Rim Rant

This was posted to the rec.bicycles.tech newsgroup a while back.

In article <4uvvuc$5mc@netaxs.com>, siu@moberg.com (Paul Siu) writes: With so many people talking about rims, I was wondering what features are actually useful. A look at a typical rim ad, you'll see:

Hard-Anodizing - claims to increase strength.

Machined side wall - concentric ring like surfaces that's supposed to increase braking.

Ceramic coating - supposed to increase braking.

Welding - supposed to be stronger than pins, and creates a smooth side wall so that brakes won't grab on the gap.

Heat Treated - ?

Double eyelets

Which of these rim features are actually functional, and which are marketing gimmicks?



KB responded one night when he had nothing better to do...

KB's response:

Good question - I'll take a stab.

1. Hard-Anodizing - claims to increase strength.

There is no evidence for this in my experience.

One can perform an crude analysis on a rim's cross section, adding the properties of the hard anodized material and aluminum together using the method of sections, as did Mario Emiliani way back when in a publication called Bike Tech. This shows some improvements in stiffness by using the hard anodizing since it has a very high modulus. However, this type of analysis can be misleading in this kind of case, especially when the properties of the different materials that make up the section are so different, and we were not able to measure these in a direct, controlled experiment. We measured a number of rims of the same cross section with very small variations in cross section and weight due to extrusion tolerances, some were anodized and some not. They were built into identical wheels. We tested several wheels in each configuration to be able to estimate the scatter in results due to variations in the builds themselves. The tests showed no advantage for the HA rims.

My guess is that the differences in the modulii of the two materials does not perform in the way that the simple calculations predict because the brittle characteristics of the AO dominate it's behavior and because lateral loading on a rim in a wheel assembly results in a very complex mix of stresses on the rim - it's not a simple cantilever beam problem. This is impossible to account for in the simple calculation.

Bottom line - hard anodizing of rims does not increase the strength or stiffness of a rim in any tangible way. It does increase (bulk) corrosion resistance, though decorative anodizing does this adequately with less cost or compromise and the corrosion of bare 6061 aluminum is not a problem in most off road environments. That is, 6061does not seem to have any structural problems when it corrodes in normal off road use. Corrosion at micro cracks probably would be a big problem with HA rims if corrosion of the substrate was a problem.

HA does reduce the rate at which brakes wear away the side of the rim, while it is there. It works in a way that is similar to ceramic coatings, though it compromises braking performance far more, especially wet braking, and so it is not the best way to do reduce brake wear at the side walls of the rims. It abrades eventually and this exposes bare aluminum.This wear improves braking performance.

It is possible that HA increases the tendency of the rim to crack around the spoke holes, one of Jobst Brandt's claims, and this follows from the problems one sees when a brittle coating is applied on a more ductile base metal. Fatigue cracks in the base metal can be initiated at the inevitable micro cracks in the brittle coating. Corrosion in these cracks can also lead to accelerated fatigue failures.

I have not studied this in detail, and most modern MTB rims have sufficient metal in this area and short life spans in the field (for other reasons) that fatigue failures around the eyelets have not been frequent. Mavic's 217 do show this characteristic though, so the topic is not dead. It would not surprise me that failures around eyelets due to fatigue still shows up on lightweight road wheels, on the rear, around the drive side spoke holes, after a few years (or less) of hard riding.

2. Machined side wall - concentric ring like surfaces that's supposed to increase braking.

There is something to this claim initially but it has little to do with the machining per se. The process for brushing bare aluminum rims we developed, a simple abrasive finish on bare 6061, is also great in braking performance right out of the box. So are the machined rims. I think braking performance has to do with the way the pads and a clean, flat, bare aluminum braking surface interact. I do not understand this in detail but I see no evidence that tool marks on the machined surface improve braking performance. In fact, there are ways that they can degrade performance, though this effect is also very small.

So, the bare brake walls improve braking, but the "concentric rings" bit is not a big deal. Keep in mind that the brakes walls of a welded rim are machined on a lathe, so the rings are the pattern the tool leaves on the rim and are the surface texture that you'd expect, or are stuck with when you think about it. This claim could easily be chalked up to "creative" marketing, (a bunch of marketing folks looking around for "bullet points", features and advantages, tools to sell with - generally not determined by their truth value or laws of nature - if you know what I mean) or misguided rumor.

And this advantage of a machined or brushed braking surface is an extremely short lived phenomenon, and will not last beyond a long, hard ride or two. The brake pads and grit wear away the small irregularities in the bare aluminum on the surface of the rim quickly, whether they are due to machining or brushing. The long term braking surface is formed by the abrasive interaction between the rims and pad.. All rims that are not coated with a very hard material, such as a ceramic, end up in the same state in a relatively short time (exceptions are hard anodized rims which take a little longer - see comments above).

