Rough idle 30 sec after cold start

[Kerry in Mpls]

My motorcycle is a 2006 R1200RT, about 67000 miles.

 

I have a rough idle after a cold start. It has done this for a long time, but I just haven't tried to find the cause until now. The bike runs great at higher RPM's, good power, good gas mileage. It idles much better (but not perfect) after the bike is warmed up.

 

The interesting thing is that the roughness does not happen immediately after startup. It starts about 30 to 35 seconds after starting the cold bike - it is very repeatable.

 

I made a quick video of it, check it out:

(you can hear it best on a device with speakers that have some bass and can reproduce the lower frequency of the exhaust. Probably harder to hear on a phone or tablet)

 

10 seconds into the video, I start the bike.

35 seconds after that (45 seconds on video timeline) the roughness starts.

 

Could a bad O2 sensor cause this? Maybe the sensors heat up at the 35 second mark, then one (or both) provides bad data back to the engine controller?

 

A GS-911 scan shows no current faults. There are some old logged faults, but nothing related to how the engine runs.

 

Ideas?

 

Thanks.

 

Kerry

 

[dirtrider]

Morning Kerry

 

Yes a bad o2 sensor could cause your problem but you would think that would show a trapped failure code.

 

Usually (but not always) when there is an engine runabilty issue, or an idle issue, AND you have no stored codes then the problem tracks back to a failing stick coil issue or an idle stepper that is not at the counts that the computer thinks it is at.

 

If the problem goes away after running for while then possibly look for an oil fouled lower spark plug or a failing lower stick coil.

 

At higher RPM's the lower spark plugs & coils basically do nothing so if it runs good at higher RPM's then that again more points to lower spark plugs or lower stick coils.

 

A leaking intake boot can also cause some idle issues but that usually increases the idle RPM's on a hot engine.

 

Your problem is going to be difficult to diagnose over the internet so about all you can do is work your way through the basics.

[Don_Eilenberger]

Morning Kerry

 

Yes a bad o2 sensor could cause your problem but you would think that would show a trapped failure code.

Afternoon DR,

 

In my experience (and it's been with multiple O2 sensors) there is no code stored when a sensor goes bad. None. The only way I found to determine if the sensors are good is to monitor their output with the GS-911 in graphic/plotting mode.

 

The heated sensors used on the hex (and cam) head bikes start working within about 15 seconds of the engine being started. The heater in them brings them to temperature very quickly. If a heater went bad, a sensor could be slow in responding and take time to be heated up by the exhaust gases. If that happened, that might account for the symptoms that are described.

 

The ECU is also designed so that if one O2 sensor fails - it applies the output from the other O2 sensor to both cylinders. I imagine if both fail it drops back to a stored fueling map.

 

Usually (but not always) when there is an engine runabilty issue, or an idle issue, AND you have no stored codes then the problem tracks back to a failing stick coil issue or an idle stepper that is not at the counts that the computer thinks it is at.

I'm probably the exception to this.. but I've never had a stick coil or stepper fail on me. I have experienced 3 O2 sensor failures. Luckily there are some generic 4 wire ones that can substitute for the rather expensive stock one.

 

If the problem goes away after running for while then possibly look for an oil fouled lower spark plug or a failing lower stick coil.

 

At higher RPM's the lower spark plugs & coils basically do nothing so if it runs good at higher RPM's then that again more points to lower spark plugs or lower stick coils.

 

A leaking intake boot can also cause some idle issues but that usually increases the idle RPM's on a hot engine.

 

Your problem is going to be difficult to diagnose over the internet so about all you can do is work your way through the basics.

 

It is rather frustrating that the bike's ECU doesn't store data like misfires or O2 sensors acting up. It makes diagnostics more of a "replace and see" sort of game, which I hate.

 

I would suggest Kerry setup his GS-911 to plot O2 sensor output from a cold startup and see if anything appears. It may be possible to spot a rich mixture caused by a stepper failure or coil misfiring by watching the O2 sensor output. It's certainly possible to watch the O2 sensors for normal output.

 

If it was possible to upload and display an image here I'd post a screen capture of a good O2 sensor output on an R1200 motor, but since we can't I won't.

[Kerry in Mpls]

I tried an experiment.

With the engine cold, I turned on the key, let the bike do its power-on dance, and heard the steppers click to what I presume was their proper starting position. Then I unplugged the stepper wires and started the bike.

The stumbling started after 35 seconds as it always has, there was no difference in behavior.

