Hub gear versus derailleur and drops versus flats

[martinf]

This started in "What's your Rohloff combo (chainring, cog)?"
http://www.thorncycles.co.uk/forums/index.php?topic=4412.msg20936#msg20936

but drifted way off topic, so I am continuing here.

I'm currently testing various things with my current bikes before ordering a new Thorn. A by-product of this is that I have got some timed test-rides with the same bike going round the same circuit in various configurations.

The circuit is slightly less than 50 kms, with several hills but no really steep ones.

Now done 3 sets of 5 rides each in 3 different configurations.

Config. 1 - drop bars and 3x7 derailleur gears, range 18" to 87". Av. speed 24.26 kph

Config. 2 - flat bars + bar ends and Nexus Premium 8 hub, range 25" to 76". Av. speed 23.61 kph

Config. 3 - drop bars and Nexus Premium 8 hub, range 25" to 76". Av. speed 24.05 kph

For the 15 rides, min speed 22.25, max speed 24.53

Conclusions:

(A) With the same bars and shifter position (bar end) I was slightly less than 1% faster with derailleurs, compared to a Nexus 8 Premium hub gear. This is a much smaller difference than I was expecting. I had a Nexus 7 a few years back and it was noticeably less efficient than my old 5-speed hubs (didn't do a Nexus 7/derailleur comparison).

With just over 1,000 kms of use so far, I am very pleased with the Nexus 8 Premium. It doesn't have the range or reputation for reliability of the Rohloff, but the range is enough for me for local use. And as the Nexus is much cheaper I won't be too bothered about leaving the bike locked up around town.

(B) Going from drops to flats made slightly less than 2% difference in speed. Also less than I was expecting. In this case I reckon I could probably reduce the difference substantially by training myself to use the bar-ends more effectively.

I have other issues with flat bars, but now I no longer consider that the speed reduction over drops is important for touring type rides.

[Danneaux]

Hi Martin!

I've been following your speed trials with greatest interest, and you've surely done a fine job of back-to-back trialing, holding factors as constant as possible.  Well done!  You've done really well, and have been very good to share the results with us.

Not intended in any way as a criticism of your work; just a general observation: It surely is hard to conduct such trials with really rigorous methodology. That is a problem that has bedeviled trainers and kinesiologists since sport began -- how can one possibly hold constant for the human factor? Yes, erg meters and strain gauges on hubs and cranks help, as does indoors testing to compensate for wind and slope and such, but mood can play a factor as well. Back in my uni days, I considered my 8.9km morning commute (76-100km coming home so I would get a good day ride in) a time trial of sorts, and I loved seeing if I could deliberately knock even a few seconds off it (side-drafting city buses helped a lot!). Oddly enough (sans bus-drafting), on exam days I was fastest; that's when my mind was occupied, and I didn't pay much attention to form and such and had the lowest point-to-point times. So, that said, I think you've done especially well on your own comparisons. Again, well done!

In some ways, I'm not to surprised at the results.

Though all my bikes have had drop handlebars and derailleur gearing and my position is the same on each, they vary greatly in weight and specific configuration. Some "feel" much faster/more sprightly than others. Regardless, I find my average speed to be nearly identical on all, given similar weather/wind conditions and terrain. My variance in speed is very similar to yours. It is only when I deliberately set out on a "speed run" that any difference emerges, and that could be due to variance in my efforts between bikes for any number of reasons.

Jags and I have talked about this in our off-Forum correspondence and have concluded we each ride about the same speed regardless of which bike we're on. The real difference is in perceived feel and comfort. After riding a loaded touring bike, a sport-tourer or racing bike feels like a rocketship. I just "know" my rando bike is faster than Sherpa...until I look at my times, which are nearly identical. 'Bar types aren't a factor for me; I still prefer drop handlebars for how my wrists, elbows, and shoulders are configured (plus they're familiar to me after decades of use), and I feel like I'm making more headway when I go onto the drops into a headwind and "ride inside my elbows" -- even if it doesn't make much difference at my usual speeds.

Martin, in looking at your data, I do notice your highest speeds were with drops handlebars. While the difference is miniscule for the tourist at touring speeds, it would surely be significant in a timed trial or perhaps at racing speeds on a road course. I fully agree, using the bar-ends on straight handlebars most effectively would effectively eliminate any practical difference between them and drops. You're moving right along at ~24kph/15mph, but don't forget drag increases as the square of velocity -- the faster one goes, proportionately more energy is required to do so -- a lot more. The differences you've found at what are generally regarded to be pretty good touring speeds would be much greater at sustained racing speeds. For example, the long (around 200km) relatively flat early stages of the Tour de France are designed specifically for sprinters and are ridden at remarkably fast speeds often averaging ~45kph/28mph.  That's nearly twice your test average and is really humming.

Extrapolating further, I had always expected to really feel a difference when riding with panniers and a handlebar bag -- I thought it would be huge, given they result in a drag coefficient approximating that of a barn door flat-face to the wind (Cd = 1.0 in that case). Yet...I must be pretty insensitive to it or somehow adjust unconsciously, because my speed on the flats doesn't drop as much as I would expect compared to riding unladen. Similarly, when I remove my heavy panniers, it feels like I have jet-assist on the unladen bike...yet my speeds remain pretty much the same.

