My personal "Leave the house riding a motorcycle" limit is 35* and rising. I try not to ride under 40* though.
I have a Turntech LiFePO in my Beta, and it starts at 40 degrees with no problem at all. And, last winter when I was busted up from a crash, it sat for a couple months with no charger... fired right up when I tried it. When I need to replace my stock TR650 battery, I'm going LiFePO. I've read stories of unreliability, but I've had a few and they have all worked better than I expected.
I have Shorai LiFe batteries in 3 bikes now (Guzzi Stelvio, Ducati Sport Classic, and Suzuki DRZ400-S) and am very happy with them. They crank the engine over very nicely and seem to hold a charge quite well. I friggen love the weight savings as most of my bikes have the battery located fairly high just under the seat. In addition, they don't ever leak acid like many others will with a canted battery position (big issue on the Stelvio). Now I do live in SoCal so very low temps are not part of my riding scenario, which I'm thankful for, so take that into consideration.
This next link will take you to a complete ECU reset and a log that was run with the o2 sensor uplugged. Roger asked for this and I presume this run is to see if we get the same oscillations in the OEM limp home mode as we did at 13.5.1 AFR. If you want to view the log you will have to download Log Works found under the "shipping" section in the link I provided. It's interesting to note that "limp home mode" set the average AFR to 13.17:1 and it is also nice to know if the o2 sensor goes south I will be able to make the bike run.
After running a log with the disconnected o2 and reset ECU, Roger asked me to run a second log with the o2 sensor reconnected and using the same settings as this past weekend (13.5:1). HERE IS THAT LOG. This is what the AFR distribution looks like. Roger says it is a good one and that distribution looks great. As soon as I speak to Roger I will know more as to why it looks good. I'm gonna plead ignorance here.
Roger is interpreting the latest reading on my recent log. His commentary copied from ADV: Below is a chart made from TRZ_Charlie's TR650, with an Innovate Motorsports LC-2 measuring AFR ahead of the catalytic converter. I will add a few more charts that expand portions of the log to illustrate how the MM ECU is managing fueling. The LC-2 is also programmed to send a simulated, narrowband signal to the TR650's ECU using one of the LC-2's Analog outputs. It has been programmed to simulate an AFR of 13.5:1, which is 7% richer than stock. It looks to me like Charlie has a successful integration of the LC-2 and his bike but I also see what look like weaknesses in the MM ECUs control of fueling. He's ridden the bike for many tanks of fuel at 13.5 and the fueling seems to have settled in. Notes on the Chart --This is a 3 minute warm-up and 6 minute conservative test ride. --If you study the chart you can see two horizontal dotted lines, one at 14.7:1 (which shows where the center of stock fueling would be) and the other at 13.5:1 (which is the new Closed Loop target AFR created by the LC-2). Looking at the chart overall you can see that the fueling appears to vary around the target AFR of 13.5. To the right of the main chart is a histogram showing how much time the bike spends at each AFR and you can see that the center of the peak is, as planned, at 13.5:1. --On the left side of the chart, the first 30 seconds shows the cold-start/warm-up enrichment fueling period. The bike is Open Loop and the ECU riches the fuel to about 11.5:1, then decreases the enrichment smoothly until it enters the Closed Loop idling interval. This is a fairly typical looking warm-up interval based on the Motronic and BMSK LC-1 charts I've looked at. --From about 30 seconds to 3 minutes the bike is idling and the ECU is running it's Closed Loop program. The AFR is centered around 13.5. Every few seconds the AFR goes up or down a little. This is the usual shape of Closed Loop fueling. What is unusual (not noted on the R1100, R1150 or R1200 from BMW) is the long term fluctuation of AFR it seems to cycle up and down such that the Closed Loop AFR target varies a few tenths of an AFR every minute. Since TRZ_Charlie is using the LC-2 for measurement and Narrowband, but BIG-T is using a stock O2 sensor, AF-XIED and separate LM-2, it appears to me that the long-cycle AFR variation is somehow due to the MM ECU. I don't have a good explanation for what is going on. If big-t were to run the same test, with the AF-XIED disconnected and the stock O2 directly connnected to the MM ECU, I think it could be concluded that source of the AFR variation was due to the ECU. I don't see this as a riding or performance issue but it doesn't seem ideal. --Next you can see a series of sharp dips to 13:1 or 12:1. These are normal and occur when the throttle is opened. These are usual acceleration enrichments. Also, you can see above that some deceleration lean-off examples. Finally you can see the AFR shoot up to the top of the chart. This is Overrun Fuel Cutoff and occurs during Closed Throttle deceleration. All usual in magnitude and shape. When I expand the chart, the MM ECU appears to lose its ability to quickly return to Closed Loop, sometimes drifting in the wrong direction. I will post an example of that later. Also in an expanded view, I believe I see that just as the throttle is Opened, the AFR leans-off suddenly before the Acceleration Enrichment kicks in. This could well be the explanation for the stalling that TRZ_Charlie has observed on his stock bike. By shifting the entire curve down to 13.5:1, the lean-offs have also been richened. --From 8:20 onward the bike is idling, and again you can see the slowly varying Closed Loop AFR. In a future post I will show some expanded sections of the above chart which will be interesting. Then later, I have an O2 sensor disconnected test ride that Charlie made and I'll post that too. Thanks to Charlie for his work to integrate the LC-2 and make dozens of test rides.
