Which brands/models chainsets / chainrings have perfect centering?

[silva]

That does sound like some sort of eccentricity. May I suggest you remove the chainring/sprocket and draw its outline on paper as well as the bolt holes. Then rotate until you see the place where it seems eccentric and draw that outline. It will show clearly in the drawing if you manage to keep the pencil vertical all the way round. I use that method to check symmetry of templates all the time.

According to the formula, 0.5 mm offcenter suffices, I don't think I can draw that accurate.

This is a very typical way that a worn chain on a good set of sprockets presents. If the chain is worn and used on a derailer bike, the derailer spring allows the chain to ride up the sprocket tooth and skate over the top. On a single-speed it doesn't have that escape route and the incoming tooth (front or back) forcibly collides with the chain roller that's in the wrong place for perfect engagement due to chain elongation. That feels like a strong vibration through the pedals.

But the chain was new/unused, just made on right length. I placed the bike upside down, took off chain, cog and chainring, put new ones, tensioned the chain with the bottom bracket excenter, and noticed the variation.

 

[Yellow Saddle]

The upper part of the chain when bike upside down is the return run, but why would that matter, I'm slowly moving the cranks by hand, nearly no force. When I do the test on the "tension" run, it gives just the same result.
And yes, the high (and low) tension point is just once per crank rotation.

It matters because the return run is subject to up-down movements related to pedal speed and road roughness. It isn't a good place to measure tension by distance inference. The return run could quickly become the tension run if you pedal the other way and I didn't want to go through a rigmarole of questions to figure out whether you were measuring a return run or tension run.

Are you in a position to trace the chainring? Failing that, if you have a vernier, could you take some strategic measurements?

There is another reason a chain "vibrates" but I won't deal with that now. The fact that you're running a 16 in the rear just about eliminates it.

 

[silva]

I have a venier, and just took some measurements, it's a 5 bolts mounted chainring, I measured the 3 distances between non adjacing holes, with venier inwards measuring tips on the hole side closest to the one of the other hole, in other words the mininum distance that can be measured. ande these were all 12.7 mm, with the lower tenths of mm scale of the venier precise matching scale lines so 12.70 cm.
This likely implies that the chainrings center / rotation point lies in the middle of the bolt holes, as they should be.

EDIT 12.70 cm instead of 12.70 mm ofcourse

 

[rogerzilla]

Sugino Zen chainrings and EAI sprockets are very round. The best Sugino track cranks (75 or better) should also have concentric drillings. The XD cranks (110 BCD) can be awful; the last pair of RD cranks (130 BCD) I used were pretty good.

 

[silva]

I have had a very good experience with a Velosolo rear cog, a 16t, it held out almost 30000 km, with most of that running on a 5 mm offcenter chainline and a 3-4 cm chain tension variation. It outlasted 2 chainrings and 2 chains.
Somewhere at 20000 km I flipped it due to hook fin shape as to start wear the other side of the teeth.
The cog has very thick teeth.
The Velosolo chainring recently mounted shares this characteristic.
The teeth are wide but short.
On a previous fixed gear, I had a TA Specialites 52t chainring, also happy about it, the teeth were different, not as wide but longer.
Question is what wears out first, a thick but short tooth, or a narrow but long tooth?
If I take a guess, I'd say thick but short last longest.
The idea behind this claim is that a thick tooth has a greater contact surface, force is spreaded over that bigger surface, so less pressure per surface unit, and wear is a function of pressure. Abit like a parallel system. A narrow long tooth can then be seen as a serial system, where one after another have to withstand all and fail to do so.
The Stronglight chainrings teeth also were quite narrow. The Velosolo chainring appears thus promising, time will tell.
My whole drive train looks impressive now, looks like a lasting one, being the goal, only that tension variation indicates the presence of a crap bicycle part, whatever it will be.

 

[Pumpkin the robot]

The 12.7cm (5 inches) figure only tells you the holes are evenly spaced on the pcd, not that they are concentric to the axle. You need to measure the holes to the centre of the crank.

 

[silva]

General tolerance in the machining I do is +/-0.25mm I find it hard to believe that they accept something being half a mm out. Pay peanuts, get monkeys I guess.
You may find that the chain ring has an error and the chainset has one, causing a compound error. You would be able to rotate the chain ring to eliminate some of that if that is the case .

