Evens or odds?

[RonS]

My daughter’s UK posting will soon end and she’ll be returning to Canada. I’m going to take advantage of her return and place an order with SJS to save the international shipping and possible duties.

Among the things will be a new chainring and rear cog to use when I finally wear out the current ones. My current setup is 38X17 which suits me fine. Just for the sake of experimentation, (I’ll take one for the team) I’m going to go with a larger combo, to see if it measurably increases chain life.

My question: Do you think chain and cog life will be better with an even-even or odd-odd combo? It looks like 40X18 or 43X19 will give me ratios almost the same as I’ve got now. 43X19 has the added pizzazz of being prime numbers. Will that make a difference? Beats me. I’m a retired trucker, not a retired engineer. I’ve read the Sheldon Brown theory of the even setup wearing the same teeth but I don’t understand how that’s better than the odd setup spreading the wear over all the teeth.

Of course, there’s also the chance that it really doesn’t make much difference what combo is used, but it’s fun to experiment.

Also, my front ring is a Thorn 104 BCD, which I am under the impression is too thick to use with a Chainglider. I have not found a “thin” 104BCD ring larger than 36T, and I don’t wish to purchase a new crankset just to fit a ‘glider.  If I am in error, I’ll welcome the correction.       

I’ve got until the end of next week to decide. I welcome your thoughts.   

[mickeg]

I use the even theory.  That in part occurred because I bought my Thorn Nomad Mk II as a frame and fork, bought my Rohloff hub elsewhere.  Thorn uses 17.  Stock Rohloffs come with 16, or at least they did when I bought mine.

But I can't tell if one is actually better by X percent of chain life.  Chains and sprockets are cheap enough that it really does not matter that much.

If you go with an even number of teeth, I suggest you mark one of the teeth so that you can always put a outer plate link on that tooth.  If you might forget that when you put the chain on and accidently put it on the wrong way, then it is settled, an odd number would work better for you because with an odd, it does not matter.

I cut a small notch in my teeth with a Dremel moto tool cutoff wheel, but once that slot got filled with grime it was hard to see.  Later added a dab of nail polish to mark it.

Bigger should be more efficient.  As the chain unwraps from a sprocket and wraps onto a chain ring, the angle that each chain link flexes is less, thus less wear.  And bigger means less tension on the chain, so less wear again.

When I built up my bike, I went with the 16T sprocket because that is what it came with, and they were expensive since they were threaded and more complicated.  Now the splined sprockets are cheaper.  Knowing what I know now and if I was building up a bike now, I would likely choose an 18T instead of a 16T.  But I already own the chainrings I want for a 16T sprocket, so I am not changing. 

Since you now are using an even chainring, even if you stick with odd in back, it might be a good idea to mark a tooth on the chainring and always put an outer link on that tooth.  If that tooth is under a crank arm, it is easier to find.  If some of your teeth are more worn on the sides, those teeth were already worn that way by having the narrower inner plate links on those teeth.

[Andre Jute]

Of course, there’s also the chance that it really doesn’t make much difference what combo is used, but it’s fun to experiment.

The only update to a hub gearbox transmission that makes a truly large difference is a Hebie Chainglider. How large precisely will depend on how successful you were before in extending open chain life. I'm a masher, hell on chains, and the Chainglider extended chain life to nearly 3x the mileage I got before. I bought my stock of chains in bulk on sale with free carriage thrown in, so it probably takes me longer to pay for the Chainglider by the saving on chains it makes possible than someone who buys his chains one at a time for the full retail price. Even so, I reached breakeven on the Chainglider somewhere between six and seven chains, and after that the apparently indestructible Chainglider was free.

Everything else is nuanced, with fractional incremental gains adding up.

Larger tooth counts on the chainring and the sprocket will clearly spread chain wear over more teeth and thus cause less wear. How much? How long is a piece of string? The exception is the Rohloff sprocket, which is of an unearthly quality of metal. I'm tied into 16T sprockets forever because I bought spares with my bike and haven't even turned the first one yet... I have used the Rohloff with 38x16 and 16x44; no experience on even-odd drive.

Also a result of larger tooth counts, the angle of entry of each roller to the tooth will be less and you get another fractional extension in chain life, and a fraction less wear on the teeth.

A Rohloff HGB is technically classed as a single-speed drive. So you can use fixie chains, which are theoretically cheaper, so that you should get a fractional advantage per dollar or pound or euro spent. In real life, common derailleur chains are cheaper because such a volume of them pass through the discounters. I have established to my own satisfaction that the KMC X8, a derailleur chain, gives me better mileage and thus a bigger bang for the buck, not only because it is cheaper than the equivalent KMC fixie chain, but because it has more flexibility from side to side, but I'd hate to put a number on it because the conflating effects are difficult to isolate. (I suspect that one of the reasons KMC's top of the range Rohloff-specific chain has failed to prove enough superior to the X8 to justify its elevated cost is that it doesn't have the same side to side flex as the X8.)

Another cost saving with the KMC X8 chain/Hebie Chainglider combination is that I run the chain for its entire life on KMC's factory lube, no cleaning, no extra lube, no nothing, zero maintenance, a priceless convenience.