 

But I'm not sure how to interpret that. Is seeing the same behavior good or bad? :-)

 

I also removed both stepper units and checked the "plungers" on each for scoring or crud - both were clean and smooth. They have spring-loaded sleeves that can be slid back -- there was grease on the shaft under the sleeves, and it was still liquidy, not dried up. There was also no gunk or wear marks in the bores on where the steppers normally sit. The O rings were also both in good shape. When installed, I can clearly hear them activating when turning on the key, and they seem to be even more active for a period after turning off the key.

 

The GS-911 that I used to check for error codes isn't mine, but I'll see if I can use it again to get O2 sensor output plots.

 

[roger 04 rt]

With the motorcycle cold, set up the gs-911 to log all realtime values to a CSV file. Then after a 3-4 minutes of data has been collected, stop logging and save the file. I'd be happy to look the file over but the idea is to see what data looks wrong, if any. If all the data looks good then you might think coils.

[Kerry in Mpls]

Will do, thanks Roger.

 

[dirtrider]

I tried an experiment.

With the engine cold, I turned on the key, let the bike do its power-on dance, and heard the steppers click to what I presume was their proper starting position. Then I unplugged the stepper wires and started the bike.

The stumbling started after 35 seconds as it always has, there was no difference in behavior.

 

But I'm not sure how to interpret that. Is seeing the same behavior good or bad? :-)

 

I also removed both stepper units and checked the "plungers" on each for scoring or crud - both were clean and smooth. They have spring-loaded sleeves that can be slid back -- there was grease on the shaft under the sleeves, and it was still liquidy, not dried up. There was also no gunk or wear marks in the bores on where the steppers normally sit. The O rings were also both in good shape. When installed, I can clearly hear them activating when turning on the key, and they seem to be even more active for a period after turning off the key.

 

The GS-911 that I used to check for error codes isn't mine, but I'll see if I can use it again to get O2 sensor output plots.

 

Morning Kerry

 

That test isn't 100% conclusive but is pretty darn close as the steppers reset at every fresh key on. (unless a totally dead stepper they usually reset correctly, at least on first key on cold)

 

If you can't easily get that GS-911 back then you might try completely disconnecting the o2 sensors then try the cold start again (that takes the o2 sensors completely out of the cold start, cold warm-up sequence. (be real careful of the o2 sensor connectors as the little lock is easy to damage)

 

If disconnecting the o2 sensors stops the problem then try reconnecting them one at a time then try another cold start. (if you have a bad o2 sensor this might tell you which one)

 

If it still acts up with the o2 sensors disconnected then leave them disconnected & remove the (-) battery cable for about 30 minutes as that will clear the fueling computer adaptives. Then try another cold start.

 

If starting with both o2 sensors disconnected & the adaptives cleared doesn't eliminate the cold start issue than your o2 sensors are not the problem.

 

[lkchris]

The interesting thing is that the roughness does not happen immediately after startup. It starts about 30 to 35 seconds after starting the cold bike - it is very repeatable.

 

This is classic ... this is when your "choke" shuts off. More to the 21st century, it's where electronic enrichening stops.

 

Usually when a cold engine won't run without enrichening it's a vacuum leak. Goes from rich to too lean ... instead of normal.

Edited October 13, 2016 by lkchris

[Kerry in Mpls]

I will be able to use the GS911 again, friend bringing it over probably tomorrow (Fri) evening, we'll log a bunch of data on a cold start.

He sent me a small sample of data collected from his 07 R12RT. I can see a few lambda data items: heater on/off, sensor voltage, and control factor. I presume these are the values that would indicate whether the O2 sensors are doing their thing.

 

[Kerry in Mpls]

This is classic ... this is when your "choke" shuts off. More to the 21st century, it's where electronic enrichening stops.

 

Usually when a cold engine won't run without enrichening it's a vacuum leak. Goes from rich to too lean ... instead of normal.

 

Aha, I was hoping that someone would recognize something special about that duration. If the data log seems to indicate the O2 sensors are OK, will probably start looking for a leak.

[dirtrider]

Morning Kerry

 

Yes, many things can be derived from the logged data but it isn't always what it seems at first look.

 

Some logged data is just commanded info (like Lambda Sensor Heating or Tank Valve Venting)-- that doesn't actually tell you that the o2 heaters or tank vent are on or working. That just tells you that the fueling computer has commanded them to be on.

 

Things like Lambda Sensor Voltage (cyl-1) or Cyl-2) can be very useful but again it needs to be interpreted by someone familiar with reading that type data as voltage shift could mean a bad or lazy sensor or could be pointing to a vacuum leak or rich combustion with the o2 sensor output reporting what it sees correctly.

 

Logged data can be a very good indicator of what's going on just be careful that you understand WHY the data that you see is what or where it is.

 

 

[Kerry in Mpls]

Follow the link to see a bunch of graphs of data from a GS-911 data capture, just under six minutes in length.

https://goo.gl/photos/hkhZb6nerAKPSWc47

You can use the magnifying glass icon and thumbnail at top-right to zoom in and out and pan around.

