|
|
|
|
Thread Tools | Display Modes |
02-18-2004, 12:48 PM | #1 |
Scooby Specialist
Member#: 24472
Join Date: Sep 2002
Chapter/Region:
MAIC
Location: Marshall, Virginia
Vehicle:2001 Miata White |
Why the heck can't UTEC tuning be automated?
I can't understand why tuning has to be such a manual process.
How does a tuner tune? He looks at the feedback from the various sensors connected to ECU, does some form of analysis in his head and makes changes to the input maps. Why can't that same analysis be performed by a computer program?
* Registered users of the site do not see these ads.
|
02-18-2004, 01:04 PM | #2 |
Scooby Specialist
Member#: 7138
Join Date: Jun 2001
Chapter/Region:
MAIC
Location: MD
|
Wow your analysis of feedback control is amazing. I mean why would they even have an entire field dedicated to the subject since its 'so easy'. Just kidding . Yes, it may seem easy to you, but the algorithms are the big deal. Who is going to write them? How aggresive can they be without destroying engines? How do you weight the variables (sensors input)? Its not an easy piece of software to just come up with (I know Ive worked on traction control algorithms). I'm sure that TurboXS is eventually going to work on a self tuning version, but I wouldnt get antsy. If you want self tuning now you are going to need a TEC 3 or a MoTec or something at that level, and even those 'self tunes' arent completely optimal. Sorry if I was harsh - I just hate when the MechEs on my team assume it's 'easy'.
- Chris |
02-18-2004, 01:05 PM | #3 | |
Scooby Newbie
Member#: 34060
Join Date: Mar 2003
|
Re: Why the heck can't UTEC tuning be automated?
Quote:
Cost is the answer. Basically .. the stock ECU is very "smart" due to the tremendous efforts afforded by large fiscal backing (FHI in this case). Engineers were able to test many different situations, develope many algorithms, and do real life testing with many different engines, to have an ECU tailer made to their specific needs. |
|
02-18-2004, 01:20 PM | #4 | |
Scooby Specialist
Member#: 24472
Join Date: Sep 2002
Chapter/Region:
MAIC
Location: Marshall, Virginia
Vehicle:2001 Miata White |
Quote:
Everything you stated either is included in the tuner's analysis or it is ignored. The end state is that he changes a value in a map. There is a specific set of decisions he made to arrive at that value, even if he's only 'guessing'. I'd be satisified if I could take some log output and my current map, pass it to a program and have the program spit out a new map. Sure, there might be some additional parameters to consider. But it's still possible. Damn, I wish I understood this stuff. I'd do it myself. |
|
02-18-2004, 01:24 PM | #5 | |
Scooby Specialist
Member#: 24472
Join Date: Sep 2002
Chapter/Region:
MAIC
Location: Marshall, Virginia
Vehicle:2001 Miata White |
Re: Re: Why the heck can't UTEC tuning be automated?
Quote:
Last edited by ITWRX4ME; 02-18-2004 at 01:38 PM. |
|
02-18-2004, 02:25 PM | #6 |
Scooby Specialist
Member#: 3841
Join Date: Jan 2001
Location: Sacramento, CA
Vehicle:2004 WRX Wagon UTEC+VF22+WI |
You're absolutely right, but what you're missing is the economy of scale. How many tunes do you have to do during the product's lifecycle until you beat the cost of doing it by hand - especially considering you'd still have to spend money on dyno time to get that data to feed into your Tune-O-Matic? I'd wager more than a specialty market like this could support. Hence, no product exists.
Your flawed assumption is that it's cheap to create software. |
02-18-2004, 02:45 PM | #7 |
Scooby Specialist
Member#: 24472
Join Date: Sep 2002
Chapter/Region:
MAIC
Location: Marshall, Virginia
Vehicle:2001 Miata White |
I get the whole economy of scale thing. Believe me, I do. I'm a programmer. I know how much it costs to create software. I do it every day.
