About to throw in the towel nailing down idle

**5_Liter_Eater** · 12-05-2007

***I don't want to read all this...just take me to the procedure!***

I am getting VERY frustrated trying to eliminate hanging idle, surgind idle, idle that stays at 1200 until I come to a stop then will eventually meander down to ~900, idle that dips when I push the clutch in then surges up to 1500, idle that won't come down below 900.

I have been trying to develop some sort of methodology for E40 idle tuning and I have been anything but successful. Everything I do seems to result in a worse idle. I'm getting ready to remove the rev limiter, put her in neutral and put a brick on the gas pedal.

I guess I am just looking for any advice/guidance for tuning idle parameters on LS2's. I've tried numerous things around the minimum RAF and idle adaptive spark and like I said, everything I try makes things worse. Gen 3 idle tuning was SO much easier and more consistent!

Since I can't upload a damn tune while at work reference my tune here: http://www.hptuners.com/forum/showpo...8&postcount=40

**LazMan** · 12-05-2007

Try these changes, pics attached. They are based on my 2006 GTO (same computer) and keeping in mind your big cam. I've had good results. Let me know how it behaves / if it helps.

I did the fancy base idle airflow changes you had (for example numbers past a certain RPM at zero) and that didn't work for me. Crazy idle surging up and down. My curve looks kind-of like the one in MOD_01.jpg, but since my cam is stock my airflow numbers are not as high as the ones I chose for you. It should help. Making the curve look like that also cured some "on and off" behavior in my DFCO.

Also, I noticed your low-octane spark table doesnt have 35 degrees for idle set in it. I'd change it to match your high-octane spark table's mods so timing stays at 35 degrees where you want it.

Maybe that is an issue too?

Good luck, and if you post your updated tune, I'll try to help. (Subscribed!)

**5_Liter_Eater** · 12-05-2007

Thanks, I'll give these a try shortly.

**BBA** · 12-05-2007

I can't believe this all over again...I tried all the same things with the big cam LS2 TBSS (as conventional wisdom says), none of it worked, so I took drastic measures and found what really affected idle. The answer is NOT in airflow, it's timing. More the effect of overlap on the timing tables.

Ended up killing the underspeed idle advance and increasing overspeed (slightly), did not move up very much on idle airflow (+6 all) and set timing below 1000 RPM's to mid 30's in both High and Low octane tables.

Idles smooth as silk at 900, no problems slowing down (except the occasional buck when slowing down in gear) and starts first time hot or cold.

Don't listen to all the advice the 'pro' tuners give about 'that' doesn't work and shit until you try it yourself.

Heck, I'll send you the tune if you want to see.

**LazMan** · 12-05-2007

I forgot about BBA ... thats right ... I think you went through this already on the "big" cam. =)

Listen to BBA. He knows. BTW, I agree with the "do what works" philosophy. 5L had his timing set to 35 everywhere that matters, but he missed the low-octane, which I imagine wont help.

I'd love to see that tune sometime too BBA.

(I'd also like to see a big bumpstick in my GTO. What cam did you put in the TBSS? Stock converter?)

**BBA** · 12-05-2007

Originally Posted by LazMan

I forgot about BBA ... thats right ... I think you went through this already on the "big" cam. =)

Listen to BBA. He knows. BTW, I agree with the "do what works" philosophy. 5L had his timing set to 35 everywhere that matters, but he missed the low-octane, which I imagine wont help.

I'd love to see that tune sometime too BBA.

(I'd also like to see a big bumpstick in my GTO. What cam did you put in the TBSS? Stock converter?)

This is what I ended up with on Matt's TBSS. The cam was an aggressive ramp 228/232 112 LSA, Kooks LT, off road Y, CAI, K&N, e-fans, 3200 stall, etc... The first tuner could not get it to idle and it took me about two days of playing to see where the real problem was, the cam was freaking out the timing by freaking out the MAF sensor at idle. I just got rid of the MAF signal from affecting timing altogether at idle. Basically, it idles on RPM based timing now on the underspeed/overspeed timing adders instead of on MAF based timing. Base overall idle timing is between 12-17º at 900 RPM.

**12secSS** · 12-06-2007

Bill, I have been able to get "big cams" to idle nicely on LS2, even when the cam was the onbly mod!

And by "big cams" I mean 242/248 duration and up, with no more then 112 LSA. I barely glanced at your tune, but I see a lot of items that need to be addressed in order for the idle to be calm. If you'd like, send me an email and I will disclose my "secrets"

to LS2 idle tuning ... and NO, it is not solved by jacking the timing up passed 30 degrees in idle regions.