I only collected data items I thought could be of interest for this specific issue.

My untrained eye says most of it looks pretty good. There is a spike in the ignition angle and also in the ignition dwell time, but only one each, so probably not important.

 

Something that does catch my eye is the graphs for additive trim bank 1 vs bank 2. I presume they are for cylinder 1 and 2. They are quite different. I don't fully understand what those are, or if there is any significance to the difference.

 

Thoughts?

 

Edited October 15, 2016 by Kerry in Mpls

[roger 04 rt]

Nice job!

 

Did the roughness start 30 seconds in as usual?

 

Bank 1 is the right cylinder, bank 2 the left.

 

Most of the charts look good but if you could say at exactly what point the roughness starts it would help.

 

Btw, your O2 sensors are working as they should and the lambda control factors and multiplicative trims are fine.

 

Additive trims can be different near idle due to canister operation. Did you record the canister valve binary? I found it helpful to always record all data.

Edited October 16, 2016 by roger 04 rt

[Kerry in Mpls]

New data capture today, this time collecting all data items.

Graphs:

Entire capture, nearly seven minutes in length: https://goo.gl/photos/tPbbSR6kHkutFcs56

Same data, but just first 70 seconds or so for better resolution near the start: https://goo.gl/photos/sxnij6wVdQHA8Hwy8

 

I also shot a video at the same time.

https://www.youtube.com/watch?v=N74pnBcpeso

 

The bike is started 5 seconds into the video. Since I started logging data a few seconds before that, this corresponds to about the 9 second point in the graphs (where engine RPMs rise from zero). The rough idle begins 35 seconds after starting the engine, which I'd say is about the 44 second point in the graphs.

 

Edited October 16, 2016 by Kerry in Mpls

[roger 04 rt]

Great job with the data collection and graphs. What took are you using to produce the graphs on one page like that?

 

Looking at the graphs very closely, nothing is going wrong in the first 70 seconds. It sometimes helps to start the RPM graph at 1000 rpm so that the scale is better but sensor-wise, everything is well behaved.

 

What it looks like is that your BMSK is going from after-start enrichment mode into active Closed Loop at the moment it starts running poorly. That means the mixture is getting leaner and harder to combust. As the engine heats it will be more able to burn the leaner closed loop mixture but at the cold Engine Temp at 35 seconds of running its hard to run smoothly. You can see the richness as the O2 sensor starts to come to life.

 

Let me be clear, it is not closed loop that is your problem, not the O2 sensors, which are fine, it's just the leanness and cold engine.

 

It's not the steppers as they stay the same, and the idle speed is fast enough.

 

Could you tell us what the dwell value is from 0-100 seconds? It looks like much less than 1 mS. I'm asking because your 2006 dwell angle behaves like my 2004 twin spark, with a very low dwell when the engine is cold. On mine it looks like a software bug in the Motronic. Later R1200s use a longer dwell time during cold starting, from memory it's between 2 and 3 mS while the bike warms up.

 

The dwell time, which is the time that the ignition coil charges, normally runs at 1 mS on the R1150 (lost spark fires twice per combustion cycle) and 2 mS on the R1200 (fires once per combustion cycle). At less than 0.5 mS, a weak coil may not fire the mixture well when the engine is cold. Your battery voltage is fine so there's enough voltage at the coil.

 

When closed loop begins, the BMSK advances the timing a little, as it should to help with combustion.

 

Based on your data and my experience, I think you have an ignition problem. One or more of your four coils are weak or you have a bad plug. During a cold start you need all four plugs firing well for smooth running.

 

The other possibility is an injector with a bad spray pattern. However, your additive and multiplicative trims (which, as a total aside, prove conclusively that boosterplugs don't do anything) look normal and fine.

 

Edited October 17, 2016 by roger 04 rt

[dirtrider]

Morning Kerry

 

I don't mean to step in between you & Roger on this so I won't address your fine looking trapped data.

 

If you think that your problem is pointing to a bad stick coil you m-i-g-h-t be able to find it with an old AM transistor radio (remember those?).

 

Put the (AM) radio on an off-station (ie just basic low background static)-- then with your engine running (especially during cold warm up) place the radio near each coil & listen for popping in cadence to the firing, or increased ignition sounding static. (if you have one coil with way more static than the others that is probably the bad one)

 

The usual failure in a stick coil is the coil arcing internally to the RFI shield & that can broadcast to an AM radio.

 

You should probably try 4 new spark plugs JUST to eliminate that possibility.

 

One other thing that I forgot to put in my first post above is: IF your problem showed up after filling the fuel tank (like wasn't there at all before fill-up & suddenly appeared next morning after tank fill-up) then possibly suspect water in the fuel.

