Idle woes, even after RTFM
 

OK guys I am having a really tough time getting my idle to work the way I would like to. I am so frustrated I am standing on the roof right now with a shotgun to my head. I'm either going to pull the trigger, or jump, or both. I need to be talked down.

Background/What I have tried so far
I have tried following Travis' procedure on tuning the idle step by step. I've tweaked it myself, I've played with the idle screw (not the butterfly stop adjustment, the screw on the side). I've tweaked the AFRs to keep them close to constant throughout the idle region (0-1500 rpm, 0 to 60 kPa).

The timing is set at 14 degrees (advanced) throughout the same range (0-1500 rpm, 0 to 60 kPa). This seems to be a sweet spot for idle timing after playing with the master trim. I tried ramping up the timing at lower RPMs like 0-500 for stability but it did not help.

The IAC works when I change settings, but does not do much for the base idle below about 55. In other words, it does not matter if it is on 45 or 55, the rpm is the same (open loop, min 20% max 100%). I have tried it with a PWM frequency of 500 Hz and 1000 Hz. I did not notice any difference between the two. I checked the resistance on the coil and it is at about 10.5 Ohms which is OK per the shop manual IIRC. So the IAC is OK as far as I know.

I tried setting the idle speed (with the IAC off) to where it would idle with the fans coming on and off. I have two Spal fans, wired in parallel, so the fan load really pulls down the idle when they come on. It will vary from about 800 RPM to 1100 RPM with the fans.

My idle AFR varies but is centered around 14-15.

I am using the 1100cc modified Bosch injectors discussed in this thread. They are new
https://www.miataturbo.net/engine-performance-56/another-large-injector-option-1100cc-48509/

Toyota COPs, dwell set at 2700, new NGK BKR7 Iridium IX plugs gapped at 0.030"

The car runs great, everything except for the idle.

Problems:
There is a problem with the transient load when the fans turn off and on. The idle will drop fast, or just stall the engine completely. Using the Recovery RPM does not seem to help. When the A/C turns on is even worse. Two fans, and the compressor hit at the same time. If the fans are already on (for water temp), and then the A/C comes on, it holds RPM fine, just a small drop.

If I am sitting still, and idling, no fans on, with the A/C and cabin fans on, I can turn on the headlights and the cabin fans on (same time) and the idle is OK. If I try going from no cabin fans, to cabin fan with A/C (compressor turns on) then it will stall the motor. Fast.

After trying several different approaches the only thing that seems to work well is just setting the idle speed higher, like 1300 rpm instead of 900 rpm. The way when a load comes on, like the fans, or the A/C, there is some buffer RPM between the steady-state RPM (say 1300) and the RPM at which it gets unstable, which is below about 700-800 RPM. I would prefer not to go around with a high idle all the time, but if this is the only way to get it to work then I will.

I have set the extra effort for A/C when in open loop only (PID 0,0,0) so that when the A/C comes on, it will return to the same idle, or slightly higher. The problem is, if I have the idle at about 1100 or lower, the engine will stall before it has a chance to recover with the A/C extra effort. The RPM drops too fast into the below 700 RPM unstable region and the engine stalls. And, sometimes, even at the higher idles, the A/C will kick off and on a few times, everything is OK, then I will be sitting at a light or something, the A/C kicks on, and the engine dies immediately. Then it is a bitch to restart.

If I shut the engine off, and restart, it starts fine. If it stalls from the A/C, it is hard to restart. It takes 2-3 seconds of cranking, maybe more, and sometimes WOT to clear the cylinders.

Also, when I get everything set up in open loop, if I try to implement closed loop, everything gets unstable. It will start lean surging no matter what I try with the PID settings. I am tuning them per the guidelines from Travis and Adaptronic's manual.

