Slope question

Post Reply
nsidirop
Posts: 28
Joined: Thu Apr 18, 2013 4:43 pm

Slope question

Post by nsidirop »

Today I rode a hill repeat ride however the slope estimation in my Newton seems to differ between the different laps (see attachement - laps 2-5), max slope varies between 15.5% and 17.6%
Anybody who has an idea what this is due to? Could it have anything to do with if I am sitting down or I am out of the saddle so that the iBike unit tilts differently during the climbs?
How does the variation in estimated slope affect the power calculation?
Attachments
iBike_2013_07_17_1838_40_km.ibr
(539.62 KiB) Downloaded 374 times
Zoltan
Posts: 213
Joined: Mon Jan 23, 2012 9:47 am
Location: HUNGARY

Re: Slope question

Post by Zoltan »

nsidirop wrote:Today I rode a hill repeat ride however the slope estimation in my Newton seems to differ between the different laps (see attachement - laps 2-5), max slope varies between 15.5% and 17.6%
Anybody who has an idea what this is due to? Could it have anything to do with if I am sitting down or I am out of the saddle so that the iBike unit tilts differently during the climbs?
How does the variation in estimated slope affect the power calculation?
Acccurate instantaneous acceleration measurement of your bike at low speed is impossible with one magnet => slope will be inaccurate, too. More explanation is here: http://www.ibikeforum.com/viewtopic.php?f=6&t=2546

Let me highlight that I still use 5 magnets evenly spaced on my first wheel.
User avatar
racerfern
Posts: 1356
Joined: Fri Jan 11, 2008 12:47 am
Location: Baldwin, NY
Contact:

Re: Slope question

Post by racerfern »

Magnets have nothing to do with it. The one moment that some max slope got recorded does not affect the average watts for that lap. If you zoom in on a lap, say #5 then select the max section, you're talking about a two second window over the time you spent on that lap. Newton happened to record the slope it saw at that particular instant that can be affected by pulling harder on the bars, being out of saddle, etc. at that instant and causing the slightest change. There is no need to analyze down to this detail. The key to this exercise is the climbing effort and the average wattage for that effort.

You wouldn't worry about the minimum slope on the downhill section and for that matter you don't press the lap button in EXACTLY the same spot so the slightest variation will affect the average slope ever so slightly. I think you're data is spot on, you're just worried about the devil in the details when it's a bigger picture that's more important.

To allay your fears, here's my intervals from yesterday with essentially the same variations you have. The difference is I hit the lap button at the top of the climb and coast down. There's one more difference, you're a lot stronger. :cry:

Edit: BTW, this is a powerstroke ride so if you're curious you can see the watts start to get wasted as the efforts get harder and harder. What I found interesting is how the first and last lap which are the same terrain have different powerstroke. After all those efforts, the last lap has a much more even stroke than the first lap. I do this exact sequence about once a month, then I sum the thirteen climbs and compare it against other rides of the same terrain. I tend to not even drill into the laps, since the first couple you're warming up and the last one or two you're pretty much cooked. I look at the sum total of the time to complete all thirteen climbs.
Attachments
FM_07_25_2013_1819_15_Miles_TCX_HiDef.ibr
(959.56 KiB) Downloaded 411 times
Fernando
Zoltan
Posts: 213
Joined: Mon Jan 23, 2012 9:47 am
Location: HUNGARY

Re: Slope question

Post by Zoltan »

Does Newton work in a different way than Gen3 as far as slope calculation is concerned?

The accelerometer data has to be cleaned from the effect of acceleration to get the slope.

The slower one rides the less acceleration data per secundum is available. On slopes like 17% or more, it is quite common to ride at 2-3m/s, meaning that iBike has a mere 1-1.5 speed data per sec (NB: the circumference of the wheel is roughly 2m). If you have such rare speed input, it means that also acceleration can be measured so rarely while the bike in reality has sudden accelerations acnd decelerations on steep ascent depending on where your legs are in the pedalling cycle.

Of course there are other factors which can result in different slope measurement and these were listed once by Travispape. The more evident of them is the eventual "shaking" of the handlebar.

Am I wrong?
nsidirop
Posts: 28
Joined: Thu Apr 18, 2013 4:43 pm

Re: Slope question

Post by nsidirop »

Nice effort racefern. Also very consistent slope readings during the 13 climbs. Your last climb is smooth, you stay seated and that's probably the reason for wasting only 4 W (1.6% of 244 W). On climb #10 your out of the saddle and your wasted watts jump to 13 W (4% of 328 W). On the other hand you raise your effort with almost 100 W so that's the price you have to pay for going harder.
User avatar
racerfern
Posts: 1356
Joined: Fri Jan 11, 2008 12:47 am
Location: Baldwin, NY
Contact:

Re: Slope question

Post by racerfern »

@Zoltan: My point is...whatever. The point is that between pulling on the bars, leaning the bike side to side and the effect of turning the fork as you fight your way up a 16+% incline can throw off the iBike calculations and store a higher value. The point is that it is only for a second and has no effect on the overall climbing results.

