Tuning with a NB o2, paging the good Dr. Hess

I was doing some research through the archives for what others have done for a/f tuning on challenge cars and came across the JAW wideband controllers I remembered reading about last year. These still seem like a great value, however i allready have the narrowband from the car that is still good. In that thread and a couple others, Dr. Hess mentioned tuning all of his MS'd vehicles with a narrowband. Now, it was my understanding that a narrowband's lowest reading before losing accuracy is in the 14.2/14.3 a/f range. being that most of us would, i would think, tune to something richer than that, i'm a bit confused. Dr. Hess, can you or anyone else explain a little more to me how you are getting useful info from the narrowband? is it that you tune till you're just past the rich end of the scale?

thanks for any insight!

Using a wideband is leaps and bounds faster and easier. For $200 you can get the Innovate LC-1 kit, complete with sensor, bung, etc. If you want to find a sensor at a junkyard and use an anti-fouler for a bung, you can do it for about $50 cheaper. In the end you'll have a better tune, a ton of time saved, and you can use the wideband for any of your future projects...even tuning carbs is easier with a wideband laying around! :)

FWIW, you can tune to be more rich/lean than the narrowband will read if you go open loop. You can dial in to stoich with the narrowband, then essentially calculate the amount of fuel you want to dump to go more rich/lean from there, but you have to either limit your correction factors or go open loop to keep it from automatically correcting for the rich/lean condition with a narrowband sensor/tune.

Bryce

You can tune on a narrowband but it takes a lot more time. The 'easiest' way to do it is to tune up to around 65-70kpa in closed loop with the wideband as high rpm as you can get it to work. Take that load row and copy it to any rows above (85, 100kpa for instance) and then multiply by the ratio of the tuned afr vs the afr you want to achieve. If you wanted 12:1 afr you'd multiply all your values by 14.7/12=1.225 (22.5% richer). This will get you close but depending on how quick your ve changes in those higher binsl DO NOT TUNE TO STOICH AT WOT, period, it's an easy way to lose a motor instead of buying a $200 wideband setup.

You want your O2 volts to be past .85 at all times at full load from what I've seen.

Well, I've never tuned with a WB. I'm sure that would be the real easy way to do it, but I didn't want to spend the bucks. Plus, on the bike, that would be tough to add on. That one was already like putting 8 lbs of E36 M3 in a 5 lb bag.

So, what I do to tune is start with some table that idles OK and at least runs. Start driving around with the auto-tune on. Burn the auto-generated table every few miles. Stop and look at what's been auto-generated. Smooth out the map, like if it goes 44-67-54, change that to 46-50-53, etc. Drive with the MegaTune set on the fuel map tuning tool with auto GoTo bucket on and shift up-down as you hold the throttle at various places and eyeball the AFR from the sensor. The first time, it helps if you do the tuning while someone else drives. You can get a good feel for what is too lean and too fat doing this. Sound of the exhaust, etc. Drive some more and keep logging with auto-tune off. Run the logs through one of the tools like megalogviewer or mstweak. In MLV's VE analyzer, check NB sensor and change the target map to how you want it, where you want it lean, where you want it fat, set the other parameters like temp or time to start looking at, etc. Eventually, you get to where you are lean at cruise, like with one or two at the most dots on MegaTune and fat when you step on it with all or almost all the dots lit up. I've found that as long as you have just a little signal from the O2 sensor at cruise, that is, not 0V, you are good. The places where it drops to 0 are too lean. Also, you might have to hack the MSQ with a text editor for the buckets on the target tables (RPM, TPS or VE). I couldn't figure out how to change them otherwise.

That's just how I do it.

How do you tune full load like that?

A NB sensor really only has two responses: richer than stoich and leaner than stoich.

(image taken from PLX website)

So you can see how it's difficult to get any sort of accuracy when trying to tune for anything but stoich. Given the price, a wideband is an excellent investment, especially if you're running forced induction. Naturally aspirated cars are much more forgiving and have a bigger sweet spot.

