I believe I heard there may be some yet unreleased software that supports autotuning for the SD '730. I know of several people who would be interested in such a thing, but I haven't found any threads by searching...

Anyone have any idea if I'm dreaming, or what software this might be? Thanks in advance,

 

It is a Unicorn. Always will be. That is why you can't find anything searching.

 

DFI has a similar routine called Autocal....

 

VEMaster with DM log files are about as close as you'll get to anything like that. VE tables only (limited as well)
Not exactly autotune either. Just automates the process of the VE tuning a bit.

 

All the hardware is in place. Software integration is coming close. Just a matter of driving it the rest of the way.

Honda guys have had this stuff for years now. Why can't we?

 

Got any good info on the process and/parameters used?
Might as well use this thread to get the thought processes moving
Obviously a VE adjuster is in order but how to compensate for spark?
Jump in with theory on how it could be done for an all around tune.

 

I was just referring to a VE tuner... I guess it's assumed that one knows what kind of spark table would work best with their setup, or can make an educated guess...

I wouldn't trust a computer to tuning my spark table when my engine is at the mercy of a working knock sensor.

 

What is a good tune anyway? Most of the import drivers I know have no idea what that is. It is all about getting it the way you like it. How can a computer program know that......fuzzy logic?
True, they maybe getting it close for the average driver, but there is more to be had.
Whenever I have tuning problems I consult the experts.....Grumpy, Rbob, and others at thirdgen.org. Maybe the Honda driver's don't have people like that.

 

I've looked at the way various combinations respond on the dyno and on the street, and in general, I'd have to say that spark can be specified 'ad-hoc' so long as knock isn't encountered. So, I'd agree that spark tuning doesn't lend itself to auto-tuning per se.

On the same note, from what I've seen, once you get some experience on this stuff, you can just about go through and say 'I want to have Z air fuel ratio at X rpm and Y map'. You'll be specifying from 12.5:1 to 15:1 for N/A across PT and WOT, and down to 10-11:1 for boosted apps at WOT. Within a +/- 0.4 AFR or so, you won't see any appreciable difference in terms of torque. And that's what we're tuning for, max torque at a given RPM/load(map). That's what the reflection feedback is on combustion efficiency, right? So if you get it in the 'zone', it'll be fairly robust. Particularly if you set it conservatively.

So given that you have a datalog which has wideband data imbedded within an array of RPM and MAP data, and if you have a realtime emulator running with which you can make changes to the VE table to achieve a 'target' AFR throught he RPM/MAP space, where is the challenge?

What the engine wants isn't some big mystery. It's not a philosophy, and doesn't require a PhD to 'nail it'. It's just fuel, air, and spark. We're burning hydrocarbons, and there's a limited number of ***** we have and a limited number of feedback variables. So long as the recipe you provide gets it in the zone, you'll be one step closer to correct. I'm not saying that auto-tuning would be the end-all be-all which provides license to bypass rational thought. On the other hand, it can save a significant amount of time in 'roughing in' a tune, especially for initial VE tuning.

That's my thoughts on the subject. Make of it what y'all will. I think there's some value in the approach, and would like to see it reduced to practice in short order.

 

What you said is totally right. I have read a lot about people trying to give it all the spark they can and it that is not a good thing. It makes zero HP gain and the engine more sensitive to bad gas. The heads, quench, and camshaft define how sensitive it is to spark changes. Most are not that sensative.

The DIY-WBO2 is one of the best things I ever got for the car. It changed the way I do AFR and now I can't tune without one. I never tried the VE program but read good things about it.

The *hard* part is getting the decel fuel cut-off, accel., idle, and part throttle (if worried about MPG) right. This is where I spend all of my time tuning. Most of the time I am not really tuning, but just trying to find what I like.

 

FWIW, I've looked at all the Hondo, tuning stuff, and I'm not too impressed. Things like AE are *hiden*, and evasive to work with.

Threshold knock timing, will get you close enough, that if you do a **quick** too much timing pass, by just a few degress, and then subtract 2-3, you'll be close, at least for the quality of gas available at the pumps today.

One of the better options, IMO, is now tuning with 89, and then going to 91 or whatever premium you have locally. Reguardless, what you do, you now have to leave something on the table, to allow for the gas.

If you want real self tuning, then you need to have some sort of feedback, for sensing torque, and then *learn* the timing and fuel corrections, based on it.

 

Yes, I recently pulled a couple of degrees out of my tune to allow for fuel quality. I didn't think I'd feel it, but I did (33 to 29 degrees). I wish all fuel was high octane and that tetraethyl lead was still legal (sigh).

I completely agree regarding the transitional tuning. Wide-open and easy-transition part-throttle can be addressed using calculative algorithm quite easily using WB as a feedback. However, getting things like crank fuel, startup enrichment, tip-in, decel enleanment, and those sorts of things takes the 'human touch' for sure. There's quite a bit of 'SOTP' feedback. Maybe we need a sensor wired up down there ;^).

We had an engine on one of those engine stand marine dynos the other day, pretty neat. Just dial in the MAP and RPM you want to hold to, and then start tweaking timing and fueling. Watch the torque and try to maximize it by tweaking the ECM. Then on to the next point in the table. Kinda neat, but you get nervous when you're doing it at 7000 rpm for extended duration, marching up from low to high load. I thought it'd be neat to have the torque readings fed back to the laptop and coupled with the RPM/MAP/knock/WB readings and then to set up a program that would automate the 'march-through' and actually control the dyno as well. The guy who runs it is a bit of a homebrew techie who's willing/able to try that sort of thing, so we'll probably put something together. Pretty slick.

 

FWIW, when I am tuning at the track, I play with PE timing in 2* increments. I also play with PE air/fuel ration in .2 increments. Eventually I will go the highest MPH. BTW, the wideband greatly speeds the process, as does part throttle BLM tuning.