It's the same effect as when you pull away from a stop and see a sudden high tilt number even though you're on a flat. It's just a number for a very brief period and has no effect on the overall results.

I'm replying also because I think although numbers are obviously the answers to our power quests, at times many people (myself included) can become obsessed with some number that doesn't fit their thinking or their experience. 99.9% of the time these are outliers and should be ignored. Keeping the bigger picture in mind can be as challenging as drilling into the minute data.
Fernando
User avatar
Russ
Posts: 370
Joined: Sun Feb 10, 2008 4:08 pm

Re: Slope question

Post by Russ »

nsidirop,
Your original question, I am strongly impressed, would be in the realm considered iBike proprietary secrets. So the best Zoltan, me and others who are more interested in how things work can do is make analysis based on physics and how the devices such as accelerometers perform in the realm that is public knowledge.

So as I see it, Zoltan, and racefern to a degree has shed light on your original question while racefern has directed the analysis to the practical and usable cases.

That said, I have been reading in the forums for do it yourself robots and drones. The limits of accelerometers, giros, electronic compasses, gps, air pressure and speed/distance sensors for robotics are being melded by a higher level approach called either sensor fusion or motion fusion. These techniques use the best of each sensor type to correct for shortcomings of the others to produce a unified more accurate 3d picture.

That said, I suspect iBike is using a home grown specialized form of sensor fusion that uses the barometric pressure and the speed/distance sensors to limit and or remove the accumulated errors of the accelerometer that would probably be useless after a short time were it all by itself. This an admirable feet in itself and unless they come out and thump the chest a little and shed a bit more light, the only thing we can do more is go along for the ride :-)

For me it has been a great ride.
Thanks,
Russ
Velocomp
Velocomp CEO
Posts: 7793
Joined: Fri Jan 11, 2008 8:43 am

Re: Slope question

Post by Velocomp »

Zoltan wrote:Does Newton work in a different way than Gen3 as far as slope calculation is concerned?

The accelerometer data has to be cleaned from the effect of acceleration to get the slope.

The slower one rides the less acceleration data per secundum is available. On slopes like 17% or more, it is quite common to ride at 2-3m/s, meaning that iBike has a mere 1-1.5 speed data per sec (NB: the circumference of the wheel is roughly 2m). If you have such rare speed input, it means that also acceleration can be measured so rarely while the bike in reality has sudden accelerations acnd decelerations on steep ascent depending on where your legs are in the pedalling cycle.

Of course there are other factors which can result in different slope measurement and these were listed once by Travispape. The more evident of them is the eventual "shaking" of the handlebar.

Am I wrong?
Your analysis is quite correct.
John Hamann
Zoltan
Posts: 213
Joined: Mon Jan 23, 2012 9:47 am
Location: HUNGARY

Re: Slope question

Post by Zoltan »

Thanks for the confirmation.

I have never given compliments too easily, but I still love the slope measurement of my iBike Gen3, and it is superior to that of any other gadgets of mine.
KenS
Posts: 141
Joined: Sun Feb 28, 2010 2:55 pm
Location: Melbourne, Australia

Re: Slope question

Post by KenS »

There's a device out there with a name disturbingly similar to Velocomp that uses two magnets near the hub and a magnetometer to measure the field between them to a claimed 1 degree or 0.01 metres precision. But it is Bluetooth rather than ANT+
Unfortunately a competitor to the Newton but if Newton were combined with a similar type of speed sensor it would no doubt open more possibilities for even better accuracy and other features.
-- Ken
Velocomp
Velocomp CEO
Posts: 7793
Joined: Fri Jan 11, 2008 8:43 am

Re: Slope question

Post by Velocomp »

KenS wrote:There's a device out there with a name disturbingly similar to Velocomp that uses two magnets near the hub and a magnetometer to measure the field between them to a claimed 1 degree or 0.01 metres precision. But it is Bluetooth rather than ANT+
Unfortunately a competitor to the Newton but if Newton were combined with a similar type of speed sensor it would no doubt open more possibilities for even better accuracy and other features.
I know the product (but I don't remember its name) and it's not a competitor. It measures speed and wheel acceleration accurately but does not measure wind speed or gravitational acceleration. So, it is incapable of measuring the opposing forces required to measure power.

