Possible bug in software

[bjrmd]

Hi
Been doing some test rides and I have noticed that despite a very close newton time advantage calc (a few secs), the isaac software puts the time advantage at 60 sec for just a 7 mile ride. The average watts were within a few % as well. Enclosing the file. I did not take a cellphone shot of the head unit display but it was no more than 1-2 sec in the time advantage screen (which I thought was impressive)

[bjrmd]

Did another short test ride today and the time advantage on the newton was negative 65 sec or so but the software shows a positive of about 5 sec. Weird

[travispape]

Hi bjrmd,

Thanks for posting your files. I took a quick look and I don't have an answer for the disconnect yet. I will be traveling this weekend and won't have a chance to look at it until I get back, but I wanted to let you know I'll try to figure out what is happening.

[bjrmd]

I was wondering if anyone else is seeing this?

[GWPOS]

What do you mean by time advantage? Maybe post pic or something to help describe issue.

[bjrmd]

What do you mean by time advantage? Maybe post pic or something to help describe issue.

here is the cda/time advantage chart--on the newton it was about 3 sec

[bjrmd]

Did another careful 20 miles out and back, wind cal was zero. The cda was averaging high on the newton the whole way and time advantage was negative about 10 sec per mile on the newton.
When I put the file in the Issac software, avg watts were 1.5 % higher than the powertap(very very close), cda graph below--shows positive time advantage.

Something is wrong here?

BTW Stats:

Cadence: Avg: 73.0 rpm (ignoring zeros)
Max: 105 rpm
HR: Avg: 0.0 bpm (ignoring zeros)
Max: 0 bpm

------------------------------------------------------------------------
IBIKE COMPARED TO ANALYTICCYCLING.COM
------------------------------------------------------------------------
Plug the following parameters(*) into the static forces calculator at
http://www.analyticcycling.com/ForcesPower_Page.html

Effective Frontal Area(#) = 0.756 m^2
Drag Coef (keep the default) = 0.5
Air Density = 1.160 kg/m^3
Weight of Rider and Bike = 99.8 kg
Coef of Rolling Resistance = 0.0055
Slope = 0.00005 rise/run
Speed = 7.906 m/s
Pedal Cadence = 63 rev/min
Crank Length(**)
Effective Pedaling Range(**)

Compare the avg power result from Analytic Cycling to the comparable
result from the iBike 3 calculation engine:

Propulsion Power = 151.32 W <---<<< compare to AnalyticCycling(##)
Rider Power = 154.41 W

* The AnalyticCycling calculator assumes perfectly calm wind
conditions. Further, it does not account for any accelerations or
decelerations.
# (Drag Coefficient) x (Effective Frontal Area) = CdA
** Crank Length and Eff Pedaling Range do not factor into the avg
power calculation
## The AnalyticCycling calculator ignores drive train losses.

------------------------------------------------------------------------
POWER METERS COMPARED TO EQUATIONS OF MOTION
------------------------------------------------------------------------
Ride File Min Pow Avg Pow Max Pow
AnalyticCycling.com 150.81 154.41 157.03
Speed Elev 0.00 172.15 593.01
Speed Wind Elev 0.00 162.65 474.61
Speed Wind Slope 0.00 166.62 546.60
Newton+ 0.00 165.68 445.06
iB DFPM 0.00 163.19 355.00

------------------------------------------------------------------------
R^2
------------------------------------------------------------------------
| Newton+| iB DFPM
Newton+| 1.000| 0.686
iB DFPM| 0.585| 1.000

R^2 calculation is row w.r.t. column.

[bjrmd]

Well, after many more rides I have a consistent discrepancy between the time advantage/cda on the newton and what is shown in the PC software. It amounts to near perfect Isaac power similarity between powertap and newton(cda about .38) but I have a time loss of about 10 sec per mile showing while riding on the road with cda near .43.

Any ideas or do I have a defective unit, bad firmware flash?

[Velocomp]

We're not ignoring this thread; it's just that we have a lot of things going on!

Travis will take a look and report back.

[bjrmd]

Thanks

[travispape]

Any ideas or do I have a defective unit, bad firmware flash?

No, I am seeing the same phenomenon as you so it's not a defective unit. There is not a quick and easy way to debug it. The FW developer and I will have to get our heads together and try to find where the disagreement is coming in. If I had to guess, we have a situation where there is a round-off error in the time benefit calculation, and since time benefit is a cumulative statistic a small error adds up to be a noticeable one by the end of the ride.

But that's just a guess. We plan to take the time to get a more definite answer when we can both get a break for doing so.

Travis

[ldmitruk]

A bit of an aside but somewhat related. Is there a good explanation of the information present of the stats page and how to interpret it. In particular what do the R^2 tell us and what kind of numbers should we looking for when comparing our iBike data to DFPM data? I've looked through the documentation and haven't been able to find anything.

[bjrmd]

Thanks for the reply--at least I'm not doing something wrong.
Will not obsess anymore and just wait for your solution.

[bjrmd]

FWIW:
Newton CSt(?friction)=17.130 but Isaac says 17.131 in edit profile
Newton CSt(aero)= .250 but Isaac says .251

[bjrmd]

Is this close to being fixed?

