Fluctuating Power

Originally Posted by aquatrain

Hello everybody. I have a question that I don't think I've ever seen asked or answered here.

First some background. I am Cat 4 and have been racing for 5 years. I am pretty much pack filler. I don't have a power meter but have occassionally been using a CompuTrainer for the last few months. I am currently not near race shape but am not in terrible shape. I climb like a rock and my power levels drop dramatically when I go uphill. I train about 5 to 7 hours per week focusing on sharter harder efforts as opposed to 4 hour base training rides. My training in December was mostly SST. In january I did mostly SST and 2 minute intervals. This month I am doing threshold intervals. I am pretty much a follower of Morris and Ross.

Here is the question. It appears as if my ftp is probably around 240 as I can ride at that level for about a half an hour on the flats. However, my power output is rarely 240. It fluctuates between 260 and 220 so that the average is about 240. The fluctuations are pretty quick. Five to ten seconds at 260, five to ten seconds at 220. It is as if my power output pulsates.

Does anybody have any insight into this? Does it show an inherent weakness? If so, how do I correct it?

Thanks

Originally Posted by aquatrain

... my power output is rarely 240. It fluctuates between 260 and 220 so that the average is about 240. The fluctuations are pretty quick. Five to ten seconds at 260, five to ten seconds at 220. It is as if my power output pulsates.

Does anybody have any insight into this? Does it show an inherent weakness? If so, how do I correct it? ...

Originally Posted by daveryanwyoming

Power does fluctuate quite a bit while riding. Less when riding indoors at relatively constant intensity, but it still fluctuates. As Robkit suggests most power meters allow you to set a display smoothing which does a running average on displayed data to reduce the fluctuations allowing for better pacing. I only use my CT to set the load and mostly use it in erg mode(F2 from the handlebar controller) where it just shows a single number but my recorded PT power still varies quite a bit.

I've attached a couple of screenshots from a recent SST effort on the Computrainer, the data is from my PT hub. The first one shows the data with no smoothing applied for with the CT in erg mode with fairly smooth cadence(green line). The second shows the same effort with 5 second smoothing applied. Notice how much the smoothing reduces the peaks and valleys for the power curve(yellow line)? Also notice how HR is relatively smooth in both graphs(red line) since HR is basically smoothed physiologically in the body, IOW it doesn't respond to instant changes the way power data does.

Anyway, jumpy power curves are normal and I set my PT display for 5 second averaging so that I don't follow every peak and valley in the power curve while out on the road. It doesn't change the recorded data so you don't lose things like peak power in a sprint, but it sure makes it easier to see what I'm really doing while training or racing.

-Dave

Originally Posted by Alex Simmons

I'd suggest the jumpiness in your power line has more to do with riding with a cadence in the low 80s which is bang in the middle of the cadence range to accentuate the PT's aliasing effect - hence your real power output is not that jumpy - it's just an artifact of the way the PT calculates and records power. I bet it would "appear" to smooth out a lot if you rode at 95rpm (even though it was most likely already smooth to start with)....

Originally Posted by daveryanwyoming

Fair 'nuff, I was avoiding all the sampling theory and nyquist rate discussions but even at higher cadences power jumps a lot even in ergo mode relative to say HR. Here's a screenshot of a steady 100 rpm effort, still pretty jumpy...

-Dave

Originally Posted by Alex Simmons

Fair 'nuff

PT files are generally pretty choppy no matter what we do!

Suppose we're just used to it.