 

When riding under power & speed the engine can digest some water mixed with the gasoline. Then when it sits overnight the water drops to bottom of the fuel tank out of pickup range. Then once started cold the fuel return flow can stir the water back into the fuel supply & cause rough idling (especially when it goes into closed loop). Kind of stretch but something to think about if the problem just showed up suddenly.

[roger 04 rt]

Interestingly, my coils passed every test I could think of but the bike idled and started much better after new coils.

 

Hopefully Kerry has a friend and can borrow them. I too would do the plugs first.

Edited October 17, 2016 by roger 04 rt

[Kerry in Mpls]

Great job with the data collection and graphs. What took are you using to produce the graphs on one page like that?

 

The tool I've been using to plot these graphs is KST2, available here: https://kst-plot.kde.org/

Besides static plots, it can also monitor a live data file, and update plots in real time, like an EKG. That requires that the application that is generating the data writes to the file in a reasonable manner, e.g. flushes or closes and reopens the file periodically. I didn't try that with a live GS911 data file yet, but I may just to see if it works. That would permit having any data items you want in nice big separate graphs crawling across a monitor in whatever range and timescale you want. Could be handy.

[Kerry in Mpls]

Roger -

Four new plugs were installed in July of 2015 when I had a few other service items performed at the dealership. The rough idle occurred both before and after that service, so new plugs didn't affect it. (I did not ask them to address the rough idle)

The same guy that loaned me his GS911 texted me this AM asking if I wanted to borrow his coils (07 R12RT). He is following this thread and must have seen your post before I did - heh. So maybe we'll get around to that this coming weekend.

 

In the meantime, here's a better graph of dwell time and RPM's.

https://goo.gl/photos/CYSGx4ZHJdPEpEa37

 

Let me know if you want a closer look at anything else.

 

D.R. -

This is a long term symptom, more than a year. That should rule out temporary fuel quality issues or water in the tank.

The AM radio idea is clever -- since I may get a chance to experiment with known good coils soon, I'll keep that as plan B.

 

Thanks both of you for your time and expertise.

 

[roger 04 rt]

Great job with the data collection and graphs. What took are you using to produce the graphs on one page like that?

 

The tool I've been using to plot these graphs is KST2, available here: https://kst-plot.kde.org/

Besides static plots, it can also monitor a live data file, and update plots in real time, like an EKG. That requires that the application that is generating the data writes to the file in a reasonable manner, e.g. flushes or closes and reopens the file periodically. I didn't try that with a live GS911 data file yet, but I may just to see if it works. That would permit having any data items you want in nice big separate graphs crawling across a monitor in whatever range and timescale you want. Could be handy.

 

Thanks for the link, I'm going to look into it.

[roger 04 rt]

Roger -

Four new plugs were installed in July of 2015 when I had a few other service items performed at the dealership. The rough idle occurred both before and after that service, so new plugs didn't affect it. (I did not ask them to address the rough idle)

The same guy that loaned me his GS911 texted me this AM asking if I wanted to borrow his coils (07 R12RT). He is following this thread and must have seen your post before I did - heh. So maybe we'll get around to that this coming weekend.

 

In the meantime, here's a better graph of dwell time and RPM's.

https://goo.gl/photos/CYSGx4ZHJdPEpEa37

 

Let me know if you want a closer look at anything else.

 

D.R. -

This is a long term symptom, more than a year. That should rule out temporary fuel quality issues or water in the tank.

The AM radio idea is clever -- since I may get a chance to experiment with known good coils soon, I'll keep that as plan B.

 

Thanks both of you for your time and expertise.

 

Based on the update plot, your dwell time is just under 2 mS, which is good. It was harder to read on the other plot and I missed the division size.

 

Still, coils and injectors seem like the place to look first. As I mentioned, the only differences before ~30 seconds and after than I can see are the leaner mixture and slight change to spark advance.

[Ponch]

Roger -

Four new plugs were installed in July of 2015 when I had a few other service items performed at the dealership.

 

 

 

Are you sure they changed them?

Edited October 18, 2016 by Ponch

[Kerry in Mpls]

Roger -

Four new plugs were installed in July of 2015 when I had a few other service items performed at the dealership.

 

 

Are you sure they changed them?

Yes.

[James L.]

You know, I've been following this thread, and not being a BMW guru, I've basically left the posts to people who know the bike better, but I do have an idea, that someone else can hash over and tell me I'm full of it.

 

Could this be a semi-clogged fuel injector?

 

Think about it, when running rich at start up the tendency to run rough would not totally be there, and when warm the fuel is evaporated by the heat, therefore easier to ignite. When cold a messy spray pattern would be hard to ignite, and would make a rough running engine.

