19

I want lower (sub 1:1) gearing on my next drivetrain but also a decent range, reasonable spacing between gears and still be reasonably light.

While looking at different combinations of components, I noticed that:

  • 1x options were not nearly as light as I expected, mostly due to the huge cassettes needed for decent range.
  • On 2x, large cassettes were killing my weight compared to "standard" road setups (50/34 compact double with 11-28 cassette).
  • Shimano GRX setups gain weight in all the components, partially negating their gearing benefits.

At this point, for a lark, I threw in some randonneur type cranks and the old heavy bottoms brackets that go with them, and noticed the numbers coming up not that much worse than modern drivetrains.

Furthermore, I found that some rando cranksets can be made to come with super small chainrings. One manufacturer I looked at went down to 42/26(!), and pairing this with an Ultegra 11-28 met my criteria and was the lowest weight combination. I'd love to buy an Ultegra 42/26! Just imagine what that would weigh!

So, why don't more people use smaller chainrings? Or the flip side to this question: why don't drivetrain manufacturers offer them?

Is this just the standard reason that "the bike industry is driven by racers"? Or am I missing some obvious physical thing that makes smaller chainrings inferior? Should I just give up and become a retrogrouch?

This answer seems claims that smaller cogs have more friction, but it doesn't seem like it would be a lot more, considering we have "small" cogs on most of our cassettes.

Ryan Witt
  • 193
  • 1
  • 2
  • 7
  • 12
    large cassettes were killing my weight Huh?!?! Are you really concerned with the weight difference between something like an 11-28 and an 11-34 cassette? Also, at 80 RPM you're only going to get 24 mph/38 kph out of a 42-tooth chainring on an 11-tooth cog. If you do any competitive group rides, you're going to get dropped on any long, slight descent - say 2-4% - unless you can sustain pedaling at 110 or 120 rpm or so. – Andrew Henle Dec 02 '19 at 11:32
  • 1
    Comparing Shimano 11-28 with 11-40 at comparable groupsets (Ultegra vs XT) I find a difference of 130 g. (CS 6700 230 g, CS M 8000 400 g). That is a significant difference, but still much less than shifter, front derailleur, double crank spider. Also note that 42:11 might be a little light for group rides in flat areas. I know hobbyists in Denmark who average 40 km/h. However, these are exceptions, as most cyclists are now. Hence the many gearing options. Ignore weight differences and get your personal optimum. – gschenk Dec 02 '19 at 11:50
  • 6
    @AndrewHenle Do you work for shimano? ;-) 90%+ of riders don't ride fast group rides, and if they hit 24mph they are delighted and stop pedalling. As it stands, available gearing is fine for the minority with ftp>3.5W/Kg, but far too big for the majority of riders who have ftp<2.0W/Kg – Andy P Dec 02 '19 at 13:31
  • 4
    @AndyP 24 mph on a 3% false-flat descent is not fast at all, especially for any type of group ride. Why do you think I picked such a low RPM - and yes, 80 is pretty low - for my example of how limiting a 42t chainring can be? Imagine how limiting a 42t chainring is for someone who struggles to maintain 60 rpm... – Andrew Henle Dec 02 '19 at 13:39
  • 4
    The answer you referenced in the last paragraph is correct. Smaller chainrings, cassette cogs, and even derailleur pulley wheels make the chain articulate at sharper angles. This does increase drivetrain friction. It's been shown that a 1x drivetrain has about 3W higher drivetrain friction than a 2x setup at 250W because the former has greater bends than the latter - you're either on a comparatively smaller cog or a comparatively smaller chainring. You're correct that the increased friction is of limited practical consequence - most riders might not even notice a 3W difference. – Weiwen Ng Dec 02 '19 at 14:38
  • 8
    I would second the recommendation not to worry too much about small amounts of weight. If you're into competitive riding, then you should already know that optimizing aerodynamics matters a lot more than saving weight. Now, I'm not talking about 5kg/10lb differences in weight. However, the weight differences between GRX and the equivalent road group or between 1x and 2x are not consequential for most riders. – Weiwen Ng Dec 02 '19 at 15:35
  • 2
    The trend is going towards smaller chain rings for nearly two decades. A 110 mm bolt circle diametre (BCD) crank (eg Ultegra or GRX) allows to mount chain rings with merely 33 teeth. GRX increased the maximum difference between large and small chainring to 16. One can ship 42/26 with those cranks. (Shimano default 46/30). We are certainly past the days of ubiquitous 53/39. – gschenk Dec 02 '19 at 16:53
  • 2
    @WeiwenNg not only are the angles sharper; smaller chainrings also require higher tension of the chain at the same torque. So the frictional loss actually increases quadratically as you make the chainrings smaller! – leftaroundabout Dec 02 '19 at 17:31
  • 1
    @WeiwenNg on second thought: actually it's probably not quadratic. I think it is only the higher tension that causes the higher friction. The sharper angle doesn't matter, because it is compensated by fewer chainlinks getting bent each second. IOW, the bending rate of the chainlink that's passing onto the chainring at a given instant only depends on the cadence, not on the size of the chainring. Therefore, it is only the higher tension that increases the friction. – leftaroundabout Dec 02 '19 at 18:04
  • 1
    I’m getting from these comments that maybe there are not a lot of riders that care about weight yet are not also into “high performance”. That could explain why this market is underserved. But I think it is growing. “Gravel” seems to be catching some of these riders and bubbling up their considerations to industry. – Ryan Witt Dec 02 '19 at 18:52
  • I wanted to add that I’m leaning toward a configuration where I’m doing 26-27 MPH @ 90 rpm, which feels plenty fast for the longer distances I plan to ride on this bike. I can always spin at 120 for a bit to catch someone. I am definitely betting that lower climbing gears and smaller steps to adjust to cadence will make me happier than downhill gears. – Ryan Witt Dec 02 '19 at 18:58
  • 1
    @WeiwenNg And if one is concerned about weight the typical place to start reducing is one's body. – Peter - Reinstate Monica Dec 03 '19 at 06:48
  • 1
    @AndyP the discussion is about road bike groupsets for road racing bikes. While I'd agree probably most road bike riders don't RACE, a lot of them do go on fast group rides. Not so fast that a compact 50 or indeed even 48 would be terminal, it is true that amateur riders don't need pro gearing. But there is a limit, and spinning out at 40km/h would be a problem for many amateur road riders. While non road bikes like hybrids and mountain bikes already have much lower front gearing anyway. – Ivan McA Dec 03 '19 at 08:34

