Pick-up truck on 3/5/2010

[Mark Debbage]

Did anyone catch the license plate of the pick-up truck on

Portola Road yesterday? I have the dangerous overtaking
and near-miss crash on video. It is a silver Ford but I
can't make out the plate unfortunately.

Mark.

[Mark Debbage]

Here's the video with the truck:

http://www.youtube.com/watch?v=rmpjs34Kv04

And here's some of the ride back on Canada Road:

http://www.youtube.com/watch?v=tmvQD1Xg6jk

Mark.

[Dan Connelly]

First, I'm glad I bailed @ Portola Alpine: yikes!

But that video overlay is very, very cool. Is the gear
determineation based on the ratio of speed to cadence? That's my
guess.

On Mar 7, 9:14 am, Mark Debbage <mark.debb...@gmail.com> wrote:
> Here's the video with the truck:
>

> http://www.youtube.com/watch<http://www.youtube.com/watch?v=rmpjs34Kv04>
> ?v=rmpjs34Kv04 <http://www.youtube.com/watch?v=rmpjs34Kv04>

[Mark Debbage]

On Mon, Mar 8, 2010 at 7:11 AM, Dan Connelly <djco...@gmail.com> wrote:

First, I'm glad I bailed @ Portola Alpine: yikes!

But that video overlay is very, very cool.   Is the gear
determineation based on the ratio of speed to cadence?  That's my
guess.

Yeah, that's right. There's some heuristics to clean up the
data and to give more plausible output. There is also a
braking indicator but that doesn't show up in this particular
video.


Mark.

[Matt Allie]

r u in love with Ken???

there's 5min of Ken's butt on the 2d video!!! :))))

not sure the wattage is accurate: it shows 450W on Canada road at Edgewood... u either weigh 290lbs or ur power tap is slightly off ;-)

cool video tho!

-Matt

--
You received this message because you are subscribed to the Google Groups "The Noon Ride" group.
To post to this group, send email to noon...@googlegroups.com.
To unsubscribe from this group, send email to noonride+u...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/noonride?hl=en.

[Mark Debbage]

On Mon, Mar 8, 2010 at 10:08 AM, Matt Allie <matthi...@gmail.com> wrote:

r u in love with Ken???

there's 5min of Ken's butt on the 2d video!!! :))))

not sure the wattage is accurate: it shows 450W on Canada road at Edgewood... u either weigh 290lbs or ur power tap is slightly off ;-)

cool video tho!

I have to balance the butt content. If my wife sees too much
girl butt, then I'm in big trouble.

I believe the wattage is accurate enough, though it might be slightly
out of sync with the other data and video. The PowerTap itself
adds a little lag, plus I don't have the right stuff to record
everything synchronously. I think the high readings are because
of the slight grade and the aero drag from being out in the
wind at speed. I'm very tall but definitely not 290lbs! That
power is a spike that I'm unable to sustain as the entire
group goes by me. Sometimes I add a little smoothing
on the power display, but this time there was no averaging
at all.

Mark.

[Dan Connelly]

Powertap is prone to spiking since it isn't synchronized with the
pedal stroke, and the power tends to vary a lot through a pedal
stroke.

See, for example, here (standing, worst case)
http://www.metrigear.com/2010/02/26/the-cyclist-as-a-two-cylinder-two-stroke-engine/

Nice work on the gear ratio! Nontrivial to handle coasting well.
You must have a cadence magnet. PT internal cadence isn't good
enough. Even then, cadence and speed need to be well synchronized for
it to work.

On Mar 8, 10:20 am, Mark Debbage <mark.debb...@gmail.com> wrote:

[Mark Debbage]

On Mon, Mar 8, 2010 at 11:26 AM, Dan Connelly <djco...@gmail.com> wrote:

Powertap is prone to spiking since it isn't synchronized with the
pedal stroke, and the power tends to vary a lot through a pedal
stroke.

See, for example, here (standing, worst case)
http://www.metrigear.com/2010/02/26/the-cyclist-as-a-two-cylinder-two-stroke-engine/

An excellent point, thanks. I presume that the PowerTap
effectively averages the instantaneous power over its
1.26s sampling period. With a 90 rpm cadence there
will be a power peak every 0.33 seconds and the number
of peaks that falls into the sample could vary from period
to period, leading to measured variation even under
a steady power output. I guess one would call this a
sampling aliasing effect. I'll go back to 3 second averaging
for my power display to hide this. One advantage I have
over a head unit is that I can average over samples from
the future too!

If you align the sample periods to the pedal stroke
you would be generating irregular sample periods.
All the downstream protocols and software expect
constant period size, and typically coerce to 1 second
periods. I suppose with fine-grained sampling, one
could do some more intelligent averaging on the
device to smooth out appropriately to the desired
sample rate.
 

Nice work on the gear ratio!    Nontrivial to handle coasting well.
You must have a cadence magnet.  PT internal cadence isn't good
enough.  Even then, cadence and speed need to be well synchronized for
it to work.

Thanks! Indeed, I started off with the PowerTap cadence
but quickly switched over to the Garmin cadence sensor
to keep the data synchronized. The heuristic gearing,
coasting and braking indicators worked out way better
than I was expecting them to.

Cheers,

Mark.

[Dan Connelly]

> An excellent point, thanks. I presume that the PowerTap
> effectively averages the instantaneous power over its
> 1.26s sampling period. With a 90 rpm cadence there
> will be a power peak every 0.33 seconds and the number
> of peaks that falls into the sample could vary from period
> to period, leading to measured variation even under
> a steady power output. I guess one would call this a
> sampling aliasing effect. I'll go back to 3 second averaging
> for my power display to hide this. One advantage I have
> over a head unit is that I can average over samples from
> the future too!

Powertap ANT+ is 1 second, pre-ANT+ is 1.26 second.

You could, if you really wanted, change averaging time with cadence.
Maybe too much work. Also there's better averaging functions than
uniform sampling windows. For example, use weights of (1/12, 1/4,
1/3, 1/4, 1/12) for a centered average.