Closed loop idle too early?
 

Ok so I might have figured out my dying issue, but not sure exactly what setting to change.

Whats happening is holding constant throttle, low load it enters into closed loop idle, and tries to drag my idle down. Then when I push in the clutch my rpms drop like a rock (because my idle valve is barely open).

I'm guessing the culprit is PID disable rpmdot, or RPMdot Threshold. What say you guys.

https://www.miataturbo.net/attachmen...ine=1449249354
https://www.miataturbo.net/attachmen...ine=1449249354

I don't have any help for you, but I want to follow all of the CL idle threads I can since mine still needs a little work.

from a glance my guess is: its activating while youre still driving with the TPS under 5%.

On that log you're at 2% TPS, then push the clutch.

Once you hit 5% and met other thresholds, it's trying to close the valve. You can see in your log where the idle speed target kicked in, so you were running CL idle while still driving. Then you actually let off throttle and it's unable to recover fast enough because the MS thought it was already idling, and no dashpot + initial duty gets added now to prevent the RPMs from crashing.

corrective action: tps activation % just above where it sits at idle, readdress initial duties table.

Ugh. My tps varies too much. I'll try and lower it some. But depending on the day my base TPS can vary by a few %

then add lag factor to the TPS to reduce noise. I see the TPS log hits -5.7%. You have CL idle running while you're techincally at 11% throttle.

a typical setting is 0.8-1.0% TPS.

I'll try that. Thanks.

Edit: Its already at 10. Which is "highly smoothed"

I set my TPS closed ADC to a few points below zero. And set "Auto-zero TPS" on. We'll see what happens.

why so noisey?

something that helps there is rotating the tps in reference to the TB itself so that at idle is not such an extreme point on the variable resistor. you can lossen the two screws holding it in place and rotate it a little so closed might actually be where 10% tps is normally now for you, after recalibration, this might help you find a better working zone for your TPS.

Its not as much noise. As changing values on startup. Its relatively stable when running. But some mornings its -3, sometimes its 2. I wonder if my throttle return spring is wack, and its actually not returning to the same place.

I've been tuning CL idle on my car for 2 days now, before that no experience at all.

That said, I looked at your log. 2 things stand out that will cause problems

1. TPS does not return to same place every time. Either throttle is sticking (that's a safety problem-fix asap), or TPS sensor is shit, or wiring to TPS sensor is somehow causing problem. My guess is throttle as the log shows RPMs dropping when TPS drops at the end when it dies on you in the log.

2. Voltage dropped when it died at end. Fix 1 and this problem will probably go away. I can say though on my car, I have a 220A alternator and it doesn't like low RPMs, so I'm going to put a smaller pulley on it so I can idle lower without my voltage dropping on me.

TPS sensor is damn near brand new. Throttle may be sticking. Whats a good way to check this? Wiring is stock. I could scope the signal and see if its actually jittery/noisy

Voltage drops below like 500rpm. The car is basically dead at that point.

Look at the throttle and cable. Either can cause it to not close all the way. On my setup I have 2 throttle cables as I built a cam device to linearize my throttle. I have to leave a decent amount of slack in the cable when cold, else when it's hot the throttle won't close.

Warm car up, drive it, check for some slack in throttle cable. Adjust if needed.

Then with car off key on, zero TPS and then hit throttle and let off, check it. Then slowly apply and slowly let off, blip it, WOT and let off quickly, etc. Try to make it stick. If it does, get out of the car and psychically push the throttle closed, see if it moves. If it does, it's sticking.

+prop. Thanks. I'll tinker with that.

On my MS3 PRO, the ECU reading lags on voltage. It's like a smoothed/average value. You might want to confirm on yours.

Also I run 50 Lag factor for TPS. If you got high smoothing that could mask your problem from being a noisy sensor or bad wiring.

Battery lag is 90, so almost none.

TPS lag is 10, so very high. I'll scope it out this weekend if i get the motivation.

Having a hotside SC setup, I've spent more hours messing with idle and low speed driveability than I care to admit..... But I can say I probably have one of the best idle/low speed stabile hotside SC setups ever just from my brute force throw hours at it.

