R1100 Fueling, Open Loop, Closed Loop & CO Pot

[Roger 04 RT]

There are several threads running at the moment that are all discussing how the R1100 (and R1150) are fueled by the Motronic. Rather than answer the detailed fueling questions on the R1100 piecemeal, it seemed like everyone might benefit from actual, measured, riding results from an R1100.

http://ukgser.com/forums/showthread.php/396345-1996-R1100GS-Oxygen-sensor-or-not
http://ukgser.com/forums/showthread.php/396350-AF-XIED-for-1150gs-fitted-and-first-tests
http://ukgser.com/forums/showthread.php/394031-rechip-bmw-1150-s
http://ukgser.com/forums/showthread.php/374172-JohnGS1100-Tuning-Chip​

A friend of mine in Vancouver, Canada with an R1100RT volunteered to install an LC-1 on his motorcycle a couple years ago and we took a lot of data which shows exactly how the Motronic MA 2.2 fuels the R1100. Michael (the owner of the bike) ran the bike in stock form (Yellow Coding Plug), richened mixture by lambda-shifting (LC-1 in his case but AF-XIED would be the same), Open Loop (no Coding Plug for the RT), without CO Pot and with the CO Pot.

The next several posts are cut/pastes from the original thread. Although they are very detailed, they answer the questions we're discussing with measurements.

Most of the charts are riding logs of realtime AFRs. This allows you to see how the mixutre varies as the motorcycle is ridden down the road. There's no guessing going on in these charts, just the actual AFR as seen following combustion. You can see cruising AFR, acceleration enrichment, deceleration lean-off and Overrun Fuel Cutoff modes. There are also a couple AFR plots that show how the Motronic fuels during a Cold Start.
RB

 

[Roger 04 RT]

LC-1 Installed on R1100RT

A friend from Vancouver who we know as Happy Wanderer agreed to be a beta tester for the BMW-AF-XIED on his R1100RT: R1100RT XIED Beta Report. The success of that inspired him to go whole-hog and install the Innovate Motorsports LC-1. GS Addict helped him with the installation and the workmanship is top notch.

The significance of this to all of us is that we will, as Happy Wanderer's time allows, start to get the first clear picture of how the R1100 makes its fueling decisions. Before going into detail, from the several test rides that HW has made to date, it's looking pretty clear that the R1100 and R1150 have very similar fueling algorithms. At first glance, it is difficult to tell the difference. Below is the first test ride made with the LC-1 set at Lambda 0.96 (4% richer or about 14.1:1). If I didn't know better, I'd say this data was taken from my R1150 it is that similar.

After riding for a few days at Lambda 0.96, HW increased the Closed Loop enrichment target by 2% to lambda 0.94 (6% total enrichment) and set up and took data from a cold-start to test ride on the road, that chart is below. A summary of what I can see from this and other charts he sent:

--Cold Start Enrichment: The R1100 has a similar start-up and cold start enrichment sequence, conducted as an Open Loop process. A difference to the R1150RT is that the R1100 seems to run a fixed time sequence from cold start where the R1150 shortens the time to Closed Loop, based on engine temperature, probably to reduce emissions. The amount of cold start enrichment seems the same as the R1150--10-15%.

--Adaptation Values: Based on several sequences HW sent me, and his reports, it is very clear that the Motronic MA 2.2 has learning adaptation, much like the R1150 and R1200. I can't say that the process is exactly the same, just that it exists. There are many who see the Motronic ECUs as simplistic, through the course of this project, I've seen many sophisticated capabilities in all models. This should not be a surprise since Bosch/BMW had had electronic fuel injection for about 15 years when the Oilheads were introduced.

--Acceleration Enrichment: Looking at the dips below the 13.8:1 line on the chart, you can see a significant acceleration enrichment, just like the R1150 and R1200. AFRs get to nearly 12:1 with a good turn of the throttle.

--Deceleration Enleanment: Likewise you can see bumps up to 14.4 or 15:1 showing that during mild deceleration the mixture is leaned by 4-8%.

--Overrun Fuel Cutoff: Just like with the R1150 and R1200, when the throttle is closed, the Motronic on the R1100 shuts off the injectors and the mixture shoots to the top of the chart, greater than 22:1.

--Rock Steady Closed Loop Fueling: HW is running E10 fuel during this test which shows that the Motronic has adapted (the fuel is 4% leaner than pure gas) its Closed Loop fueling to 6% richer than stock fueling (10% total enrichment considering the E10) and readily gets it back to 13.8:1 after acceleration, deceleration or overrun fuel cutoff.

