Idle Initial Map - Firmware Using MAT Instead of CLT?

[VcrMiata]

Now that the weather has been getting colder I'm finding that my idle has been getting higher even when CLT is up to normal temp.

I have a MS3 with 1.4.1 beta 1 firmware on a mx5 MSM.

When reviewing my logs I see that the initial IAC duty cycle is around 35 where it should be around 27 or 28 and it sure looks like the firmware may be applying MAT temps rather than CLT temps.

Log_CLI_35_CLT88.png

CLI Settings has CLT set for the initial table y-axis.

CLI Settings.png

the idle RPM target for CLT of 88 is 850 and the target for MAT of 18 is about 1100.

the initial values table

Closed-Loop Idle Initial Values.png

So why is 35.3 being selected for initial duty cycle instead of 28 + 2.7 dashpot (~30.7)? From what I can see the only way is if MAT is used.
i.e. 18C/1100RPM is around 33 and adding 2.7 which would 35.7 (with interpolation it's pretty close to 35.3)

Am I missing something or is there something weird going on in the firmware? I don't know if this was an issue earlier (prior to 1.4.1 beta 1) as I installed my MS3 in June and haven't hit really cold weather until lately.

[billr]

[VcrMiata]

I thought the snapshots would be adequate. Here are the files

2015-11-25_18.11.28_LOG0808.msl

2015-11-25_18.00.53 IgnUpd lt 4000 and gt 100Kpa.msq

[muythaibxr]

You should be using MAT not CLT. CLT is factored in by the fact that the target can be changed with CLT which then factors into the initial value lookup.

CLT is only there at this point because some people are using it and I don't want to break them when they upgrade to newer firmware revs.

[VcrMiata]

You should be using MAT not CLT. CLT is factored in by the fact that the target can be changed with CLT which then factors into the initial value lookup.

CLT is only there at this point because some people are using it and I don't want to break them when they upgrade to newer firmware revs.

Thanks, I'll redo my initial value based on MAT instead of CLT if this is the case.
Due to the proximity of the "Use Last Value or Initial Table" setting to "Use CLT or MAT in Table Lookup" setting, I assumed they were related.
To remove confusion, the "Use CLT or MAT in Table" setting should be renamed "Use CLT or MAT in RPM Target Table"

However, if the intention was to provide the option of using CLT lookup on _both_ the RPM target and the initial value table when CLT is selected for the "Use CLT or MAT in Table Lookup" - then that is not fully working as it appears to only apply MAT to the initial value table.

[Note1: my analysis seems to imply MAT is being applied to the RPM target table too]
[Note2: noticed that I had not mentioned my temps are in Celsius ]


Regards,
Tony
Edit: added Note1

[muythaibxr]

That setting only applies to the initial value table and grays out when you are not using the initial value table.

[nathaninwa]

If I read the first post correctly, your having troubles with idle not going to commanded correct? So if your target is 800rpm your idling at 900 let's say? (Sorry haven't looked at the logs)

The initial value table won't fix this issue, it's job is to set a value for the closed loop operation to start at. I'm running 1.4.1 firmware in both my cars and haven't noticed anything like that

[VcrMiata]

If I read the first post correctly, your having troubles with idle not going to commanded correct? So if your target is 800rpm your idling at 900 let's say? (Sorry haven't looked at the logs)

The initial value table won't fix this issue, it's job is to set a value for the closed loop operation to start at. I'm running 1.4.1 firmware in both my cars and haven't noticed anything like that

No, the eventual idle target for CLI is correct and the engine does converge on the proper RPM.
The problem is with the selection of the initial value, the duty cycle selected is too high - once the engine warms up the initial idle is too high when CLI first activates.
I had CLT Lookup set to use CLT but it appears to use MAT for accessing the initial values table.

In fact when accessing the initial values table it seems to use MAT for both the temp (y axis) and the RPM (x axis) lookup. i.e. I think it is using MAT to pull the target RPM from the target RPM table for deriving the RPM axis too; it's the only way I could derive an initial value duty cycle similar to the one seen in the log (~35 vs ~31 i.e. ~33 vs ~28 + dashpot(2.7)).

I'll spend some time this weekend to fully characterize the behaviour and move to MAT instead of CLT.

[nathaninwa]

I don't know the full functions of all the ins and outs of the idle setups but I have messed with it a lot. I too was using coolant and had issues with cool summer mornings to hot summer afternoons. I switched to mat and spent some time tuning the table and its darn near spot on

You can also try lower the htz of the output and maybe upping the control time. I'm using a ford valve and have had good luck at 150htz and 150ms control time. Slowing it down may give the valve time to achieve what your commanding

I also don't like using the idle timing, i just use the main timing table for that but do use the idle timing correction

I also don't have any delays in my pid, It may not be correct, but I have had good luck with waiting 0 seconds. This may mask a to high of initial value, but it's working for me. I've got return to idle pretty solid with only a partial oscialtion, like 2 less than 50 rpm overshoots when your easing your way to a stop behind traffic

I looked at the log and there is only one spot you can really look at When you log next for posting, make sure to keep for like 30 seconds to get the data needed. Your on and off the gas a lot

[jsmcortina]

Could you point out a time in your datalog where you believe this is happening?

James

[VcrMiata]

Could you point out a time in your datalog where you believe this is happening?

James

This happens whenever the the PID activation settings are met -
for example:
203.0 : MAT is 15.9, CLT is 86.9 - using CLT, initial duty cycle should be ~31 but is 35.3
237.8 : MAT is 16.9, CLT is 89.4 - using CLT, initial duty cycle should be ~31 but is 35.3
262.1 : MAT is 17.7, CLT is 87.8 - again using CLT, initial duty cycle should still be ~31 but is 35.3

Perhaps I'm not understanding how this works but it looks like it is using MAT instead of CLT.
In any event, I'm going to try adjusting the initial value table to use MAT this weekend.

Hey, I just noticed that the initial duty cycle doesn't appear to be changing when MAT changes, it seems to be stuck using 35.3 for an initial value - previously when I analysed a log, MAT was reasonably static and I didn't notice it wasn't changing. Wonder if something else is going on.

OK, I'm going to back off of this thread a bit as I have to do some more analysis - I'll see what I can do this weekend to try to figure out the behaviour.

[VcrMiata]

Quick update.

mea culpa on the MAT versus CLT question - turns out when selecting the CLT initial value table that CLT was definitely being used.
The constant duty cycle chosen when MAT was changing gave me a clue that something else was coming into play.

I found an adjustment for duty cycle in the Basic/Load settings->Fan Control dialog that provides an idle-up duty% and an idle-up target increase(rpm) - it was this setting that was screwing me up.
Once I found this everything made sense. It was a coincidence that the offset happened to be close to the MAT versus CLT temperature duty cycle difference (on the y axis of the initial duty cycle table).

So problem came down to me thinking that all of the idle duty cycle controls were provided in the idle dialogs - this is not the case; the fan control over the idle duty cycle is not provided there.

By the way, I did move to a MAT based initial duty cycle table. I've re-worked the values quite a bit and it seems to be working pretty well - idle during and after warmup has been smoother than it's ever been from MAT of 10C to normal operating temp of 35C in the winter (CLT from 10C to about 95C). I'm hoping it works as well when things start warming up in the spring.