Metrigear pedal spindle-based weenie-compatible power meter

[djconnel]

From the MetriGear pre-Interbike open house this evening:



The schematic shows the electronics inserted into the axle (which Metigear takes care of: they'll sell the power meter inserted into a new pair of pedals) and an external battery/transmitter which fits on the crank arm (on the side opposite the pedal, wrapping around to the side of the crank arm).

Strain gauges are inserted into the hollow of the Speedplay spindles, along with electronics, transmitting ANT+ data to a head unit. Weight penalty = 25 grams per pedal on top of a Speedplay stainless steel spindle. Two pedals provide better power data (left and right foot separately, or the average) but if the extra 25 grams is too much, one pedal only is an option (this relies on pedal balance being fixed, something I wouldn't personally trust: the old Ergomo problem). Most of the weight is the battery, it seems. They'll also have a heavier (+25 gram per pedal) battery as an option for longer battery life.

Even though Frogs are spindle-compatible, they didn't have any Frog prototypes yet, but I personally see no issues with those. I also asked if it would handle negative torque, as on a track bike (after I yanked up on the display pedal, and it reported negative force). Yes, it can. I asked about Ti spindles. I think there was confusion, because I was told the Ti spindles lacked wrench flats and required a hex wrench be inserted where the electronics are, but my X/1s have wrench flats, so I need to check on this. Maybe he was referring to nanogram spindles.

They claim Q1 2010. Engineers tend to forecast the time it will take to solve known issues, but tend to overlook the time it will take to solve unforeseen issues. So maybe a bit later than that. Super-impressive stuff, however.

[Raxel]

I wonder why powermeters are that expensive... after all FSRs (force sensing registers) are very cheap, and sending the signal wireless is not that expensive either.

http://www.trossenrobotics.com/store/c/ ... nsors.aspx

[djconnel]

In this instance, I think the electronics are fairly pricey: the circuit board is very small (fits into the spindle of a pedal). But beyond that, development costs need to be amortized over a relatively small number of units. No massive economies of scale here.

The estimated cost is "under $1000 per pair". Which is really two power meters.

[The Stig]

I would think that measuring power at the pedals is like measuring HP at the crankshaft in an engine...I think that the best way to measure power accurately are at the wheels or wheel horsepower as they say, minus the rotational weight/friction of the drivetrain....

[djconnel]

I think this is an excellent point: some are more interested in the power propelling the bike (PowerTap) and some are more interested in the power generated by the body (MetriGear, Quarq, SRM, BrimBrothers). For a bike power-speed modeling standpoint, the PowerTap type number is more interesting. But physiologically perhaps the other is more relevant.

More info: page 1 of the information sheet:

[nitropowered]

I would think that measuring power at the pedals is like measuring HP at the crankshaft in an engine...I think that the best way to measure power accurately are at the wheels or wheel horsepower as they say, minus the rotational weight/friction of the drivetrain....

Wouldn't a crank based PM be the same as measuring power at the crankshaft of an engine?

This system would be equivalent to measuring power at the cylinder


for purposes of a car, motorcycle, other motorized vehicle, wheel hp is most important since thats what is going to determine if you win the 1/4 mile. However on bicycle applications, You would want crank (or in this case pedal) power so you can get the most accurate amount of power data. Athletes use power data mainly for training and personal evaluation. The drivetrain losses on a bike is very similar between bikes so I think it would be irrelevant.

[The Stig]

good point....but for me, I'd rather measure net power at the wheels rather than gross power...

[Ghastly]

wouldn't it be better, for training purposes, to get the power-input (what you produce) on your screen instead of the power-output (on the wheel or on the road)?

[CharlesM]

The more important metric is repeatability / consistency of the measure rather than where the number is taken...

All of the meters will measure within a few watts of each other. the bigger thing as a training aid is that they measure it the same way every time.

[The Stig]

wouldn't it be better, for training purposes, to get the power-input (what you produce) on your screen instead of the power-output (on the wheel or on the road)?

For others maybe, but like I said, I prefer net power....

[The Stig]

But I must say, that is the cleanest looking, most compact power meter solution I've seen...

[Coolhand]

The more important metric is repeatability / consistency of the measure rather than where the number is taken...

All of the meters will measure within a few watts of each other. the bigger thing as a training aid is that they measure it the same way every time.

Exactly- consistency, reliability and solid customer support are key for Powermeters and also the reason why SRM and Saris dominate the field.

[youngs_modulus]

I asked about Ti spindles. I think there was confusion, because I was told the Ti spindles lacked wrench flats and required a hex wrench be inserted where the electronics are, but my X/1s have wrench flats, so I need to check on this. Maybe he was referring to nanogram spindles.

They claim Q1 2010. Engineers tend to forecast the time it will take to solve known issues, but tend to overlook the time it will take to solve unforeseen issues. So maybe a bit later than that. Super-impressive stuff, however.

It's great to have your report, DJ...thanks for posting it. This system looks very promising. I wonder if Richard Byrne won't try to buy Metrigear.

I just checked my 2007 Speedplay X1 pedals, and they too have wrench flats. It's possible that the Ti spindles have a smaller cavity than the stainless ones and that the guy you talked to was just confused about why the cavity on Ti spindles wouldn't work. I sincerely hope this becomes available for Ti-spindled Speedplays. Certainly, the firmware would need to be re-calibrated for the more-flexible Ti axles, but that shouldn't be a big deal. I sincerely hope that Metrigear will retrofit users' pedals. I'd do this in a heartbeat.

On the wattage list, I expressed some concern about how Metrigear was determining crank RPM. I was worried that they were using a once-per-revolution chainstay pickup or something like that. However, now that some details have come out (and we know that the pedals use both strain gages *and* accelerometers), there's an easy way to get constant crank rotation speed. Radial acceleration=(V^2)/R; R is your crank length, and we get acceleration from the accelerometers, so V is easy to determine. You can easily filter out noise from bumps by summing the signals from both pedals. (Bumps would cancel each other out, but radial acceleration would be cumulative). If you have an accelerometer in the pickup, it gets even easier...the radial signal would be: [raw pedal accelerometer signal in radial direction] - [pickup accelerometer signal in vertical direction] = radial acceleration.

While I hope Metrigear is able to deliver on time, I am even more concerned that they deliver on budget. I was really disappointed that the Quarq meter went up in price so quickly. From a business perspective, it was probably the right thing to do (i.e., we're selling as many as we can make; lets adjust the price to match demand). However, it meant that I couldn't afford a Quarq and had to go with a PowerTap. If this becomes real, I'll sell my PowerTap and buy a Metrigear setup!

Jason

P.S. There's nothing wrong with PowerTaps...I'm perfectly happy with mine. It's just that I'd like to switch wheels at whim. Also, it would be easier to move pedals from bike to bike than it would be to adapt a PT, Quarq or SRM to the track.

[djconnel]

Right: You could imagine looking at a DFT of a recent time sampling of the signal, where you'll find peaks at f and 2f, where f = 2 pi / cadence. It would take a few pedal strokes to totally lock in to this. But those guys know more about this stuff than I do....

Page 2 of the info sheet:

[HammerTime2]

They claim Q1 2010. Engineers tend to forecast the time it will take to solve known issues, but tend to overlook the time it will take to solve unforeseen issues. So maybe a bit later than that. ...

I don't know what Metrigear claimed at the open house, but p. 2 of the info sheet says "Vector is targeted for availability starting in Q1 2010." Targeting availability is not the same as claiming availability.