GS-911 Error Code Question - Lambda Sensor Related

[roger 04 rt]

Hi Trail'R, I'm going to look at the file, first thing this morning.

 

You mentioned the Bosch 15109 earlier in the thread, the connector looks wrong to me, can you show me a picture of your old sensor?

 

As a precaution I want to check two things:

-What sensor did you use as a replacement?

-Can you send a picture of the connector end of the O2 sensor?

-Are any of the wires in the O2 sensor connected to its case? (Check with an ohmmeter.

-Where are you connecting the ground of the AF-XIED?

 

This isn't making sense to me

 

I'd like to quickly determine whether we need to contact Steve at Nightrider and get him (or possibly Mike at BB) to immediately send you a replacement.

[roger 04 rt]

The O2 sensor connector for your R1150 has the following:

--Four male pins recessed inside the shroud.

--The shroud has two small ears.

--Two white wires connect to pins 1 and 2. Neither white wire has a short to the O2 sensor case.

--The Gray wire goes to pin 3

--The Black wire goes to pin 4

--Neither the Black or Gray has a connection to the O2 case.

 

Some on-line pictures of the Bosch 15109 show male pins and some other pictures show female pins. That makes me wonder if there are other differences.

 

 

[roger 04 rt]

Trail'R,

I'm sure this is a bit frustrating but we'll get there.

 

I've looked through the data set, the negative O2 sensor voltage doesn't make sense. When you can get to the other questions above, that would be helpful. There are two added measurements:

 

If you measure pins 3 and 4 on the motorcycle harness (female pins) with the af-xied unplugged you should get the following voltages with the key on:

 

Pin 3: ~140 mV

Pin 4: ~590 mV

 

If the questions from my earlier posts and and this one don't show results, I'm going to help you get a new XIED and cable. Did you buy from Beemerboneyard or Nightrider?

RB

[Trail'R]

 

Replacement Sensor: Bosch 15109

 

 

New Sensor Connector

 

 

 

Old Sensor Connector

 

 

 

Pinout of Old & New match with the same color coding. No ground to case on either old or new sensor. Male pins on the new connector.

 

Gound connection test at two locations. Once at frame to ABS pump and 2nd to negative terminal of the battery. Both produce negative lambda voltages.

 

Pin 3 voltage: 141mV

Pin 4 voltage: 588mV

 

 

Wire Color Coding

 

O2 Sensor to AF-XEID

White to Gray

White to Red

Black to White/Blue

Black to Yellow/Blk

 

ECU to AF-XEID

Gray/White to Red

Green/Brown to Gray

Black to Blue

Yellow to ---(no wire)

 

Yep, we will get there. The assistance is appreciated.

 

[roger 04 rt]

Nice work. Motronic connector voltages are good.

 

Will review and be back to you later this morning.

 

Edit: Where did you buy the XIED so I can contact them when needed?

[roger 04 rt]

TrailR

 

O2 Sensor to AF-XEID

White to Gray

White to Red

Black to White/Blue

Black to Yellow/Blk

 

I think one of the two Black wires is Gray. Could you check and revise?

[roger 04 rt]

Based on the wiring, the voltages into the Motronic with no O2, the pictures, and the logs from the GS-911, it should be working, but it isn't.

 

Before moving on: you probably noticed that the inputs to the Motronic are offset by about 140 mV. That means that when you measure a negative 50 mV it is actually 100 mV at the input to the O2+ line (which makes it 50 mV below the O2- line). So the negative numbers mean that the AF-XIED hasn't started. And your most recent data it went from -54 to +90 after 1 min. 25 sec. Normally this start up takes about 20 seconds.

 

We also have the problem of the dropouts. I thought they were due to your low O2 sensor voltage but with the new O2, that's not the problem.

 

So my diagnosis now is intermittent cable or flaky AF-XIED. The only way to rule that out is to get the manufacturer to ship you a new set of cable and AF-XIED. I've alerted him.

 

If you contact Nightrider, I believe they will ship you a new set and you can ship the old one back to them. Feel free to e-mail them or call.

RB

[Trail'R]

Interesting thoughts. I am not as familiar with the RT, but if it is an easy process I can try mounting and running the same cold start log. If it is a faulty AF-XIED, would we see the same O2 dropout on the RT as the GS? And if that is not definitive enough, I can look at mounting to the K1200RS to solidify the dropout. My only concern is that both bikes are not as familiar to myself and I have not had the time to really understand where they are in the current state of tune.

[roger 04 rt]

Trying the AF-XIED on the R1150RT isn't a bad idea if you don't mind the work. It would be interesting to see what happens. I'm pretty familiar with the results.

[Trail'R]

 

Here is an update on connections:

 

O2 Sensor to AF-XEID

White to Gray

White to Red

Black to White/Blue

Gray to Yellow/Blk

 

On the RT, I assume that I will need to pull the tank to access the connections in a similar nature to the GS?

[Trail'R]

I started a new thread on the RT. I figure I need to get the RT tuned before introducing the AF-XIED into the mix. She is up on the lift ready for the diagnosis. Let's tune this baby so I can enjoy some riding this summer. My wife is starting to get suspicious about so many bikes, but none are road worthy. She is ready to scrap the lot and use the proceeds for new kitchen counters. Let's not let that happen!

