November 5th, 2009
BikeBiz Dyno Hub Test
The folks at The Bicycle Business recently conducted a test comparing the Schmidt SON dynamo hub to the newer generation Shimano hubs (and the Novatec).
Prior to this test, the only data I’ve seen on Shimano dynamo hubs was outdated and taken from a last generation NX-30. These new tests were conducted on the Alfine, and as you can see, the Shimano competes favorably with the SON. The latest Shimano DH-3N80, DH-3N72, and DH-3D72 all have the same internal parts as the Alfine, so they should perform equally well.
The tests were conducted with a powertap-equipped bike on a set of rollers. Readings were taken at 12mph and 18mph with the light on and off. The numbers on the left are watts/kph.
17 Responses to “BikeBiz Dyno Hub Test”
Glad to see my SON is still on the top of the chart (or, I guess I should say BOTTOM of the chart).
I remember agonizing over spending the difference in cash between the SON and the Shimano and rationalized the extra bucks by thinking it would be insignificant over the life of the hub. On top of that, I knew I’d always wonder about the performance difference every time I crawled onto the bike if I went with the Shimano.
Since I have the Busch & Muller Lumotec IQ Cyo R N Plus LED light, I run with my headlight and taillight on all the time. It’s comforting to see that I’m not taking much of an energy hit over riding dark.
Thanks for the info.
Great test, I was looking out for data with the latest components. I’m still curious to find how much more energy, if any, a SON hub in the “off” position would absorb compared to a good normal hub.
The Shimano hubs are certainly comparable to the SON in terms of rolling resistance, but the SON wins in terms of durability, especially in nasty weather. We have had to warranty many of the upper-end Shimano hubs, but every SON we’ve sold has been 100% reliable.
I don’t think this data makes much sense. The vert axis is “watts/kph” but the speed is given as mph. So 12 mph is 19.3 kph and 18 mph is 29.0 kph. So the power levels for the Alfine are respectively for off12, 12, off18 and 18 are 103, 105, 157, and 162 watts. Where is the 75% improvement?
What power is being measured? Total roller propulsion power? At 12 mph the on condition drew only 2 watts more than off, and at 18 mph the difference is 5 watts. What is the spec power output? 2 watts will not develop 6 volts with a standard 6volt 3-watt load.
This data doesn’t even show that the SON is an acceptable power source, drawing only 2.4 watts at 12 mph.
Something is not right.
@Ken
I forwarded your comment to The Bicycle Business.
Here’s my take on it. The “watts/kph” measurement has nothing to do with the current the hub is generating; all dynamo hubs generate essentially the same output. The “watts/kph” reading is the amount of energy required to travel at a specific speed, as measured at the Powertap. In other words it’s measuring the energy (in watts) required by the rider to maintain a specific speed, not watts as output by the hub. They’re using this figure to determine how much drag each hub introduces into the mix. Whether it’s read as kph or mph is irrelevant because we’re only interested in looking at the comparative drag introduced by each hub, not their output (which is essentially the same or they’d be incompatible with standard headlights).
Alan
Gotta go with Ken on this one. These numbers (Watts/km/h) make no sense. Let’s take the Son at 12 mph (19.1 kph). Multiply the 5.1 watts per kph by the 19.1 kph. That gives a total drag of about 100 W. If Wikipedia is to be believed, that is equal to the total power required for a 70kg person to ride a bicycle at 25 kph on level ground. Is the Son is more than doubling the effort required to ride a bike? That seems, um, unlikely. I think that they have their units very wrong.
(As an aside, those pseudo 3D graphs are often used to obscure the meaning of data because they are harder to read and interpret than either a plain bar graph, or lines or points.)
Ken- Alan is correct. We were not measuring or concerned about the power output of the hubs at all. We only measured the drag of each hub. Bill- I don’t have an answer for you on the wattage required to ride a bike, but even if our math is screwed up we did the same math for all of the hubs. So the comparison of the hubs (our goal) should still be valid even if the numbers are off. The data generated by the Powertap was exported from their Poweragent software into Excel where we made the calculations. I then used the data in PowerPoint to make the graph. I realize that the data/graph we came up with is not perfect and will not satisfy everyone, but we are just bike geeks not engineers. The goal was to give people some idea of how the new Shimano hubs measure up to the Schmidt SON in terms of resistance. I encourage some one else with more time/resources to improve on what we did.
Bob,
It would be nice if you could have added a baseline measurement to the chart, any normal front hub; be it Shimano, Phil Wood or Formula. That way you could have figured out the increased drag (in both off and on positions) that one would incur in exchange for having the convenience of a dynohub.
@Barturle
For anything other than racing, the drag introduced by a SON or Alfine dyno hub is inconsequential. I have bikes both with and without dyno hubs, and there’s no discernable difference in ride quality related to the hubs.
Alan
I hope that it is true, as Bob says, that you can compare hubs at a given speed. However, the difference between Watts and Watts per kilometer per hour makes a BIG difference in comparing the hubs at different speeds. According to th, the Novatec hub is better than the SON for slower rides, because the Novatec has about 40% less drag at 12 mph than the SON has at 18 mph.*
I agree with Alan – I have an Alfine and I can’t feel any drag in normal use. In fact I almost always leave my light on even in daylight.
BTW — I’m very glad that Bob and the folks at BikeBiz did this test. There’s a big difference in price among these hubs and it’s good for people to know what they are paying for. Thanks.