As already mentioned, the recently placed chainring has zero clearance between its inner edge and all 5 crank spider notches, I only got it in its (sidewards) location by tensioning the bolts, unlike the Stronglight chainring that it was replacing.
So I think rotating the chainrings position wouldn't alter any of the current excentricity.
Unless there is some other excentricity cause that I'm still not aware of, it looks like the Stronglight crankset is its cause. The as far as I know other causes have been eliminated.
The previous setups were certainly compound errors, a part due to wear, a part due to chainring, a part due to crankset, aggravated by a 5 mm wrong chainline.

It's almost unbelievable how the producer and dealer of this bike have screwed me. The whole motivation of the purchase was to have a bicycle that finally solved the problems I was plagued with for many years.

The bike is a travelbike ment for internal gears (hub), choice chain or tooth belt, modified to a fixed gear.
I had demanded drivetrain components that were very robust and lasting, wear minimalized.
I received plain crap. Despite I had hinted the dealer on a motorcycle chain, I ultimately had to find the chain I wanted myself, and that on a wellknown singlespeed shop in the same country as the producer of the travelbike. So they didn't do much search effort.

I had asked a 1/8" drivetrain, the new bike was delivered with a 3/32" Surly stainless steel chainring. I wasn't told and didn't notice, only to discover it when the Surly chainring wore out in a single month.
The dealer then was unable to find a 1/8" chainring with a 110 bcd (for the Sugino crankset).
Meaning the crankset had to be replaced too in order to have the 1/8" drivetrain I had demanded and had been promised.

And then, during the cranksets mount, the spider of this Stronglight crankset turnt out to be narrower, making the chainring void the bike frames chainstay, causing another extra replacement - a longer axle.

Nearly a year later, with the Gusset "tank" chain hanging 45° tilted on two places, and with the dealer telling me he was unable to explain why, I asked the question on a forum, and was immediately told that this was very likely due to a seriously wrong chainline. I was teached how to measure it, did so, and I measured 5 mm off.

When the bike was in production, the dealers last message before delivery was that the bicycles delivery would be delayed some weeks because the chainline wasn't "100% straight" but that they found a solution.
They didn't, they (a Netherlands based dealer named "Santos") decided a deliver as is / bike sold / get paid.
When I confronted the dealer with the wrong chainline forum post, he said that "he had followed my measurements and that they were correct". Pretending alike he didn't knew from the beginning lol.

However, I was able to correct most of the problems myself, a rather shame for the 4300 euro that the bike costed, only this chain tension variation remains of the endless series problems I had, with the possible exception of a rear wheel running 5 mm aside the front wheel, because when I measured the chainline I discovered that the center of the spoke flanges was that much out of the center of the wheel mount pads, so the wheel should be spoked abit alike an umbrella to compensate for this, and I still didn't check this. I did notice that my rear tire wears out quite more at 1 side than in the middle.

This to tell the whole story, which may answer some questions before asking.

And as it looks like now, I'll have yet another further cost, a second replacement of the chainset to one that IS machined precise enough to not have a chain tension variation. Being the goal of my topics question. I found out it's smarter to ask other cyclists/mechanics than to dealers/producers whoms highest priority is a sale.

 

[Yellow Saddle]

I have a venier, and just took some measurements, it's a 5 bolts mounted chainring, I measured the 3 distances between non adjacing holes, with venier inwards measuring tips on the hole side closest to the one of the other hole, in other words the mininum distance that can be measured. ande these were all 12.7 mm, with the lower tenths of mm scale of the venier precise matching scale lines so 12.70 cm.
This likely implies that the chainrings center / rotation point lies in the middle of the bolt holes, as they should be.

EDIT 12.70 cm instead of 12.70 mm ofcourse

This measurement doesn't take into account the outside perimeter of the chainring. You have only eliminated the fact that the ring does not precess (providing the crank is perfect).

I think you'l get a better picture if you measure from bolt hole to tooth root, if you can figure out a way to make the measurement perfectly so that you measure the same feature each time.