 

Now, I'm not a BMW mechanic, or know much about the BMW, but it is one thing you guys haven't talked about.

 

James L

Edited October 19, 2016 by James L.

[roger 04 rt]

You know, I've been following this thread, and not being a BMW guru, I've basically left the posts to people who know the bike better, but I do have an idea, that someone else can hash over and tell me I'm full of it.

 

Could this be a semi-clogged fuel injector?

 

Think about it, when running rich at start up the tendency to run rough would not totally be there, and when warm the fuel is evaporated by the heat, therefore easier to ignite. When cold a messy spray pattern would be hard to ignite, and would make a rough running engine.

 

Now, I'm not a BMW mechanic, or know much about the BMW, but it is one thing you guys haven't talked about.

 

James L

 

Yes, I'd have to agree it's a possibility, but the BMSK has a pretty good ability to us the O2 sensors to compensate unequal injectors. See below

 

Great job with the data collection and graphs. What took are you using to produce the graphs on one page like that?

 

Looking at the graphs very closely, nothing is going wrong in the first 70 seconds. It sometimes helps to start the RPM graph at 1000 rpm so that the scale is better but sensor-wise, everything is well behaved.

 

What it looks like is that your BMSK is going from after-start enrichment mode into active Closed Loop at the moment it starts running poorly. That means the mixture is getting leaner and harder to combust. As the engine heats it will be more able to burn the leaner closed loop mixture but at the cold Engine Temp at 35 seconds of running its hard to run smoothly. You can see the richness as the O2 sensor starts to come to life.

 

Let me be clear, it is not closed loop that is your problem, not the O2 sensors, which are fine, it's just the leanness and cold engine.

 

It's not the steppers as they stay the same, and the idle speed is fast enough.

 

Could you tell us what the dwell value is from 0-100 seconds? It looks like much less than 1 mS. I'm asking because your 2006 dwell angle behaves like my 2004 twin spark, with a very low dwell when the engine is cold. On mine it looks like a software bug in the Motronic. Later R1200s use a longer dwell time during cold starting, from memory it's between 2 and 3 mS while the bike warms up.

 

The dwell time, which is the time that the ignition coil charges, normally runs at 1 mS on the R1150 (lost spark fires twice per combustion cycle) and 2 mS on the R1200 (fires once per combustion cycle). At less than 0.5 mS, a weak coil may not fire the mixture well when the engine is cold. Your battery voltage is fine so there's enough voltage at the coil.

 

When closed loop begins, the BMSK advances the timing a little, as it should to help with combustion.

 

Based on your data and my experience, I think you have an ignition problem. One or more of your four coils are weak or you have a bad plug. During a cold start you need all four plugs firing well for smooth running.

 

The other possibility is an injector with a bad spray pattern. However, your additive and multiplicative trims (which, as a total aside, prove conclusively that boosterplugs don't do anything) look normal and fine.

 

 

[James L.]

 

Yes, I'd have to agree it's a possibility, but the BMSK has a pretty good ability to us the O2 sensors to compensate unequal injectors. See below

 

I'm not talking about unequal injectors. I'm talking about spray pattern. These are not the same thing.

 

Equal means quantity. I'm talking about an atomized pattern igniting quickly versus a blob of fuel which ignites slower (only one injector with bad spray pattern). They can be the same quantity, and use the same amount of O2, but still burn at different rates. The rate at which fuel burns will effect the power of the pulse for that cycle.

 

But I may be full of it.

 

James L

[roger 04 rt]

You're not full of it but see my text in red above.

[Kerry in Mpls]

I still hope to swap in known-good lower coils soon and see if that makes a difference.

I'm ready for the mystery to be solved...

[Kerry in Mpls]

Well, that didn't work.

 

My friend Paul loaned me all four coils from his 07 R12RT to try in my bike.

I replaced both right side coils (with my left side coils still installed) and started the bike (cold), and the rough idle appeared right on schedule, approx. 35+ seconds after startup. I shut down the bike less than 15 seconds after it was clear that it was idling roughly, to try not to build up much heat.

 

OK, the problem is apparently not right side coil(s).

 

With Paul's right side coils still installed, I installed his left side coils, and started the bike again. 35 seconds came and went with no rough idle, and I thought, hmm maybe? But no, around 55 seconds the familiar rough idle began again, chunky as ever. I don't know why 55 seconds vs 35 seconds, I suppose it was related to some residual heat left over from the previous run. There were maybe 10 minutes between engine runs. The first run was short enough that the valve covers were still cold.

 

It was getting late, so I just reinstalled my coils and called it a night.