 

New RT link... Linky

[roger 04 rt]

Am following the RT thread too. Another rider with the same bike as you (2000 GS) has just taken before and after data on a Bosch 15109 replacement center. I will post later today.

[roger 04 rt]

The data from the new sensor looks very good. One thing I'm not used to seeing though is lean operation during warm-up (will show plot later). Now you should ride for a tank of gas or so to let the adaptives fully form. Testing during this period is fine, just try and not reset the Motronic.

 

Then after a half tank of fuel or so, take another cold start log just like the last one.

 

Edit:

 

The chart below looks good for each sensor, better than before. The O2 sensor is outputting a higher voltage (100 mV higher) and importantly, the Lambda Control line stays in Closed Loop steadily now compared to dropping out of Closed Loop in the earlier charts. (It leaves closed loop when the throttle is blipped, that's okay.)

 

However, look at the period before Closed Loop but after the Fast Idle lever is dropped (that's when the TPS signal drops from 1.92 to zero). During that time (engine temp in the 40-60C range), the O2 sensor voltage goes from rich (800mV) to lean and even bounces around. This indicates that the AFR is around 14.7:1. That's 5-10% leaner than I'd expect.

 

Let's see what adaptation does after you've ridden for a while, then take another cold start log.

 

P.S. In the state it is now, the AF-XIED is likely to work. But let's figure out the cold start leanness first if we can.

 

 

Trail'R,

 

Here is a 2000 R1150GS, running an AF-XIED on S6, with a new Bosch 15109 sensor (he replaced it after seeing this thread!). His before voltage was 700 mV like yours. His after voltage is at least 100 mV higher, like yours.

 

The big difference is that his mixture stays rich all through the Open Loop, warm-up period. Since he is running setting 6 it means he is richer than 14.4:1 (the short lean blip is him twisting the throttle, which he did several times during the plot). In your case, the data is without AF-XIED so during the warm-up period it is getting leaner than 14.7:1. His bike is what I'd expect. Either your bike has something we haven't figured out or perhaps your XIED/Cable has an issue--based on this and the won't-hold-closed-loop problem.

 

I'm still stumped.

 

[Trail'R]

Roger,

 

That is really interesting data and a good comparison between bikes. Looks like the Bosch 15109 is performing as expected, so we can rule that out as an issue. Both bikes are producing waveforms from the O2 sensor that are nice and consistent.

 

Would a CCP change the warm-up characteristics? There is no CCP installed on my bike. Maybe the other bike owner can jump in on this thread and describe his bike and and modifications made. I can make a jumper and log some data after my fuel injectors are returned from cleaning.

 

What other sensors are used initially or during open loop? Maybe there is a sensor error that is not readily obvious such as oil temp, or air temp.

 

I do have an IICE AIR that we could use to compare logs. Maybe this would help isolate what happens during cold start-up by changing one of the input variables into the ECU.

 

Final thought, is there any electrical system that feeds the ECU where a ground short would only be observed during the startup procedure? This would be like looking for a needle in a haystack, but I feel I need to ask the question.

 

All of my thoughts here are that during open loop the ECU is using a set mapping. What inputs determine that fuel map?

 

 

[roger 04 rt]

Comments:

 

I don't remember whether that bike has the stock Yellow(30-87) coding plug. You should, though, make up a 30-87 jumper and run a Cold Start log, that would be interesting.

 

The other owner has no mods to his bike other than the AF-XIED.

 

Your report on the before after flow rates on the injectors will be interesting.

 

Introducing the IICE Air won't help, for a few reasons. It doesn't provide a constant 20C like the BoosterPlug, at least not the one I tested, and there is no provision for it to measure actual intake air unless you mount it inside the airbox, and then it is slow to respond since the sensors are inside a plastic case.

 

There isn't a ground that would only be affected by cold start that I can think of, but I've been thinking about grounding issues. Your data looks okay ground-wise.

 

The Open Loop question is a good one. Below is something I've posted before.

 

Open Loop fueling calculations go roughly like this, performed on a millisecond by millisecond basis:

1) Read RPM (Hall Sensor) and TPS: Find a base-fuel-value by using the four closest cells in the fuel table

2) Measure Engine Temperature: If not warmed-up, add a correction, otherwise skip this step (warm up enrichment runs from about 15% to 3%, declining as the engine warms.)

3) Measure Input Air Temperature (from the airbox IAT sensor): Apply a correction factor from a table

4) Measure Air Pressure (from the barometer inside the Motronic): Apply a correction factor from a table

5) Measure Battery Voltage (measured at the Motronic): Apply a correction factor from a table

6) Based on RPM/TPS: Apply (up to 3) Mixture Adaptation values learned previously during Closed Loop

7) If the RPM or TPS is changing quickly or if the engine is cold, go back to step 1, otherwise continue to step 8

 

Closed Loop fueling steps

8) Apply the most recent Lambda Control Factor (LCF) (short term fuel trim)

9) Read the O2 sensor and determine whether to increase or decrease the LCF so as to remain close to Lambda = 1

10) Determine whether to update any Mixture Adaptation value (long term fuel trim)

11) Stay in Closed Loop until a change event occurs (usually throttle movement or RPM change).

