Pastabatman
Member
If that is so, it is woefully inaccurate on my bike. I *wish* I could get 15 miles on one bar! That is what that my display showed after my last full charge.
Questions about "Distance To Go" and "Ah" setting on new display unit
- Thread starter PowerMe
- Start date
Svein
New Member
when
When battery is fully charged, it shows the distance til empty. But when you start riding, it counts down to the point when you will loose the first bar.
Pastabatman
Member
I can get well over 30mi on full assist, including some big hills, so 15mi is way off. Miles to go until you lose a bar seems like an incredibly odd metric to present on the display, but I'll try to observe what it does as I drop bars on my commute home tonight.
Last edited:
Pastabatman
Member
For my 12mi commute, I started at 5 bars/100%, max assist, and miles to go was 15mi. During the ride, miles to go steadily dropped down to almost zero until I lost the 2nd bar (3 bars/55%), at which point it popped up to 19mi and began decreasing again. Miles to go did not increase when I lost the first bar (4 bars/75%).
Next theory?
Next theory?
Ebike Commuter
Member
On the Street, the "miles to go" values are wildly inaccurate starting from ~100% charge. However, after the charged decreases to ~60% or so, the sum of the total miles ridden for a given trip and the "miles to go" do not vary by more than approximately +/- 1 mile. In other words, the miles to go indicates the miles left on the battery within approximately +/- 1 mile, at least on the 1 ride where I verified this periodically for 10 or so miles.
Ebike Commuter
Member
The Miles to Go readout is not random and is in fact very ordered over a wide range of battery charge (see chart below). However, as can be observed in the chart, there is a discontinuity above approximately 75%.
If one extrapolates the data contained within the linear portion of the chart for % battery charge values between approximately 75 and 10%, the line intersects the y-axis at approximately 27 miles. The linear regression of that data yields the following equation: y = 0.2691*X - 0.045 and has a correlation coefficient of R^2 = 0.9999. In other words, the algorithm appears to be based upon the assumption that a full depletion of the battery will yield a trip distance of approximately 27 miles.
Given the extreme linear nature of the response, it appears that the algorithm used to generate the "To Go" readout is based upon the same information used to generate the Battery % Charge readout. That information appears to be contained within the electronics inside the battery enclosure, since both the "To Go," and the "% Battery" values return to the same readings when the battery is removed and reinserted into the frame. The presence of a "Fuel Gauge" on the battery supports this assumption.
If one extrapolates the data contained within the linear portion of the chart for % battery charge values between approximately 75 and 10%, the line intersects the y-axis at approximately 27 miles. The linear regression of that data yields the following equation: y = 0.2691*X - 0.045 and has a correlation coefficient of R^2 = 0.9999. In other words, the algorithm appears to be based upon the assumption that a full depletion of the battery will yield a trip distance of approximately 27 miles.
Given the extreme linear nature of the response, it appears that the algorithm used to generate the "To Go" readout is based upon the same information used to generate the Battery % Charge readout. That information appears to be contained within the electronics inside the battery enclosure, since both the "To Go," and the "% Battery" values return to the same readings when the battery is removed and reinserted into the frame. The presence of a "Fuel Gauge" on the battery supports this assumption.
Ravi Kempaiah
Well-Known Member
- Region
- Canada
- City
- Halifax
Neat analysis. You could send this to EmotionUSA tech people and educate them
Interesting discontinuity at ~75%
Bosch Nyon is coming up with a console that takes in your route .gpx data and considers your previous riding history to give miles to go estimate.
PowerMe
Well-Known Member
Note that in the graph above at no time does "mi to go" ever estimate a range higher than 20 mi. BH/Easy Motion claims the bike gets between 30mi and 60mi per charge, depending on lots of factors.
The highest mi to go my bike ever showed on this new console was 19 mi, and that was after the console started with "6.3 mi to go" on a full charge, then decreased down to zero over a few miles, then reset itself when mi to go ended at "0" and displayed 19 mi to go (similar to the graph above).
Whatever is being measured is not accurate, as far as I can tell.
The highest mi to go my bike ever showed on this new console was 19 mi, and that was after the console started with "6.3 mi to go" on a full charge, then decreased down to zero over a few miles, then reset itself when mi to go ended at "0" and displayed 19 mi to go (similar to the graph above).
Whatever is being measured is not accurate, as far as I can tell.
Ebike Commuter
Member
The data from the graph above was generated with an 11.6ah battery. Values could be different, depending on battery size. It is apparent from the data that the values from 100% charge to 75% are not indicative of expected distance. If one wanted to use the values prior to the discontinuity, one could add "20" to the number (e.g. 20+6.3 = 26.3), which would provide a reasonable estimate of range.
However, this does not mean the actual distance on one battery charge will be 26.3 miles. The actual value could be 16 miles, 26 miles or 36 miles, or some other value. The BH manual makes a statement to the effect that the actual miles will depend on terrain, speed, input from the rider and other factors, which makes sense. Nevertheless, the algorithm seems to have ~27 miles as a range assumption, which is outside the 30-60 mile range referred to above.
It also seems that the "To Go" value will most accurately predict the actual miles to go, the closer the battery is to full depletion. That is to say, the accuracy will be better the closer the batter is to empty. This would require additional data to demonstrate; but is based upon the apparent simplicity of the algorithm and direct observation.
However, this does not mean the actual distance on one battery charge will be 26.3 miles. The actual value could be 16 miles, 26 miles or 36 miles, or some other value. The BH manual makes a statement to the effect that the actual miles will depend on terrain, speed, input from the rider and other factors, which makes sense. Nevertheless, the algorithm seems to have ~27 miles as a range assumption, which is outside the 30-60 mile range referred to above.
It also seems that the "To Go" value will most accurately predict the actual miles to go, the closer the battery is to full depletion. That is to say, the accuracy will be better the closer the batter is to empty. This would require additional data to demonstrate; but is based upon the apparent simplicity of the algorithm and direct observation.
Ebike Commuter
Member
If Easy Motion plans on updating the display code, perhaps they could either eliminate the discontinuity altogether, or if that is not feasible, display no numbers in the "To Go" field for the first 25% of battery depletion. The current display generates "To Go" readings for the first 25% of charge that are confusing due to the discontinuity. The current algorithm gives the impression of being a "beta" version.
One can also imagine more advanced algorithms, that add a multiplier based upon the PAS level. Even more advanced algorithms could be envisioned, such as described by Ravi above that take into account previous or current ride data.
About six months ago a local bike store employee mentioned that a BH rep had indicated that they were working on a way to interface a cell phone with the bike. I believe he said something about an app that took into account tide terrain and adjusted power output based upon that. He was not very clear on the details. Perhaps a cell phone app could be another manner to add more advanced ride metric capability. Does anybody have any information regarding that topic?
One can also imagine more advanced algorithms, that add a multiplier based upon the PAS level. Even more advanced algorithms could be envisioned, such as described by Ravi above that take into account previous or current ride data.
About six months ago a local bike store employee mentioned that a BH rep had indicated that they were working on a way to interface a cell phone with the bike. I believe he said something about an app that took into account tide terrain and adjusted power output based upon that. He was not very clear on the details. Perhaps a cell phone app could be another manner to add more advanced ride metric capability. Does anybody have any information regarding that topic?
PowerMe
Well-Known Member
If anyone does, @Ravi Kempaiah would likely know who they were within the BH/Easy Motion company and maybe he'll ask on our behalf?
