07-31-2013, 04:29 AM
Originally Posted by Roger 04 RT
I would love to run this by a couple of my engineer friends who are heavy into programming fuel injection & turbos... Couple of industry guys... Can you provide a link to where you gleaned this information? I'm sure I will get back a detailed explanation.
07-31-2013, 09:46 AM
John, Sure. The explanation seems straight-forward: reference oxygen gets to the internal sensor through the cable (rather than a hole in the sensor body to avoid contamination. Roger
Originally Posted by jstrube
Here is a link to the consumer article: http://www.boschautoparts.com/BAP_Te...allGDWEB09.pdf.
From Bosch's staff:
You should never solder a O2 sensor wire because the sensor "will breathe" through the cable. The acceptable method of splicing or repairing O2 wiring is with crimp connectors. Here is the technical description of why you do not solder the connections on the O2 sensor side of the updated wiring.
The outside of the bulb is exposed to the hot gases in the exhaust while the inside of the bulb is vented internally through the sensor body to the outside atmosphere. Older style oxygen sensors actually have a small hole in the body shell so air can enter the sensor, but newer style O2 sensors "breathe" through their wire connectors insulation and have no vent hole.
It's hard to believe, but the tiny amount of space between the insulation/wire and through wire insulation provides enough air to seep into the sensor (for this reason, grease should never be used on O2 sensor leads/connectors because it can block the flow of air). Venting the sensor through the wires rather than with a hole in the body reduces the risk of dirt or water contamination that could foul the sensor from the inside and cause it to eventually fail.
On this manual 2/3 through on page C4: http://www.bosch.com.au/car_parts/en...r__Cat_WEB.pdf.
07-31-2013, 10:02 AM
Lorne, I've been working with Alan To, who makes a good but non-programmable product, the Spartan. It needed to be reprogrammed to work well with the Motronic MA 2.4 and 2.2. Alan did this for me. I now have a couple prototypes that work very well. There are no datalogging capabilities.
Originally Posted by R100RTurbo
For an R1200, F800, R1100 or R1150 installation there are several considerations the led me to conclude a shifted narrowband O2 would be a better upgrade for most riders: narrowband sensor requires no calibration, narrowband sensors are more rugged, the narrowband sensor would not have to be removed from the exhaust, the Spartan cannot be user programmed, a shifted narrowband would be plug and play.
Because I like Alan and his product, it remains an option after the market gets going. See picture of Spartan with my programming and R1150 connector (from earlier in this thread: http://www.bmwmoa.org/forum/showthre...l=1#post868229) below.
07-31-2013, 02:48 PM
Roger, thank you for the detailed info!
So what have people been finding is the optimal AFR to run for an RT?
07-31-2013, 06:11 PM
John, That's a very good question. We've had about two dozen units installed on many different models. There is a relatively small spread of results. A 4% enrichment makes a significant improvement in driveability and low RPM torque. By 6% the bikes usually seem transformed. So most seem to settle somewhere between 4-6% although some have added as much as 8%. The results are the same for LC-1s and BMW-at-XIEDs.
Originally Posted by jstrube
It is sometimes hard to believe that so little additional fuel can make such a big difference. It just seems like it's the boxer's sweet spot.
08-01-2013, 01:54 PM
That's very interesting Roger, & obviously a lot of detailed investigation and work.
I feel compelled to comment on how deeply that simple aspect intertwines with the provisions of "self tuning" and in particular (and my case) the MS technology.
Originally Posted by jstrube
Just to know what my old lump was doing with its original Bings (sorry for airhead content) I ran the original LC1 on the bike and documented afr behaviour at various temperatures and load/ speed conditions. To be fair, jetting and carb condition enters into results but with the bike running "very normal" I was shocked at how rich it would sit on cruise, and how lean it would transition towards on hard accel.
Everything to do with "inherent performance & behaviour" changed once the Bings gave way to injection and forced induction, but learning what AFR the bike likes is just a process of trying out different settings (on all states including idle, mid accel, hard accel, cruise, and decel) where all can be changed independent of the other. I find I can lean out to 16.5 easily in cruise with no surging or transition hiccup, as well have recently been playing with an economy tune to allow 87 octane while maintaining detonation free hard accel driving (our gas getting so expensive lately). I believe some megasquirters just seek out excuses to meddle with their tune, as it is perhaps too much fun
08-01-2013, 02:27 PM
It has long been assumed that the R1100 and R1150 surge with light throttle, in Closed Loop. It is now seeming more likely to me that they surge in an Open Loop corner that is not well documented--namely RPMs greater than about 2000 and throttle angles less than 5 degrees. This is a very problematic area documented earlier in the thread by WallyG and myself. In this area, the bike alternates easily between a slight accel enrichment, and a slight decel enleanment--and it feels like it is surging there--and it is Open Loop.
