Are you saying there is an indoor trainer that measures power as accurately as a Comcept2 ?

If so which model?

I'm looking for accuracy of +\- 0.5%. Is that defined enough for you?


What I'm asking is is there an accurate reliable repeatable trainer or do I need to get an SRM Pro?

Quote about SRM Pro accuracy from SRM site:

"+/- 1% - scientifically proven". There you go. Not even the best can provide what you want. Instead of "scientifically proven" (there is no need for any science in here, just decent set of measurements/protocols on certified equipment) what I'd really like to see is the explanation of said claim, something like "all our measurements were accomplished on such and such CERTIFIED dynamometer under such and such conditions etc. etc. Until then ...

0.5% really? Keep dreaming...

There is already a DLL that works with both the Velotron and Computrainer. It's imbedded in the RM1 package and it's quite easy to use.

   
http://www.atm.ox.ac.uk/...hysics/ergometer.htm


By measuring the damping, the ergometer will automatically compensate for any of the following:
- Opening/Closing the vents to increase/reduce resistance
- Changes in friction on the flywheel bearings with time
- Changes in air pressure, density, viscosity etc.
- Environmental factors such as proximity to walls or other ergs
Things that are not compensated are:
- Changes in the chain friction
- Changes in the tension of the return mechanism
- Manufacturing variations flywheel moments of inertia (probably negligible)
- Changes in flywheel moments of inertia (unlikely with the solid flywheels)



It isn't unreasonable to want better than is currently available.

Thought SRM Pro was more accurate than +\- 1%.
Edit: checked, the SRM Track Pro, 8 strain gauges is accurate to +\- 0.5%.

Concept2 is accurate. See above. One old study and one athlete who seemed to row in a manner which showed differences between the Concept2 and the measuring devices does not prove they are not.


Don't get me wrong. I would get a lot of entertainment from exposing the Concep2 if they were not as accurate as claimed.

If your willing to spend huge $ I'm sure almost anything is doable. The more important question is what exactly do you hope to gain by 99.99% accuracy?

SRM is "listed" as +-1%. Quarq is listed as +- 1.5%. All of these devices using strain gauges are influenced in some way by temperature, so who knows how accurate they "really" are under normal operating conditions.

I can't help liking accuracy. I gave a watch to a charity shop on Saturday because it was gaining 60 seconds a day. Really annoyed me.

If you think about it, if the power meter you are using is 2% out, that is the equivalent of over a minute an hour. You wouldn't use a watch that inaccurate to time a training session - you might as well use bloody Strava.

SRM has been around decades now, and their Pro is accurate to +\- 0.5%, the competition should try to match or improve upon that. But it seems the public just wants cheap and will accept inaccurate data- hence the daft situation where people buy a power meter which only measures the power of one leg. That sums it up really, no one cares about accuracy, all they want is features and apps and software and pretty graphs to play with on computers. All daft when the data you start out with is garbage.

There was a chap on another forum complaining that the timekeepers were giving him the wrong time in a time trial because Strava said he was faster, what a twat, sums up things today

There is a big difference between a watch and a PM.

Assuming no temperature fluctuation, PM accuracy is within say 2%, but it's consistent. I.e. your power might be 205 watts and it's reading 201, but it will always read 201 when doing 205 watts. So it's consistent from workout to workout.

Further more there is no evidence that training at exactly 201 watts as opposed to 203 or 205 will make any difference (nor would you even notice). Now if you show up at the train station at 10:00 on your watch and it's really 10:01 you most likely will miss your train.

This.

Actually it is not. It depends on one's speed. So if you want to be "accurate" you could start right with your statement.

I don't want an argument with you. I should have worded that better. The point I'm making isn't that 2% less power would cost over 60 seconds an hour. I'm pointing out that no one would put up with a watch being 2% fast or slow, or scales which weigh 2% heavy or light. So why settle for a power meter which isn't accurate?

As to the time 2% misreading of watts might cost, it depends on the power, the weight of the bike and rider, the CdA the wind direction and speed, and the terrain how much 2% of power might cost over an hour.

I can't be arsed to work out the numbers but if you take 2% 250 watts that is 5 watts. So yes I agree my statement grossly exaggerated the time difference.

My point though is you are trying to measure small improvements (hopefully) in performance. A well trained cyclist may be working his balls off to get improvements of a watt or two a month.

I think you assume I'm having a go at the use of power meters, I'm not. I'm just frustrated by the unreliability, repeatability and inaccuracy of some power meters.

If you want to measure CdA you need accuracy.

I just wish power meters were more accurate and reliable. Perhaps I'm an obsessive about these things. Perhaps I'm asking for something which isn't possible technically at the present time.