It's important to note that the accelerometer in the Newton measures total acceleration--the sum of gravitational and inertial acceleration. It does this extremely accurately, which is why power measurements are so good. Hill slope, which is a directly related to gravitational acceleration, is separated out from total acceleration by using acceleration measurement from the speed sensor. Zoltan is right that, the slower the bike speed, the longer it takes for the wheel to turn, and the less often the bike acceleration is updated. This is the reason that reported hill slope varies under conditions of low bike speed or high bike acceleration. But the variation lasts briefly, and overall iBike hill slope is very accurate--much more so than baro-derived hill slope or GPS-derived hill slope.
John Hamann
Zoltan
Posts: 213
Joined: Mon Jan 23, 2012 9:47 am
Location: HUNGARY

Re: Slope question

Post by Zoltan »

Anybody who is chasing max slope accuracy more wildly than how it is provided by default can add as many magnets as he can get provided they are evenly spaced on the wheel. I still have 5 and the set circumference is 419mm. As a sidenote iAero does not let you set 419mm as a wheel circumference, presumably because it helps the users to avoid mistyping, but I could do it via the predecessor of Isaac. And I guess it is still possible via 2.1.5.

Finally the whole thing is just for fun, moreover one should be quite disciplined to try to sit on the saddle in the very same way on very steep ascents as on flats placing the same weight on the first wheel. Nevertheless I measured the slope of some steep streets with iAero, always in a gear of 39/30 to maximize the smoothness of pedalling. Or just pushing the bike walking next to it.

This sort of voluntary "job" is required as a minimum if one has 4 gadgets on his bikes while two other on his wrists. :-) And I always named iAero in local Hungarian forums as the tool of measurement, although nobody understood what I was speaking about, because they got used to barometric slope measurement.
User avatar
Russ
Posts: 370
Joined: Sun Feb 10, 2008 4:08 pm

Re: Slope question

Post by Russ »

Kens said:
There's a device out there with a name disturbingly similar to Velocomp that uses two magnets near the hub and a magnetometer to measure the field between them to a claimed 1 degree or 0.01 metres precision. But it is Bluetooth rather than ANT+

and John said that it only measures speed/distance.

It is the Velocomputer smart sensor and they have three models planned. The entry level is planned to ship in September and only measures the speed/distance but with 1 degree precision. However they do have an accelerometer which the planned mid level will spit out x y z acceleration values in to the spreadsheet file. The pro level which is there current first offering also calculates power but only is accurate in a Velodrome or no wind condition.

Hint Hint to Velocomp. They have a developers SDK and are looking for commercial partners and so might supply it as a future iBike speed sensor - zowweee! (As Kens has suggested somewhat!)

A super cool thing about there smart phone app and sensor together (look for youtube video on it) is you can stop and move in reverse or whatever and it keeps a running total in the odometer accounting for every 1 degree movement of the wheel in either direction!

Besides great acceleration data derived from changes in speed, with proper calibration this may be able to replace the mechanical wheel rotation counter required for measuring official race courses if they get there marketing act together!

Regards - dreams of greater iBike capabilities,
Russ (Yes I am one of the measurement freaks : -)
Velocomp
Velocomp CEO
Posts: 7793
Joined: Fri Jan 11, 2008 8:43 am

Re: Slope question

Post by Velocomp »

Russ wrote:Kens said:
There's a device out there with a name disturbingly similar to Velocomp that uses two magnets near the hub and a magnetometer to measure the field between them to a claimed 1 degree or 0.01 metres precision. But it is Bluetooth rather than ANT+

and John said that it only measures speed/distance.

It is the Velocomputer smart sensor and they have three models planned. The entry level is planned to ship in September and only measures the speed/distance but with 1 degree precision. However they do have an accelerometer which the planned mid level will spit out x y z acceleration values in to the spreadsheet file. The pro level which is their current first offering also calculates power but only is accurate in a Velodrome or no wind condition.

Besides great acceleration data derived from changes in speed, with proper calibration this may be able to replace the mechanical wheel rotation counter required for measuring official race courses if they get their marketing act together!

Regards - dreams of greater iBike capabilities,
Russ (Yes I am one of the measurement freaks : -)
For those of you who who want to know how steep the grade is you're riding during the middle of severe accelerations, then this device might help, ASSUMING everything else in its design is working perfectly. What is the "everything else"? What has to work perfectly? Much, much more than meets the eye :D

Folks, there's a product that's been on the market for 8 years that measures total x/y/z acceleration 800 times a second, with accuracy greater than 99.5%. Anyone care to guess its name?
John Hamann
User avatar
racerfern
Posts: 1356
Joined: Fri Jan 11, 2008 12:47 am
Location: Baldwin, NY
Contact:

Re: Slope question

Post by racerfern »

John,

Ooh, ooh, ooh, I know the answer, teacher!

