According to the firmware release 4, the coast down range is still 18 mph to 9 mph. This is quite low to differentiate CdA and Crr. For example, assuming Crr = 0.55% and M = 75 kg and CdA = 0.35 m² with rho = 1 kg/m³, that's 4.0 N of retarding force from rolling resistance and 11.3 N of retarding force from wind resistance @ 18 mph. On the other hand, were the coast-down to begin processing data at 30 mph, which every racing cyclist should have no problem reaching, the retarding force from rolling resistance is of course the same, but the force from wind resistance goes up to 31.5 N: almost 3x as large. This would help more clearly differentiate rolling and wind resistance. Down at 9 mph, using these numbers, the wind resistance drops to 2.9 N, allowing for the reasonable determination of Crr if CdA can be well estimated from the high end of the velocity range.
This has been discussed here before, but I'm wondering why the coast-down range wasn't made variable....
coast down velocity range
-
lorduintah
- Posts: 662
- Joined: Fri Jan 11, 2008 9:37 am
- Location: Plymouth, MN
Re: coast down velocity range
I personally think it could be a little higher - but 30 mph is far too high for your average Joe. I am having little effort getting to 22 mph when I crank up for a coast down and then I have to wait for a slowdown to 18 for the data to start. Some variability would be nice for the above average biker, though. If it were really smart, you could crank up your speed to X mph and have the computer start taking data at x-2 mph or something like that as long as you got to 20 mph??? Detecting this would be as simple as zero cadence.
I also have been doing 8 mile cal rides and the plots are really clean. What would it take to get plots for the coastdowns?
Tom
I also have been doing 8 mile cal rides and the plots are really clean. What would it take to get plots for the coastdowns?
Tom
Re: coast down velocity range
It would be nice... but on the other hand, it isn't tough to figure CdA out after the fact by looking at your wattage on descents (assuming your wind scaling is spot on). Crr will always be difficult to extract accurately since a small error in slope is a pretty big error in Crr.djconnel wrote:This has been discussed here before, but I'm wondering why the coast-down range wasn't made variable....
-
lorduintah
- Posts: 662
- Joined: Fri Jan 11, 2008 9:37 am
- Location: Plymouth, MN
Re: coast down velocity range
In statistics, a simple linear or rank regression with excess data allows an estimate of the error on the parameters as well as the overall determination (R^2). So I am assuming that Crr and CdA are a couple of parameters that could fall into that category of determinations. Why not have the software (totally too much for the iBike itself) - version of the calibration provide such estimates? You get those and an error bar on wind scale (so that it is "spot on") and then the math does the talking, not a little guess from a black box.
Tom
Tom
Re: coast down velocity range
I'm not sure what you are getting at. I'm pretty sure that the software does a regression on your speed vs distance curves from the CDs. Obviously to get CdA and Crr everything else must be a known value, and the regression determines a CdA and Crr that is a best fit of the data. Slope errors will directly effect Crr and wind scaling errors will directly effect CdA... meaning that they will be indistinguishable.
-
lorduintah
- Posts: 662
- Joined: Fri Jan 11, 2008 9:37 am
- Location: Plymouth, MN
Re: coast down velocity range
So as in any regression, the error estimate of the parameters and the overall regression statistic represent how significant the parameter estimates are - done by regression, one can use that information to critique the estimates as to their worthiness. If these statistics were provided one could see the "goodness of fit".
Tom
Tom
Re: coast down velocity range
I've done this myself using Excel, and you can too if you wish. Just do coastdowns with the iBike in regular mode, and do an iterative fit on the equations of motion with CdA and Crr as variables.
I don't think the "goodness of fit" is the most important part, but rather separating the CdA and Crr. You can have a very good fit but that doesn't mean this numbers will be accurate. IME getting the Crr to a very precise degree is difficult... which is why it's usually better to just ascribe a reasonable value to it and then you only need to fit CdA. As Dan mentioned, determining an accurate CdA is easier at high speeds and 20mph isn't the best starting point. But if you are concerned about your CdA being off you can check this at any time while coasting at higher speeds (ie the closer to 0 watts the better).