* I’m estimating
SON — 5.35 W/kph OFF and 5.56 W/kph ON at 18 mph (28.97 kph)
Novatec — 5.58 and 5.79 W/kph at 12 mph (19.31 kph).
Multiplying it out
SON — 155 W (lights OFF) and 161 W (SON lights ON) at 18 mph
Novatec — 107.75 W (lights OFF) and 112 W (lights ON) at 12 mph
Which SON model was tested, the SON 28 or the SON 20? If this was the SON 28, I’ll assume that the lowest-drag setup would still be the SON 20 laced to a 700c wheel. This was the setup used by many if not most of the builders in the Oregon Manifest Contructor’s Challenge this year.
I should add that it’s great to see some data on the newer Shimano hubs.
The Schmidt web site links to a test of a german cycling magazine (but the graphs should be clear without knowing any german):
http://www.nabendynamo.de/service/pdf/ar_01-02_2009.pdf
The second graph on Page two compares power with lights on, the third graph does it with lights off, each for 10, 20 and 30 km speed. The test includes both, SON28 and SON20R.
(Watt (without /kph) should be the correct label also for Bikebiz graphs.)
Bill said – “the Novatec has about 40% less drag at 12 mph than the SON has at 18 mph.”
I don’t see how you are getting that since the Novatec has worse drag for any speed than either of the other two. In fact even the lowest for the Novatec – 12, off – is worse than the SON, 18, on!
@No – The graph is in watts per kilometer per hour. The drag we experience is total watts. You are skipping the multiplication step in the post. IF the units are correct then the graph says that the hubs produce so many watts of drag for EACH kilometer per hour. The total drag, then, would be the number represented by the bar (watts/k/h) multiplied by the number on the x-axis (k/h), which would give drag in watts. The point of my post (and Ken’s) was that the results do not make sense in terms of what we know about the hubs. The graph probably represents Watts AT a given speed, but the Powertab/Poweragent apparently gave the units as w/k/h, which is probably wrong.
Bill makes some very good points, but I’d like to expand a bit:
As far as I can tell, they’re measuring (1) rolling resistance of tires rollers, (2) resistance of the roller system, (3) electrical resistance of hub generator, (4) little bits here and there like hub bearings etc (insignificant).
By far the most significant factor in the real world is wind resistance. A bike on rollers will not encounter that, but will encounter some resistance from the rollers. A comparison is fairly meaningless until we know more about the rollers? Wind resistance goes as the square of speed: at 18 speed units you’ll see 2.25 times the drag that you will at 12 speed units. Since this is not reflected in their numbers, either the roller is not simulating wind resistance very well, or some linear-with-speed factor is overwhelming wind resistance here. That’s either horribly unrealistic or simply horrible.
The calculator at http://www.analyticcycling.com/ForcesPower_Page.html suggests that at 12mph (5.3m/s) a “standard” rider on a flat road will need to supply 40 watts, about 60% of which is due to air drag. That’s about 2 watts per km/h. At 18mph (8m/s) their standard rider will be up to 100 watts–3.5W/(km/h), 77% due to air drag. So it’s pretty safe to say that the graph above is not measuring output for any model that includes air resistance. Indeed, since their numbers are between 1.5 and 2.5 times the power output expected by analyticcycling, it’s really unclear what’s going on.
A second data point: M5 (eg. http://www.m5-ligfietsen.nl/site/EN/Models/Carbon_Low_Racer) claims that at 22mph a racing cyclist is at 250W whereas the analyticcycling page estimates 168W. Different assumptions etc, I’m sure, but at least it’s in the ballpark.
So what’s going on here? Apparently SON is the best, but I can’t figure out what this graph is supposed to tell me beyond that.
I also second Bill on the “3-D” bar graphs for graphing 2-D data. They serve only to obscure data, and should never be used.
I think I have these data figured out… none of the comments seem to draw this conclusion, and I believe it to make the most sense.
The data shown are not watts/kph, but rather watts/mph. So at 12mph, the son test was a little under 64 total watts while the Shimano was a little under 65 total watts. @ 18mph, about 64.5 watts for both off and 66.5 total watts for the son and 67.2 total watts for the Shimano.
In my experience, the difference between 12 and 18 mph on un-resisted rollers is almost unnoticeable.
The point here is that at 12mph, the son requires about a watt less than the shimano and the difference is the same or a little smaller at 18mph. we are really splitting hairs, and I question if the equipment used for the test was really good enough to detect differences of only 1 watt out of 65.
On a side note, according to Kurt Kinetic’s power curve (http://www.kurtkinetic.com/powercurve.php), a normal sized rider on a 1 percent grade with no head wind will need 100 watts to go 12mph (wind resistance is very small) and 200watts to go 18mph (wind resistance is a substantial piece of the pie). If you figure the rollers don’t include the 1% grade or any wind resistance, I could see the +-63 watts required to go 12… +-35 watts needed for the 1% grade. The +-66 watts to go 18… +-80watts for wind resistance, +-50 watts for the grade and +-5 watts for added mechanical resistance.
Looks like the folks who did this testing went to a lot of trouble to produce questionable results at best.
I just picked up a Edelux LED light and a Shimano 72 series dyno-hub… don’t notice it a bit on the commute.
My 2-cents.
Submit a Comment
Please take a moment to read our discussion guidelines before submitting a comment.
You can also follow the discussion by subscribing to the Comments RSS Feed.