[Trail'R]

 

Interesting list of steps. So is it possible for no value to be in a fueling table or a value that will drive it very lean? It appears the bike goes lean during warm-up and right after the idle lever is dropped. What I don't know is if it goes lean during other open loop situations or if it will stay rich until closed loop if the idle lever remains at mid-point. I believe that a test will require open road data logging as speeds are adjusted to force out of closed loop. If it is lean during all open loop conditions, this may be the source of the lean surge I feel at low speeds.

 

As I reflect back on the history of the bike, I believe the ECU was replaced by the dealer. I will look to see if I can find the records to make sure.

 

In regards to the TPS as an input, I have tested and adjusted via the GS-911. Would a TPS dead spot force it to go lean? I have not looked at how hard it is to pull off the bike, but I could switch the RT and GS TPS to see if there is a change. However, this does not seem likely since the GS-911 is able to read the TPS voltage at closed throttle.

 

After I receive the injectors back, we should make a single change with these ideas and test results in mind to see if we can zero in on the issue that is influencing the cold start-up.

 

Meanwhile I will mount some tires on the RS and use it for my summer ride this year as I continue my battle with the RT & GS. I have no doubt that I will have two very fine running machines after this effort.

[Trail'R]

I noticed that looking at the '00 GS and my '01 GS the injection amplitude for time is very different between the two bikes. The chart y-axis scales are the same on the chart, but it appears that the '00 looks to be adjusting quicker in response to the inputs as well as inputting more fuel for each pulse. This may be difference in fuel maps for the given inputs, but it does make me question responsiveness of the injection system. In fact the lean period right before closed loop, the injection time is perfectly flat.

[roger 04 rt]

Here is generally how the fuel table values are grouped. All of the map is Open Loop but some areas are also covered by Closed Loop. All areas are affected by adaptation.

 

1. Most of the table is lambda=1 for gasoline meaning it would achieve 14.7:1. But that means most of the table is a little lean if you run E10.

 

2. Deceleration area, small throttle angles, higher RPM. Leaner than lambda=1

 

3. High Power areas, ~6-8% richer than lambda=1

 

BTW, I've contacted Nightrider and when/if needed they will swap.

 

Get the injectors back, put in the Yellow coding plug and let's see.

 

Interesting point too about the injection times. We'll need to look at that too.

[roger 04 rt]

Trail'R, I just got an email from Steve at Nightrider. He's seen how hard you're working on this. He has sent a resplacement to you for comparison purposes--cable and AF-XIED. This will allow you to quickly rule it in or out soon.

[roger 04 rt]

Interesting list of steps. So is it possible for no value to be in a fueling table or a value that will drive it very lean? It appears the bike goes lean during warm-up and right after the idle lever is dropped. What I don't know is if it goes lean during other open loop situations or if it will stay rich until closed loop if the idle lever remains at mid-point. I believe that a test will require open road data logging as speeds are adjusted to force out of closed loop. If it is lean during all open loop conditions, this may be the source of the lean surge I feel at low speeds.

 

As I reflect back on the history of the bike, I believe the ECU was replaced by the dealer. I will look to see if I can find the records to make sure.

 

In regards to the TPS as an input, I have tested and adjusted via the GS-911. Would a TPS dead spot force it to go lean? I have not looked at how hard it is to pull off the bike, but I could switch the RT and GS TPS to see if there is a change. However, this does not seem likely since the GS-911 is able to read the TPS voltage at closed throttle.

 

After I receive the injectors back, we should make a single change with these ideas and test results in mind to see if we can zero in on the issue that is influencing the cold start-up.

 

Meanwhile I will mount some tires on the RS and use it for my summer ride this year as I continue my battle with the RT & GS. I have no doubt that I will have two very fine running machines after this effort.

I noticed that looking at the '00 GS and my '01 GS the injection amplitude for time is very different between the two bikes. The chart y-axis scales are the same on the chart, but it appears that the '00 looks to be adjusting quicker in response to the inputs as well as inputting more fuel for each pulse. This may be difference in fuel maps for the given inputs, but it does make me question responsiveness of the injection system. In fact the lean period right before closed loop, the injection time is perfectly flat.

 

Trail'R,

I think you've hit on a key point here. The 2000 R1150GS has a TPS input of 0.32 and is at 1000 RPM but the injection time (just before it goes closed loop) is 2.3 mS. Your bike at those conditions (except your RPM is higher) is 2.05 mS. That is a huge difference. Barometric pressure is 1003 vs 978 which is 2.5% difference, not enough to account for the fuel difference. Air temp is the same.

 

I can't explain the injection time difference but am curious about the ECU replacement and Coding Plug. I've asked the 2000 owner whether they are using the Yellow plug.

[roger 04 rt]

In fact, the more I think about this, looking at what happens after closed loop is the injection time climbs to about 2.2 mS. That is the fuel needed to reach 14.7:1 (lambda=1). For some reason, your ECU is not calculating the correct amount of fuel (even after reset and without the AF-XIED attached).

