Trim_Small_IPW_max_RPM:.float 4000.0 ; DATA XREF: sub_33720+Ao
ROM:000C7AA8                                         ; ROM:off_33774o
ROM:000C7AAC Fuel_Injector_Trim_Small_IPW_max_IPW:.float 600.0 ; DATA

M:0008618C Table_Fuel_Injector_Trim_Small_IPW:.data.w 8 ; DATA XREF: sub_33720+1Ao
ROM:0008618C                                         ; ROM:off_3377Co
ROM:0008618E                 .data.w h'400
ROM:00086190                 .data.l flt_C8688
ROM:00086194                 .data.l byte_C86A8      ; = 1.266
ROM:00086198                 .float 0.0078125
ROM:0008619C                 .float 0.0

nsfw · **Joined:** Thu Nov 23, 2006 2:23 am **Posts:** 2565

This just came up in a thread at NASIOC a few days ago, too. Dominic Acia (used to post here fairly often, years ago) said it might fix the low-RPM stumble that I've got.

I don't have a "tip-in enrichment compensation (RPM)" in my A2WC522N ROM though.

Are there any other tables that might show me the way?

td-d · **Posted:** Thu Oct 11, 2012 8:58 am

NSFW wrote:

This just came up in a thread at NASIOC a few days ago, too. Dominic Acia (used to post here fairly often, years ago) said it might fix the low-RPM stumble that I've got.

I don't have a "tip-in enrichment compensation (RPM)" in my A2WC522N ROM though.

Are there any other tables that might show me the way?

I'm sure I have your rom disassembled - let me have a peak.

td-d · **Posted:** Thu Oct 11, 2012 9:04 am

NSFW wrote:

This just came up in a thread at NASIOC a few days ago, too. Dominic Acia (used to post here fairly often, years ago) said it might fix the low-RPM stumble that I've got.

I don't have a "tip-in enrichment compensation (RPM)" in my A2WC522N ROM though.

Are there any other tables that might show me the way?

Right above your tip in compensation boost error table:

Code:

ROM:00083CBC Table_Fuel_Injector_Trim_Small_IPW:.data.w 8 ; DATA XREF: ROM:00030528o
ROM:00083CBC                                         ; ROM:off_30564o
ROM:00083CBE                 .data.w h'400
ROM:00083CC0                 .data.l flt_C7830
ROM:00083CC4                 .data.l byte_C7850      ; = 1.055
ROM:00083CC8                 .float 0.0078125
ROM:00083CCC                 .float 0.0

ROM:000C7228 Fuel_Injector_Trim_Small_IPW_max_RPM:.float 4000.0 ; DATA XREF: ROM:00030514o
ROM:000C7228                                         ; ROM:off_3055Co
ROM:000C722C Fuel_Injector_Trim_Small_IPW_max_IPW:.float 1250.0 ; DATA XREF: ROM:0003051Eo
ROM:000C722C                                         ; ROM:off_30560o
ROM:000C7230 Fuel_Overrun_RPM_Delta:.float -17.0     ; DATA XREF: sub_30A76+48o
ROM:000C7230                                         ; ROM:off_30C88o
ROM:000C7234 Fuel_Overrun_initial_enrichment:.float 0.0 ; DATA XREF: sub_30A76+68o
ROM:000C7234                                         ; ROM:off_30C94o

Figured I'd throw in the overrun parameters as well

nsfw · **Joined:** Thu Nov 23, 2006 2:23 am **Posts:** 2565

Oooh, bonus! Thanks!

Easty · **Posted:** Thu Oct 11, 2012 10:34 am

Sweet you've obviously got my rom open at the moment!
If you don't mind i'm about to ask you a favour over at the tau thread!

td-d · **Posted:** Thu Oct 11, 2012 10:50 am

Easty wrote:

Sweet you've obviously got my rom open at the moment!
If you don't mind i'm about to ask you a favour over at the tau thread!

Yup - I was intrigued by your cruise control find!

Sure, I'll have a look for you.

Easty · **Posted:** Thu Oct 11, 2012 11:10 am

Thanks very much!

Merp · **Joined:** Thu Jul 23, 2009 5:46 pm **Posts:** 863

Awesome stuff!!

Can't wait to try this out with some ID2000's :mrgreen:

Templates and posted roms pushed to git: https://github.com/RomRaider/SubaruDefs/tree/Alpha

dschultz · **Posted:** Fri Oct 12, 2012 3:38 am

I've been looking at this a bit further and I don't think your table scalings work for my ROM which is the same as NSFW's.
The table axis, I suspect is the current semi-compensated injector pulse width in microseconds and the data value is the compensation which will be a multiplier where 1 = 100%, 1.55 = 155% etc.