Like I said, I'd do it myself but I lack the skills in the tuning half of the equation. |
02-18-2004, 02:52 PM | #8 | |
Scooby Newbie
Member#: 34060
Join Date: Mar 2003
|
Re: Re: Re: Why the heck can't UTEC tuning be automated?
Quote:
Aftermarket tuners analyze the the data and then GUESS at what to do next. If it doesn't work they GUESS at something else. |
|
02-18-2004, 03:03 PM | #9 |
Scooby Specialist
Member#: 33410
Join Date: Feb 2003
Chapter/Region:
Tri-State
Location: wishing i was somewhere else
Vehicle:08 wrx dgm |
it's a lot more to do with guessing though. If I brought my car to a tuner and he was *guessing* about what timing to run/etc, I'd kick his ass and take my car somewhere else.
The fact is, it's not just a set limit of parameters to adjust to get the most power out of a car, reliably. It has a lot to do with experience and having a nack for it. Each engine is different even though they came off an assembly line, each has a small difference in what it can handle before it fails. Having tuned a large amount of cars, knowing AND FULLY understanding how that particular engine works is what makes a "tune-o-matic" a pipe dream. hence the reason you would trust vishnu,turbo-xs,cobb before you'd take it to bob down the street. |
02-18-2004, 03:11 PM | #10 |
Scooby Specialist
Member#: 36528
Join Date: May 2003
Location: florida
Vehicle:2002 Impreza Wrx Silver |
UTEC=User Tunable Engine Computer
If the utec did everything, it would be called something different. It's nice to want something to be easier, but it's messed up to want something to be done for you when the whole idea is that you do it yourself. Turboxs is trying to make tuning with the utec as easy as possible, and with firmware updates this will happen by adding different features and what not. Maybe some day there will be fields in the boost map to put in desired psi or bar, max/min timing for different zones in the timing map, and a/f ratios for different zones in the fuel map, but I don't think we'll ever be able to just plug the utec in and go cause that would defeat the point of what it is. It's a utec, no a unichip...... peace |
02-18-2004, 04:34 PM | #11 |
Scooby Specialist
Member#: 32576
Join Date: Feb 2003
Location: Stevens, PA
Vehicle:04 STI |
The AEM stand alone has this feature once you install a wide band. BUT, I think it only controlls A/F ratios in closed loop. Jr of Turbotrix is using this on his Supra. Haven't seen this yet on Subies, but AEM claims they are releasing their stand alone for the WRX and STI. Now when is this going to happen
|
02-18-2004, 06:52 PM | #12 |
Scooby Specialist
Member#: 9237
Join Date: Aug 2001
Location: Youngstown, Ohio
Vehicle:2008 Saab Turbo X 400+ lbft |
yeah, automatic fueling would be easy for them to do, automatic tuning, not so easy, however software upgrades with the TUNA would be easy to do, ie, tell the UTEC you want an AFR of 10.4 under load, and the utec/tuna watches the wideband and adds or removes fuel to keep it at the desired AFR, im sure that would be relativly easy to impliment
|
02-18-2004, 07:45 PM | #13 | |
Scooby Specialist
Member#: 29735
Join Date: Nov 2002
Chapter/Region:
MAIC
Location: Westminster,MD
Vehicle:2011 JSW TDI - DSG 2008 EVO X MR |
Quote:
|
|
02-18-2004, 07:58 PM | #14 |
*** Banned ***
Member#: 2992
Join Date: Nov 2000
Chapter/Region:
NESIC
Location: Connecticut, USA
Vehicle:02 WRX Sedan Silver |
The first problem is that there isn't any sensor that reports the torque produced by a cylinder after it fires. If there were, then any sane ECU would:
1) adjust A/F for that cylinder to maximize its torque contribution 2) adjust ignition advance for that cylinder to maximize its torque contribution 3) increase or reduce boost in order to maximeze the average among all cylinders. The second problem, of course, is recognizing knock when it occurs. I think most understand how hard it is to reliably recognize knock based solely on information from a block-mounted knock sensor (accelerometer). In-cylinder pressure sensors are one answer. Ionization feedback about cylinder pressure from the spark plug gap is another. These folks offer a solution based on the latter, http://www.adrenalineresearch.com, but it's pricey. |
02-18-2004, 08:28 PM | #15 |
Scooby Specialist
Member#: 9861
Join Date: Sep 2001
Chapter/Region:
RMIC
Location: Colorado Springs, CO
Vehicle:1999 2.35 RS-T w/NOS Hybrid Rally Blue Pearl |
HAHA, only $22k
|
02-19-2004, 08:58 AM | #16 |
Scooby Specialist
Member#: 24472
Join Date: Sep 2002
Chapter/Region:
MAIC
Location: Marshall, Virginia
Vehicle:2001 Miata White |
Guys, I was NOT talking about realtime tuning. I was talking about something that reads the logs from the UTEC interprets the data and makes corrections to a NEW map. That map will then be reloaded into the UTEC.
