CA18DET - Base setup help

[raddy]

the high rpm missfire will be spark uber common on CA's. Easy fix is decent coil packs (not yellow jackets or super spark as both are plain junk) and a set od BCPR7eix gapped to 0.6mm this will allow 20-30psi no problem but without the low speed issues of coppers gapped under 0.7mm

plug gap of 0.6mm will solve common issue of spark blow out at high boost/rich AFR/weak coils, but also reduce spark energy and fuel burn ratio...what will lead to reduce of produced torque. Best solution is to use WV TSI pen coil packs (easy to install with new harness and cheap to purchase), as I did recently (no need to alter any tabs in nistune). After regapping to 1.1mm (from 0.7mm) I got about 0.5point more rich AFR, just to give you idea how much fuel left cylinder unburned with small electrodes gap...

[STATUS]

the high rpm missfire will be spark uber common on CA's. Easy fix is decent coil packs (not yellow jackets or super spark as both are plain junk) and a set od BCPR7eix gapped to 0.6mm this will allow 20-30psi no problem but without the low speed issues of coppers gapped under 0.7mm

plug gap of 0.6mm will solve common issue of spark blow out at high boost/rich AFR/weak coils, but also reduce spark energy and fuel burn ratio...what will lead to reduce of produced torque. Best solution is to use WV TSI pen coil packs (easy to install with new harness and cheap to purchase), as I did recently (no need to alter any tabs in nistune). After regapping to 1.1mm (from 0.7mm) I got about 0.5point more rich AFR, just to give you idea how much fuel left cylinder unburned with small electrodes gap...

hence why you use a fine tip spark plug like i listed, we have no dramas with these at 1000hp at the qouted gaps but normally we fit a set of R7434-9 or -10s but they are not always on hand.

[louiswun]

CamsX, sorry, I've forgot lambda and non lambda ecu parts number.
23710-44FXX is non lambda ecu, no O2 sensor, the 4th pin on the AFM should connect a resistor (or variable resistor ) to adjust idle CO (A/F mixture).

[CamsX]

Thank you all very much for your input. I've learned a lot here, while also reading every single bit of information I could find on the internet.

I was finally able to get the car to idle @14.7~afr. I'm also very happy with the way the car drives under WOT, even for a street tune. Set my cooper spark plugs to 0.8mm and no more stutter. Might look into upgrading the Coil Packs in the future.

Now what I would like to do is improve overall driveability. I don't know if what I did to solve the idle is affecting the cruise parts of the map. Any suggestion is welcome.

At nico I got recommended to disable all O2 flags on my fuel map, which I did. And after adjusting the latency and keeping the MAF voltage around 1.120 like Louis recommended, no more wanting to go rich.

But now I'm having 2 behaviors that I'm not sure how to correct. When moving from a stand still the mix would lean out to 16.7-17.3afr and then progressively get back to mid-highs 15afr, even if the entire map is at 14.7 or less. Sounds like the latency is wrong, right? Should I enable o2 flags for low rpms/cruise and just skip the idle cells?

The other issue is that if I blip the throttle the mix would go all out rich, even if I'm not near the boost area.

This is correct. I've just removed that table from the address definitions for this ECU since this reminded me to look into that code when I was testing the CA18DET ECU last week
There is no check against TPS position in a table where the fuel maps are referenced in the ECU code

Matt you said that Alpha/N shouldn't have any effects if the values are changed, right? But I do notice differences in throttle response when the table is maxed out vs. default values.

Notice the 2nd screenshot. Why am I on 10afr if I'm on "TPS idle" position.

Nistune 004 by CamsX200, on Flickr

Nistune 005 by CamsX200, on Flickr

Nistune 006 by CamsX200, on Flickr

[Matt]

Now what I would like to do is improve overall driveability. I don't know if what I did to solve the idle is affecting the cruise parts of the map. Any suggestion is welcome.

Depends what you changed. Adjusting TPmin and latency should not affect other parts

At nico I got recommended to disable all O2 flags on my fuel map, which I did.

This will only affect O2 feedback trims

But now I'm having 2 behaviors that I'm not sure how to correct. When moving from a stand still the mix would lean out to 16.7-17.3afr and then progressively get back to mid-highs 15afr, even if the entire map is at 14.7 or less.

Sounds like throttle enrichment perhaps? Just remember that those numbers in the maps do not equate to actual AFRs necessarily. They are just numbers which control how much fuel is injected based on cell value. Always check with the wideband

Sounds like the latency is wrong, right? Should I enable o2 flags for low rpms/cruise and just skip the idle cells?

O2 feedback when enabled should result in 0% trims if you have your K constant in the right value

The other issue is that if I blip the throttle the mix would go all out rich, even if I'm not near the boost area.

Yes throttle enrichment. What is K like now compared to factory value?

Matt you said that Alpha/N shouldn't have any effects if the values are changed, right? But I do notice differences in throttle response when the table is maxed out vs. default values.

There is no use of that table around the fuel maps like with other ECUs. However the table is probably used by something else which I havent decoded yet. I would need to check it out. It would have an effect elsewhere in the code

[CamsX]

Hi Matt, thank you for looking at my comments. Right now the Kconstant is set at 349 (vs. 183 which is in the default bin). Notice I'm using Z32 MAF and 560cc evo8 injectors with fuel pressure set at 50psi (vs. 370cc injectors @ 36psi of fuel pressure).

