CA18DET: Batch fire changeover, ECU retards ignition?

[Buddyworm]

So, in my datalogs, I've noticed that as my injection pulsewidths ramp up towards the magical 8ms mark (roughly where the ECU switches from sequential to batch fire) my timing will also retard several degrees relative to the maps. Then, after the ECU reduces the pulsewidths the timing will move back in line to match the maps.

Is this intenional? Can anybody else confirm their ECU does this?

I see it as somewhat beneficial, because less burn time inside the cylinder would result in a "richer" mixture as detected by the wideband. I don't really experience the lean patch some people experience during the sequential to batch fire changeover either, in fact my mixtures tend to go richer, so I think my hunch might be correct.

Engine is a CA18DET

CP Pistons
HKS 256* IN/EX Cams
TD06-17C Turbo
450CC Bluetop DSM Injectors
FMIC

I've attached my ROM and a datalog showing a few good rips through the gears and some highway driving.

[raddy]

check if you have DTC of knock sensor, as this is causing mentioned timing retard in specified RPM area...

[ricardomendes]

mine doesn´t do it mate.

[Matt]

Agreed with knock sensor fault retard. It will pull timing in addition to main timing maps if out of range, or if the ECU is actually detecting knock then check the knock report panel

[Buddyworm]

Hi Matt,

Can the knock sensor go "out of range" without throwing the DTC knock code? Consult isn't showing one. I haven't seen the ECU flip to the knock maps either...

As to the knock report panel, I was an early adopter so the firmware running on my board is from 2008 or so which, I believe, was before the knock reporting firmware was released.

Guess I'd better build me some knock ears... I have all the components, just need to assemble!

[Matt]

If out of range it will have a knock fault code

If knock is detected it will retard timing based on knock counts. Only will switch maps once it hits a certain amount of counts. Timing can be pulled and then recovered when not knocking afterwards

Suggest listing with knock ears and watching the knock warning panel for CA18 since it will flash the indicator and increase the knock count in Nistune

[Buddyworm]

Update on my current progress: Built the knock ears and discovered my CAS was set to 15* BTDC after hearing some nasty crackling through the top end. I must have set it and marked it back in the day and never bothered to double check after all these years... Very embarrassing considering I'm super anal about CAS timing when I tune other people's cars. Lesson learned. Never assume I knew what I was doing lol.

ANYWAY, my timing follows the maps to a tee now, except for one region around 4800 RPM. I'll continue working on this area of the map and see if I can iron things out. I haven't heard serious detonation in that region as far as I can tell but maybe I need to get some noise isolating headphones to quieten down my screamer pipe...

[Buddyworm]

Ok, so I'm thiiiiiis close to pulling this ECU and going Haltech for managing my engine. I have COMPLETELY reworked my timing maps while tuning with my knock ears, ECU still pulls timing between roughly 57-70 TP, depending on gear.

I did the detonation sensory bypass with a 470k Ohm resistor, ECU shows no DTC code for knock sensor and it is STILL cutting my timing for no apparent reason.

I have tried shifting my TP scale to be less steep during boost transition like other posters have recommended in the below thread with no luck: viewtopic.php?f=7&t=1682&start=15



What is going on here? It's cutting me down to 0-2* BTDC in my boost transition area (even ATDC timing in some logs!!). This kills my spool and is affecting the rest of the pull, something that is absolutely unacceptable for something managing a performance engine.

Attached is a log featuring the bypassed detonation sensor and my current unfinished rom. See 00:02:50 where it cuts me down to 2* BTDC. The only way this does not happen is if I start spooling the turbo up higher in the revband.



And FYI Matt, the knock report panel does not work with my board, it displays nothing when consult is connected and the engine is knocking.

[Matt]

Okay will look through this when I get back next week

[Buddyworm]

Thank you very much Matt! Let me know if there is any way I can assist your efforts and I will do my best.

[Matt]

Looking at this one tomorrow. Sorry one week off and school holidays meant a lot of catching up!

[Buddyworm]

Cool! Hopefully this isn't too complicated a fix. The snow still hasn't fallen here so I'm hoping to take the car out for a few more rips before the white stuff flies.

[Matt]

And FYI Matt, the knock report panel does not work with my board, it displays nothing when consult is connected and the engine is knocking.

There is no "knock count" parameter in your logs which means you are using a board which does not have the knock count firmware programmed into it.

[Buddyworm]

Yes, I had identified that as the issue in an earlier post:

Hi Matt,

Can the knock sensor go "out of range" without throwing the DTC knock code? Consult isn't showing one. I haven't seen the ECU flip to the knock maps either...

As to the knock report panel, I was an early adopter so the firmware running on my board is from 2008 or so which, I believe, was before the knock reporting firmware was released.

Guess I'd better build me some knock ears... I have all the components, just need to assemble!

I had hoped we would be figuring out why the CA18 ECU keeps messing with my timing curves...

[Matt]

Well the good news is I can repeat it here. I am not sure why it is doing that but appears to occour on movement only but then 'settles' to the values in the map. So its something to do with fluctuations in MAF/RPM causing it to go astray. Looking into it further today