Notices
911 Turbo (930) Forum 1975-1989

ECU for EFI

Thread Tools
 
Search this Thread
 
Old 03-14-2007, 02:34 PM
  #16  
A.Wayne
Formula One Spin Doctor
Rennlist Member
 
A.Wayne's Avatar
 
Join Date: Jan 2005
Location: RPM Central
Posts: 20,448
Likes: 0
Received 0 Likes on 0 Posts
Default

Wow , thanks for the information , Very detailed and professional in setup and operation . I will say The setup described is not unusual for proper mapping and most professionally mapped systems and engines on an engine dyno are mapped in a similar way and do change A/F based on all sensor inputs with the lambda unit ultimately correcting the a/f to get target.and yes you would not , nor require more than +/- 7 while in closed loop operation ...after fully adjusting the fuel map. we also like to do the base map manually as oppose to closed loop , only using closed loop afterwards as previously described by you.

The setup you have described is very professional and elaborate , but most professionals will have different procedures in their attempt of achieving the same end goal, ususally 10-20 mins is correct in getting a clean base map and making power runs etc, but to completly map a turbo motor , or any engine in less than 30 mins , unless you are duplicating, some previous work and setup, is hard to accept , as an engine completly mapped.........
Old 03-14-2007, 10:51 PM
  #17  
m42racer
Three Wheelin'
 
m42racer's Avatar
 
Join Date: Feb 2003
Posts: 1,666
Likes: 0
Received 2 Likes on 2 Posts
Default

Yes you are correct I think in the time to do it perfectly. What I was suggesting is what I saw. The base map came in very quickly and the software was so good that the fine tuning neteed very little. The base line map was so good that the engine test was conducted with this map. Then any changes in the A/F showed where the different engine parts were different and how they changed the amount of air inducted into the engine. Once they found the best combinations, the mapping was done as perfectly as possible to achieve the the highest torqure etc. The base map that came in quickly was not changed much at all. This is what amazed me. This was my point about a system that works and the need to have the tuner who knows how to use it. The time required is shortened.
Old 03-15-2007, 10:22 PM
  #18  
pzull
Burning Brakes
 
pzull's Avatar
 
Join Date: Nov 2002
Location: 2 ends of the Pacific
Posts: 761
Likes: 0
Received 0 Likes on 0 Posts
Default

Wouldn't running closed loop require a very good (expensive) O2 sensor to work well for anything other than gentle driving?

My coupkle hundred dollar bosch sensor (used in my AF meter) reacts considerably slower than the thousand dollar sensor in a venturi housing (excluding meter) that my tuner used for tuning. W ehad the 2 meters side by side and even eyeballing it shows that mine had a lag before picking up AF changes. This is most apparent in big rapid changes in throttle.

Or was my tuner BSing about his sensor being much higher spec (and doesn't last as long) to justify tuning costs....I hope not cos the car runs great even though my O2 sensor sometimes records seemingly wrong AF that should manifest in visible symptoms.....but his meter reads that tuning is spot on in situations with rapid throttle changes
Old 03-15-2007, 11:21 PM
  #19  
Geoffrey
Nordschleife Master
 
Geoffrey's Avatar
 
Join Date: Dec 2001
Location: Kingston, NY
Posts: 8,305
Likes: 0
Received 12 Likes on 10 Posts
Default

Your tuner may have been using a labratory grade NTK sensor which has a response time of 100ms where a standard Bosch LSU-4 has a response time of 300ms or 3 times slower. The MoTeC ECUs can use either sensor.
Old 03-16-2007, 03:13 AM
  #20  
Raceboy
Three Wheelin'
 
Raceboy's Avatar
 
Join Date: Jul 2005
Location: Estonia
Posts: 1,631
Received 17 Likes on 15 Posts
Default

Originally Posted by Geoffrey
Your tuner may have been using a labratory grade NTK sensor which has a response time of 100ms where a standard Bosch LSU-4 has a response time of 300ms or 3 times slower. The MoTeC ECUs can use either sensor.

Geoffrey, usual NTK L1H1 or L2H2 sensor are not lab grade. We tested them thorougly in our technical university. Response time is basically the same as LSU4.
Lab grade NTK sensor cost more than 600$ and they have special calibrating circuitry to get the readings perfect all the time (not similar with rCal on standard WB sensors).
Old 03-16-2007, 09:58 AM
  #21  
Geoffrey
Nordschleife Master
 
Geoffrey's Avatar
 
Join Date: Dec 2001
Location: Kingston, NY
Posts: 8,305
Likes: 0
Received 12 Likes on 10 Posts
Default

Yep, I think that is what I said. In addition, MoTeC ECUs allow for user calibration which I do with every sensor that gets installed.
Old 03-16-2007, 09:24 PM
  #22  
A.Wayne
Formula One Spin Doctor
Rennlist Member
 