Originally Posted by R100RTurbo
To that end, I have now set my LC-1 to lambda = 1.08 which is about 15.9:1--very, very lean. There is no surging in cruise in 5th or 6th gear. As you would guess, the engine doesn't seem nearly as robust as at lambda 0.94 (13.8:1) but it doesn't run much differently from the stock 14.7:1 setting.
My next test is to see what the cruise gas mileage is with lambda = 1.08, out of curiosity.
BTW, Here were some recent fuel stats at lambda =0.94
147 miles odometer 143 miles (google map) (odometer seems 3% high)
90 miles highway (60-80 mph) est. 70 mph average
53 miles local/stop&go, speed less than 50 mph
Top case and two side cases, windshield full up on highway.
2.877 gallons fuel consumed
smooth/brisk acceleration but not aggressive
average 143.2/2.877 = 49.8 mpg about 2/3 highway, 1/3 local with stop and go (subtract a mile per gallon if you want to add some uncertainty to the tank fill volume)
08-03-2013, 04:39 PM
Here's an idea of how gas mileage is working out at lambda 0.94 (6% rich AFR). Put 143 miles on my bike the other day. Here are the stats.
147 miles odometer, 143 miles (google maps) (odometer error 3%)
90 miles highway (60-75 mph) est. 68 mph average
53 miles local/stop&go, speed less than 50 mph
Top case and two side cases, windshield full up
Air Temp 75F
2.877 gallons between fills (error allowance +0.1 gallons), 143.2 miles,
average 143.2/2.877 = 49.8 mpg (with error allowance 143.2/2.977=48.1 mpg)
Next Up: A test run of lambda 1.08 (AFR 15.9:1) 8% leaner than stock
08-09-2013, 11:36 AM
Running the Oilhead Very Lean
The Innovate Motorsports LC-1 has two analog channels. One is usually used to simulate a Narrowband O2 output. The other analog channel is usually set to drive an AFR gauge, however it is possible to program it to simulate a Narrowband O2 output as well.
With that in mind Analog 1 to lambda = 0.93 (7% rich, 13.65:1 AFR) and Analog 2 to lambda = 1.08 (8% lean, 15.9:1 AFR). By means of a selector switch you can choose which mixture the Motronic is using, even while running down the road.
You can see in the chart below how easily the Motronic moves from Closed Loop at 7% rich to Closed Loop at 8% lean, and also how well it holds Closed Loop at either setting. Also on the chart is a reference line at 14.7:1 which represents the stock oxygen sensor setting.
I had a chance today to make a 50 mile run with Closed Loop at the very lean setting of lambda = 1.08. Some observations:
--The Motronic easily adjusts to 1.08. This wasn't a very long test but under riding conditions over the 50 miles the bike seemed to be running normally, but with, for lack of a better phrase, a more anemic feeling.
--In cruise mode at speeds up to 60 MPH (top speed I could run for any distance on the roadway), there was no misfiring and no pinging. In fact I tried to make it ping with some uphill roll-ons and couldn't. The engine didn't seem to run any hotter than usual.
--Since my BBSs were set to idle at 13.65:1, I had to keep the fast idle lever up to idle at 1100 RPM. With the leaner mixture it takes more air to idle at a given speed.
--The engine performed "okay" but it took more throttle to start off than in the richer setting, the bike stalled as I let out the clutch to leave the driveway. Roll-on throttle while cruising, led to a very slight hesitation before acceleration
--At low speeds, throttle input exhibited some jerkiness during transitions from acceleration to deceleration that aren't there at all at lambda 0.93. The "hop" that can occur when the bike comes out of Overrun Fuel Cutoff was pronounced.
--The natural up-shifting point seemed about 1000 RPM higher than I grown accustomed to with richer than stock mixtures. It really did not want to be upshifted until 4000-5000 RPM. It just wasn't settled and ready to be shifted much earlier.
--Was there any surging? Yes, but only a bit and no more than I remember with a stock O2 sensor. Mostly light throttle in the range of 3000 to 4000 RPM.
--The course was 5 miles long with short stops and starts during course reversal. The 53.5 miles on odometer (51.9 corrected) was ridden mostly at speeds between 45 and 55 mph but top speeds were 65-70. Fuel consumed was 1.05 gallons +/- 5% which is about 49 mpg +/- 2 mpg.
In effect, this test was a side by side comparison of a richer fueled R1150RT and a leaner fueled bike. There's not question which is the more satisfying motorcycle to ride, the richer mixture by far.
Although the Motronic seems very flexible and able to run rich or lean of the stock setting with modified Lambda input, there is a big difference in the feeling of the engine. Richer mixtures (plus 4-6%) lead to a much smoother, slightly more powerful engine, especially below 4000 RPM, that "asks" to be shifted about 1000 RPM lower than a leaner one.
Leaner mixtures (minus 0-8%) lead to a more anemic feeling and seem to amplify whatever bad manners the engine exhibits (OFC hop, surge, roughness, hesitation). Overall it was an unpleasant ride.