4 Answers4

14

Gear ratio range.

If you decrease the chainring sizes you decrease the highest ratios available. It's not possible to make the gap between the chainrings much bigger and get decent front shifting so the large ring has to shrink with the small one.

It's easier for manufacturers to make a wide ratio cassette that retains an 11 tooth sprocket and shifts well.

Argenti Apparatus
  • 75,665
  • 4
  • 86
  • 188
  • I had not thought about the difficulty in getting the shifting “right”. That’s a great point. I had good performance on a Velo Orange 46/30 rando crankset, but it was only 8 speed. I can imagine this gets harder to feel perfect at 10/11 speed. Thanks! – Ryan Witt Dec 02 '19 at 18:34
  • 2
    @RyanWitt Another issue is the jump between ratios on the large and small rings. If the jump is very large riders will have to also change several gears at the back whenever they change gear at the front. – Argenti Apparatus Dec 02 '19 at 18:37
  • @RyanWitt Also note that this was different when triple rings were standard if you wanted a wide range drivetrain. Triple rings were paired with narrow ratio cassettes (to keep the total capacity required of the rear derailleur reasonable), but the range of the drivetrain came from the difference between smallest and largest rings not the range of the cassette. – Argenti Apparatus Dec 02 '19 at 18:41
  • I’ve noticed this, and it’s kinda a trade-off, right? It’s nice to only have to shift the rear by 1 or 2 on my 50/34 11-30 when shifting in front, but the overlap in gears is so huge it like I only really have 13 speeds. I think I’ll try out this huge FC jump and see what the downside is like. – Ryan Witt Dec 02 '19 at 18:42
  • @RyanWitt Some folks try a bigger jump in front rings 52-34 is common. Beware of your rear derailleur total capacity though. – Argenti Apparatus Dec 02 '19 at 19:30
  • @RyanWitt 46/30 = 16T drop. That's the exact same as a standard 50/34 compact crank, and illustrates perfectly his point of you have to shrink both. There is one GRX crank at 48/31 which is a 17T difference, that's the largest gap Shimano do officially. I know people who run 52/34, adding a 34 to a semicompact, and say that works OK. I'm sure it does, but "OK" is somewhat subjective and for sure front shifting gets worse the bigger the gap, compact has got pretty good but a standard 53/39 still shifts better, and with less jarring cadence difference. I have a triple 24-39-50 which shifts "OK". – Ivan McA Dec 03 '19 at 07:58
  • @ArgentiApparatus these days touring setups can have a fairly wide range at the back and a triple. I've got 30/39/50 and 11-32; I could change that to 11-34. Not a solution for the OP, but on 16kg of steel in hilly terrain – Chris H Dec 03 '19 at 14:12
  • @IvanMcA if you don't mind me asking, what is that wide-range triple? I wouldn't want to go that low but not far off, and I'm intrigued – Chris H Dec 03 '19 at 14:14
  • @ChrisH - it's a standard Shimano 105 FC-5703 road triple with the 30T granny ring replaced with a 24T. Interestingly if you want to run a really large difference on the front key is not the front derailleur but the REAR derailleur as it is the rear that takes up the chain slack. I have used 12-28 and 12-30 at the back. This used be really terrible with a road RD but since I replaced with a Deore M591 MTB it has worked pretty good. But it's nowhere near my double road bikes for shift quality. I am running 10sp- 9sp Shimano MTB RD is 100% compatible with 10sp road brifters. – Ivan McA Dec 03 '19 at 15:07
  • 1
    This is on my touring bike and has evolved out of what it originally had on it, trying to get lower gears. Starting from scratch today, I wouldn't do it this way, I would go for 11sp with a subcompact double at the front (maybe 30/46) and a slightly wider range cassette. 11-34 or 11-36 with a 30 would get me around the same place. I really don't need a 50 on my tourer, but I couldn't go any smaller or the FD would hit the 39 middle.