I don't think the "goodness of fit" is the most important part, but rather separating the CdA and Crr. You can have a very good fit but that doesn't mean this numbers will be accurate. IME getting the Crr to a very precise degree is difficult... which is why it's usually better to just ascribe a reasonable value to it and then you only need to fit CdA. As Dan mentioned, determining an accurate CdA is easier at high speeds and 20mph isn't the best starting point. But if you are concerned about your CdA being off you can check this at any time while coasting at higher speeds (ie the closer to 0 watts the better).
-
lorduintah
- Posts: 662
- Joined: Fri Jan 11, 2008 9:37 am
- Location: Plymouth, MN
Re: coast down velocity range
Yes, Crr is hard to pull out. Tilt and wind scaling seem to be pretty straightforward - I have had very clean data for that with a much longer ride of 8 miles (4/4).
Tom
Tom
Re: coast down velocity range
My experience with the iSport, if you want the most accurate CdA calibration you can get, don't bother with a coast down calibration. Instead, set Crr to a fixed value and coast down a hill at 35mph. Do multiple runs and set CdA so to get near 0 watts when coasting. Of course across multiple coasts you are going to have to hold everything as constant as possible, position, clothing, etc.
Having repeatedly done this for many months now, across various weather conditions, roads, and what not, I have never gotten to the point where I always see 0 watts when coasting at high speed. I am not sure why not. Maybe the basic assumption that CdA is fixed is simply not true in real life.
Bottom line. I am getting very consistent and reliable results with the iSport while climbing at less than 10mph. CdA and Crr matter little, and the iSport does a great job measuring riding tilt and road slope. For climbing the iSport is fantastic. But at higher speeds over 20mph, iSport power is not reliable in my opinion, no matter how much CdA tweaking you try to do. Still useful at higher speeds, just not reliable and consistent as you might expect or want.
Having repeatedly done this for many months now, across various weather conditions, roads, and what not, I have never gotten to the point where I always see 0 watts when coasting at high speed. I am not sure why not. Maybe the basic assumption that CdA is fixed is simply not true in real life.
Bottom line. I am getting very consistent and reliable results with the iSport while climbing at less than 10mph. CdA and Crr matter little, and the iSport does a great job measuring riding tilt and road slope. For climbing the iSport is fantastic. But at higher speeds over 20mph, iSport power is not reliable in my opinion, no matter how much CdA tweaking you try to do. Still useful at higher speeds, just not reliable and consistent as you might expect or want.
Re: coast down velocity range
There are a couple of things that can skew zero wattage coasting down a hill. When I do the same test I look for zero watts at about 18-24mph or so since that's going at a good clip on level road. IOW with a headwind I'll be below that and with a tailwind I'll stay above that consistently. Generally below 18mph the watt displayed are negative and above 24mph they're positive.
You also might find yourself tucking in tight or even dropping down to a real tuck position. This goes against the setup that you have and will skew numbers in this quasi coastdown mode.
For my situation, the wind is too variable flying down a hill at breakneck speed. At times I can feel gusts coming off the ocean and my wattage will jump around because of this. However, this coincides with the effort that would be required on a level into a gusting headwind.
I've got my climbing wattage nailed to the point I'll put it up against any other PM. I'm actually trying to borrow someones PT when they go on vacation.
On the flats I'm very satisfied with the results. Whether in a group or solo, regardless of the wind direction I think my wattage is consistent and reliable. I'm participating in a discussion on another forum about the same subject and I can't seem to find a way to PROVE the iBike matches another PM without actually having it on the bike.
BTW, it doesn't make any difference if you have an iSport or iAero. They all do the same job of measuring watts. The iAero has more bells and whistles and of course allows downloading, but the basic function of measuring wattage is identical across the models.
You also might find yourself tucking in tight or even dropping down to a real tuck position. This goes against the setup that you have and will skew numbers in this quasi coastdown mode.
For my situation, the wind is too variable flying down a hill at breakneck speed. At times I can feel gusts coming off the ocean and my wattage will jump around because of this. However, this coincides with the effort that would be required on a level into a gusting headwind.
I've got my climbing wattage nailed to the point I'll put it up against any other PM. I'm actually trying to borrow someones PT when they go on vacation.
On the flats I'm very satisfied with the results. Whether in a group or solo, regardless of the wind direction I think my wattage is consistent and reliable. I'm participating in a discussion on another forum about the same subject and I can't seem to find a way to PROVE the iBike matches another PM without actually having it on the bike.
