![]() |
|
|
|
Moderator
Join Date: Dec 2001
Posts: 9,569
|
Automatic tuning by EFI system possible?
Is there any EFI system on the market sophisticated enough to tune itself for best power?
Suppose you built an engine and wanted to set it up on the dyno. Working from known precedents you would try to optimize the fuel map under a wide variety of loads and conditions, likewise the ignition map. This takes a long time and is one of the biggest costs of an EFI system-- the setup and tuning. Is there any EFI system that can start out with a base map and adjust the settings dynamically as you go? Perhaps with a test program that would incorporate accelerometer data? I can imgine a test routine that would call for you to do several 5-45 pulls, 30-70 pulls, 55-80 pulls (subject, as ever to speed limits and road conditions) and be able to perform iterations within those pulls to figure out what combination gives the best longitudinal G. I wonder if MOTEC can do this? Seems to me that computing power has increased way beyond the original EFI systems, maybe this is possible?
__________________
'66 911 #304065 Irischgruen ‘96 993 Carrera 2 Polarsilber '81 R65 Ex-'71 911 PCA C-Stock Club Racer #806 (Sold 5/15/13) Ex-'88 Carrera (Sold 3/29/02) Ex-'91 Carrera 2 Cabriolet (Sold 8/20/04) Ex-'89 944 Turbo S (Sold 8/21/20) |
||
![]() |
|
Registered
|
Several system (Autronic, Adaptronic) has "auto tune" functions, at least for fuel mapping, but they tune by wideband lambda.
__________________
Magnus 911 Silver Targa -77, 3.2 -84 with custom ITBs and EFI. 911T Coupe -69, 3.6, G50, "RSR", track day. 924 -79 Rat Rod EFI/Turbo 375whp@1.85bar. 931 -79 under total restoration. |
||
![]() |
|
Registered
Join Date: Jul 2001
Location: Portland Oregon
Posts: 7,007
|
None of the current "autotune" features work well enough to be viable, IMHO.
At this juncture, its hard to beat an experienced tuner who really knows engines and how to drive the software,.. ![]() ![]()
__________________
Steve Weiner Rennsport Systems Portland Oregon (503) 244-0990 porsche@rennsportsystems.com www.rennsportsystems.com |
||
![]() |
|
Registered
Join Date: Feb 2003
Location: chicago
Posts: 1,077
|
good question, i've wondered how the aftermarket efi systems sold for american V8s are so plug and play. holley, edelbrock etc. offer these systems and i'm wondering if they can be adapted to our engines. i know that the V8 is a different animal but there is quite a range of engines that these systems need to work for right out of the box. i think the high end race systems from motec etc.. are designed for total control by someone who wants to see everything and will be able to understand how to tune it themselves. i think most get wrapped up in tuning to perfection, i would like something simple that is better than carburators and mfi but not so complex that every couple years it needs upgrading to the latest software. it would seem to me that with a throttle position sensor and a tempature sensor a simple analog controller could pulse a set of correctly sized injectors the right amount.
__________________
BMW 128i 73 rsr clone - sold 68 912 project to become 911r (almost done!) |
||
![]() |
|
Registered
Join Date: Sep 2007
Posts: 500
|
I do recall seeing something about a system whereby the engine torque is continuously monitored and the control algorithm would optimise the base data set for torque
I think torque measurement was non contact (optical or hall trigger) based on angular displacement of a transmission shaft ("twist" or "wind up" of the shaft) System was definitely not aftermarket, think F1/developmental etc. A search of SAE papers would probably turn something up. - I'll have a look. |
||
![]() |
|
Registered
Join Date: Sep 2007
Posts: 500
|
Here's a start.....not quite aftermarket, but the technology definitely exists.
Development of a Magnetoelastic Torque Sensor for Formula 1 and Champ Car Racing Applications Engine Monitoring of a Formula 1 Racing Car Based on Direct Torque Measurement "With access to the actual engine output a range of applications can be performed in terms of engine and drive train monitoring. An engine map can be acquired directly on the track and the performance of gearshifts can be measured and optimized. The torque signal can also be used for efficient traction control. " From the above reference John Last edited by jcge; 09-16-2008 at 03:15 PM.. Reason: more links |
||
![]() |
|
![]() |
Registered
Join Date: Sep 2007
Posts: 500
|
And a bit off topic, but this one is just incredible (in my opinion anyway)...a very clever way of checking out the opposition.