 

As I mentioned earlier, the plugs were replaced only last year; would there be any value in pulling them and inspecting them, and if so what might I look for?

 

Any thoughts on next steps?

 

Kerry

[roger 04 rt]

Well that's too bad but it's good to have ruled out the coil possibility. If it were me, given the low cost, I'd install four new plugs to rule that out too. After that I'd send my injectors out for cleaning, including before and after flow test reports.

 

If you'd collected all realtime data, there might be an explanation for the delay to 50 seconds. Maybe in your scenarios, closed loop occurred later after the second start.

 

Does your bike have any modifications?

Edited October 22, 2016 by roger 04 rt

[RobinW]

Does your GS-911 show the actuators going out of balance when it happens?

 

I just managed to hear the utube on a home PC with sound instead of a fone.

 

Sounds like a poor spark to me?

[roger 04 rt]

Kerry, One other question, does the bike idle smoothly when it is fully warmed up?

[roger 04 rt]

Although you've swapped coils, see what happens if you run the same test as you've been running, first only with the upper coils connected and then only with the lower coils connected. It's a shot in the dark but might tell you something.

[Kerry in Mpls]

Does your GS-911 show the actuators going out of balance when it happens?

 

I just managed to hear the utube on a home PC with sound instead of a fone.

 

Sounds like a poor spark to me?

 

The idle actuators are operating in close sync with each other, so they seem to be fine. See links to graphs posted earlier if interested (e.g. 10/16).

 

Yes maybe haven't ruled out spark yet.

[Kerry in Mpls]

Kerry, One other question, does the bike idle smoothly when it is fully warmed up?

The engine idles pretty smoothly after warming up, but I can't say it is absolutely flawless. There is a tiny burp from the exhaust every few seconds which I've always considered normal. Small enough that I can't feel it when sitting on the bike, I only notice/hear it when standing near the exhaust. But maybe it's worth asking - should a 1200 hexhead idle completely flawlessly, or are tiny burps like that to be expected?

 

Bike has no modifications other than power for some electrical gadgets (GPS and power lead for heated liner).

 

I should probably mention that after the most recent big data capture (back on 10/16) I reset the adaptives using the GS911. I figured why not let those start those over from the current state of the bike. The GS911 app indicated that some relearning would be necessary after the reset and further instructions would follow. After the reset, all it said is that I needed to operate the bike in each gear for at least 10 seconds to relearn the gear positions -- no info about any other learning. So I took a short ride to do that. The bike has probably less than 10 more miles on it since the reset. There seemed to be no noticeable change in the cold startup behavior after the reset.

 

Also, we observed that the GS911 showed that the ECU firmware has never been updated:

Factory I-Level: K024-05-08-510

Actual I-Level: K024-05-08-510

 

Could that be that significant?

 

[Kerry in Mpls]

Although you've swapped coils, see what happens if you run the same test as you've been running, first only with the upper coils connected and then only with the lower coils connected. It's a shot in the dark but might tell you something.

I had the very same thought, will definitely try it.

Edited October 22, 2016 by Kerry in Mpls

[Kerry in Mpls]

Latest experiment / results:

 

I disconnected the upper coil leads and started the bike.

The engine immediately ran about as rough as it has previously upon reaching the 35 second point on all coils. And when it did reach the 35 second point, it got even a little bit rougher. I shut it down within a minute.

 

Half an hour later, I reconnected the uppers, and disconnected the lowers and started the bike.

The engine immediately ran extremely rough, in fact was barely running.

I tried to give it a bit of gas to see if it would improve, but it stalled.

Definitely not happy without the lowers!

 

I have to correct an earlier comment I made.

After the second experiment above, I reconnected all coils, started it, and let it run until it warmed up. Probably less than five minutes. I have some clutch rattle that I was earlier confusing for engine noise. Pulling the clutch lever quieted that down, and the engine idled very smoothly.

 

Warm engine = happy engine.

 

[roger 04 rt]

Kerry, Based on the recent data, here's what I'm thinking:

 

1. When the engine is cold, it idled well with a richer mixture.

 

2. While still cold, as the BMSK goes closed loop and leans off the mixture, idling is rough.

 

3. When the engine is hot, it idles acceptably with the lean closed loop mixture.

 

4. Youve swapped coils, try new plugs, it's easy and not expensive. Hopefully someone else reading the thread can comment on the disconnected coil test.

 

5. One of the reasons for richer mixtures in a colder engine is that the fuel doesn't vaporize well, condenses on the intake and valves and is harder to ignite. Therefore the excess fuel. If an injector has a contaminated nozzle, and bad spray pattern you could get poor cold running. The additive trims are more important at idle and they're different. It could suggest an injector issue but I don't have enough experience interpreting them.