I could try setting it to idle richer all the time, like 13-13.5 AFR. However, I have noticed that after long idle times it gets less able to handle the transients from 'fan on' and 'A/C on'. I suspect it is because the plugs are getting fouled from being run at a rich idle for too long. I literally sat in my driveway for four hours yesterday messing with the idle. It was nuts. It eventually got so bad that I had to give up. I pulled the plugs and they were carbon fouled so I cleaned them. During my normal commute the plugs do not carbon foul at all. The insulators stay clean. After reinstalling and going for a drive, the idle was better but still not perfect. I still could not get closed loop to work for example.

My concern is that if I set it up to idle too rich, if I sit in traffic for too long the plugs will carbon foul and the system will get unstable, presumably from a weak spark.

So, this has me wondering if the large injectors simply cannot maintain a stable idle at an RPM below 800 or so. I have been able to get it to sit at 800 RPM in open loop, with the Recovery RPM holding it tenuously at an AFR of about 13.5. But the idle is rough. It idles at one RPM, but the engine is vibrating heavily.

Also, from what I have read, the Adaptronic will not sense the Extra Electrical Load unless it is wired into one of the inputs. So I cannot add effort for when the fans come on (without the A/C). Which forces me (in open loop) to idle higher so that when that load comes on, there is enough margin to catch the RPM drop and keep it above 1000 rpm or so. Trying to get closed loop to catch the idle when the fans or A/C is coming in is not working. It is like it cannot react fast enough to hold the idle. Either that or the way I have the low RPM cells set up (fuel and/or timing) is not working.

I think one way to fix it would be to dampen the fan start up load with a Capacitor or something like that. But, that seems like a band-aid to me. I should be able to get this to work without resorting to that agreed?

I have tried all of the normal stuff and am going in circles. I need some HELP!

Here is the latest ECU file. This one will idle in open loop at higher RPM (1100-1300 RPM), but becomes unstable when I try to add PID.

Hey John,

The reason the valve is acting funny is because it goes below its "monotonic" limit somewhere in the low 60's. At 55 the valve stops opening like it should. So if your operating the valve down there whats likely happening is that the computer lowers the value from 55 to 54, nothing happens, so their is error in the PID loop, then it lowers it to 53, and still nothing happens, so the error again grows, now the system might goto 51, and still nothing happens. That keeps feeding back until for whatever reason the idle gets too low, THEN thats when crazyness ensues because the system is going to bump idle in huge amounts because the controller has a memory on this error but when it goes above the monotomic limit, the engine surges like crazy because you don't have to change the valve PWM by 15% to do something effective.

Solution?! Crank the screw valve shut on the throttle body until you get your IAC PWM upto 70 percent or so at the idle RPM and AFR you want. Once you you get it there, try to add PID up to 2,7,2 or there abouts. Need log of idle if that doesn't work, and then I can help guide you in the rest of the way.

OK Travis, makes sense, I'll give that another shot and report back. I have not had the idle screw turned in far enough to the point where I needed a base idle setting at 70.

Also, one more question. What should the settings be for the air correction under the base idle settings (in the open loop area)? I have seen some base maps where it is flat (all set to 0) and some where it ramps slowly up to 20%. I think maybe I should try zeroing these all out while I retry the procedure you describe above.

I have always used zero, if the PID is working like it should, it should be able to correct for you without having to mess with it. In order to set it up right though, what I would do is do open loop idle tuning in the morning, and then in the hottest part of the day. Use that graph to get the same idle RPM for both conditions once warm. Then interpolate the rest of the graph.

This works for me.


Re: Logic input
« Reply #1 on: October 09, 2009, 10:12:19 AM » Quote

--------------------------------------------------------------------------------

Hi Andy,

Try this:
- Set Digital Input 9 (logic only) to 'electrical load' type
- Set Aux Output 9 (logic only) to have the same behaviour as the cooling fan
- The Logic only inputs are linked to the Logic only outputs in the firmware, so the input should be triggered when the output activates.

Cheers
Rob

That is freaking clever!

I like it, thanks for your contribution!

Best,
Travis

Agreed, that is awesome. I thought there had to be some way to make this work in software but was not having any luck figuring out how. The linked input-output logical channels is the key. I'll give that a shot. I think it will be quite helpful to add some open loop (pre PID tuning) extra effort for the fans as large as the fan load is.