[NorBike]

Maybe this can be the reason?:
http://www.ibikeforum.com/viewtopic.php?f=55&t=2534

[bjrmd]

Using the new firmware has not solved this--I am still getting a negative time advantage and higher cda on the newton screen than the isaac software shows(this is usually mildly positive)

Also the coasting cda showing in isaac for the powertap is fixed at the baseline cda(the newton does show a change during coasting). This will of course invalidate the time advantage in isaac.

I just did a 15 mile loop and the avg power between the newton and powertap with within 1 %.

[travispape]

Ok, I finally had a chance to investigate this and I found some issues with the way we are reporting Time Advantage in both the FW (what you see on the road on the Newton) and the SW (Isaac).

First of all, let me clarify that the Realtime CdA that is being reported by the Newton and in the SW are very good and there's nothing to fix. In fact, once you upgrade to OS 2 and beyond, you should notice that it is working even better. As a reminder, what you get out of this tool depends on the care you put into using it. To start, create a good profile for your initial ride position, call it ride position A. A good way to do that is to do an O&B ride on a day without crazy wind conditions and on good roads without traffic. Then download the ride and do a Analyze->Check Calibration in the software to tweak your profile (to your DFPM if you have one). Of course, be sure your DFPM is zeroed and working well. Once you have established a good profile for ride position A, then use the realtime CdA feature as a tool to test different ride positions and/or equipement that affects your aero drag. It would help to use roads of the same quality and the same tire pressure because you want the only difference between your Newton and the DFPM to be due to the aerodynamic changes you are testing. Of course, don't draft while realtime CdA to test ride position. And the less traffic the better. Also, keep in mind that the reported CdA is filtered with about a 30 sec time constant so you need to hold your position for a while before you start looking at the reported CdA. But none of this is new--if you do a little effort you can use the realtime CdA function as a tool to improve your aerodynamics and that has always worked well and now works better with OS2.

That said, I found problems with the way Time Advantage was being reported in both the FW and the SW. First of all, it is only possible to show an approximation of Time Advantage on the Newton while you are riding. I've created a numerical solution of Time Advantage and used that to improve the approximation that we are going to start using in the FW.

bjrmd, you attached a couple of ride files to this thread. The following screenshots are from those rides:

TA1.PNG (44.26 KiB) Viewed 17980 times

TA2.PNG (54.73 KiB) Viewed 17969 times

The white plot shows what the right answer for Time Advantage based on a numerical solution using the DFPM data together with the Newton data.
The green plot show what the FW has been reporting--there were bugs obviously.
The red plot shows what the SW has been reporting--there were different bugs obviously.
The blue plot shows what we are going to try next in the FW. It is only an approximation, but it should be much better than before.

By the way, I will start using the numerical solution in the Isaac SW in the future also--watch for it to come out in a future release.

The bottom line is that there is no reason to worry about the realtime CdA--it has been good and is even better with OS2. It takes some effort on your part, but it remains a useful tool. We had bugs in Time Advantages in both the FW and SW--fixes are coming. Once they come out, don't worry about small differences between what you see approximated while riding and the numerical solution that you will see in a future release of Isaac.

Travis

[bjrmd]

Thanks for the new versions and bug fixes

Noted close time advantage with newton and isaac now--great

But--The numbers may not make sense--ie--the powertap avg was reading lower than the ibike by a few watts but the time advantage was negative and cda numbers high.
Data enclosed

[travispape]

But--The numbers may not make sense--ie--the powertap avg was reading lower than the ibike by a few watts but the time advantage was negative and cda numbers high.
Data enclosed

You are absolutely correct that, in general, if the DFPM reads lower than iBike the general expectation would be that your CdA is low. If you improve your aerodynamics, then you achieve higher speeds for a given effort causing the resistance-based measurement to be higher than the DFPM power measurement. A low CdA, in turn, would result in positive Time Advantage.

But the devil is in the details--and fortunately the unit and the SW is handling the details correctly. I took a close look at your ride file and what I found is that at the highest wind speeds, where your aerodynamics matter most, the iBike power measurement was actually lower than the DFPM. It was at the slowest wind speeds, where aerodynamics are only a small portion of the frictional resistance, where the DFPM was reading lower.

It might be worth double-checking your profile. Do you have really good tires? Or do you have really good roads to ride on? It might be that your Crr needs to be lower. Also be sure you are riding with the correct weight setting.

But based on the profile that you are using and the power measurements relative to the wind, it looks like everything is working correctly.

Travis

[bjrmd]

Thanks
I reviewed the file and it also seems that the cda goes up on downhill portions to a greater degree than up/or level. I have noted that on other rides as well--is this something that should be ignored or warrant a profile change?

Tires are Hutchinson intensive tubeless (25) and the roads are asphalt

[Rick]

There seems to be much that is discussed here and assumed knowledge that goes way beyond what is in the Newton Instructions pdf. Is there another source users should be reading beyond the pdf? I.E should one read manuals for the other/earlier iBike products to be fully versed?