Rogers comments continued: In the chart below I've expanded the Test Ride data in my prior post to show the detail of Closed Loop idling and what I believe is the first acceleration, by zooming in on the interval between 2 minutes 28 seconds and 3:20. Here you can see the well behaved idling Closed Loop with a nice tight AFR oscillation around 13.5:1 with a quick 1 second period. That means the mm ecu is ramping the fueling up and down and back up in about a second. That would be fine if there wasn't also the slow variation of the AFR target that you can see in the earlier chart. Of more interest is the lean-off that occurs when the throttle is first opened, before the mm ecu can sense the throttle moving and apply an acceleration enrichment. Without a gs-911 log it isn't clear exactly what the sequence of events is, but something is causing the lean-off before enrichment. That would be interesting to examine further. Charlie has mitigated the lean-off by shifting the fueling richer overall. But if it is a response to opening the throttle, it's easy to imagine hesitation or a non-robust response to the throttle off idle. I've got another chart coming tomorrow that highlights some other throttle-change anomalies.
This brings to mind my test on the TB I have on the shelf. Almost a quarter turn of the set screw before any change in tps resistance.
So, maybe the TPS sensor itself isn't necessarily mounted optimally to the Throttle Body? When I had messed around with a Jeep TBI system, it was very important to get the TPS sensor mounted such that the output voltage was between a fairly tight range. That system had a TPS with mounting holes that allowed for some adjustment when mounting the sensor. Does the TR650 TPS sensor have any adjustability? Or is the TPS sensor itself just crappy? In other words, even at mid-throttle, does the TPS not change voltage immediately when opening the throttle slightly further?
I really can't speak to your question intelligently at this point. I haven't even inspected the TPS sensor or its peripheries. Perhaps somebody that has inspected these in depth (Magoo?) can chime in.
Rogers final comments on my charts, continued: Here is an expanded chart from the middle of the Test Ride posted earlier. It has some accelerations and decelerations, that are worth looking at. As before the upper horizontal, dotted line is 14.7:1 and the lower line is 13.5:1. These LC-2 logs that TRZ_Charlie has taken not only let him know what's going on with his bike but can be especially useful to anyone considering a Power Commander. The reason is that although a PC V can change the base fueling it won't alter how the MM ECU handles throttle transitions. Later I will post some Open Loop data Charlie logged, which will be useful to the PC V crowd too. Seen above, the MM ECU has acceleration enrichments and deceleration lean-offs, but the ECU seems to hunt for the desired AFR at times too. I'm not enough of an FI expert to say it's not ideal but I can compare it to the Motronic on Oilheads (much older) and the BMSK that is used on many current BMW models. Looking at the above chart, at 4 minutes 56 second there is a good example of an acceleration enrichment. The fueling moves quickly from 13.5 to 11.8, an enrichment due a twist of the throttle of about 14%—fairly usual from logs I've seen. At 5:03 and 5:09 it appears to my eyes (giving the benefit of the doubt) that the MM ECU is ramping its way down from about 15.4:1 after a quick lean-off; and then up from a short enrichment to 11.7. it's not clear what caused these ramps. A "hunting" behavior Appears between 5:18 to 5:23. There the mixture slowly gets richer. Following that, at 5:23 the ECU slowly leans-off from 12.5:1 to a lean 15.2:1. I can't tell without more data why that might have happened, it's as if the ECU lost its way. This could be examined further but would be a lot easier if TPS/RPM data was recorded simultaneously. By way of comparison, the chart below shows how crisply the Motronic on an R1150 handles an aggressive WOT roll-on then roll off, during the first 10 seconds. Btw, the speed at the end of the 10 second acceleration was about 90-100 mph, which explains the 3 second fuel cutoff period that follows, during deceleration. I hope that the charts here and in the two prior posts, along with those from big-t, give a clearer picture of how the TR650 is fueled. Much more experimenting can be done but based on what's been measured on big-t's bike using his LM-2 and on TRZ_Charlie's bike using his LC-2.