 

I would try new plugs, or I would try your disconnected coil test one at a time to see if one coil or cylinder has less effect when you remove the one coil.

 

Then if that didn't yield insight I would replace the plugs, then test, clean and retest the injectors. Or borrow the injectors from your friend who loaned you the coils.

[NickInSac]

Latest experiment / results:

 

I disconnected the upper coil leads and started the bike.

The engine immediately ran about as rough as it has previously upon reaching the 35 second point on all coils. And when it did reach the 35 second point, it got even a little bit rougher. I shut it down within a minute.

 

Half an hour later, I reconnected the uppers, and disconnected the lowers and started the bike.

The engine immediately ran extremely rough, in fact was barely running.

I tried to give it a bit of gas to see if it would improve, but it stalled.

Definitely not happy without the lowers!

 

Warm engine = happy engine.

 

This sounds like you have a mix of bad intermittent upper and lower coils (for instance, a bad lower on the left side, and a bad upper on the right side). But, you changed the coils from a friends bike with no changes to idle at 35 seconds. Seems really weird that they fail right at 35 seconds. I'd expect them to be more intermittent.

 

And no fault codes?!? No stored codes?

 

[roger 04 rt]

Nick, It seems like it's not the coils (maybe a connector?) and the reason it happens at 35 seconds is as I explained above, due to the lean-off associated with Closed Loop in a cold engine.

[Kerry in Mpls]

Update:

I was looking through the various GS911 data captures, and looking specifically at cylinder head temps, I noticed that cylinder 1 (right) ran slightly warmer in all cases. The difference starts out very small (as one would expect), and diverges as the engine gets up near normal running temp. It ends up being between 5 and 10 deg F difference, right cylinder higher than left. The log data from a friend's 07 R12RT shows the temps in the two cylinders are exactly the same, all the way from cold up to near running temp.

 

A difference in running temps might indicate a leaner condition on one side only. There has been some talk of possible injector issues, so I decided to see if temperature difference followed the injector. I swapped the injectors side-to-side, then did another engine run and data capture. The temperature difference remained, and cylinder 1 (right) still ran warmer.

 

Ok, while you are all chewing on that information, check out this photo of the plugs.

https://goo.gl/photos/ju1wgxte9y8FpE9i6

From left to right, their installation location was Upper Left, Lower Left, Lower Right, Upper Right.

It was tough to get the flash and focus just right to accurately reproduce the actual colors.

But you can see the two lowers are noticeably darker than the uppers, especially the Lower Left. Is that expected due to their location?

And the Upper Right is lighter than the Upper Left plug, perhaps supporting the lean-running right cylinder theory from above.

Even though these are only a year old, I have a new set ready to go in.

 

I plan to do a series of disconnected coil tests e.g. uppers, then lowers), and observe the effect on idle smoothness as well as operating temp. I did something like this earlier, but I want to repeat it in a more methodical way so that comparing results of the runs is more valid. Ideally, each test should start cold, and since the cooldown takes a while, it could be the weekend before I can put in the time to do that.

 

Edited October 25, 2016 by Kerry in Mpls

[Kerry in Mpls]

And no fault codes?!? No stored codes?

Nope!

[dirtrider]

Update:

I was looking through the various GS911 data captures, and looking specifically at cylinder head temps, I noticed that cylinder 1 (right) ran slightly warmer in all cases. The difference starts out very small (as one would expect), and diverges as the engine gets up near normal running temp. It ends up being between 5 and 10 deg F difference, right cylinder higher than left. The log data from a friend's 07 R12RT shows the temps in the two cylinders are exactly the same, all the way from cold up to near running temp.

 

A difference in running temps might indicate a leaner condition on one side only. There has been some talk of possible injector issues, so I decided to see if temperature difference followed the injector. I swapped the injectors side-to-side, then did another engine run and data capture. The temperature difference remained, and cylinder 1 (right) still ran warmer.

 

 

Morning Kerry

 

You can't compare your 06 cyl head temp data to your friends 07 cyl head temps as your 06 1200RT STILL has a cyl temp sensor on each cyl head so on your bike you are reading each cyl head temp independently.

 

On your finds 07 1200RT, that bike ONLY has one cyl head temperature sensor so the fueling computer only gets one temp input then shares that one input as a common to both sides (any of the hexheads 2007 & up only have one cyl head temp sensor & will always show the same temp on both cyl heads)

 

I have seen a slight side to side variance in the earlier 1200RTs, some is probably due to a breeze blowing more on one side of the engine, some to slight combustion differences, & some to ????

 

Your cyl head temp difference might very well be pointing to some minor side to side firing difference but could also just be within normal variances.

