A watt is a watt is a watt? Trainer vs. Road Power. - Page 2

I understand what Weisse Luft is saying, but I'm incredulous about the magnitude of error. If someone can point me to an appropriate model (analyticcycling?) for typical power/pedal angle (is this a SpinScan?), I can model this issue. I would start with input power sampled at 0.1 or 0.01s, applied against a resistance curve based on modeled road riding or a trainer. Downsampling and then recalculating power would provide an error estimate versus the input power.

I read that the PT "cleverly" measures speed using a reed switch/magnet pair, 1 piece mounted to the stationary axle, and the other to the rotating hub body. If this is the case, it functions exactly like a standard wheel magnet. The only cyclocomputer that I know of that uses a more precise system is the Avocet, but it fell out of popularity due to build quality issues.

Power is calculated from wheel speed and hub torque. Even if torque is measured continuously and integrated, any error in wheel speed will be directly passed on to power. The Power FAQ indicates that displayed values are less accurate than the raw recorded data. It also conjectures that processing time is dear, making an accurate integration implementation less likely.
http://www.midweekclub.com/powerFAQ.htm#Q10

This is an interesting theory. Regarding the effects of inertia on the physiological (and biomechanical) responses of a cyclist, our data indicate that there is some individual difference but it might be experience-dependent. t-tests between the high and low inertia groups showed no difference at all, but there were individual effects.

L.

__________________
MSc (Applied Sport and Exercise Science)
RST Associate Coach
ABCC Level 3 Coach
Doctoral Student (Physiology), University of Oxford, UK.
www.cyclecoach.com
www.science4sport.com

Me too, and me too. In particular, I'm incredulous that it's even noticeable, let alone responsible for the effects that the OP mentioned.

I'm not sure there is a typical case, as it varies by rider, by power, and probably by cadence. For the sake of your modelling, you could just assume a particular profile and have at it.

Agree.

Remember that the wheel speed is being measured between 1-2 times *per downstroke.* I have a hard time believing there's enough acceleration/deceleration between one click of the reed switch and the next to even warrant talking about it. Yes the resistance unit surges and you can hear the modulation, but each surge corresponds to 1-2 revolutions of the wheel, and is being captured by the powermeter.

I'm trying to picture this...

53x15 --> 3.53 wheel revs per crank rev
@ 110 rpms --> 388.3 wheel revs per minute or ~6.5 wheel revs per second... wouldn't 6.5 wheel revs (and presumably samples) per second yield pretty accurate results? regardless of the acceleration/decelerations?

wouldn't it seem more likely that the indoor - outdoor difference might be due to the difficulty your legs have in dealing with this variability in wheel speed acceleration and deceleration indoors... your legs need to over come greater variability in the inertia of the wheels, crank, chain... etc..

as well, if you are using stationary trainer vs. rollers your natural pedalling style (slight rolls left and right) is compromised... i use rollers but many people who use a stationary trainer tell me that when they get back on the road they need to learn how to ride again...

This must fit into the equation somewhere along the line, if it is accurate:

http://www.geocities.com/almost_fast/trainerpower/

I know this chart refers to speed, but something must account for these differences. Is this relevant to this discussion?

I have a Fluid2 trainer and a PT SL, and the differences for me are more signifcant than 10% based on my relatively non-scientific analysis. I have been trying to figure out some kind of multiplier to use to account for the differences between indoor training and the road to get my numbers squared away, but I have resigned to doing all my power-critical rides on the road, a benefit of living in Southern California and having a light that I can enjoy with rare exception.

I'd love to have closure on this issue, but even the physicists on the forum can't seem to agree on things.

So what's the real deal? Is this something that warrants more scientific study with the democratization of training with power, at least with respect to the PowerTap? Or am I making more out of this than there is? As was mentioned, when training for racing, a decrease in performance, perceived or otherwise, can be a debilitating psychological blow.

I can't be alone in coming from training by HR and cadence alone, as that's what I did with my Polar w/o power until getting the PT several months ago. Indoors, and unbeknownst to me, the workouts were actually EASIER as I've concluded my HR is usually elevated 10-20BPM on the trainer. As soon as I got the PT, everything changed. I was not only able to measure power, but speed as well. So first thing I notice is my average speed indoors is rarely above 15mph (see link above). Second, for the same HR outdoors, my indoor power numbers are much lower. Third, regardless of HR, my power indoors is much harder to maintain, to the tune of 20 - 30%. To wit, I've been using a 1.2 or 1.3 multiplier for my power numbers indoors when I record my workouts. Am I cheating myself? My RPE certainly doesn't indicate this.