OK that seemed to me that would be its intended use. There is no real discussion on it in the Adaptronic manual. I am zeroing it all out for now. Once I get everything else working on the idle, I'll come back to it later.

I had written a noob's guide of what you did here:
http://ozmx5.com/viewtopic.php?f=7&t=91

I got the extra effort to work for the radiator fans per the method above. Thanks.

However I am still having the same problems with the idle after readjusting the idle screw and having the base idle (open loop) duty cycle up to 70 from 55. Same problem. Large, sudden electrical transients or the a/c clutch kill the engine every time. If I ease into the load, it is fine. If the radiator fans are already running and I witch on the A/C compressor it is fine. If I switch on the a/c compressor when the radiator fans are not running, the engine dies. And it dies immediately, like I turned off the ignition key.

Same thing happens if all the fans are off and the headlights are off. If I turn the headlights and the cabin fans on full, simultaneously, the engine dies, instantly. It does not matter if I have PID enabled or not, or if I use high P values or not (with D to stabilize). Same problem. I spent two hours in the driveway last night working on this and basically got nowhere.

I have better things to be doing so I want to get this fixed.

Monotonic or not, the car will never have a good idle if you limit the maximum to 55. with all my loads on, I need it to be at 60. when A/C kicks on, it needs to head closer to 80 for a split second.

with all the loads off and the car warmed up, it drops to 15 (my minimum) and still idles at 1100 rpm with a target of 950.

Where is your idle screw set at? Mine is a 1/2 turn out (counterclockwise) from the full in (all the way clockwise) position. That gave me about a 70 base idle effort (car warmed up, nothing on, 14:1 afr) at about 1200 rpm. For A/C I add about 8-10. The radiator fan (with no a/c) adds about 4.

The min and max on the IAC are set at 20% and 100%.

As far as I can tell I am doing everything right here. Either there is some obscure setting buried in the ecu file somewhere that is set wrong. Or there is an issue with something in the car, like a bad IAC valve, or weak coils. But, it is all working, mostly. Everything works like it should, just the quick, large transients are the issue. AS I said they kill the motor quickly, like I shut off the ignition. But when I look at the logs, they show the voltage never going below 12V.

My ecu and log are attached. For the most part this shows where the car is idling and I am turning on electrical items like the lights, cabin fan, and a/c. When the car dies, I reset the LC-1 by powering it off so the square wave wackiness in the AFR plot is just the LC-1 warming up.

Not sure that it matters but both my triggers are set to rising edge.

As far as I know, the whole rising edge trigger thing came from that problem I was having with hot cam and crank sensors. Going to rising edge band-aided that problem.
https://www.miataturbo.net/adaptronic-63/sudden-rough-running-%5Bresolved%5D-34294/
I have since replaced both the cam and crank sensors with new ones and gone back to falling edge. I think that falling edge is more noise immune and less prone to false triggering so therefore it is preferable.

I suppose it is possible somehow this is causing it. But I did not see any RPM anomalies on the log file. It reads RPM all the way down to zero IIRC.

My best guess at this point is that even though I cleaned the plugs, there may still be enough carbon on the insulators to cause an alternate path weak spark. Carbon is hard to clean out completely. I am going to install fresh plugs and see if that makes any difference.

My idle will dip with electrical loads too, but it doesn't stall anymore with the new tune. It will occasionally stall coming down from high rpm though. I don't know what will solve your problem, but I do want to know what does

Well it was not the plugs. A fresh set of NGK BKR7E gapped at 0.030" did not change anything. All the same problems as before.