Determining engine speed, gearing, shift points, aerodynamics etc just by listening to the engine !!!! Reconstruction of Formula 1 Engine Instantaneous Speed By Acoustic Emission Analysis John Last edited by jcge; 09-16-2008 at 03:08 PM.. |
||
![]() |
|
Registered
Join Date: Sep 2007
Posts: 500
|
Back on topic - this is becoming a commercial reality
ABB's Torductor torque measurement technology "With Torductor®-S it is possible to obtain and maintain an actual map of the engine at all times." Bosch are all over this technology as well - almost certain to be integrated into future ECU's......but I would expect optimisation around efficiency and emissions (not necessarily torque..) Bosch US Patent 4976160 Daimler Chrysler have also trialled torque measurement/shift control systems on thier 45RFE automatic transmission (production car transmission) John Last edited by jcge; 09-17-2008 at 03:58 PM.. Reason: Added ref to Chrysler |
||
![]() |
|
Registered
Join Date: Mar 2003
Location: Oklahoma
Posts: 2,391
|
You are in NY? If so you have one of the best tuners in the country up there at Turbo People.
|
||
![]() |
|
Registered
Join Date: Sep 2005
Location: Northern Virginia
Posts: 1,792
|
Cool stuff.
Quote:
![]() Brian
__________________
'82 SC Targa '83 SC Cabriolet |
||
![]() |
|
Try not, Do or Do not
|
John This seems like an inappropriate topic for the engine rebuild board. Perhaps the tech board would be more appropriate. Just a matter of decorum.
__________________
Henry Schmidt SUPERTEC PERFORMANCE Ph: 760-728-3062 Email: supertec1@earthlink.net Last edited by Henry Schmidt; 09-27-2008 at 09:20 AM.. |
||
![]() |
|
Wer bremst verliert
Join Date: Jan 2005
Location: Toronto, Ontario
Posts: 4,767
|
As mentioned above Adaptronic does it, and as Steve points out you really need some dyno time and either an expert or make yourself an expert to get the last 10% out of the motor.
The autotune function does a good job getting a baseline on the fuel side. It doesnt function as effectively at spark unless you utilize knock sensing it has no way of tuning for power, just with the goal of adjusting fuel to achieve your nominated A/F goal. You need to play with spark angle on a dyno using det-cans or other knock sensing, adjusting on each pull until you see power dropping off then dialing back.
__________________
2007 911 Turbo - Not a toy 1985 911 Cab - Wife's toy 1982 911 3.2 Indiash Rot Track Supercharged track toy 1978 911 3.0 Lichtbau toy "Gretchen" 1971 911 Targa S backroad toy |
||
![]() |
|
user & abuser
|
with very little experience in the tuning world, it would seem to me like the programing required for this would not be overly difficult.
seems like running a check of a/f amd super small rpm incriminates, and have the computer make adjustments 'on the fly' would allow for some improvements. i know this isn't power maximizing, but if your afr's are spot on are you not alot closer to max power? edit: above post says it is possible, and it works about 90%... pretty cool.
__________________
vini vidi vici |
||
![]() |
|
Air Medal or two
Join Date: Jul 2003
Location: cross roads
Posts: 14,074
|
A wise man once said "When someone says it "can't" be done it just means someone else will."
Not that anyone has said that here.But rest assured this product I am sure is just around the corner as it seems only a step in computer evolution..JMO
__________________
D troop 3/5 Air Cav,( Bastard CAV) and 162 Assult Helicopter Co- (Vultures) South of Saigon, U Minh Forest, Delta, and all parts in between |
||
![]() |
|
Air Medal or two
Join Date: Jul 2003
Location: cross roads
Posts: 14,074
|
A wise man once said "When someone says it "can't" be done it just means someone else will."
Not that anyone has said that here. But rest assured this product I am sure is just around the corner , as it seems only a step in computer evolution..JMO
__________________
D troop 3/5 Air Cav,( Bastard CAV) and 162 Assult Helicopter Co- (Vultures) South of Saigon, U Minh Forest, Delta, and all parts in between |
||
![]() |
|
Wer bremst verliert
Join Date: Jan 2005
Location: Toronto, Ontario
Posts: 4,767
|
Most of our goals here would be for maximum, controllable power. In order to go beyond just gross fuel adjustments one of the torque-sensing items described above and knock-sensing would be necessary. From there on out likely a matter of programming, learning cycles and patience.
