On 18/04/2011 20:41, RussellMc wrote:
>>> However, in many cases power is
>>> averaged over a number of seconds with the display being updated
>>> periodically, so short term variations are smoothed out.
>> Yes - specifically stuff like the above, was exactly what I was
>> wondering about.
>> So in this case the 10kW for a millisecond example I gave in my other
>> mail would not be "noticed".
> It would be noticed as a contribution to mean energy in.

Yes - that's why I put the noticed in brackets :-)
I meant that peak specifically would not be noticed, although of course=20
it's energy will contribute to the reading.

> Reading too finely can cause problems.
> When pedalling a bike the power is delivered in two surges per crank cycl=
e.
> If you track this power variation and display it, a digital display
> would be a constant mess of changing figures. Even reading an analog
> display would be difficult if it tracked power surges.
>
> At one stage I had a certain update rate - maybe once per second, and
> customers asked for a much slower update rate.

I was thinking more along the lines of a setting to choose how the=20
information is gathered. I guess normally people will want to know the=20
average power over a while, so the slow update/sampling will not be an=20
issue. If you want to know peak instantaneous power it is though, so=20
maybe you could read at a higher rate, but display at a slower rate,=20
while retaining things like maximum peak power/speed attained.
You could even have a graph of power over time in more advanced machines.

>> Is this the case in most gym equipment
>> though?
> I think so.
>
>> I would have thought that measuring the delta would be more
>> accurate, as with the momentum of the flywheel you have the possibility
>> of over estimating the maximum (instantaneous) output. At least having
>> an option to set readings this way might be good.
> 1. How would you read the  delta?

With difficulty I imagine, it would require knowing about the=20
flywheel/friction, and figure out how much energy is required to keep it=20
at whatever speed it's at. Then you can (begin to) work out the energy=20
input to the system.
Either that or do something like read the force directly on the pedals=20
with a strain gauge.

> 2. As long as you can accurately measure the fluwheel energy
> contribution you are liable to get a smoother system energy level.
> Refusal of the flywheel to peak read the 10 kW / 1 mS spike would seem
> to STOP it over reading.
>

Why? If it misses a *valid* 10kW power surge (I know 10kW is very=20
unlikely but it's arbitrary), but then reads, say 2kW when the actual=20
power added to the system by the user is only 1kW, how is this more=20
accurate?
I agree for most situations people are probably only interested in power=20
over a reasonable period of time, but I am talking about the situation=20
where people want to know the peak power delivered instantaneously (or=20
as near as) plus how reliable that information is (does it include=20
momentum of flywheels etc)
I would think the best system would be the one that can sample as fast=20
as possible - you can do what you like with the data afterwards, but the=20
detail is there if you need it.

>> to make little difference (although hard to test his empirically). It's
>> one of the magnetic resistance ones, FWIW.
> MR is very good if done well.
> What brand?

Reebok - an "i.bike"

> 'Mine' were sold mainly as "Infiniti" - but also badged with other
> labels by other people.
>
>
>
> R
>


--=20
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.