A.Wayne's Avatar
 
Join Date: Jan 2005
Location: RPM Central
Posts: 20,448
Likes: 0
Received 0 Likes on 0 Posts
Default

Originally Posted by Geoffrey
Your tuner may have been using a labratory grade NTK sensor which has a response time of 100ms where a standard Bosch LSU-4 has a response time of 300ms or 3 times slower. The MoTeC ECUs can use either sensor.
Yeah!
But what if the tuner is not of laboratory grade
Old 03-16-2007, 10:33 PM
  #23  
special tool
Banned
 
special tool's Avatar
 
Join Date: Apr 2002
Location: limbo....
Posts: 8,599
Likes: 0
Received 1 Like on 1 Post
Default

Originally Posted by A.Wayne
Yeah!
But what if the tuner is not of laboratory grade


Or the engine......
Old 03-17-2007, 10:04 AM
  #24  
WERK-I
Pro
 
WERK-I's Avatar
 
Join Date: Jan 2002
Location: Wisconsin, formerly of Boston
Posts: 571
Likes: 0
Received 1 Like on 1 Post
Default

Off Topic, but............
Geoffrey, congrats on the excellent write-up/project in the latest magazine, 911(?).
Old 03-17-2007, 02:13 PM
  #25  
DonE
Burning Brakes
 
DonE's Avatar
 
Join Date: May 2005
Location: St Johns, FL
Posts: 794
Likes: 0
Received 0 Likes on 0 Posts
Default

Which mag?
Old 03-17-2007, 02:17 PM
  #26  
WERK-I
Pro
 
WERK-I's Avatar
 
Join Date: Jan 2002
Location: Wisconsin, formerly of Boston
Posts: 571
Likes: 0
Received 1 Like on 1 Post
Default

I think it was 911 Magazine. I think it was a multi-part project in which Geoffrey a project car setup.
Old 03-17-2007, 05:31 PM
  #27  
m42racer
Three Wheelin'
 
m42racer's Avatar
 
Join Date: Feb 2003
Posts: 1,666
Likes: 0
Received 2 Likes on 2 Posts
Default

Just how quick does the sensor have to be? The lag time in the sensor is still way faster than any human can map an engine. Usually it is not the sensor at fault. Its the user. I understand there are many different sensors available. The most common is the Bosch. I think price is the real reason. I know there is a OE grade Bosch sensor and a sensor Bosch recommends for engine mapping. I have no idea of part numbers. Some one her may. The NTK sensors have always been very expensive. I think the expense of these sensors is beyond most here and probably not really necessary either. Maybe spend a little extra time trying to stay in the middle if the zone or cell you are tying to map. This alone should make a big difference.
Old 03-19-2007, 10:27 AM
  #28  
Geoffrey
Nordschleife Master
 
Geoffrey's Avatar
 
Join Date: Dec 2001
Location: Kingston, NY
Posts: 8,305
Likes: 0
Received 12 Likes on 10 Posts
Default

Dave,

Thanks, it was a fun project and the magazine articles turned out great.

Simon,
If you are using steady state tuning, then the LSU-4 sensor is sufficient for tuning each cell. When doing swept runs, I always run at a rate of 300rpm per second which helps stablize the AFRs and minimize the effect of running through the cells of the map too quickly, and you can get a reading that is useful. By your comment about the sensor being faster than a human, yes, but if you are datalogging, then the sensor will be the slowest component.

Further, an engine dyno is nice, its the best scenario to map an engine, however, I have found that even on an engine dyno, once in the car, the AFR will change. There are many variables to this, but it is mostly temperature related, especially in a turbo air cooled engine where a dyno cell or chassis dyno does not have the same ability to evacuate air from the cylinder heads as efficiently as the moving air under the car. So, even in the worst case, an LSU-4 is going to be fine.

I've had both the NTK Lab grade sensor and LSU-4 on my MoTeC PLM and more recently in my racecar connected to my M800 (dual) and find that the NTK is faster. As an aside note, on most of the 911 engines I've done, the right bank runs richer than the left bank...sometimes as much as 1 point AFR...
Old 03-19-2007, 10:33 AM
  #29  
jimculp
Shop Rat
Rennlist Member
 
jimculp's Avatar
 
Join Date: Jan 2006
Location: SE of Georgia
Posts: 2,435
Likes: 0
Received 0 Likes on 0 Posts
Default

911 Magazine? Does that mean Total 911? Which issue?
Old 03-19-2007, 11:09 AM
  #30  
Geoffrey
Nordschleife Master
 
Geoffrey's Avatar
 
Join Date: Dec 2001
Location: Kingston, NY
Posts: 8,305
Likes: 0
Received 12 Likes on 10 Posts
Default

The articles were in Total 911 December, January, February, and March issues. The March issue has the MoTeC stuff in it.


Quick Reply: ECU for EFI



All times are GMT -3. The time now is 06:22 AM.