Us "pro tuners" have dealth with more radical setups that have given us years of experience in this matter.

**5_Liter_Eater** · 12-06-2007

I tried LazMan's suggestions on the way home and while out and about xmas shopping last night. It has cured any and all hanging idle problems but it made the dip way worse. So if driving in the parking lot @ ~1500-2000 and push in the clutch it dips way down, almost to zero, and actually stalled a couple times. Other times, if the RPMs are coming down from, say, 3500+ it seems to settle down without dipping. First I tried adding 1 across the board to the min RAF which didn't do anything noticable. Then I bumped up the 650 range of the min RAF to 11 with the thought that if the RPM's got that low then it would crack the TB to keep it from dropping any lower. To my surprise this didn't seem to do much either.

My idle timing is pretty consistent at 30-35*. Aside from the big dipping problem it idles pretty great. Should I ramp up the undespeed timing? Bump the min RAF up even more below 800?

I'd post the logs if it weren't for the damn firewall...

**LazMan** · 12-06-2007

I would add to the 1000-1200 range and down, the same amount. Try to keep the same shape...that shape has worked well for me.

Add to others above if needed to keep the shape ...and see if that helps.

I'll look at BBA's post tonight and see how he handles idle spark.

**LazMan** · 12-07-2007

5L, BBA's tune is full of Idle Spark GOODNESS!!! His changes to Idle Over / Under spark, particularly Over-spark, really look like they will calm down the idle.

I suggest you take a look at his entire Spark Advance page as it is very nice. Also, look at his Base Idle Airflow table which while kinda-similar in shape to the one I sent you, utilizes higher numbers to keep that air-hungry cam of his (and yours?) happier, and also uses different numbers for different gears (to keep it from bucking?).

Anyhow, I also noticed he runs closed loop ... I've given up on closed loop. I hate it.

-Laz

PS: Maybe not HATE it. lol.
PSS: I did virtually the same thing on my DFCO Exit Spark.

**5_Liter_Eater** · 12-07-2007

Last night I tried adding 3 to LazMan's min RAF table. This helped the dipping idle a lot. It only dipped down to ~600 sometimes and settled down how it should without dipping at all a lot more than it used to. Still no hanging idle although I'm afraid if I raise the min RAF any more it will begin to hang up.

Tonight I'll try BBA's min RAF, idle under and overspeed, and adaptive idle proportional tables. I noticed that BBA's idle spark advance table does not match his high octane table; It's considerably lower. I guess I don't know when the PCM uses the idle spark table VS: the normal high octane table so I set mine the same. I can see the TBSS having considerably different airflow characteristics though.

I would like to know the theories that go behind the numbers in these tables though so it's a repeatable method rather than just plugging in numbers that have worked for others.

**LazMan** · 12-07-2007

I noticed that too. I think your timing is too advanced at idle. Here is my theory on when what spark table is used.

The High (and Low) octane spark tables are used when driving (duh!). The coastdown table only is used when the vehicle is moving but your foot is fully off the pedal, and DFCO is not on. The idle table is used at idle, when not moving.

The RAF table I pretty much understand too. No matter what the PCM calculates, the airflow will never be less than that value. This is why a smooth shape seems to work better. You don't want the RPM to hang or go UP sporadically (which is scary). Try adding a bit more to my table as the RPM go down, moving it closer to BBA's "out of gear" numbers.

I think on cars the airflow in different gears won't be as critical as it is on a big, huge giant square truck (no offence, the TBSS is a nice car!)

Anyways, glad its getting "closer" for you.

**5_Liter_Eater** · 12-07-2007

I got that much timing by setting it REALLY high (40) and using VCM controls to reduce the timing and watched the ETC and MAP values. I got the lowest ETC and MAP numbers with the higher timing.

Thanks for everyone's input so far.

**Isaiah48** · 12-07-2007

Alright; so what did we end up with as to the cure to big cam idling problems?

Please speak in laymans terms if possible and provide insight as to the theory along the way...

Thanks!

**LazMan** · 12-07-2007

As far as 5LiterEater's car is concerned ... its a work in progress Isaiah.

**12secSS** · 12-07-2007

Ok, here is my way of tackeling idle tuning. This is how the idle works on GM ECM from what I have been taught and from my experience. Now I will attempt to simplify it down.

RAF Table
This the main table that let's the ECM know much air is entering into the engine. This airflow is directly affected by the overlap of the cam, the duration has less of an affect ... this is why cams with small durations and wide LSAs can idle better then cams with larger duration and narrow LSAs. Think of this as steering input as you floor it from a stand still. Stock cams will require much less steering input then wilder cams.