When you wrote...

I have other issues with flat bars, but now I no longer consider that the speed reduction over drops is important for touring type rides.

...I agree completely!

Again, very nice work, Martin!

All the best,

Dan.

[martinf]

Not intended in any way as a criticism of your work; just a general observation: It surely is hard to conduct such trials with really rigorous methodology. That is a problem that has bedeviled trainers and kinesiologists since sport began -- how can one possibly hold constant for the human factor? Yes, erg meters and strain gauges on hubs and cranks help, as does indoors testing to compensate for wind and slope and such, but mood can play a factor as well.

Yes. 5 runs is too small to be statistically significant. But better than just 1 run. And doing similar comparisons in the past I have had much bigger differences in some cases, examples are:

- New Nexus 7 about 15% slower than Sturmey 5-speed. Nexus 7 improved to about 8% slower after cleaning out the original grease and putting new grease in.
- Drop handlebars about 11% faster than flats without bar ends on a Moulton.
- Kojak tyres about 12% faster than Marathon tyres on a Brompton.  

and the obvious one:

- Knobbly off-road tyres nearly 19% slower than slicks on the same bike.

In some ways, I'm not to surprised at the results.

Though all my bikes have had drop handlebars and derailleur gearing and my position is the same on each, they vary greatly in weight and specific configuration. Some "feel" much faster/more sprightly than others. Regardless, I find my average speed to be nearly identical on all, given similar weather/wind conditions and terrain. My variance in speed is very similar to yours. It is only when I deliberately set out on a "speed run" that any difference emerges, and that could be due to variance in my efforts between bikes for any number of reasons.

Jags and I have talked about this in our off-Forum correspondence and have concluded we each ride about the same speed regardless of which bike we're on. The real difference is in perceived feel and comfort. After riding a loaded touring bike, a sport-tourer or racing bike feels like a rocketship. I just"know" my rando bike is faster than Sherpa...until I look at my times, which are nearly identical. 'Bar types aren't a factor for me; I still prefer drop handlebars for how my wrists, elbows, and shoulders are configured (plus they're familiar to me after decades of use), and I feel like I'm making more headway when I go onto the drops into a headwind and "ride inside my elbows" -- even if it doesn't make much difference at my usual speeds.

I do find differences between bikes. I've done the same circuit 5 times with my lightweight 700C drop bar derailleur bike, average speed 25.67, that's nearly 6% faster than the heavy tourer with drops and derailleurs. There's about 5 kg of weight difference, which should have a small effect going up hills. The 700C bike is a bit more aerodynamic  - British style saddlebag tucked behind the rider, whereas the heavy tourer has front and rear racks with panniers on the back. And the tyres might be more free-rolling - 700C x 28 Michelin Dynamic at 70/75 psi compared to 26" x 1.9 Supremes at 35/45 psi, although I believe the Supremes have pretty low rolling resistance.

Martin, in looking at your data, I do notice your highest speeds were with drops handlebars. While the difference is miniscule for the tourist at touring speeds, it would surely be significant in a timed trial or perhaps at racing speeds on a road course. I fully agree, using the bar-ends on straight handlebars most effectively would effectively eliminate any practical difference between them and drops. You're moving right along at ~24kph/15mph, but don't forget drag increases as the square of velocity -- the faster one goes, proportionately more energy is required to do so -- a lot more. The differences you've found at what are generally regarded to be pretty good touring speeds would be much greater at sustained racing speeds. For example, the long (around 200km) relatively flat early stages of the Tour de France are designed specifically for sprinters and are ridden at remarkably fast speeds often averaging ~45kph/28mph.  That's nearly twice your test average and is really humming.

Yes. I expected more of a difference between flat bars and drops. But bar-ends on flats can approach drops, at least for normal touring speeds.

Touring all day I am slower than just going round the test circuit for about 2 hours, for my Spain trip my lowest daily average speed was 14 kph in the Picos d'Europa - mountains and lots of interesting things to look at - and highest 22.4 kph on a 188 km ride on flat roads and sandy but flat cycle tracks near the coast in south-west France, with rather boring scenery and rain.

Extrapolating further, I had always expected to really feel a difference when riding with panniers and a handlebar bag -- I thought it would be huge, given they result in a drag coefficient approximating that of a barn door flat-face to the wind (Cd = 1.0 in that case). Yet...I must be pretty insensitive to it or somehow adjust unconsciously, because my speed on the flats doesn't drop as much as I would expect compared to riding unladen. Similarly, when I remove my heavy panniers, it feels like I have jet-assist on the unladen bike...yet my speeds remain pretty much the same.

Last year I went round my test circuit twice with the load I took for my Spain trip. I was about 6% slower than with my normal load (toolkit and raingear). I think the difference was mainly due to the uphill bits - I was consistently using much lower gears. I don't think having full panniers instead of empty ones made much difference to the aerodynamics. Anecdotally, on flat roads I go nearly the same speed loaded as empty.