BMWSportTouring BMWST DB
Previous Thread
Next Thread
Print Thread
Page 3 of 63 1 2 3 4 5 62 63
#749472 - 12/26/11 08:27 PM Re: Introduction and O2 Question [Re: dirtrider]  
Joined: Nov 2011
Posts: 2,639
roger 04 rt Offline
Member
roger 04 rt  Offline
Member

Joined: Nov 2011
Posts: 2,639
Naples FL & Cape Cod
Yes, that last chart could be confusing. It isn't a real time plot, just an ordered list of all the o2 voltages measured by the motronic as reported to the gs-911. That lets me know that there isn't an unusual signal pattern. I suppose ideally it would be a straight line.

I went back and made some more voltage measurements and sure enough the motronic puts out about 140mV on sensor low. The narrowband sim does go to high impedance and the voltage you mention is there before warming up.

Much earlier in the thread you mentioned that a lot who have tried the narrowband output of WB sensors have had mixed results. I can see why. The motronic takes a bunch of stuff into account to consider the sensor good. Also the narrowband sim is coming from what is in effect another computer so electronic noise has to be managed.

I'll rent a scope and probably look at lc-1 and stock o2 on a running bike.

As an aside, it looks like just a bit of closed loop richening (14.2) is going to yield a very smooth running bike.

Happy Holidays and Thanks

Last edited by roger 04 rt; 12/26/11 08:30 PM.
#749780 - 12/28/11 03:49 PM Re: Introduction and O2 Question [Re: roger 04 rt]  
Joined: Nov 2011
Posts: 2,639
roger 04 rt Offline
Member
roger 04 rt  Offline
Member

Joined: Nov 2011
Posts: 2,639
Naples FL & Cape Cod
A bit more info ...

More measurements. It turns out that the sensor does float on a signal 140 mV above ground. And the unwarned-up sensor voltage between sensor + and - is 450 mV. Any O2 sims like mine have to add the 140 mV to the lc-1 or other o2 computer. Although I'm still fine tuning, the bike is running strongly at 14.2:1 (13.7 to 14.7) Closed Loop. Will continue to report the clean-up work. The challenge is getting two sensitive computers, motronic and lc-1 to each love the new environment and for the motronic to be blind to it. I expect to refine:

--Wiring length
--Component positions
--Motronic heater emulation
--Electrical noise
--Simulated o2 rise time.

In spite of these planned improvements the bike continues to be smooth and powerful through all gears over a much wider RPM range.
Roger

Last edited by roger 04 rt; 12/28/11 03:54 PM.
#749863 - 12/28/11 11:00 PM Re: Introduction and O2 Question [Re: roger 04 rt]  
Joined: Sep 2009
Posts: 1,824
Peter Parts Offline
Member
Peter Parts  Offline
Member

Joined: Sep 2009
Posts: 1,824
Toronto and Delray Beach, FL
Very interesting discussion with data!!!

Big bucks were spent sorting out the map - can't be bad. Includes minor trimming for altitude, air temp, and oil temp. I'd say for sport bike riding, map is smarter than the O2 sensor.

Yes, Motronic "conserves" .45 volt stoic. So the obvious fix is not wide-band O2 sensor but an op-amp (powered by the sensor heater B+) that scales the O2 output. Simple enough.

I can see how a wide-band sensor can work with an advanced ECU to do tricks our Motronics can't. But I can't see any added smarts when it is connected to the old Motronic?

I wish all this talk about zero-crossings would go away. Yes, the Motronic output can't guess what spritz duration to use while listening to the sensor and so it is all-or-nothing, but the sensor is simply analog and never anything else, within its time-constant performance.

The other thing that really needs data is the continuing debate about just how much time a sport bike spends closed-loop. I think it isn't much for ordinary sporty riding, and that's OK with me. Data?

Ben

Last edited by Peter Parts; 12/28/11 11:03 PM.

R1100S, 1999 (don't know much about other models)
Riding 52 years but only last 47 on BMW boxers
Bike writings: http://www.tinyurl.com/6bvt23
(includes a 6400 mile trip, 20 days, with wife to LA)
#749914 - 12/29/11 03:18 AM Re: Introduction and O2 Question [Re: Peter Parts]  
Joined: Nov 2011
Posts: 2,639
roger 04 rt Offline
Member
roger 04 rt  Offline
Member

Joined: Nov 2011
Posts: 2,639
Naples FL & Cape Cod
Ben,
I have published additional data where I began this thread on another site, Here . Or here .