I love the aesthetics and longevity of the Thorn chainring, but the longevity is partly due to its thickness, which won't fit in a Chainglider. When I went over to the Chainglider I also fitted, by choice, new Japanese forged cranks (branded Stronglight) that I'd long wanted and, by necessity, a Surly stainless steel chainring (but I'm anyway a fan of stainless and even plain steel chainrings, after early expensive experience with the beautiful but disgustingly soft Shimano Nexus cranksets), but they're long since paid for by the Chainglider Saving, as is the 44T steel chainring I fitted, again by necessity (custom-made to clear the bulk of the BB mounted motor) when I electrified that bike.

The point I'm making is that insisting on absolutely wearing out components you already have just makes it more expensive at every step to get the components you really want.

I hope Dan pitches in here, because he's another who, proceeding from where he has already arrived, has a chainring mounting PCD limiting his options.

[Danneaux]

I hope Dan pitches in here, because he's another who, proceeding from where he has already arrived, has a chainring mounting PCD limiting his options.

Answering Andre's thoughtful page, yes, I am in a similar situation here...and still waiting for a Chainglider solution that is economical for me given my sunk costs include an existing chainring, sprocket, crankset and Phil Wood bottom bracket.

Background and  explanation...

Though I generally much prefer half-step-and-granny gearing on my derailleur bikes, that went by the board shortly after indexed gearing hit the market and drivetrains were optimized for crossover gearing and the needed half-step chairings and front mechs vanished over time. When contemplating my Nomad gearing, I briefly thought about mounting a rear mech on a Problem Solvers q/r hanger so I could shift between two chainrings, I soon gave up the idea and settled on duplicating the low-biased crossover gearing that had worked nicely for me on past derailleur bikes that sported it. I've attached my Nomad's gear chart below. I found it works well in practice for the full-on expeditionary use I envisioned for the Nomad, hauling 26.5l/26.5kg of water in addition to camping gear and food for 3 weeks way from resupply in the deserts of America's Great Basin.

When I spec'd the bike in 2012, I was starting with a fresh slate compared to my older bikes that used quill stems, square-taper bottom brackets, and 5-arm cranks. The external bearing BB was the same basic design as a threaded headset and had similar virtues (i.e. less need to carry larger, heavier tools on tour), so I took the plunge. I soon found however, the Shimano outboard bearings just weren't as as well sealed and durable in my use as my previous internal bearing/square taper BBs. Having gone this deep, I invested in a Phil Wood external BB and have had no problems withlongevity since. While long-term durability is still in question, I'm happy for now 13 years out. For reference, I have about 80,000 trouble-free miles on my Phil square tapers but unlike Andre, I'm a spinner and that makes a difference.

When I decided to go with an external-bearing BB, I also spec'd a matching crank from Thorn's offerings, a Deore 4-arm 104BCD model, and fitted it with a 36t Surly stainless chainring. I switched from a threaded to clip-ring Rohloff sprocket when they first came out and this is what I run to this day.

Until Surly or another maker produce a 38t stainless chainring in 104 BCD I can match with a suitable rear sprocket to fit an existing Chainglider /or/ Hebie makes a Chainglider to match my 36t 'ring, I'm stuck where I am unless I wish to incur greater cost (i.e. new crankset, BB, sprocket and Chainglider), hard to swallow while my present setup works fine albeit sans Chainglider, which I I'd really like to fit for the lowered chain maintenance. As it is, I practice good "chain hygiene" and clean and lube regularly. If on tour, I pack a small, cleaned mascara wand pinched from my sister for nightly in-camp clean-and-lube sessions, as it so neatly flosses grit from between the inner plates, better and quicker than anything else I've found for on-tour conditions.

Ron, you'll notice I went with an even chainring/odd sprocket combo because it gave me the gearing I wanted with the only stainless 'ring available at the time in the size I needed and that trumped a Chainglider. If I were starting fresh today, I would likely mix-match my sprocket and chainring to most closely approximate my present gearing while allowing a Chainglider. Hindsight is always 20/20, but I'm not totally convinced a Changlider is ideal against the incursion of talc-fine alkali dust that gets into everything through the smallest interface. It is even finer than the Mt St Helens eruption ash I encountered back in 1980 as I rode the San Juan Islands and Washington state on a tour while the mountain was actively spewing.

I pondered the even-even, even-odd, odd-odd drivetrain question in a 2012 post that may be helpful to you even now. See...
https://thorncyclesforum.co.uk/index.php?topic=4655.msg22911#msg22911

In the post just before that, I attached a photo of chain/tooth wear on the 40t Thorn chainring I briefly fitted before the Surly 36t stainless one. The oil-darkened teeth are where the inner chain links engaged. The photo just makes it a little easier to visualize George's even-even argument and why in that case it makes sense to refit the chain so it consistently registers with the teeth. Sorry the original thumbnail for that photo was lost in a Forum upgrade crash but it still downloads fine and can be seen here in my repost.