 

The value calculated is a long way off from the correct value. Without the XIED the Motronic has just enough range to make it work. With the XIED attached, it's just to far a shift.

 

That's the problem! ECU or Coding Plug or ??. Let's figure it out.

[roger 04 rt]

The earlier data is RC below. He does not have a coding plug. BTW, I have a spare Motronic for single-spark 1150.

 

Here is data from some other R1150GSs that I have for idle injection time at 60C:

 

Jim's 2000 R1150GS: 2.3 mS

Wally's 2000 R1150GS (several): 2.18 mS, 2.24 mS, 2.24 mS, 2.24 mS

Jody's 1150RS: 2.3 mS

RC's 2000 R1150GS: 2.24 mS, 2.3 mS

 

Trail'R's (several cold starts): 2.05 mS, 2.05 mS, 2.05 mS, 2.05 mS, 2.05 mS.

 

My calculations show that your 2.05 mS pluse is delivering 15% less fuel than the average 2.24 mS pulse at the end of the warm up period.

 

If you get a minute it might make sense to read the part number off your Motronic.

[Trail'R]

 

The data is starting to paint a clear picture that the lean condition is driven from the fuel side of the system. Now we just need to figure out why.

 

Bike: Mfg 03/01

Motronic: 13 61 7 658-614

 

Parts fiche indicates this is the correct part number for bikes produced before 12/02

 

The very consistent 2.05ms for each cold start is interesting in that a person would normally expect some variability in the fuel delivery as the inputs change. This may not be known data, but I question if there is a minimum default injection time built into the Motronic, almost like a limp mode until other data inputs can drive it into a correct map.

[roger 04 rt]

Okay on the Motronic for now then.

 

The 2.05 mS varies by 60 uS but I didn't show that. There may be a minimum but if there is it is in the mid 1 mS range.

 

The 2.05 value comes by looking up TPS and RPM, then correcting for Engine Temp (during warm up which is what we're talking about), air temp, barometric pressure, and battery voltage (low voltage affects the injectors). So it is calculating this number in real time.

 

When you have a chance, let's see if the number is different with a Yellow Plug cold start.

[Trail'R]

 

Injectors are back with the report below. I also have a trial AF-XIED that I am ready to try now that I have the injectors back.

 

So I will work on 3 cold starts and collect data in the same manner as before.

 

Run #8: Fresh Injectors

Run #9: CCP jumber installed

Run #10: Replacement AF-XIED (Should I use the CCP?)

 

After that we will make-up the next steps based on the data collected.

 

Happy to get started on this again now that the holiday is over.

 

Injector Report:

 

[roger 04 rt]

I would run the AF-XIED test with the correct Coding Plug installed.

[Trail'R]

New cold start log taken with the recently cleaned injectors. The results are interesting, but I did enter a new variable by accident so I will run another cold start in the same configuration. The difference this time is that I moved the idle lever from the middle position at an engine temperature of 60C vs. the previous charts at 40C. I will run another log changing at 40C to see if this is the difference for the lean start.

 

I will also note that my idle RPM bumped up a bit with the cleaned injectors.

 

Run #8a Data Log

[Trail'R]

 

I have two additional cold start runs.

 

Data log #8b

 

Data Log #8b is a log of a cold start with the clean injectors. Idle lever released from mid-point to off at an engine temperature of 40C. The resulting log looks similar to me with a lean period prior to closed loop. Interestingly lowering the idle lever at and engine temperature of 60C did not create the lean period prior to closed loop as in log #8a

 

Data log #9 consists of placing the yellow CCP with a jumper between 30-87.

 

 

 

This resulted in the following log:

 

Data log #9

 

 

I did an initial plot of the lambda voltages and it is erratic just like the previous lazy sensor.

 

 

All of the log files are located at the following location:

 

Excel Log File

 

[roger 04 rt]

Quick question, what's the fuse tap powering up? I see what looks like a relay, what does that go to?

[Trail'R]

 

The relay is tapped into fuse 7 and triggers the switched side of a Centech AP-2 auxiliary fuse panel. The AP-2 has both constant power and switched power via the relay. Easy enough to disconnect if we need to check for issues on that side.

[roger 04 rt]

I'm having a look through the data now, as usual, nice job of collecting the info.

 

What loads are attached to the Centech, and how is the control coil for the Centech relay grounded?

 

Here's what I see in the data so far:

 

Logs 8a and 8b are very good, predictable results. In 8a you actually dropped the throttle at 59 degrees C engine temperature, and just like before, the O2 sensor reading went lean. Very similar results in 8b, with the earlier throttle drop. Key point: Closed Loop Lambda control stays solidly engaged once you pass 60C engine temp.

 

In Log 9, the only difference is that you add the 30-87 jumper (right?). Now when the engine temp passes 60C, Closed Loop Lambda Control is very erratic. The 30-87 jumper connects "ground" (a special ground wire) to Motronic input 87. I'm thinking that someone grounded that relay for the Centech or something else to what looks like a ground wire but is really 31EL coming from the Motronic, a VERY sensitive ground. I can't think of any other reason why adding the 30-87 jumper would cause the erratic Closed Loop Control Line.

