CA18DET: Batch fire changeover, ECU retards ignition?

[Matt]

Okay the problem is to do with the Load Recovery (TP) table. Putting back to stock fixed it. May not be the correct table so I will need to look into it further.

Obviously pulling back timing on throttle adjustment so may be a throttle tip in table instead

Fixed file attached

[Buddyworm]

Well I'll be damned... Thanks for identifying that for me Matt!


I looked back through the different iterations of that tune and found when the TP Load Recovery table changed, and it was right around the time I first noticed the issue. I was fiddling around with the butterfly maps at the time and must have copy/pasted some values into the wrong table. I'll report back with results after I get a chance to take the car out for a rip.

Could you describe the way the ECU uses that table at all?


Thanks again Matt! Your help is very much appreciated.

[Matt]

This one was a process of elimination which was why it took a while. Gradually I copied your maps into a stock 36F05 base image and found that the TP recovery one caused the issue

Unfortunately it was a misidentified table. Sometimes similar addresses get copied from one address file to the next. However this one doesnt even have a code reference point when I was looking at the disassembly last night. In addition to this there may not be a TP recovery table for CA18DET

I noticed a question mark against this entry in the address file:

TP_RECOVERY,&H39A0,16,1,16,1,TP Recovery?

Table is actually B990 with 32 bytes length with some timing offset vs RPM. However the lookup value is then put into the ECU A/D converter. I'm unsure why it is affecting the timing

[Buddyworm]

Oh I see what happened here. That table is nearly the same address that had been labeled as the butterfly control solenoid map! And that's the one I've been fiddling with trying to manipulate the changeover!

#Add.map 1
ADD_MAP1,&H3990,32,1,32,1,Variable air close (JDM)

I made that table by combining two 1x16 tables that had been labeled as Variable Air Open and Variable Air Close. Sigh.

Has anybody actually found the butterfly control tables? There's supposedly some Aussie tuners that can tune them. There any truth to that?

[Matt]

Well it does sound close since it does control an A/D output perhaps which is could have been the butterfly control. I have not heard of details about it yet but can test it.

[Buddyworm]

Could you please? I've got a Volvo cruise control pump with an integrated solenoid and I'd love to find the tables that control the duty cycle for the CA18's solenoid output. A high degree of control over the butterfly opening is a goal I'd like to achieve.

Or could you point me in the direction of resources for decompiling the ECU code? That's something I'd like to experiment with.

[Matt]

So we are taking about pin 8 (variable intake solenoid?)

Pin 8 shows an on/off voltage which fluctuates depending on TP/RPM. It is not a duty cycle when I look at my scope

This table as I mentioned before goes to an A/D converter in the ECU code. If TP is below the selected value (based on 0-6400 rpm scale) then the output is active, otherwise inactive

I've updated the software (this did not make the 0.17 release) to add the new table

So once TP goes above the table values, the solenoid is disabled

Just check this does not affect your timing. This is added for the next release (you will need to use ADD_MAP1 for now)

VARIABLE_INTAKE_SOLENOID,&H3990,32,1,32,1

[Buddyworm]

Hi Matt,

I can confirm that my timing changes if that table is edited. You can see that the TP Recovery table and the Butterflies table change simultaneously. Correcting the TP Recover table definitely helped improve the timing, although I still am experiencing a small drop in timing over a smaller section of the map.

[Matt]

You can edit the address file and remove the TP recovery table. Latest version 0.17 has this removed now

[zbeffreyCA18]

So from my understanding, you can use the Variable Intake Solenoid Table to adjust the point where the Butterfly valves will open/close correct? Buddyworm said it affected timing but to what extent?

also, I've always ran my CA18 at 15*BTDC and never had an issue and have had no knock sensor codes ever show up...