There's always an amourous obsession with the new and untried. People still clamor for the Vector even though Garmin has owned it long enough to totally ruin it. Every new power meter out there is the latest and greatest and they all show promise. That's it, promise is all they show. The Velocomputer appears to be interesting except for the fact that it's only usable as a power meter under very controlled conditions either on a track or on a trainer and a lot of the info about it goes back to last year.

Has anyone paid attention to the price of power meters lately? It's nearly doubling the price of a reasonably priced full carbon bike and you don't even get a display head! They're all through the roof except for the Newton at under $500!

Velocomp has followed through on each and every one of it's projections and enhancements within a reasonable amount of time; not something many other companies can brag about. I happen to be a privileged beta tester of the latest development for GPS recording and have to say when it's released there will be people willing to sell their Garmin 500s. Thanks Velocomp for not accepting the status quo and for pushing the development of your products beyond what anyone ever thought possible.

Heck, you even got a compliment from Zoltan (one of the smarter people I've never met) so that's worth a feather in your cap. ;)
Fernando
Zoltan
Posts: 213
Joined: Mon Jan 23, 2012 9:47 am
Location: HUNGARY

Re: Slope question

Post by Zoltan »

racerfern wrote:Heck, you even got a compliment from Zoltan (one of the smarter people I've never met) so that's worth a feather in your cap. ;)
Indeed. :-) But it would be more political not to create any jealousness about the abilities of people around in order to keep everybody's enthusiasm to post :-) But if it happened that I also got a compliment I'd like to be sure that I really got it: does "I''ve never met" mean the same as "I've ever met" ? :?:
Velocomp
Velocomp CEO
Posts: 7793
Joined: Fri Jan 11, 2008 8:43 am

Re: Slope question

Post by Velocomp »

racerfern wrote:John,

The Velocomputer appears to be interesting except for the fact that it's only usable as a power meter under very controlled conditions either on a track or on a trainer and a lot of the info about it goes back to last year.
It works on a trainer in the same way the Newton does--through rear wheel bike speed measurements. Wonder where they will get the power curves? :D

In Velodromes, where wind speed is zero (except when there's drafting), the opposing wind speed is approximately equal to bike speed. Approximately? There's turbulence issues to deal with... And, the rider's CdA will still need to be determined. Wonder where they will get that information? :D

This product concept has been around forever, any in my view its usefulness is the same.
John Hamann
User avatar
Russ
Posts: 370
Joined: Sun Feb 10, 2008 4:08 pm

Re: Slope question

Post by Russ »

Regarding the Velocomputer smart sensor, the entry level does not use their accelerometer and basically is just a very accurate bike computer. And I am not pushing their product per say. What I am suggesting is that they may either license their 1 degree precision tech or supply the sensor as a replacement for the ant+ sensor in a future iBike, at least in my thinking on this. In this model all of Zoltan's observations of slow speed uphill data would be transformed into the same or better great performance as is at higher speeds. That problem would be solved. Of course it would take a new model iBike with bluetooth to do it, I suppose bluetooth and Ant+.
Of course I understand that the likelihood of that all coming together business wise and business plan wise has significant hurdles : -) Just a few thoughts.

I may buy one of their entry devices just so I can perform precision distance measurements. I wish I could do that easily with the iBike. I have laid out a carefully measured 1.7k course using a laser measurement tool and have found there are technical problems dialing in the iBike to the closest millimeter of circumstance.

Russ
KenS
Posts: 141
Joined: Sun Feb 28, 2010 2:55 pm
Location: Melbourne, Australia

Re: Slope question

Post by KenS »

Irrespective of what specific device one uses for measuring wheel speed/acceleration; it occurs to me that more frequent and accurate readings could open up more possibilities for the iBike. The current speed sensor only measures about 2-10 times per second versus the 800 times for the accelerometer. With two orders of magnitude more wheel speed readings it might become possible for the 5 minute auto-tilt correction to become a 3 second correction by reducing reliance on barometric altimeter readings. The cal ride could become 3 short rides: on the hoods, in the drops and out of the saddle; then during a ride the iBike might be able to detect which position you are in from the tilt calculation and adjust CdA on the fly. It might also make the iBike more accurate for MTBers.
So it isn't just about having a more accurate slope or distance reading - there could be real improvements to power readings and useability.l
-- Ken
Zoltan
Posts: 213
Joined: Mon Jan 23, 2012 9:47 am
Location: HUNGARY

