-
Has their accuracy been bought into question?
DC rainmaker said this:
“At present, based on me (and only me) it would be difficult for me to swap out my existing power meter with the Stages power meter. There’s just too much variance and fluctuations in power. Do I think that Stages can get there though? Yes, I do. But I think it’s going to take time, and likely more software work.”
However it appears that stages updated the firmware since then that seems to have fixed the issues.
http://www.dcrainmaker.com/2013/06/stages-review-update.html
-
Yes, just now, by me. I can't see how their device can measure actual power output in any meaningful way. Not that it matters, as long as it's wrong in the same way all the time for a particular rider, since repeatability is by far the most important quality for most people.
As I thought. -
How so? Is it to do with measuring single leg power and doubling or something else?
No, the doubling of single leg power is just an obvious fudge to keep the price down, it probably doesn't matter unless you have a severe imbalance which is likely to change, e.g. recovery from injury. Sure, they're not measuring actual total power, but at least they are measuring actual left leg power and letting you guess what that means for total power.
The problem I see is that it seems to treat the crank as a beam, which gets bent in one plane by pedal force. What actually happens to cranks is that they get bent in two planes and twisted, and the amount of twist changes a lot depending on the eccentricity of the applied load. The crank also stretches or compresses if the load is not precisely tangential to the circle described by the pedal. The 2nd plane bend, twist and stretch will deflect the strain gauges too, and riders with a wider stance will twist their cranks more than riders with narrow stance, for the same applied load. Also, even a particular rider will apply a more eccentric load when they lean the bike while riding out of the saddle. Unless they have been spectacularly clever with the placement of the strain gauges and the algorithm which calculates true circumferential force regardless of eccentricity and atangentiality*, they can only measure true torque for one unique value of eccentricity and tangentiality. In other words, they could make it accurate for a load applied precisely in the centre of the pedal and exactly tangential to the pedal locus by calibrating against a torque measurement at the BB axle, but they can't make it accurate for all riders in all circumstances.
*I just invented that word. It's the angle between the pedal force vector and a tangent to the circle described by the pedal. It has two components, since riders tend to push inwards or outwards as well as applying a force which tends to stretch or compress the crank,
-
Well, whatever is lost in the transmission can't be used to move you forward, so if anything it's better to measure at the wheel than at the crank. None of the power meters really measures how much power the rider is actually applying, since there are losses or unmeasured vector components in all of them. A six-axis measurement at the cleat could potentially capture all of the rider output, but some of his output can't be translated into usable power by a conventional drive train, so what would be the point?
-
According to this: http://djconnel.blogspot.co.uk/2012/09/interbike-2012-crankarm-based-power.html they could measure the twist if they measure in different locations. Do they do this?
Actually that was Rotor he was talking about.
-
Ed, definitely agree that you need to be fresh to do a 20 min or 40 min test properly.
My average hr used to be around 95 to 96% of max and I would be near vomiting by the end. I once held my absolute max hr (187) for the last 4 mins of a 20 minute test. Every second felt like a minute and it was pure agony. Watching the timer, praying for the clock to read 20:00 would only make time slow down more. I used to dread the test, genuinely fear it (I had a coach for a short while so I was at his mercy and I seemed to get a 20 minute test every month).
In fairness I think I used to go too hard in the tests as I never really managed to repeat the effort levels on race day (which is daft I know). I think I used up all my mental bullets in training. The coach was pretty shit (he was cheap and you get what you pay for) but I think I went too hard to try to prove to myself that the coach's training was working (whereas it wasn't, I was just flogging myself harder and harder), but its pretty clear to me that 80% of max isn't hard enough to give you a useful idea of FTP either.
-
According to this: http://djconnel.blogspot.co.uk/2012/09/interbike-2012-crankarm-based-power.html they could measure the twist if they measure in different locations. Do they do this?
I don't think they can instrument the crank in enough places to isolate usable torque from all the other force components which are deforming the crank within the small area of the back of the crank to which they stick their device.
