PowerTap longer sampling times = smoother graphs?

[patentcad]

If you set the PowerTap sampling option to say, 10 seconds vs. 2 or 3 secs. then I presume the resulting ride graph would be considerably less jagged, no? Is that why a longer sampling time is better for TT riders per the manual? They say it would be better, but they don't tell you why.

These are the brainstorms I get while riding when the iPod battery dies.

[AlexTaylor]

OMG! Imagine what would spew out if the PowerTap battery died! :o

How are you getting on with it?

[bdcheung]

If you're using WKO+, you can smooth your graphs yourself. pcad, there are two different data rates:

1) Recording rate. You always want this to be 1 second. There's no reason (unless you're going to be away from your computer for a super long time and are afraid of using up all the memory) to do otherwise.
2) Display rate. This is independent from the recording rate and only affects the number displayed on your PT computer. You can set this higher if you want to average it over a longer period of time - useful if you're doing constant effort.

Modifying (2) will have no effect on the graph in PowerAgent or whatever software you use to analyze your rides.

['nother]

If you set the PowerTap sampling option to say, 10 seconds vs. 2 or 3 secs. then I presume the resulting ride graph would be considerably less jagged, no? Is that why a longer sampling time is better for TT riders per the manual? They say it would be better, but they don't tell you why.

My guess is what you're talking about is the display averaging, not sampling rate (i.e. Computer Setup 2). This only affects what you see on the PT display, not the actual sampling of the unit (which I think is fixed, and is different from the *recording* rate, which you can set to either 1 or 2 seconds).

Changing the display averaging to a longer average would be helpful in any endeavor where you want to see a constant effort and are not interested in short-term spikes. So a TT or even a long hillclimb would be good places to use longer display averaging because it would tend to smooth out short bursts of power (again, on the display only -- it DOES NOT affect the measurement or ride graph). And actually they do say why in the manual, "A greater rate of display allows for a slower update of the display" -- basically, eliminates the instantaneous spikes that you might see, i.e. from a few hard pedal strokes or some such. It's just not written in patentcad-ese. ;)

[patentcad]

I have no idea what any of you are talking about. I sucked at science in HS and majored in art @ NYU. Can't you tell?

I think my PowerTap hub may need a new freehub bearing. Oh well. We'll see. I promise not to ever ask another PT question ever again. I'm limited to the Saris software because there is NO CHANCE I'll deal with Windows. Maybe I can buy that other Mac software that guy was answering questions about here, but the Saris software does work. I think it does. There's no way to really be sure of course. This is kind of like trying to operate a Space Shuttle using an owner's manual written in Turkish.

[ElJamoquio]

The manual isn't bad. The software *is* bad.

If you want a 'real' graph that you can actually use, you need a different software program than Saris. Personally I use Excel.

[patentcad]

The manual isn't bad. The software *is* bad.

If you want a 'real' graph that you can actually use, you need a different software program than Saris. Personally I use Excel.

I'd be curious how you get the PowerTap to download a graph to Excel. Can you get it to write poems to you in Word as well?

[Enthalpic]

If you set the PowerTap [display] option to say, 10 seconds vs. 2 or 3 secs. then I presume the resulting ride graph would be considerably less jagged, no? Is that why a longer sampling time is better for TT riders per the manual?

Yes.
Clear enough?

However, your post ride graphs will be uneffected; but you can smooth that raw data anyway you want... or even mulitple ways. :)


Do you at least have a spreadsheet program?

[Enthalpic]

I'd be curious how you get the PowerTap to download a graph to Excel. Can you get it to write poems to you in Word as well?

export and import *.csv

csv = comma seperated values... a data handling convention many programs use.

[patentcad]

Yes.
Clear enough?

However, your post ride graphs will be uneffected; but you can smooth that raw data anyway you want... or even mulitple ways. :)


Do you at least have a spreadsheet program?

I can spread sheets on my King Sized bed, but only if my wife is holding the other side. On the other hand, I can draw anything in Adobe Illustrator and I'm getting better @ Photoshop all the time.

Spreadsheets. Good God man, this is supposed to be cycling for fun. If I wanted to F around with spreadsheets I'd be a friggin accountant.

[bdcheung]

pcad, how about this. For $120/hr, I'll fly up to NY for a weekend and be your personal cycling computer consultant. If that's too many words for you, think of me as the Johan Bruyneel to your Lance Armstrong.

[patentcad]

pcad, how about this. For $120/hr, I'll fly up to NY for a weekend and be your personal cycling computer consultant. If that's too many words for you, think of me as the Johan Bruyneel to your Lance Armstrong.

If I wasn't fearful of your inscrutability, I might take you up on that offer BD.

[skinnyone]

This might be taking things to an un-ncecessary extreme here.

Nyquist sampling theorem dictates that your sampling freq needs to be 2x more than your max signal frequency. Often in real world applications, you sample faster than 2x for various noise and reconstruction reasons.

Lets assume a cadence of 120rpm which leads to a time period of 0.25s/pedal arm power stroke (assuming that the bulk of your power is delivered in two points of the pedal stroke which is by and large true.) So I would argue that you should have a sampling time of 0.125s in order to faithfully reproduce the power waveform. BUT, the reason this may not be applicable here is that your power output over a few pedal strokes probably does'nt vary by much and the actual variation is likely small enough to be ignored. As in you dont care if you a spent a few milli seconds at a different power level when we are talking about power graphs and trends in minutes.