If you look at the 1st image on my previous post, the AFR input trace is fairly accurate with what I'm trying to achieve with the primary fuel map. Should I be looking at different sensor for this? Or maybe I have a boost/vac leak?

[CamsX]

Me with more nonsense, but please humor my math and logic behind it.

http://www.rceng.com/technical.aspx
http://www.deatschwerks.com/resources/f ... calculator

I went to this sites to validate the flow of my injectors with the new fuel pressure, and doing so gave me the following results:

I'm using Mitsubishi Evo 8 560cc injectors which are supposed to operate at 43.5psi.

The atmospheric fuel pressure in my car was set to 50psi with the same injectors.

By entering these values in the website calculators I have an effective flow rate of 600cc.

After using the "Resize injectors" feature in Nistune to change 370cc to 600cc, I get a Kconst of 112.

And ultimately by using the "Change MAF" option to select the Z32 MAF adjustment it all ends up with a Kconst of 347!


Only 2 points away from my street calculated TP value. Does this make sense? Or is it just a coincidence? If I do the Change MAF first and then the Resize Injectors I end with a 349 value! haha.

[Matt]

Moving K constant too far from factory K can affect things dependent on it, such as throttle enrichment. I'll have a look through in the next week (working on CA18 feature pack updates at the moment) and see what I can find

[STATUS]

I find with these early ecus that it pays to not dwell on the correct math to determine k value and latency and just manipulate them until it reads correctly at the wideband.

The end injection should be perfect for normal pump fuel so that variable is removed, using the stock fuel table (flatten idle and cruise to 14.7 to make easier to target) and low boost run the car up to get your k value close to the target on the base map, usually you will then find you are too lean on cruise and idle, use latency adjustment to bring this back in line that way high speed and low speed fueling will be close to commanded base table targets. Then you can manipulate the base settings in the fuel tables to whatever you want to target be it 10 or 15:1 and it will close.

E85 or other race fuels may require manipulation of end injection tables but this is only done after a good base is found with the factory settings.

[louiswun]

Hi CamsX,

The way I adjust idle mixture is for lambda ecu without connecting a lambda sensor.
On non lambda ecu, theoretically the ecu pin 37 is for the adjustable idle mixture knob on the standard non lambda CA18 AFM, you should try to connect a variable resistor to adjust the idle A/F mixture. once Z32 AFM is installed, usually this pin left unconnected.
From what I've see on few CA18 here, the idle switch really affecting the idle mixture, even if I have adjust the VQ map lower voltage area, manually forced the idle switch to off position while throttle off, whole timing map adjust to 15 BTDC, flatten the AFR to 14.7 from 0.9V to 1.5V, after that, return the idle switch back to on position, there is extra fuel added even if no extra air is measured by the AFM, and that extra fuel cannot be remove by adjusting the injector latency or k value.
And I have notice the CA18 factory fuel map, 500rpm and 800rpm range, enable O2 flags, the target mixture is leaner than 1,200 rpm range (ie: less fuel supply).
Did you tried enable O2 flags and lower the numbers to take away fuel ?
My way to adjust idle mixture is not perfect, it is just to reduce the total fuel injected, once you open the throttle, the "idle switch extra fuel" gone, and minor lean will see, it will recover only if the AFM signal read more voltage than idle.

[CamsX]

Once more, thank you for your input. I was just looking into the math bit to see how far off I could be with my K value.

On boost AFRs are very accurate with my current map, and with some highway tests I'm close to getting my cruise and low boost afrs where I want them too.

Also Idle is almost spot on, with 14.6afrs @ 1000-1100rpms, after tweaking TTP, Latency and MAF voltage (through TB adjusting screw).

Here is my latest map:

Nistune 007 by CamsX200, on Flickr

[CamsX]

Hi guys, me again.

What mechanical issues do you guys think could be causing an instant Rich condition as the one in the image. I'm just about to hit full boost @3700rpms but the change in afrs seems so steep, specially when I'm aiming for 13-12.5 afrs on boost transition in that particular part of my map.

Boost Leaks? Bad PCV? Bad dropping resistor? I'm pretty sure it is not related to Nistune, because everything else with my tune is working great.

Nistune 008 Rich by CamsX200, on Flickr

[Matt]

Log your injection time and picture that. The increase in injection time should be consistent somewhat with AFRs

That is a bit of a drop. Monitor your fuel pressure also going on boost

[Buddyworm]

Hey Cams,

Might be a boost leak, which I suspect is why Matt wants you to check your injection times. A change in AFR without a similar change in pulsewidth could cause that rich condition.

How are the ratios through the rest of the rpm band?

Also, turn on PM's on your NICO account, I can't reply to you there

[CamsX]

Doh! Didnt know there was a setting for that in the Nico panel.

Thank you for posting. Yeah, Im also suspecting boost leaks. I have leaks on both ends of the TB shaft, but I didnt think that particular one could make so much difference.

Still need to solve it before I can confirm any other leaks.