And if I'd seen that blog post before, I could have saved myself a lot of time by just linking to it instead of essentially rewriting the same thing from my perspective :-)
-
Of the people I train with, most use a power meter of some description. Three use stages and all three has problems with battery consumption. One has problems while doing shorts efforts that the latency is so slow he'll be 5 seconds into a 200m sprint before it registers a change in power. Two have had problems with water ingress. Finally one has completely unusable data, we'll be steady climbing at 300-320w and his power meter will be reading 280, yet on the same ride he will be sat in the wheel on the flat and it will read 330w. Now I can't vouch for anyones pedalling style of how smoothly they apply their power, but this is enough first hand problems from people I actually know to put me off spending my money on them. So despite what anyone may say on the internet about sky using them I wouldn't until they're on their 2nd or 3rd generation until it has actually been proved to be reliable for some length of time.
-
No, the doubling of single leg power is just an obvious fudge to keep the price down, it probably doesn't matter unless you have a severe imbalance which is likely to change, e.g. recovery from injury. Sure, they're not measuring actual total power, but at least they are measuring actual left leg power and letting you guess what that means for total power.
The problem I see is that it seems to treat the crank as a beam, which gets bent in one plane by pedal force. What actually happens to cranks is that they get bent in two planes and twisted, and the amount of twist changes a lot depending on the eccentricity of the applied load. The crank also stretches or compresses if the load is not precisely tangential to the circle described by the pedal. The 2nd plane bend, twist and stretch will deflect the strain gauges too, and riders with a wider stance will twist their cranks more than riders with narrow stance, for the same applied load. Also, even a particular rider will apply a more eccentric load when they lean the bike while riding out of the saddle. Unless they have been spectacularly clever with the placement of the strain gauges and the algorithm which calculates true circumferential force regardless of eccentricity and atangentiality*, they can only measure true torque for one unique value of eccentricity and tangentiality. In other words, they could make it accurate for a load applied precisely in the centre of the pedal and exactly tangential to the pedal locus by calibrating against a torque measurement at the BB axle, but they can't make it accurate for all riders in all circumstances.
*I just invented that word. It's the angle between the pedal force vector and a tangent to the circle described by the pedal. It has two components, since riders tend to push inwards or outwards as well as applying a force which tends to stretch or compress the crank,
Awesome post. Repped.
Is there a massive job to calibrate for all the different cranks? or am I thinking stress/strain not out'n'out force. Thats kinda why I assumed they made asthetic copies of NDS crank arms.
-
Stages give me really strange results when I'm doing my one legged drills
You would hope that stages has been tested to see how varying the eccentricity of the load affects the reading that it produces. Very easy to do and quantify how much error it would produce, maybe its un noticable or maybe its loads and they're keeping quiet about it!
-
No point in keeping quiet.
Even amatuer blogg writers reviewing their new Stages PM, will often have access to a powertap wheel. To run a comparison. Not saying PT is absolute. But its a known quantity. Wouldnt be difficult to spot poor out-of-the-saddle measurments.
-
Is there a massive job to calibrate for all the different cranks?
No, it's pretty trivial, just hang a known mass off the pedal at a known crank length with the crank horizontal and see what number comes out of your strain gauge for that known torque. I would expect them to calibrate each individual crank like that in order to allow for manufacturing tolerances on the crank and their own mounting and production tolerances on the module. The trouble is, riders don't apply a nice neat force vector like a test mass, and they don't all use pedals with the same stance width as whatever was used during calibration.
-
No, it's pretty trivial, just hang a known mass off the pedal at a known crank length with the crank horizontal and see what number comes out of your strain gauge for that known torque. I would expect them to calibrate each individual crank like that in order to allow for manufacturing tolerances on the crank and their own mounting and production tolerances on the module. The trouble is, riders don't apply a nice neat force vector like a test mass, and they don't all use pedals with the same stance width as whatever was used during calibration.
I guess I over estimated the complexity of the system.
Then again the price reflects this somewhat.
ewanmac
Howard
gbj_tester
yetidamo
Slack
Smallfurry
daccordi
skinny
Dammit-
Post a reply
About
Power Meters / Powermeters (SRM, Powertap, Quarq, Ergomo, Vector, Stages, power2max, P2M, 4iii, InPower, Cinch)
Posted by
@hippy
Yes, just now, by me. I can't see how their device can measure actual power output in any meaningful way. Not that it matters, as long as it's wrong in the same way all the time for a particular rider, since repeatability is by far the most important quality for most people.