 

[Pumpkin the robot]

Measuring the pcd of the bolts to the root of the teeth will only work if the amount of teeth is a multiple of the amount of the bolts (5 iirc) otherwise you will need to do a bit of trig.

 

[Yellow Saddle]

Measuring the pcd of the bolts to the root of the teeth will only work if the amount of teeth is a multiple of the amount of the bolts (5 iirc) otherwise you will need to do a bit of trig.

Yup, nothing wrong with some trig on a Sunday morning.
Who doesn't love it before breakfast?
I suggested to the OP that we need a trace-out of the ring in multiple positions. That is simple and it never lies. I do this with templates I design for an obscure hobby of mine and the brutal truth of flip-flopping a template to see if it is symmetrical often stuns me and chases me to the fridge for a beer and a deep breath.

 

[silva]

I'll try that paper outline on a spare same of the mounted chainring, so that I don't have to break apart the bicycle once again lol.

edit, nah, it's impossible to judge, if I have to keep the pencil straight up, then the thickness of its tip causes it to draw several mm's away from what I'm trying to outline, and the chainring thickness is about the same as the height of the protruding pen of the pencil, causing it to not reach the paper half the time.
So what about pencil not straight up, but tilted so as to draw as close as possible to the chainrings edge?

edit2, just did the latter, drawed the inner circle, the bolt holes and the teeth outline, but now I'm not sure what to do next. If I rotate for ex 1 bolt hole, the teeth outline ceases to match, in this case there isn't even any other match than the original hole, making it unusable to compare. If I rotate around the inner circle, all bolt holes match in every position. But that doesn't exclude excentricity - the whole of inner circle and both holes could bave another center than the center that the teeth have.

By the way, the producer of the chainring states CNC machined, shouldn't that exclude any center deviation?

edit3, I think I found a method, I put a ruler so that the clockwise directioned edges of 2 bolt holes are on it, draw a line, and repeat this 6 times. I end up with lines that form a jewish david star.
Then, from the outer intersections of those lines, I repeated it to form another, smaller david star. After a second repeat same way, I ended up with a so small david star that I can visually estimate its center accurately, where I draw a dot.
Then, I did a multitude measurements of distances from the dot towards the inner circle, and found these all equal. This probably proves that the bolt holes have the same center as the inner circle.
And last, I did a multitude measurements of tooth tips towards the center dot, and these all measured 98 mm.
So if aboves method would suffice, this would prove the chainring as not the cause of my chains tension variation.

 

[Yellow Saddle]

I'll try that paper outline on a spare same of the mounted chainring, so that I don't have to break apart the bicycle once again lol.

edit, nah, it's impossible to judge, if I have to keep the pencil straight up, then the thickness of its tip causes it to draw several mm's away from what I'm trying to outline, and the chainring thickness is about the same as the height of the protruding pen of the pencil, causing it to not reach the paper half the time.
So what about pencil not straight up, but tilted so as to draw as close as possible to the chainrings edge?

edit2, just did the latter, drawed the inner circle, the bolt holes and the teeth outline, but now I'm not sure what to do next. If I rotate for ex 1 bolt hole, the teeth outline ceases to match, in this case there isn't even any other match than the original hole, making it unusable to compare. If I rotate around the inner circle, all bolt holes match in every position. But that doesn't exclude excentricity - the whole of inner circle and both holes could bave another center than the center that the teeth have.

By the way, the producer of the chainring states CNC machined, shouldn't that exclude any center deviation?

edit3, I think I found a method, I put a ruler so that the clockwise directioned edges of 2 bolt holes are on it, draw a line, and repeat this 6 times. I end up with lines that form a jewish david star.
Then, from the outer intersections of those lines, I repeated it to form another, smaller david star. After a second repeat same way, I ended up with a so small david star that I can visually estimate its center accurately, where I draw a dot.
Then, I did a multitude measurements of distances from the dot towards the inner circle, and found these all equal. This probably proves that the bolt holes have the same center as the inner circle.
And last, I did a multitude measurements of tooth tips towards the center dot, and these all measured 98 mm.
So if aboves method would suffice, this would prove the chainring as not the cause of my chains tension variation.

