Idle lean when A/C on

[jongerlings]

1996 KA24DE ECU
KA24DET
Idle AFR is great with A/C off, but when I turn on the A/C, the RPM rises about 50-100RPM (normal), but the AFR's lean out to about 17-18. After a few seconds, AFR's seem to 'self-adjust' back towards 14.7, but when it gets to about 15, the AFR quickly spikes back up to about 17. It will fluctuate like this until I turn A/C off, at which point idle drops and AFR stabilizes back at 14.7ish. Looking at consult gauges and maps, it looks to me like the ecu is pulling timing back to about 14-15 deg bdtc to get the AFRs lower, but it jumps back to 20deg when jumping columns in the "idle stabilization advance" and "idle stabilization retard" tables. I tried playing around with these two tables, but I'm afraid I don't really understand the relationship between RPM and timing on these tables. Target idle is set to 800RPM, which is perfect with A/C off, and roughly 850RPM with A/C on, but the columns being accessed in these two idle stabilization tables are labeled 400 and 500 RPM.

Changing the cell in the base fuel map where the engine sits at idle with A/C on from 128 to 129 brings the idle AFR to 14.7 as well, but this seems like a silly way to achieve what I want, since it causes some choppy transitions around this cell. Also worth noting that when I change this cell in the fuel map, timing hovers around 14-15 DBDTC as well, but stays there, rather than fluctuating.

What might I be missing here, or is this normal?
Cruise AFR is fine both with A/C on and off, my only issues are at idle.

[jongerlings]

Also, I've noticed when trying to dial in my idle AFR's, a lot of posts suggest adjusting TTPMIN values. I don't see this table anywhere in the S14 KA24DE address file. Am I missing something?

A couple other parameters I have come across that I can't find any info on: "Closed loop flag TP" and "Closed loop flag RPM"
What do these values actually control?

Thanks!

[Matt]

TPMin tables are unfortunately missing in the later ECUS (S14, B14 and ER34/WC34) so the only way to work with that is to increase the bottom end of the VQ map higher (which increases TP)

As for why your mixtures lean out with the A/C on I am not so sure. There is a FICD (fast idle control device) which switches on normally. That would increase revs. Check that is 'clicking' when you enable the A/C

Normally the ECU should adjust injectoin for the AC load, assuming it is seeing the AC switch input (check the consult view display for this)

Watch the injection time in the logs and see if that stays the same 'ms' of time before/after AC is pressed. Curious about that

Timing should not affect things. We cover timing in the 'workshop training notes' on the website. Basically AAC + timing are adjusted to target the 'target RPM table'

The O2 load/rpm tables are something more recently added in. They reference a table in RAM which holds the fuel trims. This should not normally need changing unless you are moving K constant around and have issues with closed loop functioning in particular load/RPM areas

[jongerlings]

Hi Matt, thanks for the reply. You are awesome for continuing support of your product over all these years.
I'll check to see what the actual injection times are doing when I get home today.
Consult view definitely shows both A/C switch and A/C relay turning on when they should, as well as Neutral switch and TPS idle when they should, with A/C both on and off.

Is it possible that when the A/C is turned on, and the FICD turns on, increasing RPM, the ECU is still trying to obtain the idle target RPM? This would explain what I'm seeing with A/C on, where the ECU pulls timing back. I'm guessing its adjusting timing to a certain limit based on the Idle stabilization advance/retard tables, which is when I see AFRs and timing both drop momentarily before returning to 20deg and lean AFR. To combat this, shouldn't there be a separate A/C on target idle RPM? What I'm guessing is happening based on your reply is the ECU is pulling timing to try to achieve target RPM, even though the A/C is on and FICD is open, causing higher idle RPM(normal and desired)? I think I've seen this type of situation accounted for with parameters for A/C on and off in other address files, but I don't see them here. Any reason why this address file does not have more idle control parameters or A/C on parameters?

Again, I will try to get a log together and post it here.

[jongerlings]

Okay, so I recorded a few logs, and noticed some more odd behavior at idle.
Log 1: On cold start, some weird stuff happens at around the 1 minute mark. I go from roughly 13:1 afrs to 17-18 for some reason. I don't see any map switching, so I don't know what is causing this drastic change.
Log 2: Log started after driving around for a few minutes, then sitting idle. Idle sits right around 14.7 for a bit, but leans out gradually over time. One thing I notice here is that AFRs start going lean once the O2 sensor voltage drops. I can't seem to figure out why the ecu is only using O2 feedback sometimes at idle, and how to adjust for when there is no O2 voltage.
Log 3: A/C switching on/off at operating temp idle

[jongerlings]

Oh and here is the bin file currently on the ecu

[Matt]

So I've compared your tune against stock

Blue shows differences

2021-03-12 16_32_00-Nistune [S14(KA24DE) 1996 Feature Pack 1.0 (Version_ 1)].png

(350.89 KiB) Downloaded 1600 times

Unsure why knock limit tables are lowered (may cause higher knock counts)

2021-03-12 16_32_00-Nistune [S14(KA24DE) 1996 Feature Pack 1.0 (Version_ 1)].png

(350.89 KiB) Downloaded 1600 times

Curious about end time change

2021-03-12 16_34_53-Nistune [S14(KA24DE) 1996 Feature Pack 1.0 (Version_ 1)].png

(48.85 KiB) Downloaded 1600 times

VE RPM scale - normally dont change this
Knock Timing Beta Advance - only low octane changed
Knock analysis - disabled
Closed loop ranges maxed out
Launch control enabled
Some other changes to knock leaning/switch over