 

The 31EL wire that goes to the Coding Plug socket is a Brown wire and would look like a usual Ground. Other 31EL wires are Brown/Gray and would not seem like a Ground to be used.

 

Try another Cold Start Log but unplug the Centech Relay if you can't look into the Electrical Box and see where the Relay Coil is grounded.

RB

[roger 04 rt]

Further, that Pin 2 from the Motronic 31EL Brown wire is ground for the HES, shielding of the HES, ground for the Coding Plug, and shielding for the Lambda Sensor signal wires!

[Trail'R]

Roger,

 

I have removed the relay and even went as far as disconnecting all of the electronics accessories from the bike. I have a cruise control from MCCruise installed so I will have to review the wiring to make sure that is installed correctly. The relay in the photo grabs a ground connection from the AP-2 ground bar. I do not like tapping into a stock harness so I make it a habit to route ground and accessories back to a common source.

 

I ran a cold start with the accessories removed along with fuse 7 removed. The intermittent O2 is still present. I will investigate the cruise removal and see if there is a way to easily disconnect. At a minimum I will be able to remove the cruise computer and leave just the harness.

 

Log #9c - AP-2 Removed

[roger 04 rt]

I thought we were going to get an easy win with the grounding, with your input on this and this test, I don't think so but check the cruise control ground if you can.

 

I've added 9c to the analysis and it looks better than the last one, 9, but still not quite right. Where we're at is no-Coding-Plug has a steady Lambda control but plug in the equivalent of a yellow plug and the Lambda Control bit drops out. I'm thinking more about the leanness in Cold start and idle RPM. Will be back to you on it later.

 

In the meantime, can you check and see if there are any errors in the Motronic error log?

[Trail'R]

 

Thanks for your thoughts on this. I was hopeful this would become more obvious and I shudder at the thought of chasing down electrical grounds.

 

I just checked and there are no error codes showing.

 

As I looked through the bike records, I verified the ECU was replaced. Back in '05 I purchased the bike in LA and drove it back to St. Louis. The very next weekend, there was a major short in the electrical system that required replacement of harness, ABS pump and sensors and ECU. All the work was performed by a BMW dealer and it looked to be resolved. Not sure if it is related, but as we chase electrical that could be an item to consider. All electrical items on the bike have worked for the past 9 years since the repairs and I have had not visible issues due to electrical.

 

Looking at the shielded ground, I would expect to see other issues if the HES was starting to short. Again just throwing it out as a thought if the symptoms align.

 

The bike starts and idles easily, but it has always run lean. Are there any other CCP configurations that would provide a clue?

[roger 04 rt]

Based on the 8a, 8b and 9 data sets I have some thoughts that don't involve grounding. But I think we should keep the harness replacement in the back of our mind.

 

While I'm working on the write up of what I saw in the data, you mentioned that the TBs have the original paint. I'm curious, could you check the number of turns each TB screw is out?

[Trail'R]

Hi Roger,

 

Thanks for hanging with me on this one! I see that you are not one to back away from a good challenge.

 

As far as the BBS goes: Left is 1 7/8 out and the Right is 2 1/4 out.

 

I am pondering the electrical and working to remove loads via fuses to see if there is a rogue ground issue. I will have some additional logs available for review after a bit.

[roger 04 rt]

I do like a good problem a lot, kind of like a puzzle. At the moment, it seems like a lot of things have been ruled out.

 

I spent some time today looking over the three cold start logs 8a, 8b and 9. Every way I look at it the sensors on your bike all appear to be working, stable and delivering reasonable values. There don't appear to be any intermittent signals nor are there any really noisy measurements.

 

I looked at the lamda sensor voltages as measured by the Motronic in three conditions: cold sensor, open loop and closed loop. I don't think that there is any reason to doubt the O2 sensor or the Motronic's ability to measure it accurately. This is important to what comes next because the lambda/O2 sensor is still seeing a couple unusual conditions.

 

1. Below 60 degrees C engine temperature, during Open Loop the bike is running LEAN if the throttle is Closed. It is also surprising that your bike idles well at this point with a closed throttle but it does.

 

2. Above 60C you bike idles well but we saw that if you install the yellow coding plug jumper the Motronic loses Closed Loop frequently. This is unusual. I've looked at all the sensors before, during and after the Closed Loop dropout and as far as I can see the sensor values are stable and reasonable. Some internal Motronic calculation is leading to the dropout of Closed Loop. However, I saw one spot where you opened the throttle during this condition--with a slightly opened throttle, Closed Loop holds. This is leading me to believe that the leanness (seen during Open Loop Closed throttle) and the Closed Loop drops are related. Therefore let's see if we can solve problem 1.

 

The leanness measured below 60 C with a TPS of 0.32 degrees seems real to me. Since the throttle stops are factory set (right?) and your BBS screws are about two turns out, at first look, things seem normal. However the logs show that your bike idles higher than I'm used to seeing with the fast idle lever in the detent position, suggesting more air than usual.

