PI : DI Ratio Discussion

[thambu19]

Originally Posted by ztan

Interesting. My calibration is A01G. First second of cranking is full PI, then DI gets added, full DI after around 6-8 seconds. Are later calibrations starting up with DI?

When engine is cold PFI helps because upon firing cold fuel on a cold port the fuel gets more time to evaporate in a PFI. Normally the injection timing for a super cold start is much earlier than a warm PFI start. This is also applicable for E85 which takes more time to evaporate.

Once engine is up and running they would be switching to DI for late ignition to fire up the CATs. With DI they can create rich spots around the plugs for this condition.

[thambu19]

Does anyone know if the ecu uses MAP sensor reading to calculate load at any time? Is is just the MAF or is the MAP used as well or only if the MAF is faulty?

[Kodename47]

Originally Posted by thambu19

Does anyone know if the ecu uses MAP sensor reading to calculate load at any time? Is is just the MAF or is the MAP used as well or only if the MAF is faulty?

It's MAF based load.

[ztan]

Originally Posted by shr133

1 of my goals was to get rid of the dip and the dip starts when they turn off PI and change the cam timing and a few other things....
I don't know why you would turn PI on, off and on again.... Running both will smooth it out... At mid load I have no dip in acceleration.... I think turning off PI and the exhaust cam timing is for emissions not performance...

This will never be a drag car so I'm tuning for drive ability and passing power... So I am currently running this on my butt dyno for now till I get the maps sorted out... It doesn't take a genius to know if something works or not...

I'm getting to the point I have to dyno it and start doing 0-60s to sort out the small details... But it runs way better than stock at all rpms...

I don't know if running PI all the time will help HP on a dyno but it will make the engine run better and eliminate 2 transitions that help to cause lag and hesitation...

Just flash it and drive for a day if you don't like it just flash back...

Now this one is interesting...

So I tried having PI active at least partially all of the time except for idle.
The result of my logging PI duty is below - PI not active where I set my PI ratio at 20% or 24%. Before you ask: yes, I did set all 3 PI ratio tables the same.

I've gone and trawled through the disassembly for the PI/DI ratios and come to the following conclusions:

1. PI ratio (Hot) map is not used based on the A01G disassembly. The threshold for switching between (Cold) and (Warm) maps is ECT <55 or ECT >60 (5 degree hysteresis).

2. If the RPM is over 5000RPM or Load over 1.5 g/rev, the following two thresholds are not applied:

a. If calculated PI_Ratio is over 0.75 and the RPM/Load thresholds are not exceeded, the PI_Ratio is set to 1.00.

b. If the calculated PI_Ratio is under 0.35 and the RPM/Load thresholds are not exceeded, the PI_Ratio is set to 0.00.

ROM addresses for the thresholds in A01G ROM:

0010C70C 3E B3 33 33 PI_Ratio_Threshold_full_off:.float 0.34999999
0010C70C                                                     ; DATA XREF: sub_58F8E:loc_59024o
0010C70C                                                     ; sub_58F8E:off_59178o
0010C710 3F 40 00 00 PI_Ratio_Threshold_full_on:.float 0.75  ; DATA XREF: sub_58F8E+78o
0010C710                                                     ; sub_58F8E:off_59170o
0010C714 00 00 00 00 flt_10C714:     .float 0.0              ; DATA XREF: sub_58F8E+A6o
0010C714                                                     ; sub_58F8E:off_5917Co
0010C718 00 00 00 00 flt_10C718:     .float 0.0              ; DATA XREF: sub_58F8E+7Co
0010C718                                                     ; sub_58F8E:off_59174o
0010C71C 45 9C 40 00 PI_Ratio_RPM_Threshold:.float 5000.0    ; DATA XREF: sub_58F8E:loc_59096o
0010C71C                                                     ; sub_58F8E:off_5918Co
0010C720 3F C0 00 00 PI_Ratio_Load_Threshold:.float 1.5      ; DATA XREF: sub_58F8E+110o

[ztan]

Def for above A01G:

    <table name="PI_Ratio_Thresholds" storageaddress="10C70C" />
    <table name="PI_Ratio_RPM_Threshold" storageaddress="10C71C" />
    <table name="PI_Ratio_Load_Threshold" storageaddress="10C720" />
    