There are so many variables that affect drivetrain wear and longevity, it is really hard to come up with a prescription. Chain make/model, terrain and load demands, weather, environment, maintenance intervals and lube choice and rider pedaling style all figure in among myriad others. It has been proven a well-lubed drivetrain that is protected from weather will outlast one that is not, but whether it makes economic sense will depend on your sunk costs, preferences and priorities. All those are so much easier and less expensive to accommodate at time of build rather than later, unless you find yourself in need of a wholesale replacement due to wear.

Best of luck, Ron. Please do let us know what you choose and update us from time to time, as this is just sort of thing we find fascinating and enjoyable beyond all measure.

All the best, Dan.

[Andyb1]

I don’t think side wear on a chainwheel will effect it’s life.  Wear to the teeth profile is the problem.

Do chains elongate equally across both the narrow and wide links?   Someplace I have read that elongation only happens is only across one of them while the other maintains the same pitch.

[mickeg]

I don’t think side wear on a chainwheel will effect it’s life.  Wear to the teeth profile is the problem.

Do chains elongate equally across both the narrow and wide links?   Someplace I have read that elongation only happens is only across one of them while the other maintains the same pitch.

Agree on the side wear, it is only worth mentioning because that is visual evidence on which links were used on which teeth.


Yup, every other link gets longer over time with pin wear and wear on the plates where the pins rotate in the plates.  Pins are mounted on the outer plate links, so outer plate links are the links that elongate.

More on that here:
https://www.sheldonbrown.com/chain-life.html

In the big picture, I do not think that even/even [sprocket/chainring] makes a chain wear very much slower than an odd/odd or odd/even combination, it might not change the rate of chain elongation at all.  But since the chain and sprocket teeth wear together over time, you can push a chain much further than the recommended 0.75 or 1.0 percent elongation for derailleur bikes without trouble.

That said, if you have a spring loaded chain tensioner, you might not be able to push a chain that much further.  It is the same as a derailleur bike, eventually the sprockets you use the most allow the chain to skip when you pedal hard on a derailleur bike.  The spring loaded derailleur cage will allow the chain to jump over the teeth because the spring will allow the chain to get a bit longer where it is wrapped around the sprocket.  But on Thorns with an eccentric, the chain can't skip over the teeth unless those teeth are really worn.

I was reminded of this about a month ago, on my rando bike which uses a Sram eight speed cassette (11, 12, 14, 16, 18, 21, 26, 32).  I use the 16 and 18 tooth sprockets the vast majority of the time, so those sprockets wear out first.  And on a long exercise ride both my 16 and 18T sprockets started to let the chain skip.  The chain is still below the 0.75 percent elongation point, but those two sprockets are too far gone.

With an even number of teeth, if you push a chain far beyond the 1.0 percent elongation and then put a new chain on, that sprocket will start to wear the other teeth rapidly because chains and sprockets wear together with even numbers of teeth.

It gets complicated.

[Andyb1]

So if it is the outer plates that elongate and increase pitch then with an even cog (chainwheel or sprocket) the wear from the chain elongation will be concentrated on 50% of the teeth.

With an odd cog the wear is evenly distributed between teeth as the narrow /wide links alternate.

Which is better I don’t know.


Once an even cog starts to wear is it best to maintain the narrow / wide chain link alignment….or is it better to intentionally change that alignment so wear is more evenly shared, as with an odd cog?

[mickeg]

So if it is the outer plates that elongate and increase pitch then with an even cog (chainwheel or sprocket) the wear from the chain elongation will be concentrated on 50% of the teeth.

With an odd cog the wear is evenly distributed between teeth as the narrow /wide links alternate.

Which is better I don’t know.


Once an even cog starts to wear is it best to maintain the narrow / wide chain link alignment….or is it better to intentionally change that alignment so wear is more evenly shared, as with an odd cog?

Since the sprocket and chain wear to match each other, I think that with an even system, you are constantly wearing on all teeth, but half of those teeth wear a bit faster as long as you run the same chain.  The reason I say a bit faster is that the chain with wear will continue to elongate more.  On an even tooth sprocket, I can't see any reason to want to change the setup after it is worn one way. 

I think it is best if every chain roller is wearing on every sprocket.  That said, when you put on a different chain with less or no wear, that upsets the apple cart.  Your sprocket will start to wear to match that chain.

But with the odd number, since the teeth spacing is designed for a new chain with no wear, I would think that you are wearing every other tooth as you pedal when you are using a worn chain. 

***

Different topic.  If you are asking your daughter to bring this stuff to Canada for you, do not forget that she probably will be paying VAT if she buys it there.  Not sure if she can get VAT refunded upon exiting UK, I do not know how the rules work.  My point is that there might not be any great savings compared to buying from a Canadian supplier.  But I do not know if there are any Canadian suppliers of Rohloff sprockets.

Before Brexit, I placed several orders with SJS and some other UK suppliers, VAT was not paid.  But I am in USA, at that time USA did not charge tariffs on small orders.  I have no idea how it works to order from the UK or from continental Europe for shipping to Canada.  I did place an order with a different seller in UK in June 2022 for a Carradice bag, but I have no records on what I actually paid.  That said, I paid less than I would have paid from a seller in USA.