I feel your pain

My aussie 05 SE has exactly the same issue when AC and both fans fire at the same time.

come throw some support for the AC request delay at http://adaptronic.com.au/forum/index.php?topic=1232.0


heres how i made it better (not perfect though)

my idle like yours is 1100rpm solid and idle air is 70 at running temp

try these settings

richen up the zones directly behind and below etc (all the zones it "falls" into)
note my highlighted zone is where i idle at

also did similar with timing...

set closed loop fuel as per my settings (with a 14.7 afr target)

set closed loop idle (1100pm) and all settings as per mine

if you use all of these together, you will idle nicely at 14.7 at 1100rpm and when it falls from the hit, it goes open loop fuel at 1050rpm on its way down, and then at 1000rpm, the closed loop idle recovery also helps soften the blow.

after all this happens the richer open loop fuel and timing bumps you back up into closed loop until the next hit. hope this makes sense!

you end up with a clean idle, with only occasional rich moments

worth a try. but will be so much nicer once the AC request delay is available

OK I added my request in at the Adaptronic thread you linked to, thanks.

I see what you are getting at and am willing to try it. I had tried before richening up the map below 1000 rpm but was not having much luck there. But then again I did not have the rest of it set up like yours. What would happen if I remember correctly is it would hang at the lower RPM, around 500-700 then eventually stall. I am guessing all of the extra fuel was bogging it down.

But according to my log there is a definite lean spike when the a/c comes on so it makes sense to try this. Check out the attached image. This is from cycling the a/c on and off.

Some questions about your settings...
- You have your minimum value in the overall control box set at 69%. What is behind this? Just trying to figure it out. I think I know but just want to be sure rather than assume incorrectly.
- Your extra effort for a/c is a lot lower than mine. Yours is two, mine is more like 5-10 depending on where I have the idle screw set at. Can you tell me how many turns out from full clockwise you have your idle screw set at?

Thanks

OK Bitrusty, I just want you to know, at this moment, in my opinion...






























YOU ARE THE MAN

At lunch I warmed up the car and tried your suggestions. It is much, much better. It is still not 100% bulletproof but it is a great improvement. It holds a steady 14.5-15 AFR at idle, and does not stall when I engage the a/c compressor. It could still use a bit of tweaking but the preliminary results are great.

Thanks for the input. I'll post up snapshots of my settings later. Bitrusty has my nomination for Most Helpful Noob Post of the Year.

i dont get why bitrusty does not have any additional RPM for A/C in teh closed loop idle section.

also, so you guys know... the settings in open loop idle are "baseline" values for those conditions. If you put 20 in the extra effort for A/C, it will instantly jump +30 over your base idle for those conditions to get closer to operating values for the PID.

for example: if your base idle is usually around 50 with nothing on and 60 with the A/C on, you could set A/C effort to 10.

On mine, I use that value to "catch" the hit from the a/c coming on. I think it's set to 24 or so. When a/c kicks on and load jumps to 40 or more kPa, the idle valve is wide open and revving the engine. Then the PID system corrects it as soon as it catches up.

Neither do I, but I tried it that way and it works. But then again, correct me if I am wrong here, that setting in the closed loop idle section is for deadband, not additional effort, right? That is the way I read it in the Adaptronic manual. But then again, if you open up the deadband, and you are biased (with the base idle settings) for higher RPM (than the closed loop target) then it would ACT like extra effort anyway. In other words, the idle would drift up to RPM = target idle + deadband (assuming RPM is less than what the baseline idle would be without PID).

My head hurts...

Yes I know. But good clarification for those who do not.

You know I think I tried that too (among 100 other things). If I understand you, the idea is to put in a huge base idle correction for A/C to avoid the stall (just like opening up the throttle, but a lot) and let the PID loop keep the RPM from grossly overshooting. I might not have tried nearly as much as 24 however. Good to hear it works after all. Maybe I should try it again. I can stop the engine from stalling if I stab the throttle when I turn on the A/C. This is effectively the same thing, but with the PID controlling overshoot.

