Self Tuning

THE BIRDMAN

Registered User
Just curious with all the technology we have now why can they come out with a program where the car auto tunes itself. I.E you program what air/fuel ration you want and then the computyer maps it using wide band o2 to constantly stay at the program ratio. So then all you'd have to do is program a timing curve and the required ratio and your done lol. I don't see why it's not possible.



Jay
 
A closed loop system like that is only going to work if you have the correct transfer functions for all your sensors (such as the MAF). If the sensors are reporting skewed values, you'll never achieve the ratios or settings you command. Simply put...garage in=garage out.

With that being said, self tuning with the hardware/software that we have now is pretty much what you describe. The majority of the time is spent making sure that the MAF transfer function is accuarte based on airflow mods and combos on your engine. Once the MAF is right then all it is, is enter what you want to see as a ratio and closed loop and adapative control makes corrections from that.
 
It is quite possible for a car to self tune itself with the QH, and a program written to do it.... to a point. The Honda lads have Chrome that can do it for them, and requires 2 or 3 logging sessions to dial it in.

You already have fuel tables that define what your AFR is going to be, and by monitoring commanded fuel in Open Loop and compare that to a Wideband and "tweak" the MAF to bring it close. The biggest problem though is the amount of variables that have to be considered in the process, because at one RPM and load you'll see a MAF voltage reading, but change the RPM and keep the same load and you can have the same MAFV BUT a different AFR.

As for real time monitoring and adapting like closed loop does... again totally possible and another experiment I'm going to work on eventually on my car ;), when I get time to write the code for it. I'm hoping to get to a point where I dump the narrow band O2's and run widebands on both sides with the EEC full time, and not in narrow band emulation either.

Fraser
 
In the past I've used logworks to plot out a table of my acutal wideband vs. load and rpm (just how it is in the fuel tables) and then compared that to my manage air/fuel. I'm sure something can easily be written in that fashion where it just compares the cell values and adjusts the MAF curve by a percentage.
 
My AEM has autotune capability, I know Megasquirt has a utility that will alter the map based on desired A/F ratios and a couple logs. Same goes for Autronic and believe Haltech and Motec ($$$$). Not sure about the other standalones.
 
Yeah just curious thats all. I figure with all the advances in computers if recent years someone would have developed a true plug and play system that any idiot could use and have almost a perfect tune everytime lol.



Jay
 
As for real time monitoring and adapting like closed loop does... again totally possible and another experiment I'm going to work on eventually on my car ;), when I get time to write the code for it. I'm hoping to get to a point where I dump the narrow band O2's and run widebands on both sides with the EEC full time, and not in narrow band emulation either.

Fraser

Can you actually change the stock computer "Strategy" then? I've thought that there might be improvements that could be made to the strategy of the EEC-IV. My idea was to change the knock strategy to richen the fuel mixture instead of pulling, or maybe do both. My thought for this is that all the datalogs I have seen on both my car and others cars is that every time the EEC pulls timing, it pulls a full 8 degrees and seems to stay there. It doesn't seem like it can just pull 4 degrees or 2 degrees to get rid of the knock, which makes me believe that pulling timing isn't really elimating the detonation, but richening the AFR, might...
 
If I wanted to I can completely rewrite all the code that is in the EEC, and I may even do that for the fun of it and create a race only setup with none of the emmisions or close loop stuff in it.

The knock strategy on our cars is actually very amazing. You are talking about it pulling 8 degrees of timing, and your right but its actually a gradual pulling per engine cycle with a maximum of 8 degrees retard. You have to take into account everything that is going on and not just one component.

High ACT's will pull timing as well and that starts at 120F and will pull up to 8 degrees of timing as well by 180 or so on a stock tune. The only way to get that timing back is to cool down the ACT. Now high ACT's will cause detonation as well which will trigger the knock sensor which will pull its timing amount per cycle. On my car the knock sensor started kicking in at about 142F, so not only was I getting hammered with act spark pull, I was getting it from the knock sensor and by the end of a 1/4 mile I was seeing about 12 degrees being pulled for an overall timing of 14 degrees. That was with 5% OD on a stock 94 blower with a single IC with a fan running 91 octane fuel (see why I discourage folks from OD'd their blower so much ;) ).