Excellent. I would have been very surprised if that had been the case. Now for the crank. see if you can somehow create a centre point in the bolt that holds the crank on the BB. Perhaps, remove the bolt but don't pull the crank. Now use a standard common compass to draw a circle bisecting the chainring bolt holes. See if the crank is concentric around its centre point.

I'll also suspect the freehub - what brand/type is that and the sprocket. Off you go, you have plenty of homework to do.

 

[Pumpkin the robot]

CNC machining is not foolproof. Drilling a hole, I would expect a deviation of about 0.1mm unless the hole is machined by a smaller cutter, rather than drilled.
As the chainring would be machined in one operation, I would expect it to be pretty concentric.
I think the issue you have would be the centre of the chainset not being concentric to the bolts that secure the chainring.

 

[silva]

Excellent. I would have been very surprised if that had been the case. Now for the crank. see if you can somehow create a centre point in the bolt that holds the crank on the BB. Perhaps, remove the bolt but don't pull the crank. Now use a standard common compass to draw a circle bisecting the chainring bolt holes. See if the crank is concentric around its centre point.

I'll also suspect the freehub - what brand/type is that and the sprocket. Off you go, you have plenty of homework to do.

Heh, I'm a bicycle tech rookie, the very reason I went to fixed gear was to simplify the common / high frequent maintenance and replacement jobs so far that I would be able to do these themselves, so I wouldn't depend anymore on dealers that in the end took 6 upto 9 months for a revision, causing my spare bike drivetrain to be worn before the other got repaired.
So I will certainly not try to remove a crank bolt, if I screw something up I can't even get at a repair place with it, requiring a pickup by the repairer and a delivery later on.
But I'll try to figure out a centre point in that bolt. I don't have a compass but I think that between the old tools my dad had there are some metal compass alike tools that can be fixed under a certain angle, so if I fold it open so that it ranges the distance crank and chainring bolt centers, then fix the angle, I can check the other chainring bolt distances.
But I have doubt about the accuracy. 0.5 mm equals 2 cm vertical tension variation, and I can't see how I can visually be that accurate.

There is no freehub / freewheel, it's a fixed gear.
The sprocket is a Velosolo bolt on cog 16t see http://www.velosolo.co.uk/shopdisc.html
It's mounted on a "Surly Ultra single speed disc silver 13x10mm 36G"
And as already said here somewhere, between the cog and the disc mount, there are 2 x 2 mm spacers to correct the chainline, purchased and mounted myself. As also said, this is a potential cause for excentricity of the cog, because the spacers (see CNC 6 Bolt Cog Spacer on same page) are flat thus do not "pass" the centering ring through to the cog.
(I also found out that there are hubs with IS disc mount that don't even have such a centering ring).

But if that was the reason for the excentricity then I would see 3 peaks and valleys in the chains tension, and that is not the case. Also, the degree of excentricity would be 1/3 of the chainrings.
So I would say that the rear of the bicycles drivetrain can be eliminated as cause.

Since the chainring has been eliminated today as cause (IF I can assume that the mounted one is perfectly the same as the spare one that was delivered the same time), all that remains is the Stronglight chainsets spider/whatever closer to the axle, meaning that if I want to get it solved, I'll have to spend bucks once again.

And that was the reason for my topic here, to avoid a third not well machined component purchase, by asking others if they know brands/models that proved themselves as machined with a smaller tolerance than that 0.5 mm.

 

[silva]

Btw, Yellow Saddle, since my search for a longer lasting bicycle chain I developed an interest in roller chains generally, this appears as being a specific field of interest of you, and therefore you may be able to give some definite answers on some questions I found different answers for.