    OR would look at what works in terms of gravel or MTB cranks with road brifters, I believe with SRAM you can mix road/MTB which you can't on Shimano since 11sp.

     – Ivan McA Dec 03 '19 at 15:15
  • @IvanMcA 9sp Shimano MTB RD is 100% compatible with 10sp road brifters. It "works", but in my experience the shifting quality is pretty poor compared to a recent recent Ultegra or even 105 setup. FWIW, the best 10-speed shifting IMO is from old Dura Ace RD7800s. – Andrew Henle Dec 03 '19 at 16:34
  • @AndrewHenle in my experience shifting with the 9sp Deore RD was substantially better than it had been with Ultegra 6600 or 105 5600 10sp RDs and the 24T at the front. With the road RDs, the chain kept falling off, and I needed a catcher. I think key here is if you are running a MTB RD with brifters you are doing something way out of spec for a road group at that point, it's a hack. You certainly can't compare to a modern 11sp road double, or a 10sp road double, or even a standard 10sp road triple, all run in spec... of course shifting is worse. But the MTB RD is better with that capacity. – Ivan McA Dec 03 '19 at 16:48
  • @Ivan, thanks, I'm starting from 9-speed Sora everywhere with a long cage RD (not noticeably different from the MTB 9-speed RD on my hardtail, which I put on the tourer for a while). That sounds like when I replace the chainrings anyway I could consider a 28. – Chris H Dec 03 '19 at 17:28
  • 1
    @ChrisH if it is a Sora FC-3403 or FC-3503 crank that's a 74mm granny and you can go all the way down to 24T. Then it is down to capacity. RD-3400-GS = 37, RD-3500-GS = 41. RD-M591-SGS capacity is 45, so it gives you a bit more room. Capacity on mine, 50-24=26 + 30-12=18 = 44. So I need the MTB RD. 50-28=22 so with the 3500GS you'd have 19 left for the cassette, so an 11-30 would work fine with the 3500-GS. If you wanted more than that, either front or back, you'd be better with a MTB (SGS) RD. – Ivan McA Dec 04 '19 at 06:20
13

There is a limit to the amount of tension a given chain should be put under. Smaller chainrings increase that force - the pedal arm and chainring form a lever, and the smaller the chainring (and longer the pedal arm) the more force will be applied to the chain given a fixed force on the pedal. What you might be gaining in terms of chainring and cassette size and weight would come as a tradeoff in wear and tear on the chain and decreased reliability.

For instance, most bicycle chains appear to support 1,000kg of static load before breaking. This is new, and not at the up to 5 degree angle the chain may be between the cassette and chainrings, so you should apply less force than this, but it gives an upper limit.