[Jynx]

1) Recording rate. You always want this to be 1 second. There's no reason (unless you're going to be away from your computer for a super long time and are afraid of using up all the memory) to do otherwise.
2) Display rate. This is independent from the recording rate and only affects the number displayed on your PT computer. You can set this higher if you want to average it over a longer period of time - useful if you're doing constant effort.


What are the options for recording rate and display rate? 1 sec, 2 sec, 5 sec, 10sec?

Also at a 1 second recording rate how long is it capable of recording before it fills up all the memory?

[asgelle]

So I would argue that you should have a sampling time of 0.125s in order to faithfully reproduce the power waveform.

Good thing the power tap samples at 60 Hz.

[ElJamoquio]

I'd be curious how you get the PowerTap to download a graph to Excel. Can you get it to write poems to you in Word as well?

Love poems. I'd share them but they get rather personal. Crank length, chains, lube, that sort of thing.

[kukusz]

You are looking for page 17 of your manual. The "storage rate".

http://i214.photobucket.com/albums/cc24/kukusz/Picture1.png

Set that to 1 second like people said. The display rate does not effect the data on the computer.

I found this somewhere too.

http://i214.photobucket.com/albums/cc24/kukusz/Picture2-1.png

[skinnyone]

Good thing the power tap samples at 60 Hz.

Interesting choice.. Given it being power line freq..

[Kachunk]

hah. Illustrator. you're a fraud. You should call yourself patentart. If Illustrator is CAD I'm a llama.

[ElJamoquio]

pcad, how about this. For $120/hr, I'll fly up to NY for a weekend and be your personal cycling computer consultant. If that's too many words for you, think of me as the Johan Bruyneel to your Lance Armstrong.

Hmm, a reverse auction. I'll do it for $75/hr plus expenses. I'll set up a training plan for you. I'll setup your PT. I'll do as many graphs as you want; Excel, Matlab, whatever. I'll even throw in some (real) CAD if you so desire. I'll build you some wheels or maintain your bike. Hell, I'll ride behind you screaming at you in a vain attempt to get you to hear me over your iPod.

Allez, Pcaddy. Allez.

My time is your money. The only limit is the bottom of your credit cards.

[patentcad]

hah. Illustrator. you're a fraud. You should call yourself patentart. If Illustrator is CAD I'm a llama.

Correct. You can call me the Folly Llama. Pcad's Folly has been paying the mortgage now for nearly 17 years. You better not tell my clients (the ones I've had for the better part of two decades). They'll wonder how the hell I did all those drawings for them. Besides there's some other mope who uses the incredibly cheesy name 'Patentart'.

[patentcad]

Hmm, a reverse auction. I'll do it for $75/hr plus expenses. I'll set up a training plan for you. I'll setup your PT. I'll do as many graphs as you want; Excel, Matlab, whatever. I'll even throw in some (real) CAD if you so desire. I'll build you some wheels or maintain your bike. Hell, I'll ride behind you screaming at you in a vain attempt to get you to hear me over your iPod.

Allez, Pcaddy. Allez.

My time is your money. The only limit is the bottom of your credit cards.

Pcad does not use credit cards. They collect dust in his wallet, their vast credit lines untapped.

Besides it will be much more fun annoying you racer weenies with my annoying threads. Just think, it might take me years to wrap my head around training with a PM, and I'll drag you losers into the abyss with me, your heads bouncing off every friggin step.

http://farm1.static.flickr.com/123/348659570_334fd66e77_b.jpg

Download that to your stupid laptops and graph it.

[ElJamoquio]

Download that to your stupid laptops and graph it.


http://www.nada.kth.se/~gerd/Images/spiral1mesh.jpeg


Damn it, I shouldn't have added that ellipse-term in the z axis.

[patentcad]

http://www.nada.kth.se/~gerd/Images/spiral1mesh.jpeg


Damn it, I shouldn't have added that ellipse-term in the z axis.

See? I knew somebody would take a stab at this. Well done ElJ.

This isn't the Road Racing Forum. This is the NASAvelo Forum.

[bdcheung]

/lost interest.

ooooh hey look, a new function on my Ergomo!

[patentcad]

/lost interest.

ooooh hey look, a new function on my Ergomo!

How come I keep reading Ergomos break a lot?

[kukusz]

Pcad, didn't you recently tell the racing weenies to stick our Powertaps up our tuckus and just ride our bikes? I need to find this post somewhere...

[skinnyone]

http://www.nada.kth.se/~gerd/Images/spiral1mesh.jpeg


Damn it, I shouldn't have added that ellipse-term in the z axis.

The tubes man..

http://tubesdance.ytmnd.com/

[patentcad]

Pcad, didn't you recently tell the racing weenies to stick our Powertaps up our tuckus and just ride our bikes? I need to find this post somewhere...

I'll stick my own PM up my own bunghole and ride my bike. Frankly this Power Meter siht is less up the bunghole than down the wormhole.

[bdcheung]

How come I keep reading Ergomos break a lot?

Because my cottage of wattage keeps bringing the roof down.