Joined
·
541 Posts
Instructions in post #37
BLUF: 6.7% Speedometer overread on 2016 GSX-S1000F, US model.
Not yet familiar enough with the forum to know if this is the best place for it, if not, sorry.
I'm planning to install a Speedohealer. I don't like speeding for my commute, but here in Utah, it's dangerous to drive below the speedlimit, especially on a sportbike; seems to be people expect you to be speeding and if you're not, don't expect even a modicum of respect for your personal safety. Friday I got passed (on a four lane road mind you, with me in the right track of the right lane) so close I could have reached out and touched the car. And in my experience here, even if there's a cager near by going faster, it's the biker that'll get pulled over. So, I'd like to be able to know how fast I'm actually going.
For now, I'm calculating the percent error. I took my phone and a GPS logging app for a ride, about 20 minutes sustained at 60 MPH indicated. I trimmed off the initial acceleration and deceleration data, then deleted obvious outliers with data in chronological order. Next I reordered the data ascending with measured speed. I used a linear fit, and trimmed the leading and trailing data until the R-squared value was over 0.99 I then switched back to chronological and the linear fit had a slope of .0001 I considered this sufficient. All told I had 832 remaining data points for 20 minutes at sustained speed. My standard deviation is 0.44 MPH, given the speedometer readout is to the nearest whole MPH, this is within the instrument error. I got 55.97MPH actual for 60 indicated, a 6.71% overread, calculated by ( actual/indicated – 1 ) x 100
For 75 indicated, after analysis, got 70.00 actual, standard deviation 0.34, slope of linear fit for chronological order -0.00005, R-squared for linear fit on ascending order 0.9948, but only 437 data points and 12 minutes. Total overread, 6.67%. So still rounds to 6.7%, and the speedohealer only increments by 0.1%.
I got a handle bar mount for my phone, and put the GPS app in speedometer mode. With 6.7%, some speeds I took a look at, expected Actual/Indicated: 14.00/15; 27.99/30; 41.99/45. It's hard to hold this bike at 15MPH, but the other two checked out just fine, rounded to the nearest whole. Further, drove past a speed limit sign with integrated radar and readout, sign said 28, bike said 30.
Overall, seems the speedometer error is linear, 6.7%. I've got a speedohealer in the mail, should show up today. I haven't been able to find anyone claiming for certain that the bike does/doesn't have a speed sensor independent of the wheel tone rings. I'll be doing some wrenching around looking, barring that, I also have an FSM on the way. Hopefully there is a way to patch it in without potentially impacting ABS/TC. Granted, this is the first bike I've ever had with them, and I've been riding in cold rain through a canyon to work for the last week without either system activating (TC at 3) so I'm not convinced they're necessary, but I'm not too keen on degrading even unnecessary safety features.
-Jaren
P.S. Excle spreadsheets with analysis
https://drive.google.com/file/d/0BxxoMzdrnYyjblZmZGNEZjlqV2s/view?usp=drivesdk
https://drive.google.com/file/d/0BxxoMzdrnYyjLTMyTjl6UWlMeU0/view?usp=drivesdk
BLUF: 6.7% Speedometer overread on 2016 GSX-S1000F, US model.
Not yet familiar enough with the forum to know if this is the best place for it, if not, sorry.
I'm planning to install a Speedohealer. I don't like speeding for my commute, but here in Utah, it's dangerous to drive below the speedlimit, especially on a sportbike; seems to be people expect you to be speeding and if you're not, don't expect even a modicum of respect for your personal safety. Friday I got passed (on a four lane road mind you, with me in the right track of the right lane) so close I could have reached out and touched the car. And in my experience here, even if there's a cager near by going faster, it's the biker that'll get pulled over. So, I'd like to be able to know how fast I'm actually going.
For now, I'm calculating the percent error. I took my phone and a GPS logging app for a ride, about 20 minutes sustained at 60 MPH indicated. I trimmed off the initial acceleration and deceleration data, then deleted obvious outliers with data in chronological order. Next I reordered the data ascending with measured speed. I used a linear fit, and trimmed the leading and trailing data until the R-squared value was over 0.99 I then switched back to chronological and the linear fit had a slope of .0001 I considered this sufficient. All told I had 832 remaining data points for 20 minutes at sustained speed. My standard deviation is 0.44 MPH, given the speedometer readout is to the nearest whole MPH, this is within the instrument error. I got 55.97MPH actual for 60 indicated, a 6.71% overread, calculated by ( actual/indicated – 1 ) x 100
For 75 indicated, after analysis, got 70.00 actual, standard deviation 0.34, slope of linear fit for chronological order -0.00005, R-squared for linear fit on ascending order 0.9948, but only 437 data points and 12 minutes. Total overread, 6.67%. So still rounds to 6.7%, and the speedohealer only increments by 0.1%.
I got a handle bar mount for my phone, and put the GPS app in speedometer mode. With 6.7%, some speeds I took a look at, expected Actual/Indicated: 14.00/15; 27.99/30; 41.99/45. It's hard to hold this bike at 15MPH, but the other two checked out just fine, rounded to the nearest whole. Further, drove past a speed limit sign with integrated radar and readout, sign said 28, bike said 30.
Overall, seems the speedometer error is linear, 6.7%. I've got a speedohealer in the mail, should show up today. I haven't been able to find anyone claiming for certain that the bike does/doesn't have a speed sensor independent of the wheel tone rings. I'll be doing some wrenching around looking, barring that, I also have an FSM on the way. Hopefully there is a way to patch it in without potentially impacting ABS/TC. Granted, this is the first bike I've ever had with them, and I've been riding in cold rain through a canyon to work for the last week without either system activating (TC at 3) so I'm not convinced they're necessary, but I'm not too keen on degrading even unnecessary safety features.
-Jaren
P.S. Excle spreadsheets with analysis
https://drive.google.com/file/d/0BxxoMzdrnYyjblZmZGNEZjlqV2s/view?usp=drivesdk
https://drive.google.com/file/d/0BxxoMzdrnYyjLTMyTjl6UWlMeU0/view?usp=drivesdk