Does a tuner change map settings while the car is on the dyno and the accelerator is floored in fourth gear? No. He looks at the readings on the UTEC dashboard or the logs to see what's going on. He looks at readings on the dyno. Then he does some mental gymnastics and modifies the map. Then he does another dyno run to see if it worked. The thing I neglected to add here is that I'm coming from the perspective of someone that's just trying to keep my engine from self destructing. I'd be happy with map adjustments that keep my engine from knocking under added boost levels. Taking it to the next level, where you're trying to squeeze out as much performance as possible, definitely requires measurements that are beyond the scope of just the engine sensors. My original post was out of personal frustration. I can't floor my accelerator because I get knock events and the UTEC pulls timing. I don't understand things well enough to make adjustments on my own. I was thinking, "Gee. Wouldn't it be nice if there was a program that could 'look' at my logs, interpret the data, and make appropriate map adjustments so that the knock events go away". That's all. |
02-19-2004, 09:59 AM | #17 |
Scooby Guru
Member#: 7922
Join Date: Jun 2001
Chapter/Region:
MWSOC
Location: Tucson
Vehicle:04 CTS-V b/b/b 99 GMC Highrider |
The things you are bitching about are only because you lack the knowledge for a product that you bought. Why don't you just tune it yourself to get rid of knock like everyone else? There is no such thing as middle of the road automatic tuning. It either none (like the Utec currently) or all out like on a Pectel or Autronic.
p.s. The AEM is a sweet unit. 2 guys here in Detroit have them on their WRXs and it is amazing. It runs MAP based instead of MAF and the fuel control is sick. |
02-19-2004, 10:43 AM | #18 | |
Scooby Specialist
Member#: 24472
Join Date: Sep 2002
Chapter/Region:
MAIC
Location: Marshall, Virginia
Vehicle:2001 Miata White |
Quote:
And BTW, where did I bitch about anything? I'm just asking questions and putting forth an idea or two. |
|
02-19-2004, 01:51 PM | #19 |
Scooby Newbie
Member#: 2678
Join Date: Oct 2000
Chapter/Region:
SCIC
Location: San Luis Obispo, CA 93401
Vehicle:2004 Black WRX STi (wingless) |
It seems like a viable idea to me too. But I'm likely a dumba$$ too...
|
02-19-2004, 01:52 PM | #20 |
Scooby Guru
Member#: 8512
Join Date: Jul 2001
Chapter/Region:
South East
Location: St. Pete, FL
Vehicle:2002 WRX chassis... stage-infinity.com |
It's definitely possible.