Thanks HW for the big effort you and GSA made to get this installed. I know as time allows you intend to take data with the Coding Plug out. When you do, the last of the Motronic's secrets will be exposed and we'll all know once and for all just what the R1100s and Motronic MA 2.2s do when the Coding Plug is removed.

Great work!

 

[Roger 04 RT]

R1100RT OPEN LOOP FUELING, No Coding Plug (no CO Pot)

In the spirit of a picture is worth a thousand words, here is a chart I've been hoping (and waiting) to post for a year. You can look back one page for the Close Loop charts and commentary: HERE.

Michael in BC, Canada, who has installed the LC-1 on his R1100RT was kind enough to reset his Motronic, remove his BoosterPlug, remove the Yellow Coding Plug (left the O2 sensor attached), and log a couple 40 minute test runs logging AFR data with his LC-1. There is a wealth of information in this small diagram.

First, the conditions:
Temperature: 70F
Coding Plug: None
CO Pot: None
O2 Sensor: Installed & Connected (but being ignored by the Motronic)
Motronic: Reset
Fuel: Premium with 10% Ethanol (4% lean compared to gasoline)

Looking at the chart above it is very clear how the R1100RT fuels without a Coding Plug. Since the CO adjustment target, if a CO Pot were installed, is 1.5% carbon monoxide (implying a gasoline AFR of 14.0:1) we can make an educated guess about how this chart will look with the CO Pot connected and adjusted, a test that will be made soon.

Although it would be more precise if I wrote these descriptions in terms of Lambda, but most readers are familiar with gasoline AFRs, so I have put everything in those terms.

First, I've carefully looked over two charts from this test run. Although the O2 sensor is installed and connected, it is clear that with NO Coding Plug, the Motronic does not make any use of the O2 sensor. This means that No Coding Plug is a fully Open Loop fueled R1100 motorcycle. I believe that a 30-87a (beige) Coding Plug will yield the same results.

Next, the idle AFR with no CO Pot connected is a very rich 12.4:1, were this motorcycle running pure gas at the time of the test the AFR logged would have been 11.8:1. At start-up on a 70F, the fueling is about 5% richer than that for several minutes while the motor warms up. Although this bike runs well with no CO Pot, it would be better to have it. An idle AFR of 13.8 to 14.0 would be fine.

Next, take a look at points 3 & 4 on the chart. Point three is the AFR while cruising at 40 mph and point four is while cruising at 65 mph, with AFRs of about 13.8:1 and 14.8:1 respectively. If this engine were running pure gas (what the fuel tables were designed for) that would mean AFRs of 13.2:1 and 14.2:1 respectively.

When the CO pot is adjusted to 1.5% CO, I expect those cruise AFRs with pure gas to be 14.0 and 14.4 respectively, effectively compressing the AFR spread. From this data one could conclude that the Oilheads were designed by BMW to cruise with an AFR in this range. This is well supported by the results several of us have seen with LC-1s and BMW-AF-XIEDs on our R1150s.

Looking closely at the AFR spreads between idle, 40 mph cruise and 65 mph cruise it isn't too hard to infer what the Motronic does with the CO Pot signal: it adds or subtracts an amount of time to the Injection Time calculation. In this way, it has a lot of effect at times when the Injection Pulse Times are short (like at idle and light cruise) and much less effect when Injection Pulse Times are long (like during acceleration and high-speed cruise).

In the No Coding Plug configuration, the Motronic still displays the same array of acceleration and deceleration enrichments and enleanments, still shows Overrun Fuel Cutoff and still shows a Warm Up enrichment sequence.

There's a lot to see and think about in this chart and sometime soon we'll have one run under the same conditions but with a CO Pot installed and adjusted. In the meantime, this points the way to a surge-free, best running Boxer Motor (R1100, R1150 and even R1200). Fuel it at 14:1, one way or another.

RB

 

[Roger 04 RT]

Thanks for the excellent analysis of the data I sent you Roger. Have to admit I was confused at what I was looking at and for a while was getting convinced that the highway cruise section of my test ride (part 4 in the graph) was actually closed loop. After reading your report I now know it's just the ECU using the fuel map to get to it's prime directive; 14.7:1 AFR.