BTW, it doesn't make any difference if you have an iSport or iAero. They all do the same job of measuring watts. The iAero has more bells and whistles and of course allows downloading, but the basic function of measuring wattage is identical across the models.
Fernando
Re: coast down velocity range
I did a weekly mtn climb race this summer and noticed that the iBike varied in it's determination of the avg grade ~3%... so I started normalizing the results to a fixed grade. On shorter climbing intervals (5 min) the variation was a bit greater than this.mds wrote:Bottom line. I am getting very consistent and reliable results with the iSport while climbing at less than 10mph. CdA and Crr matter little, and the iSport does a great job measuring riding tilt and road slope. For climbing the iSport is fantastic. But at higher speeds over 20mph, iSport power is not reliable in my opinion, no matter how much CdA tweaking you try to do. Still useful at higher speeds, just not reliable and consistent as you might expect or want.
At high speeds and fluctuating winds, the iBike has a challenging time with accuracy on short time scales, but it seems to even out if you have your CdA and wind scaling properly set.
Re: coast down velocity range
Good discussion! If the model was perfect, and measurements were perfect, you could extract two unknowns (wind and rolling resistance coefficients) from two measurements (for example deceleration at 18 mph and 17 mph). However, none of these things are perfect. The key is to avoid small errors in the estimation of wind resistance being compensated for by errors in rolling resistance of the opposite sign, or vice versa.
To some degree, it is probably better to use rolling resistance data (for example, adding +15% to Al Morrison's rolling resistance data on http://biketechreview.com to account for rough roads versus the smooth rollers he used, an approach recommended by Kraig Willett) and just solving for wind resistance. However, if the coast down starts as a sufficient speed that the wind resistance truly dominates, then the CdA will be basically locked down by this higher velocity range, and the algorithm will therefore reserve lower end speeds for strictly estimating rolling resistance. The goal is basically what Flannery and Press (Numerical Recipes) would call a "conjugate parameter set"; one in which errors in one parameter don't muck up the estimation of other parameters.
My concern about descents: ideally if you use descents for estimating CdA at least imaging your pedaling. Adopt a pedaling position and even turn the cranks a bit while avoiding engaging the freehub. If you tuck, CdA will be substantially underestimated for when it counts.
To some degree, it is probably better to use rolling resistance data (for example, adding +15% to Al Morrison's rolling resistance data on http://biketechreview.com to account for rough roads versus the smooth rollers he used, an approach recommended by Kraig Willett) and just solving for wind resistance. However, if the coast down starts as a sufficient speed that the wind resistance truly dominates, then the CdA will be basically locked down by this higher velocity range, and the algorithm will therefore reserve lower end speeds for strictly estimating rolling resistance. The goal is basically what Flannery and Press (Numerical Recipes) would call a "conjugate parameter set"; one in which errors in one parameter don't muck up the estimation of other parameters.
My concern about descents: ideally if you use descents for estimating CdA at least imaging your pedaling. Adopt a pedaling position and even turn the cranks a bit while avoiding engaging the freehub. If you tuck, CdA will be substantially underestimated for when it counts.
-
lorduintah
- Posts: 662
- Joined: Fri Jan 11, 2008 9:37 am
- Location: Plymouth, MN
Re: coast down velocity range
dj -
You have a good idea - to expand on it. With several coast-downs originating from different velocities you have a capability of multiple data sets that in theory have to lead to the same answer. So you could use a set from 18 - 25 mph as max velocities (4 or 5 individual CDs at least). So instead of reducing data over the same range, you would be doing so over a different type of data - but one where all parameters would have a better chance of being determinate.
Tom
You have a good idea - to expand on it. With several coast-downs originating from different velocities you have a capability of multiple data sets that in theory have to lead to the same answer. So you could use a set from 18 - 25 mph as max velocities (4 or 5 individual CDs at least). So instead of reducing data over the same range, you would be doing so over a different type of data - but one where all parameters would have a better chance of being determinate.
Tom
Re: coast down velocity range
I'm not totally sure I understand your point.....
The data fitting procedure to a coast-down uses all of the data values, not just the start and stop speeds, or at least I think so! So if you do multiple coast-downs, each should include as many speeds as possible. In other words, if you start one coast-down @ 25 mph, then are going to do a second on, you're better off starting the second @ 25 mph than starting it @ 20 mph.