Idle Spark Advance
We all know what this table does, but it directly affects the idle torque. Think of this as the grip of the tires, you only need so much.

Proportional and Integral Adaptive Idle
Many of you do not know what this is, correct? This is what tells the ECM how much error to expect for the idle to vary. Now we all know we would like to have a 250/250 cam idle the same as stock ... that isn't going to happen, nor should you expect it to. Obviously with larger cams (overlap) we need to expect the idle rpm to fluctuate a certain amount of rpms. This is for idle correction with either and IAC valve or ETC. For this I have made an Idle RPM Error (RPM) custom PID [Desired Idle RPM - Engine RPM = Idle RPM Error], it shows me how much the idle rpms vary from my target rpm. Think of this as the lines down the road that outline the lane you are in (middle lane of a three lane road, for example).

Idle Adaptive Spark Control - Overspeed/Underspeed
This is the last resort to keeping the idle where we want it, since it is a faster correction it should only be used as such. The main idle correction should come from your RAF table (STIT on IAC equipped cars). Think of this as the concrete barrier walls keeping you from crossing onto opposing traffic lanes.

These are the main items for idle tuning, for IAC equipped cars there are other tables for off idle transitions (Follower and Cracker), and for ETC equipped cars those would be Adaptive Idle (Proportional and Integral) Airflow table as well as Idle Adaptive Spark Control (Overspeed/Underspeed) Gear and Coast tables.

The way I start idle tuning is by setting a realistic Desired Idle RPM and increasing the RAF table to a value I think (from experience, you will need to experiment with this

) is suitable for the overlap of the cam and based on my "tuned" VE/MAF airflow values (Dynamic airflow or MAF airflow). I then increase the Proportional and Integral values to what I would consider acceptable idle rpm error from the cam. Now one thing I do notice that many do is crank up the timing to beyond 30 degrees. That is a big no-no. You want to log Delivered Engine Torque to see where max torque is at, relative to timing. You may see that at 30 degrees of timing you are at max torque. That would be perfect for idling right, using max torque. Wrong! Because if you should need idle spark correction, it will not help. You want the torque at idle to be well below the max value, this is achieved by (again, from my experience) using up to 26 degrees of timing at idle. Next I also adjust any idle adjustment table that help keep the idle at the desired rpm. Normally I will zero out the idle spark tables here, the reason for this is if your idle can stay within the rpm limits (Porportional and Integral) with no spark correction, your RAF table is set right. You can check it by turning off idle spark correction in the scanner and see the results. Actually on the scanner it need to be switched ON is off and OFF is on, so use the ON button.

If your idle jumps up, you RAF is too high, if your idle drops and possibly stalls, your RAF is too low. Once the RAF is dailed, I will then bring in the spark adjust to within the rpm limits (Poportional and Integral) and slowly increase (or decrease for overspeed) the spark advancement based on idle rpm error, to a max of 10-14 degrees are the extreme end (500 rpms). You want the timing to be at the upward slope of the torque curve, not near the top. Because too much timing will put you over the top then you are on your own for torque. I also ensure that the Porportional and Integral Airflow is moved outside the rpm error limits.

Now to explain it in simple terms, as you floor the throttle on your car (our example here) you use steering input (RAF) to keep the car going straight within your lane (Porportion and Integral). Now if you give it too much steering input (RAF) you run the risk of hitting the retaining walls, those will do a much harsher job of keeping you on the road. Also too little tire (Timing) will keep you from moving or cause a violent situation, on the other side too much tire (timing) and you will bog and stall the car because the engine can't produce enough torque to overpower the tires. You want enough grip where the tires can still spin, but with enough grip to be controllable.

I hope I have shed some light on idle tuning and helps you guys straighten out your idle issues.

I would highly suggest taking Greg Banish's (eficalibrator on here) class to better understand idle tuning, or at least pick up his book.

**5_Liter_Eater** · 12-07-2007

George,

This does help to be able to think about these different parameters in this way and know which ones to tune first with a rough idea of how and which ones to rely on for different situations. I still have some questions, obviously.

That is the difference between adaptive idle proportional and integral? Why would they have different RPM Error levels? - Answer:
http://www.engin.umich.edu/group/ctm/PID/PID.html#pi - Be ready to break out your old engineering books boys.
http://en.wikipedia.org/wiki/PID_controller - A little less engineering based.
After reading both of these I have a much better understanding of what these tables are. I can't say I know why they would have different enable RPM error values though.

