tag:blogger.com,1999:blog-1564958057737541664.post676057397145511329..comments2024-02-14T17:11:22.168-08:00Comments on On Bicycles, and.... what else is there?: power accuracy and the uniform cadence approximationdjconnelhttp://www.blogger.com/profile/01484858820878605035noreply@blogger.comBlogger7125tag:blogger.com,1999:blog-1564958057737541664.post-77350716335141777512011-08-20T07:52:49.608-07:002011-08-20T07:52:49.608-07:00Good question! It's been discussed on wattage...Good question! It's been discussed on wattage, and I've discussed it with people in the business, but the best reference I can find might be here, in NYVelocity's typically excellent <a href="http://nyvelocity.com/content/equipment/2009/tech-talk-quarqs-jim-meyer" rel="nofollow">interview with Jim Myer</a> (definitely read more of this -- it's great when the interviewer understands the physics):<br /><br /><i><b>AS:</b> In other words, you're only measuring once per revolution, so you have to assume that rpm is constant through that whole rev.<br /><br /><b>JM:</b> You have a time for each revolution, 600 milliseconds or whatever.</i><br /><br />The multple sensors are for extracting strain, which is used to derive torque.djconnelhttps://www.blogger.com/profile/01484858820878605035noreply@blogger.comtag:blogger.com,1999:blog-1564958057737541664.post-42144894272634219132011-08-18T07:47:28.261-07:002011-08-18T07:47:28.261-07:00Do you know for sure that the Cinqo measures caden...Do you know for sure that the Cinqo measures cadence only once per revolution? I remember Jim saying that it has something like 8 reed switches in the crank, so in theory it could be measuring cadence up to 8 times per revolution.CLBhttps://www.blogger.com/profile/15893767402284471871noreply@blogger.comtag:blogger.com,1999:blog-1564958057737541664.post-14018727940178654892010-02-02T14:31:16.365-08:002010-02-02T14:31:16.365-08:00The work I was involved with a while ago with acce...The work I was involved with a while ago with accelerometers and also piezo's tended to drift a bit and needed frequent zero correction. Drift was measurable over a 15s sample.<br />Foil gauges as used in the PM you listed are much less prone to drift if mounted correctly and temp compensated in my experience.<br /><br />I am not an expert in this area by any means, I do the mechanical hardware not the data aquisition, but this has been my experience from systems I have been involved with. There may be some new technology I am not aware of in this area and the Metrigear guys might have this under control. I should ask them and see if they tell me!Unknownhttps://www.blogger.com/profile/03412561521490228698noreply@blogger.comtag:blogger.com,1999:blog-1564958057737541664.post-14625358421005612542010-02-02T05:30:17.025-08:002010-02-02T05:30:17.025-08:00Which sensors? The force sensors in the Vector ar...Which sensors? The force sensors in the Vector are piezoresistive, which have improved stability relative to the foil sensors which, I believe, are in the SRM and PowerTap and Quarq. Correct me if I'm wrong here... that's what I was told.<br /><br />It makes sense foil would drift. Metal fatigues and oxidizes. Piezoresistance is measured with a crystal, which is more stable.<br /><br />There's other forms of drift, however. What if the mechanical properties of the spindle change over time? A lot of that depends on cleat maintenance, to avoid the cleat spring rubbing against the spindle.djconnelhttps://www.blogger.com/profile/01484858820878605035noreply@blogger.comtag:blogger.com,1999:blog-1564958057737541664.post-12126021570915034322010-02-02T02:08:48.826-08:002010-02-02T02:08:48.826-08:00I designed a force pedal system for the Aust Insti...I designed a force pedal system for the Aust Institute of Sport using an encoder in the pedal axle. I haven't seen any data from it, but it did have 1 deg resolution. The problem with SRM is that a reed switch is not the most accurate in terms of timing 1 rev but works well because it cancels out any error in the next rev.<br /><br />Some questions I have relate to power accuracy at high versus low cadence at high versus low KE.<br /><br />Overall I think the SRM does a reasonable job as long as you are aware of its limitations and nuances. I am very sceptical of people who claim absolute data from any power meter.<br /><br />I hope the Metrigear stuff delivers its promise, I am still concerned about drift of their sensors.Unknownhttps://www.blogger.com/profile/03412561521490228698noreply@blogger.comtag:blogger.com,1999:blog-1564958057737541664.post-49496024803362290652010-02-01T02:22:53.425-08:002010-02-01T02:22:53.425-08:00Fantastic!
A more general result is that cadence ...Fantastic!<br /><br />A more general result is that cadence is correlated with pedal rotation in some way: <br />ω = ω₀ ‒ Δω for half the stroke duration (time, not angle), while τ = τ₀ ‒ Δτ. Then for the other half of the stroke duration, ω = ω₀ + Δω <br />while τ = τ₀ + Δτ. Δτ might be positive or negative or zero. Then <br />average power = [ (ω₀ ‒ Δω)(τ₀ ‒ Δτ) + (ω₀ + Δω)(τ₀ + Δτ) ] / 2 = ω₀ <br />τ₀ + ΔωΔτ. Meanwhile the Average Cadence Approximation (ACA) is ω₀ τ₀. So this says if Δω and Δτ are positively correlated, the ACA will under-report, while if they are negatively correlated, the ACA will over-report, while if they are uncorrelated, the ACA will accurately report (in this trivial example, "uncorrelated" would mean either Δω or Δτ are zero, but more generally, this it may simply mean the sign of one doesn't predict the sign of the other). <br /><br />This little analysis has changed my perception of whom to more believe during examples where PowerTap reports higher than SRM or Quarq for AP over moderate-duration efforts (crank phase aliasing causes disagreements on very short efforts).djconnelhttps://www.blogger.com/profile/01484858820878605035noreply@blogger.comtag:blogger.com,1999:blog-1564958057737541664.post-29941676674959460692010-02-01T00:56:07.210-08:002010-02-01T00:56:07.210-08:00This is something I have been working on for some ...This is something I have been working on for some time, we are looking at mounting a rotary encoder to a bb axle and integrating this with an SRM, should be interesting!Unknownhttps://www.blogger.com/profile/03412561521490228698noreply@blogger.com