1)
My last bicycle, where all above is about, also suffered an annoyant initial (new chain) of the bottom bracket eccentric chain tensioner: nearly at the end of its range, inflicting me the job to remove a couple chain links during the life span of the chain, quite shortly after placement new chain.
My move from an 48t to a 47t chainring had, aside being a solution for the integer gear ratio caused wear concentration within the chain, a second reason: it allowed the chain to be 2 links shorter, and brought that initial eccenter position close to the begin of its range.
Now, there was an alternative solution, the insertion of a type chain link that is known as "halflink". And chains exist that are fully built with such halflinks. Of course, the halflink has to be available for the chain, but supposing it is, I have read on several industry oriented websites that halflinks (also named "offset links") would form weak points in the chain.
For ex https://www.diamondchain.com/frequently-asked-questions/

However, due to the reduction in the chain’s working capacity, offset links, of either type, are not recommended in performance oriented drives.

or https://tsubaki.eu/catalogs/Tsubaki-CATM-EN-2016.pdf

Although offset links can be used when there is an odd number of links in the roller chain, it is better to use a design that requires an even number of links

Offset link plates are bent at the center, and the resulting concentration of stress at the bend can cause a fatigue break. Avoid using offset links in high-stress applications.

An offset link is used when an odd number of chain links is required. The pin and two plates are slip fit. The fatigue strength
is 35% (applicable to ANSI chain) lower than that of the chain itself.

35% is a serious reduction, but would this matter for a bicycle application in the case if the link plates were 3/16" - double as thick as a common chain? Often I see wide chains advertised as higher strength, but I cannot see how, since a chain is as weak as its weakest part, and if the internal width and thus the rollers/bushings stay same sized, regardless how thick the plates are, the chain won't be "stronger" than before. I see as big benefit the increased contact surface between pens and plates, spreading the force over a bigger area and thus less force per surface unit, and thus less wear. Since an offset link has same room for plates / thickness, they wouldn't void that wear benefit, so that they would give me more gear choice options without shortening need within chains lifespan.
What are your thoughts on this?

2) My current Gusset "tank" chain, see https://www.tartybikes.co.uk/chains/gusset_tank_chain_18_inch/c7p331.html has internal width of 1/8" but the plates of a 3/16" chain (a combination of 2 standards that is typically called "heavy duty version").

- Fed up of breaking chains? Then you should be looking at this beast!
- Super Heavy Duty 1/8” chain.
- Pin Length 11.6mm.
- Pin riveting power 300Kg (standard chains are around 150kg, KMC Kool chain is 250kg)
- Tensile strength 1500kg (KMC Kool chain is 1300kg).
- Plate width 1.4mm (standard chain is 0.8mm, KMC Kool chain is 1.1mm)
- Includes split link, but we do not recommend using this for trials.
- Weight: 550g

There are contradicting / wrong specifications about amount links but since I purchased a dozen of these, I know for sure: 102.

Recently, I discovered this: https://www.tartybikes.co.uk/chains/kmc_k910_316_inch/c7p12836.html

Ultra strong chain from KMC!
- Super wide 3/16" construction with extra thick plates gives this chain a whopping max. load capacity of 1500kg!
- 13mm pin width. (EDIT: 12.9 mm also listed and probably more correct)
- The chain tool we recommend for this chain is a Park CT-7, alternatively we are happy to split the chain to your desired length before dispatch - just put a note at checkout. Please remember the split link will add extra length to the chain (1/2" in total) when fitted!
- Suitable only for single speed setups.
- Mushroom head pins to eliminate splayed links.
- Supplied with split link.
- Please note the colour of this chain is Silver, there is an anti-rust coating applied which makes it look a little brown in the photos.
- Weight: 586g (full length, 100 links)

I wonder what (wear related) benefits the KMC may have relative to the Gusset.
The KMC chain appears to be a full 3/16" chain. Meaning the internal width is 3/16" instead of 1/8".
From experience, I know that 1/8" is hugely beneficial relative to 3/32", very likely due to the mechanical contact surface being 50% bigger.
But so far I was unable to source any 3/16" bicycle sprockets, yet, KMC produces such a chain, alike they are available.

Now, imagine the KMC used on my current sprockets. So a 3/16" internal width chain on a 1/8" chainring. This gives the links some room to move sidewards, or, reduces the sideways contacts between the sprocket teeth and the chains inner plates.
And most of all, the contact surface between pen and bushing, 1 of the 2 causes of elongation under tension and thus wear on the sprocket teeth, increases a further 50% to double the one of a 3/32" chain.

But, do you see drawbacks too?
Weight is for me non important.
And second, do you happen to know any companies producing 3/16" chainrings for bicycles?

Regards and your knowledge in this matter appreciated.