If there is enough interest I could add the charts here too. Roughly 45% Of the time during a normal ride the bike is closed loop, including to my surprise, during acceleration.

The lc-1 that I used has a simulated narrowband output that is connected to the motronic directly. Then I programmed a lower target AFR. It makes a big difference. So the added smarts come from the innovate motorsports lc-1.

Roger

Last edited by roger 04 rt; 12/29/11 03:35 AM.
#750069 - 12/30/11 12:29 AM Re: Introduction and O2 Question [Re: roger 04 rt]  
Joined: Sep 2009
Posts: 1,824
Peter Parts Offline
Member
Peter Parts  Offline
Member

Joined: Sep 2009
Posts: 1,824
Toronto and Delray Beach, FL
That Innovate gizmo sounds pretty helpful.

How do you know when closed-loop? Even on the highway, best place for surging, you find just 47% closed-loop and "only" 80% with a steady hand. I leave it to others to decide if that sounds more like dirtrider's point of view or mine.

Interesting data on the other link. The Motronic is open-loop when you blip the throttle (any idea how much???), But if you hold it steady WHILE accelerating, it will go closed-loop. And, the dumb ECU will be falsely lean: that's why I don't think the feedback Motronic is helpful for sporty riding.

You found the Motronic totally zaps some injections (not smart enough to be proportional, just all-or-nothing). The Motronic acts like a switch and the O2 sensor is tied to it, not the other way around as certain authorities have argued in the past.

Yes, it is a paradox that these boxers need a richer mixture to burn stoich-like. The old Motronic system with a narrow-band sensor has no way to discern what to do, except to have a scaling-circuit like your gizmo or as I posited above. Or.... if Bosch were to make a 14.2 sensor!

Thanks a million. Very glad to have your great work to learn from.

Ben... soon 51 seasons



Last edited by Peter Parts; 12/30/11 12:30 AM.

R1100S, 1999 (don't know much about other models)
Riding 52 years but only last 47 on BMW boxers
Bike writings: http://www.tinyurl.com/6bvt23
(includes a 6400 mile trip, 20 days, with wife to LA)
#750142 - 12/30/11 12:27 PM Re: Introduction and O2 Question [Re: Peter Parts]  
Joined: Nov 2011
Posts: 2,639
roger 04 rt Offline
Member
roger 04 rt  Offline
Member

Joined: Nov 2011
Posts: 2,639
Naples FL & Cape Cod
51 seasons, wow, congrats! (I've worn a lot of tread off my tires too.)

The way that I know that it's gone closed loop is after examining the data. I'm riding with a GS-911 connected to the service dongle under the seat, and with a small netbook computer strapped behind me. Then I download the data into Excel and go from there.

It may also be that 14.4 works and I'll be leaning upward to see how close I can get to 14.7:1.

I was looking back at some old R1150RT promotional material. It refers to, and I quote, MOTRONIC with FUEL CUTOFF OVERRUN.

All in all I'm not surprised that 14.7:1 wasn't the best place for this model, it's pretty far off the best power mixture. BMW and users tried a lot of things to improve the bike between R1100 and R1200, careful TB balancing, valves, tuning, Lentini's Zero=Zero tune, fooling with CAT Code plugs, and even second spark plugs. There was also the class of BoosterPlugs, Techlusion, etc. While they can help, those solutions don't change Closed Loop.

When you look at the R1150RT map that comes with a PowerCommander (it has a Wideband Sensor), it starts with Closed Loop set to 13.8:1! For most PowerC users, my data suggests that its richer closed loop is one of its best fixes. Also, with it you can richen the fuel tables by the amount needed to fix the Motronic maps for E10 (or any other Exx) Ethanol-spiked fuel.

The Innovate LC-1 or MXT-L is a good product but NOT plug and play. I will try to write up the easiest way to install and locate it when every thing is done.

Fun Stuff ...

#750148 - 12/30/11 12:55 PM Re: Introduction and O2 Question [Re: roger 04 rt]  
Joined: Jun 2010
Posts: 13,993
dirtrider Offline
The Oracle
dirtrider  Offline
The Oracle
Member

Joined: Jun 2010
Posts: 13,993
Ohio
Originally Posted By: roger 04 rt
--

Although things are working really well, I am not convinced I am getting a good rich/lean toggle yet so I will probably go and get an oscilloscope so that I can look into all this with a bit more certainty.


Morning Roger

First off, thanks for taking the time to document & share your wide band project with us. This has been interesting reading & I'm sure an interesting project for you.

