The 10T cog revolution might be delayed for a bit....

[yinzerniner]

.... that is if you value ultimate drivetrain efficiency....

https://www.velonews.com/2019/05/bikes-and-tech/gear-issue-friction-differences-between-1x-and-2x-drivetrains_493185

But a couple of thoughts on the test -
-Why use different chains? It just adds another variable that could have easily been controlled and thus made the comparison even more valuable
-On that point, why not include a standard SRAM 53/39 setup as well? It's well known that Shimano's chainring design is superior when it comes to FD shifting, it could also be possible that the chainring differences also contribute some to the efficiency gaps.
-They set the chainline of the 53T Chainring and 48T chainring at the same point, but SRAM 1x Chainrings are supposed to be centered between the two chainrings. This probably would have provided even worse results for the 1x system, but again it introduces another variable that could have easily been removed if they had made the setup to manufacturer's specifications.
-There's no indication of what the virtual chainstay or actual chain length was for the test setup. Since chain articulation and chainline are the biggest factor when it comes to friction losses it would make sense that they would state these numbers

That being said, a 6% difference at 250W is very noticeable on the bike. So for anyone who's a serious roadie there seems no reason to use 1x over existing 2x - in other words, what's been said again and again and again.

[Blown Reek]

Yes, but it's 12-speed which is better than 11-speed.

[bicycletricycle]

small cogs have always been crap and nothing is going to change that.

[bikinchris]

small cogs have always been crap and nothing is going to change that.

This ^^^^ 11 tooth cogs were a pretty bad idea, 10 is worse.

[shoota]

This ^^^^ 11 tooth cogs were a pretty bad idea, 10 is worse.

But what about 9T cogs?

[parris]

I've gone to smaller chain rings in response to the push to smaller sprockets out back AND my lower fitness these days.

I don't see the use for me in the 11 tooth sprockets that are increasingly common. More offerings in the 12 and 13 tooth for a small sprocket with smaller jumps further up the cassette would be welcome.

[marciero]

Overall the effects of articulation and speed seem to be confounding much of the results, though these are cited and discussed.

Hard to make sense of statements like "2x has lower friction losses in every gear" and "the outer ring of the 2x is more efficient in any cog than the 1x is in any cog. Their graph does not show this. It seems to show given ratios at the same speeds, though I have difficulty reading it. Does the text make clear that they compared every gear of 1x to every gear of the 2x, at the same speeds, or even did this limited to the best chainlines for the 2x?

This is all n=1 here, except for the averages across all the 11 and 15 gears, which is not too meaningful (nor statistically significant, I would bet, given the high variance and low n)

On the other hand, it is interesting that the 2x has less friction in the "ideal" situation -48 X 18-tooth vs 53 X 19-tooth- assuming same speed (which was not clear) , esp. since the 18t and 48t would have more friction, resp, than the 53X 19, due to articulation, and additionally, the higher cadence of the 48t has more friction than the 53t. On the other hand the lower rotational speed of the 18t would have lower friction than the 19t. So still some confounding here.

[soulspinner]

Yes, but it's 12-speed which is better than 11-speed.

:p

[stien]

I can see it now....13 speed 53/36 and 13-33 cassette.

I'd probably buy it.

[bicycletricycle]

Parsing the data that fine really isn’t required (although always interesting)

Smaller cogs = more friction
Bad chain line = more friction

Single ring set ups are currently increasing both of these factors.


Overall the effects of articulation and speed seem to be confounding much of the results, though these are cited and discussed.

Hard to make sense of statements like "2x has lower friction losses in every gear" and "the outer ring of the 2x is more efficient in any cog than the 1x is in any cog. Their graph does not show this. It seems to show given ratios at the same speeds, though I have difficulty reading it. Does the text make clear that they compared every gear of 1x to every gear of the 2x, at the same speeds, or even did this limited to the best chainlines for the 2x?

This is all n=1 here, except for the averages across all the 11 and 15 gears, which is not too meaningful (nor statistically significant, I would bet, given the high variance and low n)

On the other hand, it is interesting that the 2x has less friction in the "ideal" situation -48 X 18-tooth vs 53 X 19-tooth- assuming same speed (which was not clear) , esp. since the 18t and 48t would have more friction, resp, than the 53X 19, due to articulation, and additionally, the higher cadence of the 48t has more friction than the 53t. On the other hand the lower rotational speed of the 18t would have lower friction than the 19t. So still some confounding here.

[redir]

My mountain bike is a 3x drive train... Woohoo! I'm even more efficient!

[unterhausen]

this isn't going to stop anyone. For the longest time, racers used single speed because of efficiency. Tourists were using derailleurs long before.

I don't like how the chain sounds in the 11 tooth cog. I'm sure the 10 tooth is even more grating (literally).