 

I have some GS-911 trapped data of earlier 1200RTs & most do have a slight side to side cyl temp difference at hot curb idle. (this is why I like GS-911 Excel data better than graphs as you can compare the data line by line with good number resolution)

 

On the spark plug pictures-- those are difficult to tell much from as they are from a cold start idled engine. To read much into them you need to do a 50+ mile high speed run then do your cold start until bad idle THEN remove & read the plugs. Or install new spark plugs then do your cold start until bad idle then read THOSE plugs.

 

I see a lot of 1200RT spark plugs that look like yours if the bike is cold started & idled just before the plugs are removed (even from good running good idling engines)

 

You can probably use your cyl head temps to tell you something if you run the engine on just the upper coils then take cyl head temps --then run the engine on just the lowers & rear cylinder head temps.

 

You might get a decent cyl head temp base line by reading the cylinder head temp delta after a good bike ride & short idle (that should give you the cyl head temp delta on a decent idling engine (do it away from any breeze or fan air flow)

 

 

[roger 04 rt]

There is nothing unusual about 5-10 degree CHT differences, we see that all the time on aircraft.

 

As DR said, there is nothing unusual looking about your plugs. As I've suggested a couple times I'd replace them before doing any more testing. It may show nothing but the cost is low and it would rule that out.

Edited October 25, 2016 by roger 04 rt

[Kerry in Mpls]

OK, good to know the plugs are probably fine, and the cylinder head temp difference is probably not important.

 

Onward...

I installed new plugs today (probably not necessary, but why not), then tried the next experiment: I disconnected both O2 sensors and started the bike, stopwatch in hand.

The "magic" 35 second point came and went, and the idle stayed smooth.

One minute... two minutes... at three minutes I shut it down.

 

Well, I seem to finally have altered something that "fixes" the specific behavior I have been chasing.

The easy answer seems to be to replace one or both O2 sensors, but I'm wondering if disconnecting the sensors could have altered something else, and the original problem is still there, just masked.

 

As always, input is appreciated.

 

[dirtrider]

Well, I seem to finally have altered something that "fixes" the specific behavior I have been chasing.

The easy answer seems to be to replace one or both O2 sensors, but I'm wondering if disconnecting the sensors could have altered something else, and the original problem is still there, just masked.

 

 

Evening Kerry

 

That sure is possible-- if the engine doesn't have active o2 sensors it can't go into closed loop so stays a bit richer at the 35 second mark.

 

But you do have a smoking gun so-to-speak so maybe try working that angle for a few more tests.

 

You might try only disconnecting (1) side o2 sensor but connect the other side. Then run your 35 second test.

 

Then reverse that & reconnect that side & disconnect the other & run your 35 second test.

 

See if this points to one side or the other o2 (or o2 related) as the cause.

 

[roger 04 rt]

As DR said, disconnecting both O2s left the mixture rich, most likely masking the issue. However trying that test one side at a time would be interesting. If one side was worse than the other, that would be telling. If that happened, then you could swap O2 sensors and see if the side changed. You could also change injectors side to side.

 

I forget, have you done a compression test?

Edited October 29, 2016 by roger 04 rt

[Kerry in Mpls]

OK, two more runs done.

With only the left side O2 sensor disconnected, the idle is near normal.

With only the right side O2 sensor disconnected, the rough idle is back.

(there were a couple of hours between tests, so both cold were startups)

 

I'll try swapping and repeating.

 

I have not done a compression test, Roger.

I did swap injectors side to side last weekend with no effect.

 

Edited October 29, 2016 by Kerry in Mpls

[roger 04 rt]

Good work. Since you get a rough idle with the O2 sensor on the right side connected, that sounds like the problem side--it's the side getting leaner. I would now swap left and right injectors and repeat the O2 tests and see if the problem moves to the left side.

 

If that has no effect, I would then swap the O2 sensors side to side.

 

Edited October 30, 2016 by roger 04 rt

[Don_Eilenberger]

Good work. Since you get a rough idle with the O2 sensor on the right side connected, that sounds like the problem side--it's the side getting leaner. I would now swap left and right injectors and repeat the O2 tests and see if the problem moves to the left side.

 

If that has no effect, I would then swap the O2 sensors side to side.

Roger - right oh, except I think you got the sides mixed (I did too when I first read it.)

 

One thing to point out - on a R1200xx when an O2 sensor fails - the mixture for both cylinders is set to the mixture determined by the working O2 sensor. With both disconnected (or failed) - as noted - the mixture defaults to a map in the ECU, and the maps tend to run rich to protect the engine.

 

Same thing as the CAT-CODE plug did on the R1150xx O2 sensor equipped engines. Pull the CAT-CODE plug and it took the O2 sensor out of circuit and the bike ran under a mapped mixture. Surge usually eliminated.