If you take a 150Lb rider, whose legs can push down and pull up with a force of over 200Lb each (downward will be greater than upward, but many people can bench press over 400Lb so I suspect it's not an unreasonable estimate), and put them on clipless pedals (so they can pull up) on a 170mm crank, with a 26 tooth chainring, then not moving but simply holding still, applying the force, they can put over 1,000Lb of force on the chain.

That's just under half the 1,000kg limit of the chain. Now if you place that rider on a bumpy trail, standing, sprinting, then you may be nearing or exceeding it with dynamic forces.

If they have a gearing that puts the chain at an extreme angle, and the chain is worn, it's going to fail sooner or later as it continues to wear under the extreme stresses. If you try and change gears while under this much pressure you may break even a new chain.

I don't think it's going to be a major consideration, but the above hypothetical rider isn't an elite athlete - most regular cyclists that use clipless pedals could produce exceptional stress on the chain with too small of a chainring if they really try.

But beyond that, it's not so much a question of failure but of wear and tear. The higher the force you place on the chain, the more quickly it stretches and distorts, and the more quickly the chainrings and cogs wear.

It's also one big reason to prefer a high cadence to a strong pedal.

If you choose to go this route, simply be aware of the additional maintenance that the additional stress causes, and the decrease in reliability under strain that could lead to a failure.

Adam Davis
  • 612
  • 3
  • 8
  • 1
    That totally makes sense! I hadn’t thought of the chainring as a lever in that way. This is a good consideration. – Ryan Witt Dec 02 '19 at 22:23
  • 2
    While it's true that smaller chainrings mean more tension on the chain, chains can handle much more tension than you can put on them while riding. You cannot break a chain by using a small chainring. However, the added tension and the reduced chainring/sprocket size does significantly increase the wear of the chainrings/sprockets. – cmaster - reinstate monica Dec 03 '19 at 07:21
  • 1
    If this were a problem, then mtb chainsets (such as 36/22), or road triples would never have existed – Andy P Dec 03 '19 at 09:10
  • 1
    @cmaster 'You cannot break a chain by using a small chainring' well, not a new one. Lower gear ratios and higher force on the chain must increase the probability the chain snaps as it wears. – Argenti Apparatus Dec 03 '19 at 10:35
  • Wouldn't it be correct to say that in order to put a certain amount of torque on the rear wheel you have to put that amount of pressure on the chain? To stay it differently, at a given gear ratio you will put the same tension on the change for a given amount of force regardless of the sizes of the front and rear sprocket. – Bill K Dec 03 '19 at 17:09
  • @BillK No. Definitely not. Torque is force time lever arm length. To deliver the same torque to the rear wheel with half the sprocket size (= lever length), you need to put twice the force on the lever. You deliver that force with the chain. – cmaster - reinstate monica Dec 03 '19 at 18:30
  • @BillK Or, put another way: Power is force times speed. If you have the same gear ratio, but half the sprocket/chainring size, the chain moves at half the speed. To deliver the same power to the rear wheel, there needs to be twice the force on the chain. No matter which way you look at it physically, you always arrive at the conclusion that the chain's tension is inversely proportional to the sprocket/chainring sizes. – cmaster - reinstate monica Dec 03 '19 at 18:33
  • Force x speed totally makes sense! Thanks, I stand corrected. – Bill K Dec 03 '19 at 19:19
  • @cmaster With regular chainrings it's unlikely that an average rider will cause a failure, but this question asks about unusually small chainrings, which produce an exponentially higher force than average chainrings. Further, it's less about failure than wear and tear, and reliability. I've added an example and further explanation to the answer, I hope it clarifies my assertion. – Adam Davis Dec 05 '19 at 19:02
  • @AdamDavis Nitpick: It's inversely proportional, not exponential. Torque is a simple multiplicative value of lever length and applied force. --- Concerning your example, that's a bit extreme, imho. If you try to sustain such extreme forces, you are gambling with your knees, hard. I'd always use the weight of the rider as a sensible upper limit: It's what any rider can easily reach when starting from a stop, and people who use clipless pedals usually know to prefer a cadence around 90rpm to limit the force they need to use. – cmaster - reinstate monica Dec 05 '19 at 20:57
  • @cmaster Yes, I shouldn't have used exponential. As far as whether the example is extreme, yes, its is. It's intended only to provide an example or way of thinking about smaller chainrings to demonstrate the possible issue. Again... putting that much force on it isn't the point, nor is immediate failure. The fact is that it will put more force on the chain than designed for and is one of many reasons why smaller chainrings aren't common. They do increase wear and tear, and do decrease reliability. If you do this, take it into account in your maintenance plan. – Adam Davis Dec 06 '19 at 20:53
11

You mostly answered your own question: the racing market drives the industry, sometimes to the detriment of the availability of real-world gearing.