I've barely spent any time on it. I just pop the clutch out and bump start it again, kinda become habit when it dies.

Sweet, thanks for that, just saw you can adjust battery lag. Never noticed. Will bump that to 100. I run 100 on MAP and AFR for the quickness. 100 for map helps with low speed throttle response for me.

Also your idle setting are a lot different than mine. I'll post mine in case they help you out. I spent about 2.5hrs coming up with these last night. This works really well on my setup, the only problem I have now is my alternator not keeping up below 1,200 with all electrical stuff on (headlights, power inverter, stereo, a/c compressor and fan, two fuel pumps)

This is with a 99 idle valve, 4001 diode across it, 306hz.

https://www.miataturbo.net/attachmen...ine=1449258407

https://www.miataturbo.net/megasquir...it-miss-84668/

Max decel load is wacky. You basically are going into closed loop as soon as you start decelerating, even if its quickly.

The rest of my settings are slightly off, because I've been battling this dying issue, and so i bumped the ramp to target times and stuff.

you all should be using adaptive timing as well. and battery votage idle % compensation.

I do. Though I'm just using basemap values. It helps a lot.

Thanks for the link, interesting 1st post about need 2 or higher for the PID Delay due to a code issue. I'll try bumping it up to 2 and see if that changes anything, I did notice it not kick in with very light throttle inputs.

I dunno if it's right, but I setup the setting in the lower right box so it would go into closed loop quickly when I let off the throttle.

Hmm, I now see the battery voltage idle% compensation, I'll mess with that.

What is the adaptive timing, just the Idle RPM Timing Correction? I do use that actually.

Aidan,

Try changing max decel load to a value closer to what you normally hit on overrun. I get down to 16 or so, so I've got it set at 18. This way, if for some reason CL idle activates while cruising, either PID disable RPMdot will kick it out, or the super low overrun map value will kill PID if RPMdot misses.

Also, set your TPS way negative if you need to, just try and get it so that you never start with positive TPS values. Everything else doesn't really matter, MS don't care if you are idling -2% TPS, because that's less than 1%.

I agree that it is a TPS issue.

Thanks guys.

I'm pretty sure it is a sticky throttle issue. When I slowly let off the gas it reads right around 5%. But if I blip the throttle and snap it shut then it goes back to 0.

Ok I'm stumped.

Slowly let out throttle. TPS stuck at 4%. Can't close it any further.

Blip throttle hard, TPS back at 0%.

This only happens when the engine is running. When it is off TPS always goes to 0%.

Easy. Does it idle at same RPM, or higher when TPS reads higher? I'm sure you can logically figure out if it's a sticking throttle or TPS sensor after asking yourself this question.

It idles higher because closed loop idle doesn't kick in when its stuck at 4%. But otherwise its the same. I pushed as hard as I could to close it and no change. Until I bliped the throttle.

I'm going to scope voltage tomorrow.

Well yeah have idle control off to test.

But if you say it doesn't change then TPS is your problem.

try adjusting the TPS like i suggested. it could verywell be the tps needle thing is getting stuck and not the TB.

Adjusting it how.

go find the damn post. hint post #8

:fawk: its too early.

I'll try that. It's already adjuste d in the middle somewhat. About 360-900 on the ADC.

I turned lag off and there is barely any noise when just sitting.

I think there is a bug in the MS 3 code. It doesn't register negative rpm changes when looking at the PID disable.

Can someone confirm that it's different in ms2 code.

Test:

Set idle activation threshold to like 50% so you can get it to activate while driving.

Then get it to idle at part throttle. Set PID disable rpm dot to 300-400. Increase revs slower than your disable rpmdot and watch idle stay active. Increase faster than rpmdot and watch it disable. Now do the same with negative rpmsot. I did it by getting idle active while driving in a low gear, and then push in the clutch. My revs dropped and idle stayed on. Shouldn't it turn off?

Again I'm new to this, but why would you want closed loop to register negative RPM changes to disable it? If it did that, it would turn off closed loop if the RPMs drop and it tries to die on you- the opposite of what you want!