To add to the madness, adjusting the resistance knob on the trainer DOES make a difference in the power numbers. Don't ask me how to account for it, but all other things being equal, it changes.

I don't think I'm helping with the solution here other than furthering the awareness of these differences and perhaps preserving someone else's sanity along with mine if I can get an "Amen, brother!"

I hope I'm not completely off base here. This has been a source of great frustration for me lately with the time change and cooler temps. I have started to write this post on a few occasions but have hesitated because there always seems to be some smart guy (literally, not sarcastically) minimizing this issue. Graber/CycleOps talks about inertia when I call them. Others have mentioned body temps, and the psychology of training indoors, and add to the stack the rather scientific debate running on this thread. Perhaps these are all factors, but it doesn't seem to account for such a large difference.

I'm going to post this now before I change my mind again.

Cheers,

Ethan

That explains the speed difference you'll see between road and trainer, but not the power difference.

Have you seen this thread? http://www.cyclingforums.com/t279842-.html
You're not alone, and it's a widely recognized phenomenon that *initial* trainer power vs. RPE can differ greatly from road power vs. RPE (see post #2). Andy suggests in that thread that after an adjustment period, however, that cyclists are typically able to produce about the same power on the trainer as they were able to on the road. If you live in SoCal, I'm guessing that a majority of your rides are done on the road? Perhaps you just haven't ridden the trainer enough to get over that adjustment period.

If you look back to that thread from last August, that 1 trainer ride absolutely killed me. Now that it's winter and I'm doing almost all of my riding on the trainer, I feel like I can produce pretty close to as much power on the trainer as I can on the road. The factors you mention above are valid, but you're right that they don't add up to the 20-30% decrease that you're seeing. The rest is the adjustment process. What kind of trainer are you using, BTW?

Ethan,
I train on the CT from Nov-March and am lucky if I can get out once a month with the snow/ice/temps/wind here. I'm pretty familiar with indoor training and have been using power feedback since fall 2002.

I find my sustainable power is consistently lower indoors ... let's say anything from 20 minutes to 3hrs (which is about the longest I ride anyway). But it's nowhere near 20-30%. If I drew a Power-duration curve of indoor versus outdoor power, it would be a hyperbola which would overlay my outdoor power at short durations, start to drop in the 10min region and droop on a modestly increasing basis until hitting 3hrs.

At the 30-minute mark I'm down around 3%, at the hour mark about 6% and I would guess around 10% at 3hrs.

How is your indoor set-up? Cool air and plenty of ventilation? A fan blowing directly on you? These things can make a huge difference.

FWIW, despite my best efforts to minimize power 'droop', my HR indoors drifts up a lot over a long/hard workout. For example 200W warming up will be 105-110 bpm but 200W cooling will be 125-135 bpm. That really hasn't changed over my three seasons.

I do NOT drop my FT power in CPS for indoor workouts. To my mind, that's making these somewhat artificially restricted workouts appear more effective than they ought to be. It can be a little depressing to finish a tough 2hr session and find you've only done 140-150 TSS but ... dem's the breaks.

ymmv,
rmur

I'll be pretty pleased next March if I discover that my FT has been depressed by 30% all winter.

To be more precise, power is calculated from average hub torque over one revolution and time to complete one revolution. These boxes are running in the MHz range with an E-clock of at least 1mhz which would put resolution at oh, at least 0.000001s. That's pretty darn decent.

Weisse Luft didn't quite have it right. The power measurement on both powertap and SRM fully account for micro variations in wheel speed and torque, it's just that in the case of PT, the resolution limit is the time it takes to complete one wheel revolution and in the srm, the time it takes to complete one crank revolution.

That is to say that if the SRM reports that you put out 250w over the last few seconds, you can be assured that your average power over the last few seconds was indeed 250w. What you can't tell is the instantanous power demands but....who cares?

Wow, thanks for that link with a wealth of well-presented info. I had done some searches on this very subject, but had not come across this thread, which is the most enlightening I have seen. Gotta love Andy Coggan - 20 years of extensive trainer experience - yikes! Becoming a morning person and riding the the 'cold' (if you call high 30s - low 40s cold) is suddenly much more appealing. If I ever move to colder environs, I may have to face this beast again. But for now, I think I'll stick to avoiding trainer workouts for power critical rides if at all possible, and work 'harder' indoors than I have been, but only when necessary. I'm trying to build fitness, not character. Call me a wimp, but if you have the means, I highly recommend it.