Roger, Thanks for your in depth analysis. For me, I have learned a lot about this bike and in particular fuel injection in general. As I delved into this project I found myself straying off to read other articles and information. We discussed lambda spread in our last email. Do you feel that if the spread was further that would help the MM stabilize and do fewer oscillations? Failing a working GS911 for our bikes it looks like we have likely hit the limit as to what else can be done regarding what the MM is doing. Having said this I think that everyone that is following this thread wants to know what I am going to say about the bike and how it responds now. First I will say that the AF-XIED device gets the bike fairly close to proper driveability. It worked for me but I wasn't satisfied with the occasional stumble I still got. The major difference is that the AF_XIED allows "approximate" AFR targeting and the LC2 addresses targeted AFR values very precisely and that difference led to what I believe to be, the "optimal" fueling mixtures for the TR650 (YRMV). My conclusion regarding the two devices is hands down in favor of the LC2. This may change if the GS911 device becomes available so that further analysis can become available. Why not a Power Commander? The Magnetti Marelli ECU exhibits characteristics unlike other ECU's exhibit. It does learn like others and seems to exhibit some advanced characteristics in that regard. So until the hex code is broken throttle position values used in the PCV are really sort of mute. I will attempt to chart the differences between the the AF-XIED and the LC2 AF-XIED 1. Plug and play and easy to use on our bike, basic knowledge required and very little expertise required to use it. 2. Adjustability for increasing fueling is somewhat vague because you cannot target specific AFR's, change o2 sensor voltages or Lambda spreads. 3. Provides some remedy to lean stumble and may work well for some of our bikes. Mine didn't happen to be one of them. 4. What did it do for my bike? It stopped lean stumble on all but the hottest days. When I hit temps of around 95* the lean stumble returned (I don't fully understand why). It also gave me a taste of the real torque this bike produces. If you put one of these devices on your bike you will notice it almost immediately. LC2 1. Plug and play (kinda). Yes, it is a standalone plug and play unit, but NOT on our bikes. To make it play nice the the MM ECU it was necessary to build some components that are added into the total integrated installation to keep the ECU from throwing codes. Notably the resistor bank shown in my build log. This is not difficult but it is time consuming and as such the LC2 should not be done by someone who is not familiar with electronics and soldering. IE: You can't end up with cold solder joints and shoddy workmanship. You must also find an old narrowband bmw o2 sensor and scavenge plug parts from it to complete this build. The build log begins with post #90 found HERE. As you will read and see it takes some gyrations to make it all work. This is especially true of finding a place to stash all the components. 2. Adjustability for increasing fueling is precise. You can target specific AFR's, change o2 sensor voltages or Lambda spreads. You can also download this information for analysis and make adjustments (to a point) for enhanced driveability and performance. 3. Does it address the lean stumble and off idle stall? Your results could vary but In my application and in my opinion, yes it stopped the stall and stumble quite handily. 4. What did it do for my bike? It stopped lean stumble and lean stall altogether. Since installation this has not happened once. It woke the bike up! What I mean by that is that, the entire experience of riding this bike changed upon installation and proper calibration of the LC2. When I first got this bike and started riding it I was disappointed that the engine didn't feel like my friends XRL 650. I was expecting lots of torque and a machine that would transport me comfortably to the worst gravel I could throw at it. I was disappointed. Sure it had torque but not what I expected. It also had a snatchy throttle that was very non-linear. With the LC2 the bike turned into a gravel chewing machine! This thig is powerful now. It is full of torque from first gear off the line through 5th gear. Torque everywhere across the entire power curve. Conclusions First, I am not an expert by any stretch of the imagination. I am a neophyte at FI and as such these are the observations of a rooky. Until we have a hex code broken on the MM or until we have a working GS911 there is no way to know what the MM is doing. If you are saying to yourself "the LC2 and AF-XIED are just masking the problem" then you would be about half right. Yes, they do mask the problem, kinda. They both address fueling issues, they both address driveability issues and therein is the fix, for now. At this point we don't know anything of significance about the MM ECU. If you spend the extra money on a PCV (IMHO) you are throwing your money in the wrong direction (at least for now). Why? You are paying to do the some of the same things that this project has addressed and you are doing for nearly $300 more than I paid for this project. Again, this is just my opinion. Addressing the very specific issue of lean stumble and off idle stalls the AF-XIED did abate them some. And for someone not looking to do a "project" and is an occasional rider it will provide you much improved driveability. If you want to increase the overall performance of your bike please take the time to learn from this project. I am pleased for now. I will play with o2 sensor voltages and lambda spreads in the near future but my fueling will stay and 7% richer than stock. This is the right fueling for this bike. The Logs from my runs prove that. At seven percent extra fueling the bike just plain wakes up. It makes it feel like a tall sport bike when you ride on the highway. It turns it into a fire breathing gravel chomping beast on two track. Lastly I cannot make these comments without acknowledging my friend and mentor Roger. He has been golden. He has put up with many stupid questions from me and has remained consistent in his support and patience while I muddle through this project trying to make sense of something that looked and felt completely foreign to me from the start. Thank You Roger! I hope this helps to clarify what the project was/is and will help others in their quest to make the bike run like it should have out of the box. I'd also would like to get some comments and discussion from the other inmates following this thread. It would be valuable to get some banter regarding the project and toss around ideas about where to go next.