    <table type="2D" name="PI_Ratio_Thresholds" category="Fueling - Injector Ratio" storagetype="float" endian="little" sizey="2" userlevel="4">
      <scaling units="PI Ratio %" expression="x*100" to_byte="x/100" format="0.0" fineincrement="5" coarseincrement="10" />
      <table type="Static Y Axis" name="" sizey="2">
              <data>PI Ratio 0% under</data>
              <data>PI Ratio 100% above</data>
       </table>
      <description>If PI Ratio RPM and Load Thresholds are not exceeded, PI Ratio is set to 0% if this value is not exceeded</description>
    </table>
    <table type="2D" name="PI_Ratio_RPM_Threshold" category="Fueling - Injector Ratio" storagetype="float" endian="little" sizey="1" userlevel="4">
      <scaling units="Engine Speed RPM" expression="x" to_byte="x" format="#" fineincrement="250" coarseincrement="1000" />
      <table type="Static Y Axis" name="" sizey="1">
              <data>PI Ratio thresholds not applied above</data>
       </table>
      <description>If this RPM threshold is exceeded, PI Ratio full on and full off thresholds are not applied</description>
    </table>
    <table type="2D" name="PI_Ratio_Load_Threshold" category="Fueling - Injector Ratio" storagetype="float" endian="little" sizey="1" userlevel="4">
      <scaling units="Load g/rev" expression="x" to_byte="x" format="0.0" fineincrement=".05" coarseincrement=".2" />
      <table type="Static Y Axis" name="" sizey="1">
              <data>PI Ratio thresholds not applied above</data>
       </table>
      <description>If this Load threshold is exceeded, PI Ratio full on and full off thresholds are not applied</description>
    </table>

[Kodename47]

@ztan this is strange.. I have the warm to hot ECT thresholds in my definitions (80/75), are you suggesting they aren't used? I guess an obvious test would be to set the 3 tables with flat values and see what happens.

So to sum up, at RPM less than 5k and load under 1.5 then if the ratio is 75% PI it will output 100% PI. This is almost never the case, I doubt anyone sets the PI as high as 75% as you'd simply be under using the DIs.

Anything under 5k RPM and 1.5 load will be 100% DI if the PI Ratio is below 35%. That would certainly explain some behaviour I've noticed before.

I wonder if ECUtek custom maps get around these conditions?

[Kodename47]

Originally Posted by Kodename47

I wonder if ECUtek custom maps get around these conditions?

Seems like it does. Just looked over 2 logs from the last few days having set the whole PI Ratio table to 25%. On the standard table I get no PI until over 1.5 load, on the ECUtek custom map I have PI just off idle, where the minimum IPW threshold is met.

Adding and checking these parameter into my A02G definition.

[ztan]

Originally Posted by Kodename47

@ztan this is strange.. I have the warm to hot ECT thresholds in my definitions (80/75), are you suggesting they aren't used? I guess an obvious test would be to set the 3 tables with flat values and see what happens.

So to sum up, at RPM less than 5k and load under 1.5 then if the ratio is 75% PI it will output 100% PI. This is almost never the case, I doubt anyone sets the PI as high as 75% as you'd simply be under using the DIs.

Anything under 5k RPM and 1.5 load will be 100% DI if the PI Ratio is below 35%. That would certainly explain some behaviour I've noticed before.

I wonder if ECUtek custom maps get around these conditions?

The second ECT threshold (80/75) is defined, but use of that switch in the code is bypassed, at least in A01G stock. Activating that switch requires adjustment of a single float value (at 591AC) in the ROM to greater than 100.0 - Subaru engineers disabled the switch by changing a value in the code rather than re-writing it.

Having said that, I am not sure that there will be any major benefit from having different DI/PI settings at different engine temps.

The subroutine is located at 58F8E if you want to have a go at the disassembly.

[steve99]

Originally Posted by ztan

Now this one is interesting...

So I tried having PI active at least partially all of the time except for idle.
The result of my logging PI duty is below - PI not active where I set my PI ratio at 20% or 24%. Before you ask: yes, I did set all 3 PI ratio tables the same.

I've gone and trawled through the disassembly for the PI/DI ratios and come to the following conclusions:

1. PI ratio (Hot) map is not used based on the A01G disassembly. The threshold for switching between (Cold) and (Warm) maps is ECT <55 or ECT >60 (5 degree hysteresis).