I've not yet gotten an iBike, but I admit to suffering data envy when I see what riders who do have them show me. So in March, when the "confidential news" about new developments is released, I'll be very interested.
The data fitting procedure to a coast-down uses all of the data values, not just the start and stop speeds, or at least I think so! So if you do multiple coast-downs, each should include as many speeds as possible. In other words, if you start one coast-down @ 25 mph, then are going to do a second on, you're better off starting the second @ 25 mph than starting it @ 20 mph.
I've not yet gotten an iBike, but I admit to suffering data envy when I see what riders who do have them show me. So in March, when the "confidential news" about new developments is released, I'll be very interested.
-
lorduintah
- Posts: 662
- Joined: Fri Jan 11, 2008 9:37 am
- Location: Plymouth, MN
Re: coast down velocity range
I would respond in this manner:
1 - It is more difficult for some to get to a higher speed and then coast. So have a range available.
Why fight to get to the same speed - just let her rip and get above x - then coast. Let the software start data collection at 2 mph less than that (or some value) which offers enough time to get into "riding position."
2 - I do not see the need for exactly the same speed range to be applied several times.
CDs that have a different overall range will be affected slightly differently by Crr (as a MAGNITUDE of the overall impact on the rate of speed loss). This has the opportunity to provide different starting points for the determination and thus might very well lead to a better estimate of the effective "intercept" along with the slope.
3 - Multiple data sets of essentially the same observable (rate of deceleration) like many other linear observations, are better determined if each individual sample is unique in some fashion - removing bias in the data. By not using the exact same speed window, some data bias is minimized.
Tom
1 - It is more difficult for some to get to a higher speed and then coast. So have a range available.
Why fight to get to the same speed - just let her rip and get above x - then coast. Let the software start data collection at 2 mph less than that (or some value) which offers enough time to get into "riding position."
2 - I do not see the need for exactly the same speed range to be applied several times.
CDs that have a different overall range will be affected slightly differently by Crr (as a MAGNITUDE of the overall impact on the rate of speed loss). This has the opportunity to provide different starting points for the determination and thus might very well lead to a better estimate of the effective "intercept" along with the slope.
3 - Multiple data sets of essentially the same observable (rate of deceleration) like many other linear observations, are better determined if each individual sample is unique in some fashion - removing bias in the data. By not using the exact same speed window, some data bias is minimized.
Tom
Re: coast down velocity range
Not sure what you are getting at Tom, but I'm almost certain that the software is doing a curve fit on the entire data set... so a coastdown from 30-5mph will contain as much data as a coastdown from 30-20, + 20-10, + 10-5.
Honestly though, it is pretty easy to get the right values without doing the coastdowns at all... do some wind scaling runs, set your Crr to a reasonable value, and then check your CdA via wattage while coasting.
Honestly though, it is pretty easy to get the right values without doing the coastdowns at all... do some wind scaling runs, set your Crr to a reasonable value, and then check your CdA via wattage while coasting.
-
lorduintah
- Posts: 662
- Joined: Fri Jan 11, 2008 9:37 am
- Location: Plymouth, MN
Re: coast down velocity range
Ron/DJ -
I am looking at this from a pure statistical sampling view.
I think you appear to not have much data sampling nor analyses experience - even with regression. If I am incorrect on this account, sorry. Multiple determinations from separate data sets - each coast down should be a different data set - regardless of the number of points - in a regression treatment can minimize errors, this is true in the case such as cdA and Crr where there is correlation of the two in that we cannot develop an experiment or set of experiments that allow one to be fixed and the other to vary (because in actuality there are more than two variables) - orthogonal treatments, such as one finds in properly run Design of Experiments (DOEs). Consequently, by taking data several times and over a different range (speed windows), one can expect any well coded regression process to be more capable of deterministic solutions to the variable coefficients. The mass of an electron was not determined as the result of one experiment. One of many benchmarks that can be used to assess the merit of a determination is the overall R^2, another is Goodness of Fit as well as the MSE and significance or chi-square values of the coefficients and estimated errors at some significance level - we don't see these reported in the iBike software, so it is difficult to know how good a job we did with the cal ride and coast-downs. The only way we can infer is by getting similar values after repeated trials and then making further adjustments like you suggest to the wind scaling and cdA.