A major compounding factor is that there are a lot of hoops a person has to jump through to get smaller rings on their road bike, starting with buying new, weird, mostly old or retro cranks. Making things work well with small rings plus road shifters and front derailleurs is often not straightforward. Notably, STI pairs badly with mountain FDs, and most existent road FDs pair badly with smaller than normal outer rings. Using an extra wide gap between the two rings of a double (to avoid mismatches with the profile of something like a 42t with a road FD) is an approach that can work and many vintage bikes did so, but making it smooth enough by modern standards under indexed shifting seems to be either impossible or something manufacturers are unwilling to pursue.

46/30 is gaining traction on mainstream bikes and is really a very good baseline.

The Herse and White cranks are both very good. If that kind of gearing is what you think will work for you, go for it.

Nathan Knutson
  • 84,914
  • 3
  • 90
  • 214
  • 5
    This reads more like an opinion piece than an actual answer. You can't blame racing for everything. Chainring, cog, and wheel sizes got pretty much determined more than a century ago when bicycles moved from penny-farthings to the safety bike. Hundreds of millions of commuter and utility bikes used close to the same wheel sizes, chainrings, and cogs, and that market dominates in volume by far. – R. Chung Dec 02 '19 at 15:54
  • @RChung sounds like your answer is “industry inertia”? Sounds plausible to me. – Ryan Witt Dec 02 '19 at 18:48
  • 3
    @R.Chung this is completely off base, the ratios we have today are actually quite different and relatively recent. A 5 speed freewheel like was common up to the 1980s typically started at 14T. There was a physical limit to how small cogs could go. 11T only really came in with cassettes in the 1980s, and even up to 9 and even 10sp in the 00s 12T was a very common small cog. It's only really 11sp in the 2010s that 11T became ubiquitous as the default small cog (and now you have 10, with SRAM XD, etc). 14T to 11T is a 27% difference, that's huge. Same as going from 40 to 52 at the front. – Ivan McA Dec 03 '19 at 08:14
  • 2
    @IvanMcA The question was really about low gears, and why chainrings don't have 26 teeth rather than 52, or 21 rather than 42. That smaller cogs have been introduced in the last 30 years doesn't address the original question. – R. Chung Dec 03 '19 at 16:21
3

For road bikes the subcompact cranks have a bolt circle of 110 mm diameter. The smallest chainring that fits is 34 teeth. There are cranks with smaller bolt circles and small chain rings available, but the selection is limited. Your proposed 42/26 is available in mountain bike cranks, but matching that with road cassettes is a challenge. The chainlines and pull ratios differ between road and mountain bikes. Many people feel 42-11 is not high enough for the highest ratio. I don't need 50-11, though the manufacturers seem to think everybody does, but I would not like my top gear below 46-11 or 50-12. There are a lot of riders faster than me that want 50-11.

Ross Millikan
  • 2,698
  • 13
  • 19
  • Yeah, I think the Rene Herse cranks are 70mm bcd vs the standard 110mm or mountain 74mm, letting you go down to 24T. I think you could consider these cranks “road” due to the low q-factor. Velo Orange has a 50.4mm bcd crankset, but I don’t think they sell small chainrings to go with it, just replacement 30T. – Ryan Witt Dec 04 '19 at 13:48
  • 1
    You're right that a 34t (round) chainring is the smallest that will fit on something with a 110 BCD. However, I believe that this BCD is called a compact crankset, not a sub-compact. I believe sub-compact cranksets either have a smaller BCD, e.g. the Shimano GRX has an inner BCD of 80mm, or sometimes manufacturers can modify the chainring or the spider to accommodate a smaller chainring on a 110mm BCD (Praxis cranksets, and in some sense oval chainrings would also fit this description). – Weiwen Ng Dec 05 '19 at 19:15