[fiamme red]

But what about 9T cogs?https://www.competitivecyclist.com/ethirteen-components-trs-race-11-speed-cassette

https://content.competitivecyclist.com/images/items/900/ETR/ETR000M/BK.jpg

https://us.3t.bike/en/accessories/cassette-579.html

https://3tusabike-blulabsrl.netdna-ssl.com/2198-thickbox_default/cassette.jpg

[unterhausen]

I have had occasion to wonder if the long-standing objection to cogs with less than 14 teeth would be lessened if the tooth shape was changed. That 9 tooth looks horribly worn right out of the package. My guess is they need such radical shapes to make it shift properly.

[benb]

A lot of this is common sense.

I feel like you can feel it riding 2X... cross chaining feels subjectively worse than non-cross chaining.

Keeping it in the big-big combos feels subjectively better than riding in the equivalent ratio small-small combo. I have long biased myself to the big-big combos and almost always avoided the small ring with the smallest 2-3 cogs.

I feel like some of the "feel" might be hearing as well... the drivetrain makes more noise in the less efficient selections.

The weird thing about 11 vs 10 is a lot of the time you don't need the 11 either! I've spent large portions of my riding life including racing with a 53x12 as my high gear. It definitely never cost me anything.. I was never a good sprinter but having an 11 or a 10 cog would not have made the difference in a million years. Being better at positioning myself and being fearless about getting caught up in a sprint crash would have been 1000x more important. (I was always happy to just help out with a train)

[arik]

this is a bigger deal if you spend a lot of time in your lowest gears, i.e. the 9, 10, or 11 tooth cogs, but I think few people spend a lot of time in these gears, and when you are in these gears, you usually have gravity to make up for the loss of efficiency. If you are on a steep descent and you can't keep up no matter how fast you spin, then your efficiency is 0%! Better to have a 9 tooth cog back there then nothing in that case.

[Mark McM]

this isn't going to stop anyone. For the longest time, racers used single speed because of efficiency. Tourists were using derailleurs long before.

I think you mean racers used single speed because they thought it was more efficient. Or more commonly, because derailleurs weren't allowed. The Tour de France famously didn't allow racers to use derailleurs until 1937, because founder Henri Desgrange purposely wanted the race to be a sufferfest. The rumor is that the Tour finally allowed derailleurs after it became embarrassing that the spectators were riding up the climbs (with derailleurs) faster than the racers.

[bigbill]

My new MTB will have an 11-46 eleven speed with a single 32 chainring.
For road, most of my bikes are campy 11 with standard cranks and 12-27. On my ten speed campy bike, it has a 13-26 which in my opinion is the most useful ten speed cassette ever for the road. My single shimano road bike has an 11-28 with a 36/52. The 11 is untouched.

[benb]

Singlespeed is probably most efficient if it's the exact right gear.

But being in the wrong gear that is vastly worse than any efficiency loss in the different drivetrains.. and a single speed is in the wrong gear almost all the time!

[Jaybee]

this is a bigger deal if you spend a lot of time in your lowest gears, i.e. the 9, 10, or 11 tooth cogs, but I think few people spend a lot of time in these gears, and when you are in these gears, you usually have gravity to make up for the loss of efficiency. If you are on a steep descent and you can't keep up no matter how fast you spin, then your efficiency is 0%! Better to have a 9 tooth cog back there then nothing in that case.

Agree with this. I've used the 10t on my mtb (which only has a 32t front ring) maybe twice or three times on paved road transits between trail systems. It was nice to have it then, but I think I'd take an 11 or 12 small cog in exchange for another gear in the upper half of the cassette - there are some long grindy climbs here where it's nice to find a cadence you really like.

I guess SRAM wouldn't be able to print "500%" on the cassette then, though.

[Mark McM]

this is a bigger deal if you spend a lot of time in your lowest gears, i.e. the 9, 10, or 11 tooth cogs, but I think few people spend a lot of time in these gears, and when you are in these gears, you usually have gravity to make up for the loss of efficiency. If you are on a steep descent and you can't keep up no matter how fast you spin, then your efficiency is 0%! Better to have a 9 tooth cog back there then nothing in that case.

At the highest speeds (steepest slopes), you can often go faster by pulling into a tighter tuck (too tight to pedal), than you can by continuing to pedal. Obviously, this speed depends on several factors, but with a 53/12 gear, I don't "spin out" until well over 40 mph. And even at those speeds, I can pedal in short bursts at even higher cadences if I need to do a quick acceleration. I rarely get left behind on a descent, and when I do, it was not due to a lack of 9 or 10 (or even 11) tooth sprocket.

[bikinchris]

But what about 9T cogs?

BIKE Friday has an 8 tooth setup for their folding bikes with tiny chainrings.

[marciero]

Parsing the data that fine really isn’t required (although always interesting)

Smaller cogs = more friction
Bad chain line = more friction

Single ring set ups are currently increasing both of these factors.

True enough. But does not explain the claim about the comparison between the "ideal" situations, where the 2x has both smaller cog and ring, or the two "better in all the gears" claims, or the which presumably includes the ideal chain line cases for each.

[Mark McM]

True enough. But does not explain the claim about the comparison between the "ideal" situations, where the 2x has both smaller cog and ring, or the two "better in all the gears" claims, or the which presumably includes the ideal chain line cases for each.