As far as the amount of data - if the interval is every second, then the data from a CD of 10 to 5 mph will NOT have as much as that of a 20 to 5 mph CD. We get around part of this by repeated trials - whatever the speed window is.
I cannot figure out a way to convey experimental error propagation and statistical inference to you that seems to ring well. My shortcoming.
Tom
I am looking at this from a pure statistical sampling view.
I think you appear to not have much data sampling nor analyses experience - even with regression. If I am incorrect on this account, sorry. Multiple determinations from separate data sets - each coast down should be a different data set - regardless of the number of points - in a regression treatment can minimize errors, this is true in the case such as cdA and Crr where there is correlation of the two in that we cannot develop an experiment or set of experiments that allow one to be fixed and the other to vary (because in actuality there are more than two variables) - orthogonal treatments, such as one finds in properly run Design of Experiments (DOEs). Consequently, by taking data several times and over a different range (speed windows), one can expect any well coded regression process to be more capable of deterministic solutions to the variable coefficients. The mass of an electron was not determined as the result of one experiment. One of many benchmarks that can be used to assess the merit of a determination is the overall R^2, another is Goodness of Fit as well as the MSE and significance or chi-square values of the coefficients and estimated errors at some significance level - we don't see these reported in the iBike software, so it is difficult to know how good a job we did with the cal ride and coast-downs. The only way we can infer is by getting similar values after repeated trials and then making further adjustments like you suggest to the wind scaling and cdA.
As far as the amount of data - if the interval is every second, then the data from a CD of 10 to 5 mph will NOT have as much as that of a 20 to 5 mph CD. We get around part of this by repeated trials - whatever the speed window is.
I cannot figure out a way to convey experimental error propagation and statistical inference to you that seems to ring well. My shortcoming.
Tom
Re: coast down velocity range
I guess my point is that each run is a new independent data set... and the greater the speed range, the more data we have... and also the better chance we have of separating the two variables. The reason this works at all is because the power to overcome aero drag is proportional to V^3 and the power to overcome rolling drag is proportional to V. Therefore, the greater the range of speed during the CD, the better chance we have of separating them. Aero will dominate at high speeds and rolling drag at low speeds.
One important issue I've found is that the tilt fluctuates and varies at least on the roads around here... meaning that each time you do a CD run the observed tilt will be slightly different... and it only needs to be "off" by .1% to equal ~20% error in Crr. Add that to the difficulty of extracting two variables from the CD curve, and I feel that pegging the Crr to a reasonable value is likely the best bet for most people.
One important issue I've found is that the tilt fluctuates and varies at least on the roads around here... meaning that each time you do a CD run the observed tilt will be slightly different... and it only needs to be "off" by .1% to equal ~20% error in Crr. Add that to the difficulty of extracting two variables from the CD curve, and I feel that pegging the Crr to a reasonable value is likely the best bet for most people.
-
lorduintah
- Posts: 662
- Joined: Fri Jan 11, 2008 9:37 am
- Location: Plymouth, MN
Re: coast down velocity range
Ron -
Cannot disagree - especially without knowing the details of the data reduction used in the iBike analysis. That and minor displacements of the orientation of the iBike affect the response of the unit to the external world and also that each rider has different wheels/tires and tubes let alone the aerodynamic cross-section of the bike and body.
Tom
Cannot disagree - especially without knowing the details of the data reduction used in the iBike analysis. That and minor displacements of the orientation of the iBike affect the response of the unit to the external world and also that each rider has different wheels/tires and tubes let alone the aerodynamic cross-section of the bike and body.
Tom
Re: coast down velocity range
Here is one of my experiments in soft peddle data collection and analysis.
My wind cal is dialed, my wind offset is fresh.
Notice that the selected portion has a tail wind that does not go fully to zero, as a portion to the right does. I guess that would make the tail wind effect suspect. Anyway the net result with this selection is that aero is reduced nearing minimum with still a bit of bike speed to generate rolling resistance power.
I have tweeked the Crr up from .005 to .0051 making .4 watts of average power appear. With the .005 value, the min avg and max power numbers are all zero. High speed portions with a head wind make the CdA have a similar advantage.
What do you all think?