Now think of it this way. In order for the spark retard to cancel itself with the knock sensor, we have to get rid of the source of detonation which would mean we need to suddenly lower our ACT...... problem here, thats not something that happens right away normally when your racing down the 1/4.

Will fuel enrichment help? I would venture to say yes, BUT how much fuel enrichment would be needed? We're already looking at 11.7 typically for our cars at WOT.... do we need a 10.0 AFR to cancel the detonation caused by 150F ACT temps at say 24 degrees of timing? How much power do we lose with an AFR and timing like that? How much power do we lose when 8 degrees of timing is pulled running an afr of 11.7? Oh to only have unlimited time on a dyno LOL!

You need to look at the whole picture, and with the QH and a wideband and GOBS of datalogging you can start to see what is really going on and optimize your tune to your car, and what you expect from the car. I'm at the point now I'll be running 2 tunes soon. One for street driving and getting best performance for the lllloooonnnggggg runs to keep the least impact of power loss from spark pulling and AFR and another tune to optimize the car for the short runs like the 1/4 mile.

So what we need to do is to do everything to prevent detonation from happening in the first place. Things like better IC's to keep the ACT's down longer to prevent detonation. Getting tunes done to optimize your motor to fit what you have. If your really after the best performance, then getting the tools to find out the limits of your motor and to allow you to experiment with settings in the tune to get the most for what your trying to do with it, and now we have access to those tools.

Its funny really, I was told a little while ago that tuning a car is nothing and very simple to do. A tune can make or break a motor. This same individual is going to have someone at a dyno shop tune his car in an hour..... yup, he's going to get max performance ;).

Sorry for the length.

Fraser
 
I been told from some of the top tuners. Its cake once you understand the logic. Rewriting code however is something all of its own...And for me..I do neither!
 
I been told from some of the top tuners. Its cake once you understand the logic. Rewriting code however is something all of its own...And for me..I do neither!

Unfortunately, there is very few tuners out there that really understand the logic, and most of what they do is a "guessing" game and ends up with fudged numbers to beat the tune into behaving ;). I've seen some interesting solutions to some problems.

Not that I know it all, but I know alot of the logic but am stilling learning how to apply it :eek:.

Fraser
 
Yea, there is a lot to tuning and in no way do I consider it simple. If my goal was just horsepower, I would have had my car dyno tuned by an experience SC tuner a long time ago and never messed with it from there, but to me, the learning experience is the most fun part of all of it.

I really don't know if fuel enrichment at knock would stop the knock faster than pulling timing. From what I've seen (which is pretty limited) whenever someone's knock sensor starts pulling timing, it pulls all 8 degrees. That makes me think that the strategy of pulling timing on an SC isn't really helping. Could I be completely wrong about that? Yes, definitely. For me, with running half E85, I would think that richening up the mixture would really help fight detonation because E85 helps lower the air temp and it can run pretty rich. But with half E85, I don't detonate anyways...And I can't go richer because my injectors are maxed out... It was more of a theoretical question because I would love to be able to change how the computer actually acts, rather than just the parameters as we do now. I'm so far from knowing how to do that now, but it's fun to dream about.

I know on my last drag strip run (stock blower, stock pulley, stock IC w/o fan) I was at 120 degrees before I even started the run and that was on a cool day. Since then I have added a 10% pulley, Magnaport SC and an IC fan and my temperatures actually seem lower.

I'm already developing a race only tune for my car. Basically a tune that keeps the high speed fan on at all times to keep the car cool in the lane, adaptive off, slightly richer mixture, 1st gear truncation off, etc...And I have a seperate tune for pure 93 octane. And another tune for half E85/half 93. I think I need to get the switch from moates for the QH so I don't have to have my laptop with me all the time...

Thanks again for the help.
 
Back
Top