Change injeciton adjust contstant from 102 to 128

Log 1. Idle is pretty high. Neutral switch on, TPS idle, but 1200rpm. Something not right here. Timing is 22-23 deg BDTC with coolant temp between 25-35 degC

Injection time, looking around 2.23ms... until 1:19 temp goes from 36 degC to 37 degC and injeciton time drops to 1.71ms. This seems to be injection time only. Unsure of the cause of this (MAF and load stable)... Also timing seems is just all shaky after this. Like something changed

Other idle log, timing a bit jumpy but injection time sitting around 1.7ms still

AC on and it jumps to 1.95ms so a bit more fuel

What I would do... start with the 72F0F file again (factory) and copy in just the maps you need (from the Comp image slot, using the COPY button to do the copy maps from Comp to Main)

Fuel, timing, fuel scale, Max, load, Load (TP) cut, Crank enrich, MAF translation, K Constant, TIM and Latency

Just copy only those maps into the fresh image and let me know how you go with that. I've attached a copy I've done for you

[jongerlings]

Matt,
Thank you for the reply.
The first log is a cold start, so the idle sits a bit high until the engine warms up a bit. I'm pretty sure this is normal? With that log, I mainly wanted to show what happens when the idle AFR during warmup goes from 13-ish (temp enrich for warmup), to 17-18. The change seems to happen after the o2 voltage steadily rises until about .7-.8V, then o2 voltage drops to 0 when mixture goes lean. The RPM's lower at a steady rate as the motor warms up until it reaches the target idle rpm. What you are seeing at that 1:19 mark is exactly where I am stumped. Timing starts to jump around all over the place.

As for the knock tables, this is something my tuner did when dyno tuning my car, because forged internals, larger cams, and firmer engine mounts were apparently causing lots of false knock readings. I'm guessing he played around with knock tables until deciding to just disable knock detection.

I will try out your suggestions and get back to you when I can. I'm currently driving the car to set OBD2 readiness codes, and I have a few monitors set (the drive patterns to set these are insanely hard), so I don't want to risk burning a new image to the ecu and possibly losing readiness monitors right this second.

Thank you again so much for all your help!
-Jon

[jongerlings]

Matt,
I passed OBD2 emissions testing today, thanks to you and Nistune! Thank you so much for offering what nobody else in the world does: a tunable OBD2 KA24DE ECU. Without you, my goals of having a legitimately passing obd2 KA24DET would not have been possible. Smiles all around!

That being said, I tried out your suggestion this morning, uploading the tune you provided. I disabled knock detection and EGR system, since I know my knock sensor is finnicky and I don't need EGR monitoring anymore. I also bumped idle target RPM to 800 at the lowest values because with larger cams, 750rpm gives the car a bit more vibration when sitting still than I would like. First impressions:
-Idle AFR is definitely more stable both with A/C on and off. Sweet!
-I did have to increase my injector latency quite a bit to achieve desired idle AFR though. I went from 740us to about 940us (still playing around with this to get it just right. Is it just me, or does the ecu only accept increments of 10us?) I will continue driving with this tune and make sure there are no abnormal characteristics with idle AFR like I was experiencing before.

One last thing I notice is when shifting under light-med load, what I would consider normal traffic driving characteristics, I get a bit of a lean spike, causing slight hesitation when applying throttle in the next gear. For example: 2nd gear, accelerate to ~2500rpm low load, closed loop area of map. Clutch in, shift to 3rd, clutch out, back on same throttle, afr goes lean after clutch in there and takes half a second or so to come back to stoich when back in gear and on throttle, but causes noticeable hesitation as load increases. Any ideas what to adjust here?

[Matt]

Good stuff!

I went from 740us to about 940us

If your AFRs are stoich 14.7:1 at cruise then latency is the right table to correct, since there is no TTP Minimum table in this ECU

Yes the converted > ECU internal values are in 10us steps, so this is a limitation of the ECU

I get a bit of a lean spike, causing slight hesitation when applying throttle in the next gear

Would need to see a log of this if you can post one

It seems maybe acceleration response. I would need to relook at the tune, but is your 'Add fuel' table maxed out?

[jongerlings]

is your 'Add fuel' table maxed out?

I don't see this table in the address file...

Would need to see a log of this if you can post one

I will get one posted up here when I am able to get a log

Yes, AFR is 14.7 at cruise in closed loop, drives like a champ! I was mainly just wondering if that latency value seemed too high to you.

[Matt]

It is a bit on the high side, but some injectors need larger numbers

Sorry Add Fuel => AlphaN increase fuel (enrichment tables section)

[jongerlings]

The AlphaN increase fuel table is set at factory right now, I will increase this and see how it affects things.

When copying the basic tables into the factory tune, I noticed you left out the Fuel throttle enrich (VE) table, and kept it stock...the original tune had this table modified in a few areas, could this be why I'm getting this hesitation?

Also, I can't find anywhere an explanation for what "Base Injector Time (end injection)" is, or why my tuner would have changed it. Can you tell me what this table normally is used for?

[Matt]

When diagnosing problems, I start with the factory base image and gradually pull the changes in

The throttle enrich map will add extra enrichment based on throttle position. We don't normally need to change it unless there is a flat spot in a certain area. It is interesting where the tuner has changed this, being mainly the higher RPM low load, which is would only be used when accelerating whilst RPMs still high following deceleration

Base injection time (end injection) - Came from the Japanese definitions. From what I can work out it is when the injection pulse finishes relative to the CAS signal. So injection time is the duration, whilst this controls the timing. Not something I normally change