Russ
My wind cal is dialed, my wind offset is fresh.
Notice that the selected portion has a tail wind that does not go fully to zero, as a portion to the right does. I guess that would make the tail wind effect suspect. Anyway the net result with this selection is that aero is reduced nearing minimum with still a bit of bike speed to generate rolling resistance power.
I have tweeked the Crr up from .005 to .0051 making .4 watts of average power appear. With the .005 value, the min avg and max power numbers are all zero. High speed portions with a head wind make the CdA have a similar advantage.
What do you all think?
- Clipboard_Paste.jpg (51.93 KiB) Viewed 22091 times
Re: coast down velocity range
Here is a high speed selection from another ride.
Both of these were on smooth pavement.
Russ
Both of these were on smooth pavement.
Russ
- Clipboard_Paste.jpg (53.15 KiB) Viewed 22060 times
Re: coast down velocity range
My first thought is that the .4 watts appears because the software is referencing the result (zero watts) rather than the original data. If you are always seeing zero while coasting, these would indicate that your overall drag numbers in the iBike are too low... ie you are putting out more power than it shows.Russ wrote:I have tweeked the Crr up from .005 to .0051 making .4 watts of average power appear. With the .005 value, the min avg and max power numbers are all zero. High speed portions with a head wind make the CdA have a similar advantage.
I think you need to put the unit on the "coast" screen... this way you can see negative values. It won't record them however... you'll just be able to see it on the screen.
Re: coast down velocity range
Thanks for your feedback rruf!
Well, I actually run a profile with .0057 Crr currently. The zero value, appears at .005 on very smooth roads when conditions allow. I did pick two 'perfect examples' but these are also the
majority and also I get the same with both my iPro and my iAero gen3's. Now with colder weather, I saw a higher Crr work out on my ride of .0057.
The Crr is below the 'factory' measured and also another published value for this tire of .0058. My profile from Boyd was last March so also colder than the .0050 value during my sample rides with temps in the 80's F. The March CD's were not on quite as smooth road and at 50F.
I think this method is proving out as good, at least for the Crr and seems to me a good tailwind with a ride speed a few MPH faster coasting and soft peddling should give a great result for Crr providing your tilt is good and the wind is not too gusty. With the higher speed and tailwind, the rolling resistance power goes up and the tailwind would suppress the CdA to the point any error due to the CdA become quite low!
Now, my CdA may infact be a bit low with this method and my iAero frequently will report higher but sometimes lower too on smooth roads. This CdA is a bit lower than Coach Boyd came up with from my coastdown data. But that CD set was done at 50F and in my winter kit. I haven't bugged Boyd with my summer kit stuff but .321 down from .347(Boyds) ..... seems at least in the ball park,
What do you think?
Russ
Well, I actually run a profile with .0057 Crr currently. The zero value, appears at .005 on very smooth roads when conditions allow. I did pick two 'perfect examples' but these are also the
majority and also I get the same with both my iPro and my iAero gen3's. Now with colder weather, I saw a higher Crr work out on my ride of .0057.
The Crr is below the 'factory' measured and also another published value for this tire of .0058. My profile from Boyd was last March so also colder than the .0050 value during my sample rides with temps in the 80's F. The March CD's were not on quite as smooth road and at 50F.
I think this method is proving out as good, at least for the Crr and seems to me a good tailwind with a ride speed a few MPH faster coasting and soft peddling should give a great result for Crr providing your tilt is good and the wind is not too gusty. With the higher speed and tailwind, the rolling resistance power goes up and the tailwind would suppress the CdA to the point any error due to the CdA become quite low!
Now, my CdA may infact be a bit low with this method and my iAero frequently will report higher but sometimes lower too on smooth roads. This CdA is a bit lower than Coach Boyd came up with from my coastdown data. But that CD set was done at 50F and in my winter kit. I haven't bugged Boyd with my summer kit stuff but .321 down from .347(Boyds) ..... seems at least in the ball park,
What do you think?