By the end of the 50 mile test I was very happy to flip the switch back to lambda 0.94 where it's going to stay for the rest of the summer.
08-09-2013, 02:59 PM
Good Test Roger
Thanks for that info Roger and kind of neat to be able to switch (A-B) between AFR mixtures. Today I'm going to switch to a pink CCP and see how she runs at sea level at 13.1:1 AFR then maybe increase to 13.5:1.
Really like the switch idea.
08-09-2013, 05:36 PM
For sure some diligent testing and documentation.
At the risk of offending (and I'm a high risk offender when it comes to discussing bike tech), you can have your cake and eat it too
With Microsquirt for injection and ignition management, its a full meal deal where you get the best of all conditions and operating states to dial up optimum values that the bike and the rider want to enjoy (assuming you know what you're doing - many will attest to potential for knowledge overload or small detail snags that can hamper successful projects). In other words, your choice as to what afr at any rpm or load state, as well ignition timing. One has both 16x16 and 12x12 tables to work with.
I have an "in progress" tune where I might utilize 87 octane (has up to 10% added ethanol) at times, I have confirmed this to be very close to detonation free when I have water injection turned off - one just can't rely on finding the commonly available B.C. 94 fuel when on trips. That takes the stress out of heavy throttle roll ons, for more "no strings attached" riding. (& no, I don't sell the product).
08-10-2013, 02:42 AM
The megasquirt would be a great project. And you could get a native Wideband o2 capability. It seems that the amount of work required to get all the timing and fueling parameters worked out would be Herculean.
08-10-2013, 03:25 AM
Hold on a sec, if I remember this post correctly you have a background in aeronautics, and obviously display a significant talent to figure things out - heading down that pathway wouldn't be all that challenging for you if I might render an opinion, but I don't want to be pushy.
I am a one bike owner at this point, but have a wish to push further into projects, I would love to MS a bike that starts out already being fuel injected and has the vast content of required components and sensors already in place and proven. The one thing that MS is good for is its adaptability - and when things surface that are somewhat off the page, there are highly educated and knowledgable forum participants that offer tuning help or program modification recomendations, even code changes and new parameter development just to make something work with the perhaps unique arrangements presenting.
In contrast, getting that all happening on a vintage airhead, with a beautiful fairing system that has to be left intact and can't be hacked - that is the crazy project to take on (and has been frequently pointed out by my patient family lol).
Who knows, I might be able to find a suitable donor bike, I'd love to get into a GS - so perhaps if one follows me home one day and I can hide it in the workshop for a while it might weather the storm of "You have two bikes on the go now!"
The old adage, "Nothing Ventured, Nothing Gained" is engrained into my project portfolio rational.
08-10-2013, 11:18 AM
The Microsquirt/Megasquirt products seem very well documented, like you I suspect, I've read most of the site and even downloaded the source code. It looks to me like most of the project is straightforward. Our sensors would work, sometimes with coding changes, and I guess a connector to mate with the Motronic connector could be found. What I think I lack are the time and resources to develop the complex timing matrix that most FI vehicles use.
Therefore is seems to me that since BMW has a well balance VE map and extensive dyno experience with the motor to evaluate its timing needs, that I would be better leaving the timing alone and fix the fueling problem by lambda-shifting the O2 reference, in effect adding fuel everywhere. I suppose a hybrid approach would be to let the Motronic handle timing and develop the fueling with Microsquirt.
The biggest benefit to that would be additional fuel could be added into the Open Loop areas that naturally surge, more than just an amount created by the lambda shift.
If I did anything now, I might like to get one of BMWs newest models and see what it took to fix the fueling on it. The GSW seems like a really neat bike but after test riding it, it seemed that it to is affected by the mandates of the EPA and could be improved. It's got lots of HP but at small throttle angles the throttle & engine get very twitchy/touchy/frisky.
08-12-2013, 08:12 PM
Were on a bit of a vacation and I'm not spending much time online.
In actual fact, creation of the various tables for MS has become rather easy, certain Tuner Studio features make possible VE table refinement for example while driving, or from logs taken and run through programs on lap top after the fact. The ignition table albeit is the hardest to achieve maximum power/ efficiency from, when you are attempting to tune without provision of a dyno. It can be done, and in many cases provides very good performance to stock even in roughly formated examples that surface on MS forum discussions. However, just as I documented afr values prior to embarking on a MS conversion, the microsquirt module (or mega) could be interfaced on a bike either in a piggyback fashion on ignition sensors, or an alternate added connection - to fully document what the original BMW timing was over very broad operating ranges and such. Then those values could be used as starting values for the MS project. I suspect however that those values would be able to be edited for further improvements as things got underway. Its not really that complicated.
I believe what you are helping me to do however, is recognize what the next project I take on might be - an oilhead conversion.