 

Continuing, there is also enough air at 50C engine temp for your bike to idle at 1200 RPM with the throttle closed, suggesting excess air. Then further, when the engine temp passes 60C, and Closed Loop kicks in, the Motronic richens the LEAN mixture a bit and the idle speed picks up about 80 RPM, this confirms the lean O2 reading as due to a slight excess of air. The added fuel combined with the excess air picks up the RPM.

 

Are you certain there is sufficient slack in the throttle cables for the throttles to fully close? What we're seeing could be explained by the the throttles hanging 0.8 degrees off the stops which is only about 0.014". I'm not saying that's the case, but if it was, the Motronic would be calculating air for 0.32 degrees and the actual angle would be 1.02 degrees. (I may simulate this with shims on my bike but I've got to pull the plastic to do it.)

 

It also seems possible that the fast idle lever could somehow be keeping the throttles from fully closing.

 

If you are sure the throttles are fully closed, I suggest taking a ride until the bike is fully warmed up, and then adjusting your idle speed from 1250 rpm to 1100 rpm. After that, let the bike cool down and run another cold start log and see how the leanness looks with the throttle closed between 40 and 60C. Slowing the Closed Loop idle will reduce the air at small throttle angles during Open Loop.

 

 

 

[roger 04 rt]

To add a little support for the excess air theory, I measured my RPM today at 40-50 C engine temperature. In the full up Fast Idle lever position, TPS was 4.16 degrees, same as yours. My RPM was 2400, your RPM is a little higher, around 2850.

 

However, in the mid detent position, we are both about 1.92 degrees but you idle at 2000 RPM and my bike only idles at 1500, indicating to me that your bike is getting more air.

RB

[Trail'R]

Evening Roger,

 

Thanks for taking the time to think this out and I understand where you are headed with your thoughts. Tonight I took some time to inspect the air intake system with the following actions.

 

1. Removed both TB's and inspected butterfly position (both are clean and look to fully close)

2. Check cables for appropriate slack and no tension at close

3. Confirmed TB stop paint is in place on both TB's

4. Checked TB at stops w/ feeler gauge to confirm fully seated

5. Removed air filter and inspected lid and airbox, looks tight.

6. Confirmed both TB's "snap" closed to the stop at the same time.

7. Installed new TB o-rings for a tight seal.

8. Removed TB vacuum hoses and used Twinmax to watch balance and to remove chance of vacuum line leak.

9. Cleaned TPS electrical connections w/ electrical cleaner.

10. Reset TPS via GS-911 - several throttle twists confirmed adjustment.

11. Checked idle lever at all positions for constancy watching TPS angle. All is consistent with repeatable results.

12. Checked for loose butterfly's or bad shafts in TB's. Did not appear to be worn and both are tight.

13. Checked airbox drain to confirm it is tight. Drained excess oil while doing this.

 

With that I ran a new cold start. Here is where it get's interesting. The cold start started out as normal as we have seen in the past. I dropped idle lever at 40C and the bike went lean until closed loop at 60C. It did appear to take longer to warm to 60C, but we are at 56F tonight, so it is cool. It was raining, so I was not able to ride the bike to warm to full temp so I decided to adjust idle speed via BBS once it was warm and while I was logging data. I started to screw in each BBS at around 77C to lower idle to 1100RPM. I did this by adjusting 1/4 turn each screw keeping the TB's close to sync while watching the GS-911 RPM reading. This resulted in a lock on the closed loop once the RPM's dropped to around 1100RPM and no erratic open/closed loop action as in the past.

 

New BBS Settings:

Left - 1 1/4

Right - 1 1/2 (Just less actually ~1/16 under)

 

 

Link to data log #9g along with a master file with some other runs where I was trying to isolate any electrical gremlins. I have an index sheet at the beginning to explain each log.

 

Data log #9g

 

Master Excel File of Data Logs

 

 

 

[Trail'R]

 

Took at warm start log. Pipes were cool to the touch but with some heat present in the engine heads and engine oil was warm.

 

I believe this is the type of log you are use to seeing on the R1150.

 

I need a cold start to confirm.

 

Data log #9h

[roger 04 rt]

Trail'R,

You've nailed it! Congratulations! This took just over two months to figure. You get an A+ for effort.

 

I've scrolled through the data on my ipad and will look at it in excel tomorrow. But from the numbers, it looks very clear to me that you're working, and after one more cold start just to be sure, you can install the original AF-XIED on setting 7.

 

I'm kind of stunned but it all makes sense to me now. I will reflect on it for a couple days and then offer up an explanation for the O2 sensor errors you were getting, especially with the XIED connected. If the next cold start is good, which I expect it will be, your efforts and staying with this have advanced our understanding of the Motronic.

 

The bottom line is that it is key that idle speed, set with the bike warm and in Closed Loop, must be at 1100 RPM +/- 50 RPM for the Motronic to function correctly (and with sufficient Mixture Adaptation range) at small throttle angles.

 

Nice work.

RB

 

[AndyS]

So if it is sorted, can someone some up simply what the problem was, and how it was fixed...because I have not got my head around this thread, but I want to.