If it drops too fast I want it to exit closed loop and go back to my initial value table+dashpot. Because something bad is happening

Swing and a miss.

Aidan's problem is closed loop turns on while he's crhsing at 1800 rpm. Closed loop can't bring 1800 rpm down to idle, because its not idle. When it actually goes to idle the idle valve duty is way too low because its been trying to slow down 1800 rpm when it shouldn't.

Pid disable rpmdot would kill closed loop if it accidentally engaged like this, and if the initial value table were tuned right then it wont stall and'll be close to the proper idle value.

I will check to see if my ms2 will disable on negative values here in a bit.

Pat, I do understand your take on it. And that very well may be the developers idea. Just wanted to check with the older code to see if it used to be different.

On ms2 I can't set pid disable rpmdot lower than 200 rpm. If rpm drops harder than 200 rpm/sec near idle values shits about to stall unless the valve is opened.

I'm nearly certain that's what it's for.

In my mind its this.

Set the PID disable to just above what your normal idle variation is, when turning on loads and whatnot.

If it goes above normal, shit has hit the fan, and add dashpot and go to initial value.

Ah, makes sense. I'll try messing with that on my car and see how it works. Can you post your MAT initial value table?

I can, but its barely tuned at all. Works somewhat decent.

https://www.miataturbo.net/attachmen...ine=1449344042

For reference my min and max are 29 and 68

Thanks. I gotta fix what I believe is a failed clutch master cylinder, but then I'll mess around with some idle settings later. If I figure anything out I'll post it here. I have it using last value, not MAT table so I'm gonna set that up first.

The problem with use last value is that it would still stall Aidan's car. It thought 29% idle valve were proper at 1800 RPM, then he stalled. If It weren't for the startup idle valve adders the thing would try and pull 29% DC on entry next time, which we know can cause stalls. (say he saved it stalling, now the car thought it idled fine and just left CL idle, and it's going for 29% next time first ting) I think it also causes the problems we saw Curly mention the other day, that is stacking dashpot adders if you blip the throttle many times in a row.

It basically exits closed loop when you blip the throttle, then goes back in while the RPMS are on the way down. If you blip the throttle again before dashpot is gone completely you enter too high after the next blip.

Tuning the initial value table is a pain, but CL idle ain't worth anything without it. My strategy like I said earlier is never closing the valve, it always tries to apply an idle valve value based on the initial table if CL is engaged or not, that way if my dashpot isn't completely out of the way I won't be adding to it, I only ever add to the initial valve table, crazy consistent values!

I wonder if this is caused by my TPS not returning all the way to zero, because its a wells tps that doesnt fit the throttle body perfectly?

On the scope there is about .02v p-p noise. And when tuner studio reads different values when i close the throttle the scope reads different voltages too. Fuck.

Deezums let me know when you check the negative RPMdot. I want to report it as a bug to Ken/James.

So it's a TPS issue. Try all the suggestions already posted here. Replace it if they don't work.

I played with the adjustment some. No go. I'll try again in the other direction. Ugh, do not want to replace this thing again. I'm on like the 3rd one.

My issue would be solved if the rpmdisable worked both ways.

I think my disable RPMdot is working on negative RPMdot values.

I also incorrectly assumed MS2 wouldn't let me set less than 200rpmdot max, I can go down to 100, and I've had it set as low as it'll go.

I dropped to around 650rpm in this log around 100 seconds and recovered using fuel and timing alone, the idle valve didn't move the whole way through. You might be able to fix the stalling, while still having dips at least.


https://www.miataturbo.net/attachmen...ine=1449356259

This is my valve at the min duty cycle in 4th gear, I just pushed it into neutral and coasted. Idle DC came up before passing my CLidle target.

I don't use PID either, it suuucks. I love this dummy mode, if it's low it adds, if it's high it takes. If I tell it to slow the fuck down and let things settle between adjustments, it does! I run EGO correction the same way.

https://www.miataturbo.net/attachmen...ine=1449356259

Is there a flag or ADC for CLidle active in a MSL? I can't find it.