Yes, I would be interested in seeing that scope data also. From the limited plots I have seen of your 02 as filtered by the LC-1 it looks like there are long periods of the signal going to the extremes with no cross counts for that period. If the Motronoic sees that maxed out 02 signal with no cross counts long enough it will pop open loop & use base mapping trimmed by the sensors. Your LC-1 might be forcing longer open loop periods than the stock narrow band did.

Same with your acceleration closed loop periods. From what I have seen using a duty cycle meter on the (stock) 02 the system goes open loop on throttle movement (difficult to use narrow band 02 input on quickly changing load/rpm) but as the TPS signal goes steady but vehicle still accelerating the fueling quickly catches up to the 02 then the system starts reacting to the 02 again.

It would be nice to see your same route ridden with the stock 02 sensor feeding your Motronic but the wide band watching your fueling control, then the same route ridden again with no 02 input & operating open loop, then ridden again using your wide band system for fueling control.

If you have some riding down time this winter & want to remove your Y pipe let me know as I have some weld in threaded stainless steel 02 sensor "bungs" I machined up for a wide band project I did few years ago. You are welcome to one to weld into your Y pipe so you can add the stock 02 narrow band back into your system along with keeping the wide band in place for A/B comparisons.


D.R. ___
Sent from my rotory dial wall phone!
#750178 - 12/30/11 03:10 PM Re: Introduction and O2 Question [Re: dirtrider]  
Joined: May 2007
Posts: 591
TestPilot Offline
Member
TestPilot  Offline
Member

Joined: May 2007
Posts: 591
Williamsville, NY
Originally Posted By: dirtrider
From the limited plots I have seen of your 02 as filtered by the LC-1 it looks like there are long periods of the signal going to the extremes with no cross counts for that period.


DR,

Pardon me for jumping in, but I think you are misreading the GS911 plots of RPM, TPS and Lambda. I believe the Lambda traces are not the voltages from the O2 sensor (real or simulated), but simply an indicator of whether the system is operating open loop (output = 0) or closed loop (output = 1).


Karl
2007 R1200ST
#750187 - 12/30/11 03:36 PM Re: Introduction and O2 Question [Re: roger 04 rt]  
Joined: Jun 2010
Posts: 13,993
dirtrider Offline
The Oracle
dirtrider  Offline
The Oracle
Member

Joined: Jun 2010
Posts: 13,993
Ohio
Morning Carl

Well that's sure possible. It does look pretty strange for 02 traces.

I would like to see a plot of LC-1 output spoofed 02 voltages as well as the cross counts vs injector on times, vs seconds (or some reference time period).

Hopefully he can get a scope on it then we can see unmolested signals.


D.R. ___
Sent from my rotory dial wall phone!
#750216 - 12/30/11 05:33 PM Re: Introduction and O2 Question [Re: dirtrider]  
Joined: Nov 2011
Posts: 2,639
roger 04 rt Offline
Member
roger 04 rt  Offline
Member

Joined: Nov 2011
Posts: 2,639
Naples FL & Cape Cod
D.R.
Karl,

Lots of good ideas and comments in the last three posts which I completely understand.

Karl is right, the GS911 is reporting 1=Closed Loop, 0=Open Loop. And to make matters worse in that sense, it can only take a sample every 650 milliseconds, or so. There are a couple charts that D.R. asked about that are voltages but they are merely a plot of all the voltages captured. I did that to make sure there were no "holes", that is voltages that were never measured. It is a way to check things in a environment where the sampling rate leads to "undersampling".

However, the GS911 has a function that you can use in the garage which shows sequential O2 points. I used that to temporarily fine tune yesterday. On the one hand, I now have the O2 sensor toggling much like it was with the Narrowband or the other hand, I now realize that I could have engineered the installation for less electrical noise and want to do that. Also, I want to see a real time oscilloscope output.

In the meantime, I will try and screen capture the sequential plots.

I can say with pretty good certainty now that I have a functioning solution. I want to work more to make it the best it can be and am thinking about welding another bung into the exhaust system so I can try some of the ideas I've had and the ideas D.R. mentioned.

When I finish, I think this will be a very good tweak for O2 switching point. And as D.R. mentioned, a very fun project.

Page 3 of 63 1 2 3 4 5 62 63

Powered by UBB.threads™ PHP Forum Software 7.6.0
Page Time: 0.019s Queries: 14 (0.003s) Memory: 2.8212 MB (Peak: 3.0066 MB) Zlib enabled. Server Time: 2018-06-24 03:20:47 UTC