Andy

[roger 04 rt]

Good point, after two months and a hundred posts there should be a summary. The next several steps should look like:

 

--Confirm by test that his Cold Start sequence is normal with the Yellow Plug jumper installed. (rich lambda signal to 60C even with a Closed Throttle, no Closed Loop Control signal dropouts, and no O2 sensor faults in the Motronic error log)

 

--Make a riding test and make sure there are no errors in the Motronic log.

 

--Reinstall AF-XIED on setting 7 and confirm proper Cold Start log.

 

--Test ride bike on AF-XIED setting 7 and confirm better driveability than stock (surging reduced or eliminated on setting 6, 7 or 8).

 

Trail'R can then summarize what his problems were and I'd like to take a crack a explaining how the Motronic mishandled this and produced cold-start leanness, Motronic errors, and Lambda Control signal instability, which I should be able to do after the next two tests.

 

 

[Trail'R]

 

Cold start log #9i is below. Nice beautiful waveform with no drops AND no lean start.

 

Too early to claim victory without a few more tests and then we will summarize findings and hopefully put this thread to rest. A few more days and we should be there.

 

Data log #9i

 

[roger 04 rt]

Hi Trail'R,

I've had a good look through the data. Your engine and Motronic are operating exactly as they should: strong O2 signal, good rich Cold Start at any throttle angle, stable Closed Loop Lambda Control, and all sensors operating in the proper ranges. Well done! I will try and post a chart a little later.

 

At this point, you could take a test ride and make sure that you don't get any O2 faults, but I don't really think it's necessary. Your call, I'm happy to look at more data sets.

 

Assuming you've reattached all your electrics that were disconnected, you could attach the original AF-XIED on setting 7 and we could then start to see what AF-XIED settings produce the best combination of smooth running and economy.

RB

[roger 04 rt]

Below is a pair of charts showing Trail'R's R1150GS performance following a Cold Start using the GS-911. This data was taken after installing a new O2 sensor, a Yellow Coding Plug jumper, and then with its Idle Speed adjusted from 1250 RPM to 1100 +/-50 RPM. This is a good example of how an R1150GS (and basically any R1150) should perform from a cold engine to warmed up.

 

The charts are technical, but show what is going on, in detail. This has allowed a diagnosis of the problems we've been working on since May. The charts proceed from left to right. On the left is the data immediately after starting, on the right is data after about 12 minutes of operation. Below is a description of the key observations.

 

Engine Temperature (Engine X10): Starts at 27C on the left and ends at 93C on the right. To make it easier to see on the chart I have multiplied it by 10 so it reads 270 on the left and 930 on the right. This shows the engine warming up.

 

Lamda closed-loop control: When it is at a value of "0", the Motronic is running Open Loop. When it is at a value of "1" the Motronic is running its Closed Loop program and using the O2 sensor as the primary determinant of fueling. It transitions to Closed Loop fueling at 60C engine temperature after a cold start. (Before fixing things, this signal was often unstable above 60C.)

 

Throttle position (TPS): The TPS starts at about 4 degrees open due to the Fast Idle Lever being held fully up. The it is moved to the mid (detent) position which is about 2 degrees open. Then the lever is fully down and the TPS reads 0.32 degrees (normal reading). On the far right of the chart the fast idle lever is briefly moved back up to 2 degrees, then 4 degrees, then back to 0.32 degrees.

 

Injection time: On the left, cranking injection time is about 5 mS per revolution, firing both injectors. After starting the injection time smoothly drops to 2.5 mS and then finally to about 2 mS. When the fast idle lever is lifted on the right of the chart the idle time increases slightly--all normal. (Note: Before adjusting the idle speed down, the injection time was increasing when the Motronic commanded Closed Loop--not normal.)

 

Ignition angle: Starts at 22 degrees on the left when cold and ends at about 10 degrees on the right when warm. This is a normal progression. Notice that when the fast idle lever goes back up on the right, the Motronic responds by increasing the advance to about 15 degrees.

 

RPM: The RPM versus fast idle lever and Closed Loop is now consistent. After starting with the TPS at 4 degrees, the RPM increases from 2000 RPM to 3000 RPM during the first minute after starting. Then when the TPS drops to 2 degrees, the RPM drops to 1700/1800 RPM (it was too fast before adjustment). The the fast idle lever is dropped and the idle speed is about 1100 RPM increasing slightly (with a rich mixture now) to about 1200 RPM until Closed Loop and then settling down slightly to 1100 RPM in Closed Loop operation as the mixture richness is reduced by Closed Loop. (Before adjustment, the idle speed was going UP when Closed Loop started. This was not correct.)

 

Lambda sensor voltage: This line on the chart is the Motronic's measurement of the O2 sensor voltage, measured in millivolts (a value of 1,000 equals 1 volt). This is the key indicator of combustion quality. The Motronic relies on it exclusively for long term fueling. Now the chart is exactly as it should be for a cold start sequence.