Here's what it looks like to me.

Attachment:

lowIPW.png

Code:

    <table type="2D" name="Low Pulse Width Fuel Injector Compensation" category="Fueling - Injectors" storagetype="uint8" endian="big" sizey="8" storageaddress="0xC7850">
      <scaling units="%" expression="x*0.78125" to_byte="x/0.78125" format="0.0" fineincrement="0.1" coarseincrement="1" />
      <table type="Y Axis" name="Injector Pulse Width" storagetype="float" storageaddress="0xC7830">
        <scaling units="msec" expression="x/1000" to_byte="x*1000" format="0.000" fineincrement="0.001" coarseincrement="0.05" />
      </table>
      <description>If the RPM and the current injector pulse width is below the thresholds then this compesation is applied, otherwies a compensation of 1 or 100% is used.</description>
    </table>
  <table type="2D" name="Low pulse width fuel injector compensation maximum RPM" category="Fueling - Injectors" storagetype="float" endian="big" sizey="1" storageaddress="0xC7228">
   <scaling units="RPM" expression="x" to_byte="x" format="0" fineincrement="5" coarseincrement="100" />
   <table type="Static Y Axis" name="Enabled Below" sizey="1">
    <data>Maximum</data>
   </table>
   <description>This is the maximum limit to use the Low Pulse Width Fuel Injector Compensation value.</description>
  </table>
  <table type="2D" name="Low pulse width fuel injector compensation maximum IPW" category="Fueling - Injectors" storagetype="float" endian="big" sizey="1" storageaddress="0xC722C">
   <scaling units="msec" expression="x/1000" to_byte="x*1000" format="0.000" fineincrement="0.05" coarseincrement="0.1" />
   <table type="Static Y Axis" name="Enabled Below" sizey="1">
    <data>Maximum</data>
   </table>
   <description>This is the maximum limit to use the Low Pulse Width Fuel Injector Compensation value.</description>
  </table>

I'm going to scale down the IPW axis and the max threshold for my DW740. Otherwise this compensation gets applied when it shouldn't be with the larger injectors.

Merp · **Joined:** Thu Jul 23, 2009 5:46 pm **Posts:** 863

dschultz wrote:

I've been looking at this a bit further and I don't think your table scalings work for my ROM which is the same as NSFW's.
The table axis, I suspect is the current semi-compensated injector pulse width in microseconds and the data value is the compensation which will be a multiplier where 1 = 100%, 1.55 = 155% etc.

Those scalings are the same as td-d's and those in the repo, with the exception of the percentage offset in the compensation.

In all previous ECUFlash definitions, "compensation" has been treated as such:
display: 0% <=> multiplier: 1.0
display; 55% <=> multiplier of 1.55

td-d · **Posted:** Fri Oct 12, 2012 5:42 am

Merp wrote:

dschultz wrote:

I've been looking at this a bit further and I don't think your table scalings work for my ROM which is the same as NSFW's.
The table axis, I suspect is the current semi-compensated injector pulse width in microseconds and the data value is the compensation which will be a multiplier where 1 = 100%, 1.55 = 155% etc.

Those scalings are the same as td-d's and those in the repo, with the exception of the percentage offset in the compensation.

In all previous ECUFlash definitions, "compensation" has been treated as such:
display: 0% <=> multiplier: 1.0
display; 55% <=> multiplier of 1.55

Yup - that's exactly how I set it up - 26% in the table is effectively a 1.26 multiplier to the base pulsewidth.

td-d · **Posted:** Fri Oct 12, 2012 5:43 am

Merp wrote:

Awesome stuff!!

Can't wait to try this out with some ID2000's :mrgreen:

Templates and posted roms pushed to git: https://github.com/RomRaider/SubaruDefs/tree/Alpha

If I had my car

I would too - had so many issues with my 1200's (granted, cr@ppy make...).

td-d · **Posted:** Fri Oct 12, 2012 5:50 am

Looking at Dale and NSFW's rom - same scaling - the table scaling, as Merp said is just with an offset (i.e. 26% = 1.26X).

Interesting to see the difference between the two - much smaller pulse width on the WRX and much bigger compensation.

Might be worthwhile to chat to some of the injector manufacturers (like ID) that publish the low pulse width calibration how to incorporate their published rates into this table.

RomRaider

Injector Scaling at low pulse width / non linear scale

Who is online