Now my head really hurts... I need a beer.

glad to hear it give you at least some improvement :)

i have spent so much time fiddling with this in my garage my neighbours think i'm trying to gas myself lol

its certainly no solution to the problem but it helped a LOT with my ride

from a cold start, and any time up to 55c WT (open loop fuel), my idle is rock solid and the AC hit is bearly even noticable but my guess this is just the extra trim propping it up

I dont have any AC rpm in CL idle because it seems to have no effect on "the hit". it just raises the idle after recovering from the hit so i dont see the point.

same with my low extra effort idle numbers. nothing seems to work fast enough to catch the hit. delaying the onset of AC or the fans separately seems to be the only real way to fix

I thought 70ish was the recommended minimum. Something about anything under 70 odd was not enough for our idle air?

I used to try idle at 69 but its just crept up from much fiddling. I'll raise the min to 70 and see if its any different.

I'm screwed 3/4 - 1 turn back from closed.

Is it just this extra effort thats giving you the lean shot?

Ill post a log of my AC idle dip soon. Heres my whole tune if it helps.

running DW 600cc

this got me fiddling some more...

I get an even bigger improvement "capping" the max idle control

my settings were 69min 100max. now, they are 60 and 80

along with the settings i posted earlier, cap the max idle valve to say 80 and set the min to 60 to begin with and give it a go

your log shows the idle flatlining at 100% from all the extra effort


also, check this out! http://adaptronic.com.au/forum/index.php?topic=1232.0

I added in the extra effort for the fans and now the idle doesn't dip like it did before. I didn't do any of the other settings though because I didn't feel like messing with it. I get an occasional low idle and once in a blue moon it will stall on me.

So I was still having some stalling problems just with the fans coming on. I richened up all of the cells around the 1000 RPM 40 kPa cell and that seems to help with stability. It helps with hunting, and when the idle settles the aforementioned cell still keeps the idle lean when closed loop works back in, about 14:1. Also, I have tried a LOT of extra effort on electrical load 1 (fans) on the oder of 15, and very little, on the order of two. It did not seem to make much difference on the RPM dip. I might have had a tad more stability with the lower setting but I cannot be sure since behavior can be inconsistent from one fan cycle to the next.

PID is 4,4,2 right now. I have tried 2,7,2 and some other I-heavy settings. It does not seem to work so well for me.

I think the 800 ms delay on the A/C clutch will be a big help. I wonder if they will add that to the electrical loads as well?

BTW thanks for chiming in all. This thread is quite useful IMO.

I hope not. We'll be back to square one i would think.

I just hope the 800ms is enough and is only delaying the AC clutch

Why? I'm interested.

I am thinking it would help. The fans are almost as big a hit as the a/c clutch, right? At least mine are, in terms of initial surge, not so much steady-state load.

I think the idea with the delay, and correct me if I am wrong here, is to allow a split second (literally) for the RPM to build with the additional IAC help. Thus, when the delayed load hits, there is more energy available (RPM) to offset the load. Plus if PID is being used, it is trying to lower the idle quickly instead of raising it, which would seem to be inherently more stable. Seems like the delay could be balanced just right by tweaking the extra effort amounts, and would work equally well for most any load type, especially since the extra efforts for the loads can be independently adjusted.

I am assuming the delay would be a total of 800 ms for fans alone or fans with a/c.

I have just been assuming its just an AC request delay, but if they have expanded on that, even better.

They have been talking about it for a while
http://adaptronic.com.au/forum/index.php?topic=226.0

FWIW, The hit I get when the fans kick in alone (from WT) is not even worth writing home about. Smooth as...

This could be how saboteur set up my loads for me at install :winner:
http://ozmx5.com/viewtopic.php?f=7&t=91

Are your electrical loads setup like this?

My fans kick in together as load 2 and the small amount of extra effort seems fine for this here.

Currently if my fans are on due to WT, I can turn the AC on and off and hardly feel a thing. My issue has only always been the stumble from the onset of both fans and AC at the same time.

shit, maybe i'm just lucky?

The way I am hoping it will behave is, fans kick in...800ms... then AC kicks in (after the AC effort has been added)

One difference in our setups may be that I have the fans wired in parallel. Coolant temp or a/c turns them both on.