Russ
Re: coast down velocity range
Are you using the zero watts result from coasting (soft pedaling) at 40km/hr to verify your Crr? AFAIK that is impossible. For one thing the zero power reading only means that your actual iBike derived power is *less* than zero, since negative values are always reported as zero. If you look at the "coast" screen you will see the real number. You will notice that this number will fluctuate a lot. Besides that the dominant drag is aero at 40km/hr not rolling, so you can't determine if the Crr is good at that speed.Russ wrote:Well, I actually run a profile with .0057 Crr currently. The zero value, appears at .005 on very smooth roads when conditions allow. I did pick two 'perfect examples' but these are also the
majority and also I get the same with both my iPro and my iAero gen3's. Now with colder weather, I saw a higher Crr work out on my ride of .0057.
Like I mentioned earlier, if you are always (or usually) getting negative power values when you coast, this indicates that one or more of your drag values are too low... probably CdA and/or wind scaling. Wind scaling is easy to check independently, so get that one nailed first.
Re: coast down velocity range
I have a question/comment:
I notice that Russ' total riding weight is 108kg. I'm typically at 100kg. Assuming (yeah I know what happens when you assume) we're more or less typical height and weigh more than we should for cycling your CdA seems low.
I am 100kg and use .381CdA and .0055Crr
My friend is 103kg and uses .389 and .0057Crr
Russ is 108kg and uses .321CdA and .0060Crr
The Crr seems fine but the CdA seems very low. Again this is just trying to fit average people into an average set of numbers. Obviously different positions, different bikes, different everything; but that different? Also Russ you mentioned .005 and .0051 as Crr yet in the screenshots it appears the Crr is .006.
I notice that Russ' total riding weight is 108kg. I'm typically at 100kg. Assuming (yeah I know what happens when you assume) we're more or less typical height and weigh more than we should for cycling your CdA seems low.
I am 100kg and use .381CdA and .0055Crr
My friend is 103kg and uses .389 and .0057Crr
Russ is 108kg and uses .321CdA and .0060Crr
The Crr seems fine but the CdA seems very low. Again this is just trying to fit average people into an average set of numbers. Obviously different positions, different bikes, different everything; but that different? Also Russ you mentioned .005 and .0051 as Crr yet in the screenshots it appears the Crr is .006.
Fernando
Re: coast down velocity range
racefern and rruf, thanks for the responses!
My weight is about 93kg, 205lbs and the bike about 13.6kg, 30 lbs and my height is 6'1" and I am using prodesign aero bars. I agree that my .321 CdA may be a bit low but the main reason I posted these was to show that it seems a good way to do Crr especially,
not really to hijack the thread into my numbers. I am more interested in what you think of the technique? Now that cold weather is back, I will switch to Coach Boyd's profile anyway
I am not confident in .321 as being my current summer kit CdA but I am confident that it is no lower than that.
The fascinating result I keep getting is that .321 and .005 kept coming up when I did get the straight baseline in the power, at least for a good number of seconds. Exactly what best to do with the tweek of CdA with this method I am not sure, like try to split the power baseline when it is more bumpy?
The cut and paste is a bit fuzzy the Crr was .005, here is another below that I cropped to hopefully make more readable and this time I tweeked the Crr up to .0052 so that you can see a flat baseline of 2.5 watts of power, nearly flat all the way across. I do not always get the flat results like this. Many times I get quite variable power curve and certainly above zero.
My weight is about 93kg, 205lbs and the bike about 13.6kg, 30 lbs and my height is 6'1" and I am using prodesign aero bars. I agree that my .321 CdA may be a bit low but the main reason I posted these was to show that it seems a good way to do Crr especially,
not really to hijack the thread into my numbers. I am more interested in what you think of the technique? Now that cold weather is back, I will switch to Coach Boyd's profile anyway
I am not confident in .321 as being my current summer kit CdA but I am confident that it is no lower than that.The fascinating result I keep getting is that .321 and .005 kept coming up when I did get the straight baseline in the power, at least for a good number of seconds. Exactly what best to do with the tweek of CdA with this method I am not sure, like try to split the power baseline when it is more bumpy?
The cut and paste is a bit fuzzy the Crr was .005, here is another below that I cropped to hopefully make more readable and this time I tweeked the Crr up to .0052 so that you can see a flat baseline of 2.5 watts of power, nearly flat all the way across. I do not always get the flat results like this. Many times I get quite variable power curve and certainly above zero.