2. If the RPM is over 5000RPM or Load over 1.5 g/rev, the following two thresholds are not applied:

a. If calculated PI_Ratio is over 0.75 and the RPM/Load thresholds are not exceeded, the PI_Ratio is set to 1.00.

b. If the calculated PI_Ratio is under 0.35 and the RPM/Load thresholds are not exceeded, the PI_Ratio is set to 0.00.

ROM addresses for the thresholds in A01G ROM:

0010C70C 3E B3 33 33 PI_Ratio_Threshold_full_off:.float 0.34999999
0010C70C                                                     ; DATA XREF: sub_58F8E:loc_59024o
0010C70C                                                     ; sub_58F8E:off_59178o
0010C710 3F 40 00 00 PI_Ratio_Threshold_full_on:.float 0.75  ; DATA XREF: sub_58F8E+78o
0010C710                                                     ; sub_58F8E:off_59170o
0010C714 00 00 00 00 flt_10C714:     .float 0.0              ; DATA XREF: sub_58F8E+A6o
0010C714                                                     ; sub_58F8E:off_5917Co
0010C718 00 00 00 00 flt_10C718:     .float 0.0              ; DATA XREF: sub_58F8E+7Co
0010C718                                                     ; sub_58F8E:off_59174o
0010C71C 45 9C 40 00 PI_Ratio_RPM_Threshold:.float 5000.0    ; DATA XREF: sub_58F8E:loc_59096o
0010C71C                                                     ; sub_58F8E:off_5918Co
0010C720 3F C0 00 00 PI_Ratio_Load_Threshold:.float 1.5      ; DATA XREF: sub_58F8E+110o

HaHa us NA guys dont ever make 1.5 load :-) , but nice to know, maybe they were plannibg a turbo ?

[thambu19]

Originally Posted by ztan

The second ECT threshold (80/75) is defined, but use of that switch in the code is bypassed, at least in A01G stock. Activating that switch requires adjustment of a single float value (at 591AC) in the ROM to greater than 100.0 - Subaru engineers disabled the switch by changing a value in the code rather than re-writing it.

Having said that, I am not sure that there will be any major benefit from having different DI/PI settings at different engine temps.

The subroutine is located at 58F8E if you want to have a go at the disassembly.

Generally it helps to go PFI when temps are low to avoid wall wetting and wash down and also for better mixture prep. Unless engine is knock limited there is no real need for DI at low loads other than for emissions and to make sure the injectors are not carboned up by those guys who baby their engines. Since they are inside the chamber they have to make sure they are fired more frequently than the PFI injectors.

[Kodename47]

Originally Posted by steve99

HaHa us NA guys dont ever make 1.5 load :-) , but nice to know, maybe they were plannibg a turbo ?

Or maybe just ensuring that parameter is not exceeded.

[ztan]

Originally Posted by shr133

1 of my goals was to get rid of the dip and the dip starts when they turn off PI and change the cam timing and a few other things....
I don't know why you would turn PI on, off and on again.... Running both will smooth it out... At mid load I have no dip in acceleration.... I think turning off PI and the exhaust cam timing is for emissions not performance...

Also interesting that even though the thresholds are in place and PI is off in the same regions of the map, the car does feel smoother at part throttle and part throttle acceleration with the ratios set between 20-24% as @shr133 noted. Possibly due to the way the values are interpolated first before the thresholds being applied making for a smoother transition between combined operation and full DI.

[thambu19]

Originally Posted by ztan

Also interesting that even though the thresholds are in place and PI is off in the same regions of the map, the car does feel smoother at part throttle and part throttle acceleration with the ratios set between 20-24% as @shr133 noted. Possibly due to the way the values are interpolated first before the thresholds being applied making for a smoother transition between combined operation and full DI.

This is interesting that now two people feel the same way. Are you sure that PFI is not being commanded in both cases?

[ztan]

Originally Posted by thambu19

This is interesting that now two people feel the same way. Are you sure that PFI is not being commanded in both cases?

I'm sure from my logs that PI is not active and I've re-read the ROM to make sure those table changes were committed (because it didn't make sense when I looked at the logs). Not sure what ROM calibration @shr133 is running and whether he has logged the PI pulse or duty and whether the same logic is in his ROM calibration as mine.