__________________
2007 911 Turbo - Not a toy 1985 911 Cab - Wife's toy 1982 911 3.2 Indiash Rot Track Supercharged track toy 1978 911 3.0 Lichtbau toy "Gretchen" 1971 911 Targa S backroad toy |
||
![]() |
|
Registered
Join Date: Oct 2001
Location: Momence, IL 60954
Posts: 1,911
|
The Autronic's autotune most definately works given you have no intake or exhaust leaks and you have to realize that there are areas you have to lock to keep it from autotuning, like idle, WOT, etc. It's good for getting a rough map when trying to tune with the engine in the car, on the road. Not ideal conditions, but that's where I've used it.
__________________
Charles Navarro President, LN Engineering and Bilt Racing Service http://www.LNengineering.com Home of Nickies, IMS Retrofit, and IMS Solution |
||
![]() |
|
Registered
|
The Autronic Autotune is not a bad system. Shame you can no longer buy these boxes as they were some of the best bang for your buck out there. There are a few ways you can use the Autronic autotune and some other ways a computer can tune a motor by itself. The Autronic worked by using wideband lambda sensors and self adjusting fuel and ignition to meet a "target" A/F ratio that the user would input for a given load. Very neat and simple. There are other algorithms that use a host of different inputs to self tune a motor. Lambda, TPS, air temp, exhaust temp, injector duty cycle, projected hp and tq output, these are from mainly higher end and custom one off software. I have seen one program which measures axle twist as a means of tuning. At the end of the day, nothing beats finding a tuner that is competent with the system you are employing and spending some time on an engine dyno or chassis dyno.
__________________
74 911s neverending story. two feet and a jetta for now. |
||
![]() |
|
Registered
|
I am surprised that no one has mentioned Megasquirt yet. While I am far away from being an experienced engine tuner with the MS-II software I was able to get a rather difficult engine to run more or less properly:
We started with a CB Performance turbocharged T1 VW engine with fuel injection and ITB's. It originally came with some sort of Delco GM ECU and a funky tuning console. In essence the tuning console was a "faking the sensor" type of adjustment for the mixture. We decided to yank the crap and start over with a MS-II system since all the sensors and the fuel injection was already there. We just needed something more flexible to control it properly. The auto-tune feature based on a wideband LSU 4.2 O2 sensor and a target A/F part throttle map was used to generate a first driveable part throttle fuel map. For the ignition we used good old manual labor. But if you implement knock-retart you could even get this semi-automated. There was a hole lot more work and some hand tuning involved but it was impressive how much can be done with an open-source system. By selecting smart target A/F ratio's for cruise and regions where power is required we got a pretty nice running package. Most of the issues I encountered were stupidity and lack of experience. I can't count how many times I tried to "tune away" trouble spots and no-start problems that turned out to be mechanic failures (air leaks, no compression, stretched head studs, poor wiring, poor MAP signal, etc) just to name a few..... Still, in my opinion there is hardly any feature that MS-II is lacking. Maybe its moderate price makes it appear to be an inferior system compared to the more established engine management solutions out there. But as soon as you use high-end components for the connectors and wiring I don't think it needs to hide behind a Bosch Motorsport custom ECU that costs thousands of $. Ingo
__________________
1974 Targa 3.6, 2001 C4 (sold), 2019 GT3RS, 2000 ML430 I repair/rebuild Bosch CDI Boxes and Porsche Motronic DMEs Porsche "Hammer" or Porsche PST2, PIWIS III - I can help!! How about a NoBadDays DualChip for 964 or '95 993 |
||
![]() |
|
Registered
Join Date: Oct 2001
Location: Momence, IL 60954
Posts: 1,911
|
CB's old FI setup was known to be crappy :-) That said, I wouldn't recommend megasquirt to anyone. Having seen a fair number of customers have unexplicable fuel system related problems with MS, I can't give it a firm recommendation. Although you don't need motec, you have other systems out there that are excellent and a bargain too - SDS and Link both come to mind.
__________________
Charles Navarro President, LN Engineering and Bilt Racing Service http://www.LNengineering.com Home of Nickies, IMS Retrofit, and IMS Solution |
||
![]() |
|
![]() |
|
Tags |
torductor , torque |