 

Starting on the left of the Lambda sensor voltage chart it begins at 450 mV because the O2 (lambda) sensor is cold. When cold it appears absent to the Motronic (due to high resistance). At about the two minute mark, the Motronic turns on the O2 sensor heater (not shown) and a combination of the heater and exhaust heat get the O2 sensor going and it registers about 860 mV, indicating a good RICH mixture. Then, even though the throttle is reduced to 2 degrees and then 0.32 degrees, the O2 sensor output stays "high", meaning rich, right up to 60C degrees engine temperature. This is what is expected.

 

Next the lambda sensor voltage enters a period where you can see the voltage varying between about 750 mV and 100 mV. It changes from high to low every 3 seconds and low to high 3 seconds later for a total Closed Loop period of 6 seconds at slow idle. This is fairly long, but normal for the R1150GS because the O2 sensor is located inside the catalytic converter due to the exhaust pipes coming into the cat separately. (On the 1150RT the sensor is mounted where the two pipes join, ahead of the catalytic converter and the period is under 2 seconds at idle.) During this time, note that the binary Lambda closed-loop control stays at a "1" value. (Before adjusting the idle down and replacing the O2 sensor it was unstable.)

 

Lastly notice that the O2 sensor voltage starts varying more quickly for the time near the end of the chart when the fast idle lever is lifted. Exhaust gas is passing more rapidly through the system and the O2 sensor changes value faster at about 4 seconds period 2000 RPM and 1.5 seconds at 2800 RPM. This is because the Closed Loop process in the Motronic can run faster as the exhaust gas moves faster.

 

Summary

The above description tells us what is happening. Later, after Trail'R makes some added measurements, we can explain what he fixed and why the O2 sensor faults, and low mileage were happening, at the start of the thread, 2 months ago.

RB

[Trail'R]

 

Roger,

 

There is some great knowledge sharing in your last post. Thanks for taking the time to document your interpretation of the charts.

 

The cold starts are a precious commodity and I did not want to waste one this evening without installing the AF-XIED, so that is what I did. This is the same unit I installed earlier. I have not opened the replacement unit should this one work just fine, as I expect it will.

 

I am unsure how to interpret the open loop portion of the start. Is this an expected lambda voltage chart? I did note that the bike remains in closed loop without the drops previously witnessed.

 

Data Log #10 w/ AF-XIED

[roger 04 rt]

Now, relative to your new lambda target set by the AF-XIED, your open loop mixture is lean. I've only had a minute to scroll through the data but open loop will look about like that for a while until mixture adaptation takes effect. Closed loop looks quite good. More tomorrow after I've digested it. Overall it looks fine for the first run after installation.

[roger 04 rt]

Hi Trail'R,

I've had a good look through the data, plotting and analyzing all the same signals as usual. Your bike and AF-XIED are working together just as expected. Now, you need to ride for a while and wait for Mixture Adaptation to do its thing. That will take a couple tanks of gasoline or so. The more steady speeds/RPMs you ride in varying gears,the quicker that process will happen. However, after an hour or so of riding, I'd like you to check the error log.

 

Regarding your data set #10, the Lambda sensor voltage as measured by the Motronic and reported by the GS-911, is no longer your O2 sensor's output. As I know you know, now it is the output of the AF-XIED. In order to do its transformative magic, the AF-XIED has to monitor/filter/etc. (the etc. being proprietary to the designers) the O2 (Lambda) Sensor output voltage. So what you're now seeing reported as Lambda sensor voltage (LSV), although it looks very different to the last run without the AF-XIED, is what I expected.

 

(I'm tempted to just say trust me but let me explain the slightly negative LSV voltages you're seeing just after starting. A Lambda Sensor is what's called a floating, differential source. It has two output leads that float and become referenced to the Motronic's low input lead. As you measured earlier in the thread that low leads is offset from Motronic ground. The AF-XIED is a ground referenced, single-ended voltage source. The result is that when the O2 sensor is in its high resistance mode, the AF-XIED outputs a signal around 100 mV referenced to ground. The low-input to the Motronic is higher than that so the Motronic reads that voltage as negative. Once the O2 sensor is heated and the AF-XIED is doing its thing, it outputs a signal that is above the Motronic's bias.)

 

Once the sensor is warm, when you read the output of the AF-XIED measured by the Motronic, the high voltages (e.g. 750 mV) mean "higher than your AF-XIED setpoint AFR" and the low voltages (e.g. 100 mV) mean "lower than your AF-XIED setpoint AFR". Therefore, during warm-up you see RICH for a while and then LEAN. Lean now means leaner than roughly 14:1, based on AF-XIED setting 7.

 

I watched for Closed Loop adaptation as soon as your Motronic went Closed Loop. It began immediately, adding about 0.3 AFR to your mixture. That means that just prior to Closed Loop your AFR was 14.4:1. That is RICH referenced to the O2 sensor but LEAN referenced to your new target AFR--this means this are working just as planned.

 

I'm very glad you had the patience to work your way through the very subtle issue you uncovered within the Motronic. For now, ride and enjoy, check for errors occasionally, and maybe make a Cold Start log again after a few tanks of gas. Now the thing is to determine which AF-XIED setting give you the best driveability. Start of 7, after a while try 8 and 6, let us know.

RB