I don't see why wanting things improved should attract ill feeling and I'm sorry if my comments annoyed you.

The point is that you have so many external factors affecting your performance at particular situation that 0.5% accuracy of power meter is statistically meaningless. The way you bounce of the saddle on this particular day at this particular grade/cadence/gear/temperature/etc/etc will likely affect the efficiency of converting the energy your body has to useful mechanical energy way more than 0.5%. So your 1-2 watt "improvement" is not statistically meaningful.



Of course it is possible. But be prepared to pay astronomical sum. I already showed the link to industrial the grade torque meters. You can see the price. And they're big and heavy and not very energy efficient. If they have to adapt it for use on a bike their already high price will skyrocket. So who is going to spend huge wad of money to develop product which is going to be purchased by whole 2 persons in the world?


You probably misunderstood me. There are no ill feeling at all. I am just trying to explain that your frustration might not have much real basis.

Yep see your point.

I get frustrated because the sat nav invariably tells me I'm going slower than the speedometer in the car. That really winds me up.

Car speedometers are legally required not to show 100 km/hr when in fact you're doing 105 for example. So they have safety margin built in and normally the speed they show is a bit FASTER then the real one. So again no need to get frustrated ;)

Might be a challenge on OS X and Linux !
And assume there is no documentation ?

Well I assume you know the car's speedo will always be wrong due to legislation requiring speedos to never report a speed under actual speed in many jurisdictions. How much it over reports speed depends on the manufacturer's calibration margin allowance to ensure they comply and of course the circumference of the tyres on the wheel used to measure speed. It's often over reports speed by as much as 5-10%.

But then that just brings up another point with your accuracy comments.

While your car's speedo may be inaccurate, it is very consistently so, always over reporting by a fixed percentage. If it does change, it does so only marginally with the evolutionary change in circumference of your tires with wear and inflation levels, and in a step change if you change tires and/or wheels and tires which have a different circumference.

Hence, provided you inflate your tires, you are able to consider your speed reading to be quite precise to whatever the unit limit on reading your speed is, and on a bike many a speedo with wheel speed sensor will read to within 0.1km/h or less than 0.5% precision, and do so with excellent consistency.

So you can say, provided you don't change tires/wheels, that making relatively comparisons with the speed measurements is still a very valid thing to do, even if the absolute accuracy sucks.

In this case, attaining accuracy is only a calibration adjustment setting away (e.g. a correct wheel circumference).

So the issue with power measurement is assessing whether you are dealing with an accuracy issue, and/or a precision issue. With high end power meters, they are generally good with precision/consistency, and also enable you to adjust the calibration so you can attain excellent accuracy as well.

But we also need to consider under what conditions of use and riding they are precise/consistent. Some struggle in some circumstances while being fine in many others.

It would be a challenge for OS X and Linux. Usage though is very straightforward as the interface is exposed.

LOL. It would be a hack, and redistributing copyrighted material would be illegal. And it will only work on Windows.
But aside from that it would be great.

How about you lose the attitude? You are coming across as extremely rude. First of all It's not a hack, it's the very same DLL that they use and can be used in the very same way that they use it. Who said anything about redistributing it? If you have their demo and you have it. Yes it will only work with Windows. That is the cost of Linux and OS X. So what is your proposed solution? .

If you don't redistribute the DLL your code won't work, how else would you make and sell software?
These days, actually for the last few years, multiplatform has become vital, not optional ; iOS, OSX, Android, RT, Linux .. Windows is the least concern.
Oh, I'm not being rude, you just don't like what I'm saying.


Well, the key point here is the CT is a closed system.
Avoid it if you can. Its old and tired technology with an unjustified premium price.

Mark

That would be a fair analogy if Apple were still selling these pieces of s*** and telling you they were cutting edge:



;)

Mark

Well your first point I agree with, it is currently a closed system and shouldn't be. Left over policies from a forgotten era. They can't possibly consider themselves a "software" company that they need to lock down the interface. If anything, that policy is costing them a lot of sales.

I don't agree though that it's tired technology. My CT is over 15 years old and is still working perfectly. Back when I got it, there was absolutely nothing remotely like it. Today there is plenty of competition, but just how durable are these new trainers? I know mine was good for over 15 years of use and likely many more. You can pick up used ones for a very reasonable price. I agree their prices for new are a bit high, although they will tell you they are assembled in the U.S if you are into that sort of thing.

So avoid? No, I wouldn't recommend avoiding them as long. Not till we see just how durable some of the new trainers really are. In the long run they could end up costing you more.

At least with Apple it is closed from a hardware perspective but their API's are available to anyone to write software. One could argue that's a selling point for them as you don't have to worry about hardware and drivers being properly designed.

Just dual boot with a windows partition.