Hey drzcharlie, Very grateful for the enormous effort put in by you and Roger. I'm keen to adopt the LC2 solution when I get some time to put it all together. I have been following your project throughout to the extent that my limited technical understanding allows. A question re comparison - if the LC2 adds approx 7% extra fuel (adjusted through the rev range) and the the AIT spoofers add roughly 6% (unadjusted) - is there still a significant improvement in your opinion over the spoofer option? Sorry if you've already addressed this this question previously. I recently conducted my second ECU reset to erase the adaptions to the spoofer and was reminded of how impressive this bike can be with a little extra fuel in the right place. But if the LC2 is another level higher - I'm definitely interested! Cheers
To answer your question, yes, there is a significant improvement. I ran the AF-XIED for a couple of months prior to the LC2. The torque and the power band have significant differences to the better on the LC2. First you can adjust the o2 sensor voltages, you can adjust the AFR spread and you can adjust the AFR to precisely where you need it. I am currently running 13.5:1 or 7% richer coupled with settings developed by Roger and myself. If you can achieve the above AFR target using the device of your choice you should realize similar results. My bike just "woke up" when we hit the right settings. Is it the final answer? No, it isn't. But it is what I can live with until we get a better understanding of the MM ECU. If you are going to do the mod make sure you start by finding an OEM o2 sensor and the plug. The plug is the part you need so that you can adapt the LC2 into the ECU wiring harness without too much fuss. This is the part number both for our OEM replacement and BMW: http://www.autopartoo.com/oem/bmw/7674452.html
Ok, I'm going to call the LC2 project a success. My latest settings are great and I have experimented with many settings over the last three months. Roger suggested several different Lambda Spreads that I have tried. The latests being Low setting 0.925 and High setting 0.93 . I am set at 13.5:1 ( 7% more than OEM). The entire set up is working very well. My fuel mileage is at 57-58 miles per gallon, the bike has a linear throttle in all throttle positions, no lean stumble, no lean pop on deceleration, torque is equal and powerful across the entire throttle range and from first through fifth gears. It's a great seat of the pants experience knowing the bike is safe from stalling. It works equally well in mild single track to highway speeds and in 90* weather as well as 20* weather. The LC2 works. It makes the bike a smile machine. I love it.
Dave, I would love to do that but I don't have a dyo within 100 miles of me, plus I am retired so I have to target my expenditures and the next big one will be suspension upgrades. Truthfully it runs so good now I am going to leave it as is, with the exception of adding Brisk Plugs .
Thank you very much, drzcharlie Nice when someone shows examples of new things with. Talks openly about issues and share information. Encouraged by this example I bought Innovate LC2 and removed boosterplug, Boosterplug is not suitable for winter use anyway, in our latitude with frost in winter. Installation was easy, after all I did have good instructions on use Only difference drzcharlies installation is resistors, I used only one "power" resistor 120ohm 3w.Fault codes are not yet able to read, fault light does not come. I suppose if all is well The test phase going, I would like to try different settings to work here in the cold. The first impression is a positive. I only drive 200km yet. With Boosterplug 7000km in summer The story continues in the future
Tokor, pm me and I will give you the settings I found best. I am done tweaking mine and it is nearly perfect. As for the fault codes, the resistors/resistor are only there to keep the wide band sensor from throwing codes. The only concern in using one resistor is the dissipation of heat. Better when using one to use a resistor that has it's own heat sink. No codes = A good thing.
