Exporting smoothed power data.

Post Reply
blp
Posts: 164
Joined: Mon Mar 26, 2012 12:06 am
Location: Vancouver, BC, Canada
Contact:

Exporting smoothed power data.

Post by blp »

I can't seem to find this topic on Newton 6 and anomalous power spikes.

I find that with one bike, my preferred Look 695 SR, that every ride I find very high power spikes. On some occasions over 1200W when 600-700W is expected. When the data is smoothed to 5", the power trace
appears to correlate to the data described in the detail box on the left. I suspect these fields in the box use 5" smoothed data for display.

When I export the ride file, these spikes are exported as well, as one would expect. Of course these show up in Golden Cheetah and WKO4 as true data points. Since GC and WKO4 don't smooth data for analytics, which Isaac seems to do, this skews the data resulting in incorrect power profiles.
Is there a way to identify and remove these spikes in the raw data before or at export? Is there a simple way to remove these anomalies in Isaac? Is there a way to determine why this happens with one bike and more than the other? Is it the out-front mount? I thought there was a low pass filter added in one of the FW updates, perhaps this was for road surfaces only and these spikes are something outside that algorithm.

I realize I can edit the file, File>Combine/Split>Delete Selection, but that can be a) tedious and b) occasionally I trim more than I need.
In GC, I can edit the ride, but again tedious, but much less so than in Isaac.

This does not happen to the same extent, still spikes, but less dramatic, when my N6 is used with my winter/gravel bike.

The N6 is mounted to a out-front mount on the Look, and the std handlebar mount of the Brodie.

I have been thinking of getting a PowerPod, and I understand that the PP is a Newton 6 in a different body (I read that in some post on this site) which I would
assume means that all the enhancements, improvements and updates made to the N6 would be embodied in the PP. But that would also imply that any issues, such as these spikes,
would also be present in the PP.

I have supplied a ibr file to illustrate the spikes, and a screenshot of how these spikes are nullifying the overall performance charts.

Thanks.

Thanks
Attachments
WKO4.jpeg
WKO4.jpeg (92.69 KiB) Viewed 6889 times
iBike_07_10_2016_1356_62_km_HiDef.ibr
(1.79 MiB) Downloaded 319 times
Velocomp
Velocomp CEO
Posts: 7793
Joined: Fri Jan 11, 2008 8:43 am

Re: Exporting smoothed power data.

Post by Velocomp »

I looked at your ride file.

The cause of your power spikes are speed spikes from your speed/cadence sensor. You can see these spikes by turning filtering to zero.

The easiest fix to this problem is to upgrade your sensors to the magnet less types. I recommend the Garmin sensors, with speed sensor mounted to your front wheel hub.

If you're exporting the file, the only expedient fix would be to remove the spikes from the .csv file.
John Hamann
blp
Posts: 164
Joined: Mon Mar 26, 2012 12:06 am
Location: Vancouver, BC, Canada
Contact:

Re: Exporting smoothed power data.

Post by blp »

Thanks for your insight. Now that you have pointed it out, I see the cadence spikes.

Looks like this, combined with Tom's assessment of a problem in my previous, indicates my GSC-10 may be on it's last legs.
Also explains why different occurrences in power spikes on different bikes.

I think most magnet-less type of sensor don't have a static reference point, ie above or below chainstay
which Isaac uses for PowerStroke. Am I correct in my understanding that Isaac only uses sensor for PS?
What do you suggest as a cadence sensor?

Thanks John
Velocomp
Velocomp CEO
Posts: 7793
Joined: Fri Jan 11, 2008 8:43 am

Re: Exporting smoothed power data.

Post by Velocomp »

You can use a magnet less cadence sensor, and set it for above the chain stay. This works just fine.
John Hamann
blp
Posts: 164
Joined: Mon Mar 26, 2012 12:06 am
Location: Vancouver, BC, Canada
Contact:

Re: Exporting smoothed power data.

Post by blp »

Thanks John.

PowerStroke will work as expected I assume.
Velocomp
Velocomp CEO
Posts: 7793
Joined: Fri Jan 11, 2008 8:43 am

Re: Exporting smoothed power data.

Post by Velocomp »

blp wrote:Thanks John.

PowerStroke will work as expected I assume.
Yes
John Hamann
blp
Posts: 164
Joined: Mon Mar 26, 2012 12:06 am
Location: Vancouver, BC, Canada
Contact:

Re: Exporting smoothed power data.

Post by blp »

John, I have taken a closer look at one of these spikes from the ibr file I attached previously.

I certainly see the power spike in tandem with cadence anomalies, but I also see other artifacts, such as spikes/depressions in the wind trace, slope, etc.
I understand the relationship between speed/wind with the Newton. But the deflections don't appear to be organic and out of character with the adjacent parts
of the trace. Same with slop trace.

Gear ratio trace is null for about 20 seconds as well.

I have also uploaded a second closeup where cadence trace indicates zero, but the power trace is over 400W and the info bar at the bottom of the graph
tells a slightly different story. (For a shot duration the cadence trace indicates I stopped pedaling for a brief moment, seldom do I stop pedaling on a 6-7% slope, and certainly not for 6 seconds.
But if the Garmin GSC-10 sensor is failing, I could see that being falsely reported.)

I have provided a closeup, 2 minute view where I have placed the curser over the 438W peak and 0 cadence.

Is the slope, speed/wind and gear ratio anomalies consistent with speed sensor issue?

Forgive me if I am misunderstanding what I am seeing or I'm overthinking this. It's in my nature to understand things when I should simply accept. :roll:

Thanks
Attachments
Glitch_at_zero_cadence.jpeg
Glitch_at_zero_cadence.jpeg (185.15 KiB) Viewed 6877 times
Glitch.jpeg
Glitch.jpeg (215.08 KiB) Viewed 6877 times
Velocomp
Velocomp CEO
Posts: 7793
Joined: Fri Jan 11, 2008 8:43 am

Re: Exporting smoothed power data.

Post by Velocomp »

Speed spikes, not cadence spikes, are the underlying problem. They cause power, slope, and other errors.
John Hamann
blp
Posts: 164
Joined: Mon Mar 26, 2012 12:06 am
Location: Vancouver, BC, Canada
Contact:

Re: Exporting smoothed power data.

Post by blp »

Got it, thanks for you patience John.

Got order in for new Garmin speed and cadence pods.
blp
Posts: 164
Joined: Mon Mar 26, 2012 12:06 am
Location: Vancouver, BC, Canada
Contact:

Re: Exporting smoothed power data.

Post by blp »

I just wanted to add that you, Tom Gore, and I'm sure everyone else at Velocomp, are great at dealing
with your customers.
Velocomp
Velocomp CEO
Posts: 7793
Joined: Fri Jan 11, 2008 8:43 am

Re: Exporting smoothed power data.

Post by Velocomp »

blp wrote:I just wanted to add that you, Tom Gore, and I'm sure everyone else at Velocomp, are great at dealing
with your customers.
Thank you for your kind comment!
John Hamann
Post Reply