I didn't take it as bitching, but your answer is above. It's a big undertaking to do safely. |
02-19-2004, 02:38 PM | #21 | ||
Scooby Specialist
Member#: 24472
Join Date: Sep 2002
Chapter/Region:
MAIC
Location: Marshall, Virginia
Vehicle:2001 Miata White |
Quote:
Quote:
But what about the majority of end users out there that feel confident enough to modify their maps, yet don't have access to a dyno? They're getting the job done somehow. Aside from feeling more or less punch in the seat of their pants, what information do they use to determine whether to change a map? I'll continue to read as much as I can and try to understand. Who knows? Maybe it'll all come together and I'll arrive at the answers I'm looking for. 'Til then, or until I can afford the time and money for a dyno tune, it's back to map 0. |
||
02-19-2004, 02:40 PM | #22 |
Scooby Guru
Member#: 7977
Join Date: Jul 2001
Chapter/Region:
MAIC
Location: Sterling, VA
Vehicle:100% Abuse |
Just stick to your horses, there, sparky!
|
02-19-2004, 02:52 PM | #23 |
Scooby Specialist
Member#: 24472
Join Date: Sep 2002
Chapter/Region:
MAIC
Location: Marshall, Virginia
Vehicle:2001 Miata White |
Who lifted the rock and let you out?
|
02-19-2004, 03:49 PM | #24 |
Scooby Guru
Member#: 6124
Join Date: Apr 2001
Chapter/Region:
MWSOC
Location: Glendale Hts, IL, USA
Vehicle:2000 NFR AP1 S2000 '07 Honda FIT sport (5MT) |
Tuning is not hard. There are some simple rules one needs to understand about why a motor works and makes power. Then there are simple things to understand about the subtle changes that a EMS system will impose on power production.
I.E. What -.02 in a fuel column will really do. That's done by going slow, with the right tools, with trial and error. But it's defineately something that's doable for all but those who lack the desire to learn. If I were you. I'd do this. Load up a base map, pull out 3 points of timing around the map, then start logging and learning. The best way to learn, is to understand why the engine wants to run timing like it does. Learning how different conditions will affect the timing you can, and will run. Looking at AFR's with a good wideband meter (a wonderful investment for any person wanting to tune). It really isn't hard. You just have to understand (to quote Clark on this one) that fuel and timing aren't knobs for power and safety. There's one correct timing curve, and one correct fuel curve for the mods/fuel you run. I'd do logs, then post them up and ask questions about stuff you don't understand and slowly learn. Jorge (RiftsWRX) www.ProjectWRX.com |
02-20-2004, 02:17 AM | #25 |
Scooby Specialist
Member#: 16779
Join Date: Mar 2002
Chapter/Region:
NESIC
Location: Yarmouth, Maine
Vehicle:2021 Toyota 4Runner Black |
Some of you guys are missing his point. He's not wanting to learn to tune, he's just posing the question about automated tuning.
Any numbnuts can learn to tune given the time and desire. (no offense intended there to anyone). I don't see why an algorithm couldn't be written to take into account specific sensor inputs and adjust accordingly to a predetermined "safe" maximum. It's not like asking a computer to recognize a human face, it seems like relatively simple feedback. Why hasn't anyone done it? The combination of a really smart programmer that knows a WHOLE lot about internal combustion engines doesn't occur too often. And the testing necessary would be insane. But it could be done. No doubt. |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Why the Heck Won't.... | OriginalSin | Off-Topic | 21 | 07-19-2006 04:00 PM |
Cobb Stage 2, stock BPV, why the heck am I getting compressor surge? | ElBartoRex | Factory 2.0L Turbo Powertrain (EJ Series Factory 2.0L Turbo) | 11 | 05-10-2006 12:36 PM |
Why the heck are kids wussies nowadays. | Vostok 7 | Off-Topic | 23 | 06-07-2004 12:17 PM |
so why the heck was nabisco down all last night? | blue-sun | Mid West Subaru Owners Club Forum -- MWSOC | 10 | 03-21-2004 12:07 AM |
('93-'01) Why the Heck Can't I Sell My Car? | jzinckgra | Impreza Forum | 27 | 04-16-2001 09:47 PM |