Hmmm.... Good point. Looking more closely at the data in the Velonews article, something seems fishy. The middle (6th) sprocket SRAM 10-42 cassette and Shimano 11-34 cassette in the test are the same size (21 teeth). For the 1x drivetrain, the chainline in the middle sprocket should be perfectly straight, and for the 2x drivetrain there should be minimal chainline offset in either chainring in the middle sprocket. You'd expect that the drivetrain losses in the middle sprocket with the 1x (48x21 gear) to be a little lower than the 2x in the small chainring and middle sprocket (39x21 gear), due to the larger chainring and less cross-chaining. And yet, the test shows that the 1x 48x21 has larger losses than the 2x 39x21. There is definitely something fishy going on with the test.

This isn't the first fishy test that has come out of the Ceramic Speed labs recently. In this on whether clutch derailleurs increase drivetrain friction (https://cyclingtips.com/2019/01/do-clutch-rear-derailleurs-add-drivetrain-friction/) from earlier this year, chain losses were compared between a standard and clutch derailleurs. The bottom run of the chain oscillated up and down to simulate riding over rough surfaces. The data showed that as the clutch derailleur cage extended and retracted, the clutch friction increased drivetrain losses. Sadly, this was not a good simulation of reality. The whole point of a clutch derailleur is limit chain oscillation. By artificially forcing the chain to oscillate the same amount with both types of derailleurs, it biased the test against the clutch derailleurs.

Personally, I'm not sold on either 1x or clutch derailleurs for road riding. But neither of these tests provides any meaningful data against them.

[yinzerniner]

Hmmm.... Good point. Looking more closely at the data in the Velonews article, something seems fishy. The middle (6th) sprocket SRAM 10-42 cassette and Shimano 11-34 cassette in the test are the same size (21 teeth). For the 1x drivetrain, the chainline in the middle sprocket should be perfectly straight, and for the 2x drivetrain there should be minimal chainline offset in either chainring in the middle sprocket. You'd expect that the drivetrain losses in the middle sprocket with the 1x (48x21 gear) to be a little lower than the 2x in the small chainring and middle sprocket (39x21 gear), due to the larger chainring and less cross-chaining. And yet, the test shows that the 1x 48x21 has larger losses than the 2x 39x21. There is definitely something fishy going on with the test.


As I indicated earlier they seem to have misaligned the 1x drivetrain 48t chainring since they centered it at the same sprocket as the 53t chainring. But your point is well taken, although a possible determining factor might be the amount of teeth on the pulley wheels having a sort of "smoothing" effect with regards to frictional losses at the largest cassette cogs. All three setups show minimal percentage changes at the largest cassette cogs and then ramp up once it hits about midpoint. In other words, the pulley wheels have an effect at the cassette while chainrings don't have such an outside influence.

But it's tough to fully dissect the numbers because of the other uncontrolled variables. Until more information is divulged from Friction Facts themselves the outcome and conclusions are opaque.

[benb]

Isn't a major confounding variable in the Velonews test that the two drivetrains tested are different in more ways than # of chainrings?

They have different rear derailleurs with likely different cage geometry & pulley geometry. The 1X is probably a clutch derailleur and the 2X is probably not. The 1X might have a Wide-Narrow-Wide ring & the 2X does not. Shimano is known for really good rings, could be a slight difference there.

Are they taking out variance in BB bearings? What about differences in the chains?

Any of these things seem like they could explain a variance in the two systems when they're in the same gear ratio and both should have a straight chainline.

Again we have this great technical discussion though and all this is probably over fractions of a watt and 5W in the worst case. This stuff is only relevant in racing and only when all other way more important variables are dialed in perfectly. Fitness, rider weight, bike fit, helmet aero, clothing aero, bike aero, wheel aero, etc..

[Mark McM]

Isn't a major confounding variable in the Velonews test that the two drivetrains tested are different in more ways than # of chainrings?

They have different rear derailleurs with likely different cage geometry & pulley geometry. The 1X is probably a clutch derailleur and the 2X is probably not. The 1X might have a Wide-Narrow-Wide ring & the 2X does not. Shimano is known for really good rings, could be a slight difference there.

Are they taking out variance in BB bearings? What about differences in the chains?

Any of these things seem like they could explain a variance in the two systems when they're in the same gear ratio and both should have a straight chainline.

Again we have this great technical discussion though and all this is probably over fractions of a watt and 5W in the worst case. This stuff is only relevant in racing and only when all other way more important variables are dialed in perfectly. Fitness, rider weight, bike fit, helmet aero, clothing aero, bike aero, wheel aero, etc..

That pretty much sums it up. All this test shows is that one particular 1x drivetrain setup one particular way (and used in one particular way under one particular environment) tends to have more losses than a 2x drivetrain setup on its one particular way. We really can't extend the conclusions much beyond that, so we can't make a general statement comparing 1x and 2x systems just from this one test (despite the Velonews article's attempt to do so).

For reasons of a combination of gearing range and size differential between gears, I'll be sticking with my 2x drivetrain on the road, but others may have other reasons for selecting a 1x drivetrain.