Re: Slope question

Post by Zoltan »

KenS wrote: The cal ride could become 3 short rides: on the hoods, in the drops and out of the saddle; then during a ride the iBike might be able to detect which position you are in from the tilt calculation and adjust CdA on the fly. It might also make the iBike more accurate for MTBers.
It seems to be a good point, but I feel that one should raise two things:
1. I am not so sure that result of tilt calculation are so different from each other, that you can detect on the fly which mode happens? Hoods, drops or out of the saddle. It is easier to distincet out of the saddle would from the other too. Maybe. Or do you mean that adding a new sensor would make it possible due to the extra accuracy?
2. The surface impacts on auto tilt, doesnt it? How could you consider this factor when selecting from the three positions.
KenS
Posts: 141
Joined: Sun Feb 28, 2010 2:55 pm
Location: Melbourne, Australia

Re: Slope question

Post by KenS »

Zoltan wrote:It seems to be a good point, but I feel that one should raise two things:
1. I am not so sure that result of tilt calculation are so different from each other, that you can detect on the fly which mode happens? Hoods, drops or out of the saddle. It is easier to distincet out of the saddle would from the other too. Maybe. Or do you mean that adding a new sensor would make it possible due to the extra accuracy?
2. The surface impacts on auto tilt, doesnt it? How could you consider this factor when selecting from the three positions.
The iBike accelerometer measures the X and Y forces but there are an infinite number of combinations of Slope + Tilt + Wheel Acceleration that can produce those. An accurate wheel acceleration measure can resolve that to an accurate Slope and Tilt.
I'm pretty sure there is enough difference between hoods and drops to be detectable. When you do coast-downs even the tilt difference between pedalling and coasting is detectable.
Surface will affect the tilt but then it comes down to timescales. With many more wheel readings you could tell the difference between a speed bump or pothole versus a consistent change in tilt due to positioning. The current wheel sensor can't distinguish anything smaller than 2 metres since that is how far you travel between readings.
-- Ken
Velocomp
Velocomp CEO
Posts: 7793
Joined: Fri Jan 11, 2008 8:43 am

Re: Slope question

Post by Velocomp »

I've been reading this thread with interest and admiration.

There are a lot of good ideas floating around here, reflecting a solid understanding of things.

But keep pushing and you might end up in a different place.

The devil's in the details; ask Garmin/Vector... :D
John Hamann
texmurphy
Posts: 122
Joined: Fri Jan 29, 2010 4:16 pm

Re: Slope question

Post by texmurphy »

Velocomp wrote:...The devil's in the details; ask Garmin/Vector... :D
Garmin Investor Relations 31 July 2013 wrote:From http://www8.garmin.com/aboutGarmin/invR ... script.pdf
The most anticipated new product deliveries in the third quarter is the Vector power meter which is expected to accelerate revenue growth in the back half of the year. Vector has entered mass production and we expect to deliver the product to retailers soon. We are thrilled to finally cross the finish line with this new product and anticipate our customers will appreciate the unique features and value proposition that Vector brings to the cycling market.
Maybe in time for Xmas?
Velocomp
Velocomp CEO
Posts: 7793
Joined: Fri Jan 11, 2008 8:43 am

Re: Slope question

Post by Velocomp »

OK, after you buy new shoes/cleats, and buy a bike computer to read its signals, it will be $2K.

That's 4x the Newton price.

Wow, what a deal!!!
John Hamann
User avatar
lorduintah
Posts: 642
Joined: Fri Jan 11, 2008 9:37 am
Location: Plymouth, MN

Re: Slope question

Post by lorduintah »

All in fun ---

I think Zoltan might be suggesting a seat sensor - the butt compression equalizer?? for a new sensor. Then the iBike can analyze cheeky stress vectors which affect your comfort level in riding. :roll:

Tom
Zoltan
Posts: 213
Joined: Mon Jan 23, 2012 9:47 am
Location: HUNGARY

Re: Slope question

Post by Zoltan »

I guess y'all started to predict my way of thinking quite well.

Actually I was thinking about to propose multi temperature sensors placed at different places on the bike and to use the minimum value of those measured by them to get an even better temperature reading. But it would not be revolutionary, so I risked my credibility and proposed the following:
"My other idea is to add a sensor which measures not the temperature, but the luminosity to make corrections if it is more than some predefined luxes" :ugeek:
It seems to be revolutionary. OK, maybe not sensible ;)
Post Reply