Rain and possible thunderstorms were forecast up here today but reality is things look pretty good outside at the moment. I am really curious to see what is going to transpire when I re-connect my CO trim pot so I think I'll get started on the next test! :dance

Installing and testing the Innovate Motorsports LC-1 wideband O2 sensor is really an eye opener. Over the past couple of years I and many others have chased down all sorts of suggestions on how to get these big twins to quit surging, end snatchy throttle syndrome and run like a big twin should. Theories and myths abound and I've tried most of them going back to Rob Lentini's work and documentation on up to the Booster Plug with a host of other things in between. The list is too long to get into here all over again. I had incremental successes but no lasting solution. Till now.

The LC-1 and the more user friendly plug and play solution the AF-XIED actually work unlike anything else I've done to date which only chipped away at symptoms. Best of all is having actual data from real on the bike testing and being able to log exactly what is happening in real time.

If we can debunk some myths and confirm some information about R259 engine fueling along the way I'm a happy camper. errrr wanderer.

You mention about the ECU using its map to get to the prime directive if 14.7, and what is clear from the data is that the Motronic Open Lopp fuel maps of NO CODING PLUG or BEIGE (30-87a) have no directive regarding 14.7:1 because these maps weren't designed to keep a catalytic converter happy!

I believe that the R1100RT no coding plug map, with stock intake tubes and pure gasoline and the CO Pot adjusted to 1.5%, were BMWs design for a great running boxer. So I'm going to go over paragraph 4 which I've pasted below some more.

The data as present showed 13.8:1 for 40 mph cruise and 14.8 for 65 mph. Let me adjust those numbers for two conditions: ethanol and the misadjusted CO pot. E10 is 4% leaner than gasoline so if you were running pure gas the 40 mph cruise AFR would have shown 13.2:1 and the 65 mph cruise 14.2:1. That is the boxer design cruise range, except we have to make a correction for the unadjusted CO Pot.

At idle, injection pulses are 2.1 mS, but 1 mS is dead time so the actual on-time is about 1.1mS. That is supposed to create an AFR of 14.0:1 at idle (1.5% CO). You now idle at 12.4:1 so your bike is idling 14% richer than the 14.0 spec. That mean the injection time is 0.14 mS too long due to CO not yet adjusted. Taking that 0.14 mS and subtracting it from the 40 mpg gasoline measures AFR of 13.2:1 would bring the 40 mph AFR to 14.1:1, likewise doing the same to 65 mph cruise would move about 14.4.

With these adjustments, it strongly appears to me that BMW designed the boxer engine to run with an AFR between 14.0:1 and 14.4:1 in the cruising range and 12.8:1 when accelerating aggressively. (14.7 is only to make the catalytic converter happy. It must have been a sad day at BMW.)

The good news is that fueling the Boxer at 13.8 to 14.1 can be achieved easily now on the R1000, R1150 and R1200 through Closed Loop Lambda control.

 

[Roger 04 RT]

Part 1 of 2

Michael has been at it again. He got out yesterday for another R1100RT test ride. He kept the bike in the No Coding Plug, Open Loop configuration. He now has a CO Pot installed and on this test ride he adjusted it to the BMW specified 1.5% CO. Looking at the first chart below, you can see that the idle AFR before adjustment (point 1) was in the vicinity of 12:1. At this point the motorcycle was warm and he adjusted the AFR, using the LC-1, to 14:1 (point 2) an AFR of 14:1 equates to a Carbon Monoxide level of 1.5% (BMW spec is 1.5% +/- 0.5% for Euro bikes).

Everything to the right of the point where the CO was adjusted is representative of how BMW designed the R1100RT to run when it didn't have to meet the needs of a catalytic converter. The only difference is that Micheal is running E10 fuel so the whole curve is about 4% leaner than it will be when he runs pure gas--a test he will make later. The important thing to see is the average cruise fueling the top marker line at 14.6:1 cruising is just above and below that level.

It seems clear to me that the BMW Boxer motorcycles were designed to run with a cruise afr of about 14:1. When you subtract the 0.6 AFR leanness of E10 fuel, you can correctly infer the BMW fueling target for the R1100RT, when they didn't have to meet the EPA conditions of the catalytic converter, were an AFR of about 14.0:1 (14.6 seen minus 0.6 AFR for E10 fuel).

Test Conditions
No Coding Plug
CO Pot (adjusted during test ride)
O2 Sensor installed and connected (but being ignored by Motronic)

(The chart below is a moment by moment recording of AFR (Air to Fuel Ratio) over the course of about an hour, during a test ride.)