Is there any reason to have the idle spark and main spark tables differ?

For what RPM range do the idle under and overspeed spark tables work?

What is the purpose of the coastdown table and any recommended tuning theory for this?

How about this for an approach:

- Zero out the RAF table, disable idle adaptive spark via zeroing the tables or using VCM Controls. (now the only thing controlling idle is the PID tables.
- Set the idle base spark table very high (40), log calculated torque and use VCM Controls to see where the peak torque lies (hint: you can only reduce the spark by I think 8 degrees via VCM Controls).
- Say peak Tq is at 37*, set idle spark region to say 30.
- Log flow against the RAF table, commanding 650, 800, 1200 RPMs to find the average flow for each
- paste each flow into each gear on the RAF subtracting maybe 10-15%
- Log idle RPM error and use to set the RPM error level for idle proportional and integral
- Set the enable rpm error level for prop and int found above. Set the proportional table to be a straight line from 0 g/s at 0 rpm error to the max amount of flow difference between RAF cells. IE: these tables can account for up to 512 RPM's worth of RPM error. If the flow difference between 1200 RPMs and 800 RPMs is 3 g/s then thats .24 g/s error per every 32 RPM (PID table resolution). I'm not sure what approach to use to tune the integral table.

Am I on track at all?

**BBA** · 12-07-2007

Originally Posted by 12secSS

Now one thing I do notice that many do is crank up the timing to beyond 30 degrees.

So you say the way I made the LS2 idle by setting high and low octane timing vs airflow static was wrong?

If you noticed the idle timing did little to nothing, but the dip in timing due to airflow was what was killing the engine and making it hard to start when hot. Removing the effect of airflow on timing at idle RPM ranges really did stop all the problems. No it did not increase idle timing very much at all, even setting it to 34º or more below 1K RPM's, overall timing was still between 12-17º at idle, but it did eliminate the timing from bouncing all over the scale negative.

Like I said before, big cams confuse the MAF sensor and the MAF sensor kills timing, so at idle, you just don't need the MAF freaking out and reducing timing on normal cam surge airflow fluctuations.

Anyway, if there is a better way, I'm all ears but you will have to give something that works better than what we ended up doing, although we do not have an idle problem anymore setting it this way.

**12secSS** · 12-07-2007

Originally Posted by BBA

So you say the way I made the LS2 idle by setting high and low octane timing vs airflow static was wrong?

If you noticed the idle timing did little to nothing, but the dip in timing due to airflow was what was killing the engine and making it hard to start when hot. Removing the effect of airflow on timing at idle RPM ranges really did stop all the problems. No it did not increase idle timing very much at all, even setting it to 34º or more below 1K RPM's, overall timing was still between 12-17º at idle, but it did eliminate the timing from bouncing all over the scale negative.

Like I said before, big cams confuse the MAF sensor and the MAF sensor kills timing, so at idle, you just don't need the MAF freaking out and reducing timing on normal cam surge airflow fluctuations.

Anyway, if there is a better way, I'm all ears but you will have to give something that works better than what we ended up doing, although we do not have an idle problem anymore setting it this way.

That's just it, you are not all ears. I looked at your timing table and if you log Adaptive Spark Idle, you will see the Overspeed Idle Correction pulling timing, bringing the spark idle timing down to around 12-17 degrees keep the idle rpm near the desired. It is because you have way too much idle airflow on your RAF table, the primary source for idling. I bet you that if you try my little test, by turning off the adaptive idle through the scanner (you actually have to hit the ON button to disable, it is a bug in the HPT scanner), that your idle will jump up in rpms. This means you have too much RAF. One way to vefiry is by logging Spark Adavance, Adaptive Idle Advance, Desired Idle RPM and Engine RPM. Try it and post the results, if you are right, the idle should not jump up in rpms and the spark advance should stay at 12-17 degrees.

Cams with more then 20 degrees of overlap start having issues with reverse flow on the MAF, something like your cam will not. Hell my cam has 7 degrees of overlap and it idles perfectly fine at 800rpms and 24 degrees of timing, with little to no idle spark correction ... your cam with 6 degrees is much easier.

You need to realize that the spark is not the main function to setting idle, it is the RAF table. Timing is afterwards. Yes, you can eliminate airflow influences to the timing table by doing what you did, but the way you reached your target idle rpm is totally wrong ... as well as your reasoning that you needed to jack the timing beyond 30 degrees to achieve 12-17 degrees of idle timing.

**12secSS** · 12-07-2007