I have a Fluid2 trainer. This is the first year I've ridden 'seriously' through the winter and have limited experience indoors. Due to this, I haven't yet invested in a fan, so I'm relying on, er...natural ventilation. I guess I've just been grititng it out, seeing my pools of sweat as badges of valor. Evidently ventialtion will make a considerable difference, so I should probably stop trying to impress myself and get one now. Anyone have suggestions on a good fan, while we're on the subject?

Thanks to all for the great info. I'm going to go get on my unventilated trainer now for some recovery miles - those are doable.

Cheers,

Ethan

I lived in San Diego before I moved here 6 years ago, so I can relate. A regular box or oscillating house fan is a must, doesn't matter what kind as long as it moves the air across your body.

Trainer riding is a way of life for me in the winter, since I don't ride in the 30's or in the dark. I feel like I get great interval workouts on the trainer since I can control power very easily without having to worry about hills or wind on a route. Also, the trainer workout tends to be very vigorous because there's no coasting, stoplights, etc. Ride early if you prefer, but after a couple rides on a nice trainer like the Fluid2 it'll feel more natural and you'll be able to generate close to the numbers you'd expect on the road. Good luck.

yeah... i think it's pretty clear that both the PT and SRM have the fidelity necessary to capture enough of the micro variations of wheel speed and crank speed to accurately capture the power reading... but what I'M saying is, isn't that's a good thing? the micro variation in wheel speed are REALLY there and they REALLY do affect the torque that is applied to the pedal and to accurately capture power, the power meter REALLY does need to capture those variations...

that is why i think it's also pretty clear that the indoor - outdoor difference is not about the power meter reading wrong, it's about your legs that are not used to dealing with this type of wheel-speed variation... or as was basically concluded in the other thread... having to deal with pushing through the dead spots that makes riding indoor more like riding a false flat than riding on the flat... i think it's pretty clear that the indoor - outdoor diff is because you are using muscles indoors that you just don't use all the time outdoors (unless you are use to doing hour long climbs) and your muscles need to fire at slightly different times and for slightly different durations indoors and outdoors... it takes time to get used to riding indoors and outdoor because riding indoors and outdoors are REALLY different... what's happening is that at the end of your dead spot while riding indoors the wheel spins down faster so your legs need to exert a larger torque than they are used to in order to maintain your rpms but since it's easier to accelerate indoor than outdoor (same intertia reasoning) your legs exert less torque than they are used to on the down-stroke... imagine, indoors, if you mapped your torque you'd get a sine wave with a greater amplitude than you'd get outdoors, but on average the torque would be the same...

i would bet though that you actually REALLY do burn more calories dealing with the high amplitude indoor torque curve than the low amplitude outdoor one though... i would think that they physiological demands of dealing with indoor training with the faster wheel spin-down is harder than the smoother outdoor riding... just like it's harder to ride rolling terrain than flat terrain... gravity is against you going up but is with you going down but it's still harder to ride rolling terrain than flat...

if you want your indoor and outdoor number to be closer you need to A) ride more indoors or B) get a trainer with a larger flywheel... you could actually use this to your advantage in that you could simulate a hill workout by riding without a flywheel and simulate riding the flats by riding with a big bad flywheel...

shouldn't the PT be more inherently accurate since it's fidelity is 2 to 3 times greater than an SRM (ratio of wheel rpms to crank rpms)?

[Edit] correction: the torque to time or pedal angle graph will actually be flatter indoors than outdoors[Edit]

Theoretically, the PT could have a finer time resolution than the SRM, but in terms of accuracy, they're the same. In practice, both are about the same because the power reporting is based on time, not on crank or hub revolutions (except for the ergomo which is rev based).

Every 1.26 seconds (or something like that for the Powertap) the hub sends out the average power for all of the hub revs in the prior 1.26 seconds. I don't know what it does with the half finished ones, it probably sends them with the next packet. One interesting thing about this is that the time base doesn't always line up with cadence so you can see your power oscillate between a number below and above where you're actually riding.

The funny thing is that SRM lets you set recording rate which is quite a bit faster than cadence (e.g. if you're pedaling at 60rpm, the SRM should only be able to calculate power every second. It however lets you go to I think .5sec or perhaps even faster). So, SRM's either got more than one crank position sensor or they are simply synthesizing a crank position based on the last crank velocity to come up with those power numbers. From what I've read elsewhere, this is a pretty safe thing to do.

I got my PT at the end of the season and only got to ride it a couple of times outside so I haven't really got a feeling for this, but is HR generally elevated while climbing as opposed to riding on the flat for a given power?

Riding out of the saddle will cause HR to rise because muscles in the back and upper body are being utilized and subsequently demanding greater blood flow. Is that possibly what you're noticing?

I can't say I've really tried to identify a HR difference for a given power between riding flat and climbing in the saddle.