- Clipboard_Paste.jpg (72.79 KiB) Viewed 21916 times
Re: coast down velocity range
rruf said:
"Are you using the zero watts result from coasting (soft pedaling) at 40km/hr to verify your Crr? AFAIK that is impossible. For one thing the zero power reading only means that your actual iBike derived power is *less* than zero, since negative values are always reported as zero."
Well, I actually was using the lower speed (first picture) for Crr and the high speed one to see how CdA looked. I know that when cadence is zero that power is forced to zero. During the soft pedaling section, the only negative power is on the gravity line for the down hill section and is positive for the uphill part. I figure that tweeking the CdA or the Crr up (like the .0052 in my third picture) would simply shift the proper amount into rolling resistance power. This, in my thinking adds in a bit extra to raise my baseline for visibility. I can do the same ride with an intentionally high Crr and then tweek the baseline downward, when I get a good flatline result. I am fairly sure I have done this and that this is what caught my attention in the first place about this method!
And again, I do not allways get the (looks good to me) flatline, many times it is wavy and does not seem to work with this method.
Again, my wind was dialed and freshly offset checked for these two (third being closeup of second one, but with .0052crr).
Thanks,
Russ
"Are you using the zero watts result from coasting (soft pedaling) at 40km/hr to verify your Crr? AFAIK that is impossible. For one thing the zero power reading only means that your actual iBike derived power is *less* than zero, since negative values are always reported as zero."
Well, I actually was using the lower speed (first picture) for Crr and the high speed one to see how CdA looked. I know that when cadence is zero that power is forced to zero. During the soft pedaling section, the only negative power is on the gravity line for the down hill section and is positive for the uphill part. I figure that tweeking the CdA or the Crr up (like the .0052 in my third picture) would simply shift the proper amount into rolling resistance power. This, in my thinking adds in a bit extra to raise my baseline for visibility. I can do the same ride with an intentionally high Crr and then tweek the baseline downward, when I get a good flatline result. I am fairly sure I have done this and that this is what caught my attention in the first place about this method!
And again, I do not allways get the (looks good to me) flatline, many times it is wavy and does not seem to work with this method.
Again, my wind was dialed and freshly offset checked for these two (third being closeup of second one, but with .0052crr).
Thanks,
Russ
Re: coast down velocity range
True, but not the point I was making. When *power* is less than zero, it is forced to zero. So if you are soft pedaling and the power reading is zero, this means that the iBike is actually calculating a negative power.Russ wrote:I know that when cadence is zero that power is forced to zero.
Re: coast down velocity range
rruf,
"So if you are soft pedaling and the power reading is zero, this means that the iBike is actually calculating a negative power."
I guess I see your point. Essentially there is more gravity than aero and rolling power added ?
When I tweek the CdA upward, with the zero line, aero power goes down and avg power goes up, providing I tweek it up enough for average power to show up. I have seen cases where the amount of upward tweek required to get an average power reading is large (much larger than one click upward), that case would indicate a significant amount of 'negative' power to me. When one click makes power show up, seems to me it is in fact in balance. Is this wrong?
Are you saying that the iBike throws away some data? When I am soft pedaling, there is no power applied positive or negative, not that we can apply negative power. So the whole equation for those seconds of data are filled with aero, rolling and gravity only. So if power shows up during soft pedaling, where does it come from?
If I intentionally set my CdA and Crr high in a profile so that soft pedaling reports power, and then tweek it downward to zero, would this not be valid?
Thanks,
Russ
"So if you are soft pedaling and the power reading is zero, this means that the iBike is actually calculating a negative power."
I guess I see your point. Essentially there is more gravity than aero and rolling power added ?
When I tweek the CdA upward, with the zero line, aero power goes down and avg power goes up, providing I tweek it up enough for average power to show up. I have seen cases where the amount of upward tweek required to get an average power reading is large (much larger than one click upward), that case would indicate a significant amount of 'negative' power to me. When one click makes power show up, seems to me it is in fact in balance. Is this wrong?
Are you saying that the iBike throws away some data? When I am soft pedaling, there is no power applied positive or negative, not that we can apply negative power. So the whole equation for those seconds of data are filled with aero, rolling and gravity only. So if power shows up during soft pedaling, where does it come from?
If I intentionally set my CdA and Crr high in a profile so that soft pedaling reports power, and then tweek it downward to zero, would this not be valid?
Thanks,
Russ