I would definetly go WB on forced induction. Too dangerous there. But, otherwise in a NA motor (my 3 MS'ed), I tune full load by logging while doing a full load run, then playing the log back and changing the map. There's a long, very steep onramp leaving work. I used to nail it going up that. That's a big full throttle run up hill, about as much load as you can put on a Rolla. Also, one or two other spots I would go WOT in 2nd or so. Redline in 3rd pushes 90, so you really can't do that in many places. Play it back and you can see what bucket you're in and where the AFR was then. Fiddle with it some, try again tomorrow. Also, keep the rev limiter set to a safe place at the beginning and after everything is OK there and you've guessed a little on the next step, then bump it up some.

One thing to think about in terms of tuning- assuming you give it pretty safe spark- don't be too aggressive on how advanced you run it- a NA motor isn't going to get hurt if you run WOT stoich, or even slightly lean. Unless you are running a complete POS engine with bad valves, seats, and pistons. I know an enigine I'm working on now is very happy running stoich with 4-5 psi of boost (I kind of have to, for other reasons).

More later...

Eric

Ok- had to move to the test site, and I have a couple of min...

Anyway, Hess has a good procedure- I like the suggestion of smoothing things out- all the work I've done has shown that enignes run with smooth and continuous calibrations. That's really important.

Darn- I've got to break again. But more again later. About going from the base cal to the desired fuel for best power cal...

Eric

ok, this is great info guys. thanks so much!

Dr. Hess, I had thought at least one of your vehicles had forced induction. Being that i'm looking at a supercharger setup, I think i'm just going to bite the bullet and go with a JAW setup. seems like I can get it up and running with a sensor for around $100 which really isn't that bad.

This will be my first forced induction vehicle so any other tuning advice is very welcome! Target is only 5psi but on a relatively high stock compression ratio (10.5:1).

Thanks again everyone!

~Josh

My Esprit is turbocharged, but that runs the factory GM EFI system. All my MS'ed vehicles are NA. While I'm sure you could tune a boosted motor this way, you would have to be very careful. 5PSI would probably be OK to work with as that's not getting too wild. But if you can get a WB for a bill, I'd go that route on a SC'ed motor.

Ok, should be final note- test is done, got my computer back...

How rich do you want to be? For the most part, and this is about 98% of the engines I've ever seen, 12-12.5:1 will make best power. Vs stoich, it's only about 5% better if only looking at fuel change, but running that rich gives you a lot of space to add spark in for performance.

And many cars I've worked with- they loose power around 11:1. So what I would do is- start with the stoich cal that Hess describes how to run- where you want it to run rich, add 10% more fuel. Test, add 5% more and repeat. When you notice the power dropping off, I would suggest averaging the 10% add and that- probably a little on the rich side, but not much harm there.

NOW you can start adding the spark, and find MBT vs. detonation.

Tuning is either one of two things: 1) very iterative, or 2) collect a HUGE amount of data, and process. OEM's generally choose 2 to start with, since we all have large buildings filled with dynos. But 1 is commonly done until the very end. Be iterative, and have a plan.

Eric

my plan is very simple. pick a few goals (car, engine, forced induction target) and use the K.I.S.S principle to get it done in time and under budget for the challenge

If this is for a challenge car, IMO the wideband would hardly be worth leaving on the car during the challenge. I'd dial it in with the wideband and run without in the challenge. If your tune is working well there's little need for the O2 for short runs like the drags and autocross. Consider that when you go WOT just about all injected cars you've ever driven ignore the O2 sensor, then consider that the amount of WOT you'll be doing in the challenge. The wideband is useful for TUNING, but not an absolute necessity for $2008 purposes. If your challenge budget is tight (when isn't it?), I'd go with the ready-for-action Innovate LC-1 to tune the car and then run without the wideband in the event. Others might say I'm crazy.

Bryce

Well, we're all crazy, aren't we

I'm pretty set on the JAW at this point and just discovered that a brand new bosch lsu4 can be purchased online from advance auto for 43.33 + shipping just need to google the 15% off online purchase discount code

Boosted the best advice is start way rich and then back it up. 9:1 a/f doesn't hurt anything but 14:1 could. Also keep timing WAY conservative until you get the fuel dialed, and creep up on that only on the dyno if you have no other experience.