 

[Roger 04 RT]

Part 2 of 2

To wrap up the basic R1100RT fueling, I'm adding three histogram plots from three separate hour-long rides. What a histogram chart shows is how many occurances there were on the test ride of each AFR level. So a tall bar means that the motorcycle was at that AFR level, many times during the hour. A small bar means it was at that AFR less times.

If you look at the chart below, you can see the AFR distribution of three different fueling styles. Remember, the Open Loop charts are 4% lean because the bike is being fueled with E10 fuel.

Looking at the Closed Loop chart where the target AFR has been set to 13.8:1, you can see that the fueling is tightly around 13.8:1. (If this were a stock bike the chart would look very similar but the center of the tall bars would be 6% leaner at 14.7:1)

Looking at the Open Loop chart with no CO Pot adjusted. You see several peaks. The idle peaks are at about 12:1, the cruise peak is at about 14.4:1.

After adjusting the CO Pot to BMW's spec of 1.5% CO you see the cruise peaks in the same 14.4:1 region but the idle spikes at 12:1 are now gone. To me this looks like a better fueling distribution than No CO Pot.

Lastly, and it's not shown, you could move the Open Loop fueling distributions to the left (richer) by running a BoosterPlug or by running pure gasoline with no ethanol.

 

[Roger 04 RT]

This is a very detailed document on how the R1100 is fueled by the Motronic. It is the type of analysis that you should get from anyone who re-chips your bike. You, as the rider/owner, shouldn't have to but the vendor should.
RB

 

[(RIP) Mark H]

ZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZ Mine runs fine bog standard

 

[Trullion]

Thanks for that Roger. Basically it confirms what I thought all along - the UK spec 1100 and 850s without O2 sensor and cat code plug but with a CO pot are the best fuelled and don't need additional fuelling correction gubbins. Certainly that is my experience of the two I've owned. Neither has ever surged or had fuelling issues, other than when an oil temp sensor went bad.

 

[bart2181]

Thanks for that Roger. Basically it confirms what I thought all along - the UK spec 1100 and 850s without O2 sensor and cat code plug but with a CO pot are the best fuelled and don't need additional fuelling correction gubbins. Certainly that is my experience of the two I've owned. Neither has ever surged or had fuelling issues, other than when an oil temp sensor went bad.

Mine runs like a bag of shit and sounds like a bag of spanners.

 

[Roger 04 RT]

Thanks for that Roger. Basically it confirms what I thought all along - the UK spec 1100 and 850s without O2 sensor and cat code plug but with a CO pot are the best fuelled and don't need additional fuelling correction gubbins. Certainly that is my experience of the two I've owned. Neither has ever surged or had fuelling issues, other than when an oil temp sensor went bad.

There are a couple very easy ways to boost the fueling of an r1100 without a lambda sensor but when you're using 5-10% ethanol ,which results in 2-4% leaner fueling.

--change the fuel pressure with a 3.5 bar regulator 8% more fuel for a net 4-6% richer mixture

--add a boosterplug for a net 2-4% richer mixture

No other device or chip needed. Never learned away by the Motronic. Just remember to reset the CO to 2% or so.

 

[fatnfast]

[QUOTE

add a boosterplug for a net 2-4% richer mixture

No other device or chip needed. Never learned away by the Motronic. Just remember to reset the CO to 2% or so.[/QUOTE]


But this does not apply to the 1150?

 

[Roger 04 RT]

add a boosterplug for a net 2-4% richer mixture

No other device or chip needed. Never learned away by the Motronic. Just remember to reset the CO to 2% or so.

But this does not apply to the 1150?

Because the r1150 only has lambda-sensor fueling, you can't richen the mixture by boosting fuel pressure or adding a boosterplug. Mixture adaptation quickly negates them.

The r1100 has at least two coding plug configurations where it ignores the lambda-sensor, hence those changes do work on those models.

 

[MasaTku]

Is this assumption about "learning" in MA2.2 still true after eleven years?

"Adaptation Values: Based on several sequences HW sent me, and his reports, it is very clear that the Motronic MA 2.2 has learning adaptation, much like the R1150 and R1200. I can't say that the process is exactly the same, just that it exists. There are many who see the Motronic ECUs as simplistic, through the course of this project, I've seen many sophisticated capabilities in all models. This should not be a surprise since Bosch/BMW had had electronic fuel injection for about 15 years when the Oilheads were introduced."

 

[Wrigsby1]