I tried the extra effort for fans and I do not like it. It will still randomly low idle but now it will hold like 1400 rpm for a minute when the fans are one and before the pid kicks in

How about using AC MIN RPM? Located in the special funtions tab.

The AC won't engage until the rpm target is reached.

Good idea but I tried that one too. I could not get it to improve the problem.

I also did the max idle value cap at 80 and it's heaps better, nice find!

Yeah that would make a difference. You could stage them within a few degrees of each other if you revert the wiring back to stock. Just replace Aux Out 7 with a logic OR of A/C and the water temp you want. Here's an example:

http://sites.google.com/site/timkent...ic-aux-out.png

Good idea on staging the fans. I may have to try that.

I staged the fans originally. It helped a lot.

Well I figured out why the idle up for fans didn't work well for me. I decided to use a logic channel that was already in use for the ac fan..:facepalm:

V6.000 has the A/C delay for better idle control. You guys might want to check that out. Should reduce the double hit of fan and A/C coming on at the same time.

Not just yet it has a bug!

subscribing

Firmware v 7.0 was just released. http://adaptronic.com.au/forum/index...15.0;topicseen

wow, I'm still on V2.0

so how do we use 7.0 so it doesnt die everytime my fans kick in? :vash:

lets not even talk about rising both windows at the same time here...

7.0? what the hell? I thought 5 was the newest. They are coming out with new ones like it is their job

well i dont know wtf is going on but my idle is back to the 11s :(

Playing in the tuning tab does nothing for a/f at idle, wtf???

hhmm i've been looking at mine and they are somewhat different than what you guys have any ideas? I'm afraid to start changing them.

Attachment 194549

My idle a/f's were running a little rich this weekend. I think it has to do with the cooler temps and running the car in open loop. I never had any a/f problems at idle or cruise in closed loop fueling

well over here it has been anything but cold and my a/fs are crazy rich, if i lean them out to where they read 14-15 on my tuning tab, once in a while whatever is adding fuel at idle turns off so it reads whatever is in the map and it dies because it has almost 0 fuel at that point.

Anyone else get random fuel changes in the other section? Its been a while since I looked at it, but I remember an other change of like 300% that was counteracting another change of almost the same value.

Is it possible the air temp correction values are not what they need to be? I'm using what I've seen in every other map, but maybe we actually need something different than what is theoretically right?

today my a/fs were 11 in the am, and on the way back home just an hr ago they were switching between 12 and 14 at idle.. i dont get it. Something else is when i am giving it some gas at idle and slowly increasing it increases slow, as soon as it hits 2k, it revs up quickly. Does not make any sense at all. Everything over 2k is perfectly fine.

Auto what temp sensor are you using? the GM one or the stock push in one?

well after some more driving today i kept my eye on those fans, when the 1st fan turns on my a/fs go up 3-4 points so if its 10-11 it goes up to 13-15. When the 2nd fan turns on my a/fs get even leaner and the car dies at that point and wont start unless i 50% gas it and keep it above 2.5-3k rpm until the 2nd fan turns off as at that point (only) at idle i'm at 19-20 a/f. Makes no sense to me as its not reading whats on the tuning map.

same thing again, fan was on, a/fs were in the mid 14s...

Post your map. It might be something simple you're missing?

Thats my current ecu file.

Open loop idle effort should taper down to 0 once the engine is warmed up. Also noticed you map prediction gain is set to 0 - should be 150-200 from what I've read.

If you haven't already check out Travis's idle tuning guide. Step by step makes it pretty easy to get it close.

well i've kinda experimented a bit. both windows at the same time up/down, a/fs go lean and it dies...

it appears that its going the opposite way of how its suppose to be. lean vs rich...

as far as the open loop idle efford you're talking about the graphical representation in the idle tab right? The only way to lower these things that i know of is by turning all the base idle value all to 0s after let say 30 degrees? or is there another way i'm not familiar with?

i'll try 150 for the % gain in the map prediction.

Disregard the open loop idle effort comment, I had coolant enrichment map on my mind.