The as machined or brushed aluminum braking surface does perform slightly differently than the worn in surface in the way it interacts with the brake pad. Keep in mind, there are no smooth surfaces if you look closely enough, so it is only the scale of roughness that affects this. There's more to it, but it makes no sense to grind on it (mentally) since it's a short lived, cheap thrill.

The long term issue, interaction between the pads and bare aluminum, is more interesting. This interaction seems optimal when both surfaces are relatively clean. This can be maintained by using a light abrasive (ScotchBrite) and mild solvent (watch out for aggressive solvents if they get on the tire's side wall) on the rim surface every now and then. Clean the slots in the pads regularly, and sand a fresh surface in if you hear grinding noises when you apply the brakes. Soft pads seem to embed particles sometimes, and these make the pads work like grinding stones when you put the brakes on, so a trip trough a muddy puddle is not quite the same as the careful maintenance steps.

Bottom line: A machined surface is flat and gives good braking performance from the beginning. It is not a long lasting effect.

3. Ceramic coating - supposed to increase braking.

This has not been our experience in dry conditions. In fact, the surfaces can often lead to less consistent braking performance because they give less "feel". By this I mean that the braking forces you get from standard pads and ceramic rims do not seem to be as smoothly proportional to forces applied to the lever as the same pads on bare rims. It depends a bit on the details of the brake's set up and the state of the rim's surface (how long it's been in service), so this is not a principal that's etched in stone. Ceramic surfaces do not, in general, noticeably improve the power or feel of properly adjusted cantilever or V brakes in dry conditions.

Some might argue to the contrary from their personal experience, but we did the experiments carefully. The results are not based on comparisons of old parts vs. new ( the typical brake product endorsement story goes - "I used STX cantis on a stock rim, but when these wore out, I replaced them with a ceramic rim and cool expensive after market pads and they were much better! Definitely worth the $400 I had to pay, blah, blah, blah"). We compared all new parts and set each up carefully. There were no measurements made, so the tests were subjective, but the results were not difficult to interpret.

Wet braking on ceramic rims is improved, but only slightly. The ceramic coating makes the initial braking power slightly better. When the brakes are applied in wet conditions, braking force is very low if the rim is very wet or muddy. But the pads wipe the goop off the rim surfaces as they go around. When the surfaces are wiped sufficiently, braking forces start to improve. The ceramic coating makes "wiped sufficiently" happen a bit faster. Once cleaned the performance of the two surfaces are not that different.

The real advantage of ceramic coatings is that they stop side wall wear due to abrasion during braking. The rim can be designed with thinner brake side walls because there is no allowance for wear required, or it will last longer at a given wall thickness. Then the limitations on rim wall dimensions are based on what can be extruded and what can be joined (pinned or welded), and rims can get a little lighter this way, in principal. In continuous wet conditions (like they have in England, Germany, and Seattle all year round) ceramic coated rims would give you a small advantage in braking performance and a big advantage in reduced rates of rim side wall wear.

The real disadvantage to ceramic coatings is that rock dings in the side walls of the rim (like what happens if you smack a square edged rock and bottom the tire, pinch the tube, and doink the rim) cannot be repaired easily. If there is a high spot on the braking surface where there is ceramic coating it really upsets braking performance. The pads grab it every time the wide section of rim comes around.

It is far worse than a bad joint in a pinned bare aluminum rim because there is nothing you can do about it. A bare rim can be tweaked or filed back to a flat state if the damage isn't too bad. This is not generally possible with a ceramic rim. You can't file the wall flat if it bulges out from a ding. And if you try to tweak it, and the ceramic coating flakes or chips, there is a fluctuation in braking force every time the pad passes over the damaged area. In either case, the change in braking performance can be large and you'd get tired of it soon. The rim has to be replaced.

Ceramic surfaces also can affect the rate at which brake pads wear I think, though I have not measured this. It is not a given, since the claims (A typical pad wear claim goes like this - "My pads wore out in three rides; these pads suck!") are often based on rides in winter conditions, and pads can be expected to wear at this kind of rate in slush and slurry rides no matter what kind of rim is used. I am also dubious about claims about pads that are "formulated" for ceramic rims for a number of reasons. It is not a simple materials problem, and I have seen no evidence that indicates a big difference in wear rates or performance, though I have not looked closely - yet.

Ceramic rims can also make brake set up trickier. There is an increased tendency for the brakes to squeal. All the usual squeal reduction tricks still apply.

Bottom line: If you are a smooth rider and don't ding rims, you ride in perpetually wet conditions, and anticipate wearing through the rim's brake walls instead of smashing them on a rock or tacoing them in a crash, ceramics are not a bad way to go - if you've got the money. If you can't afford to risk the replacement cost because you ding rims every now and then, or you think the cost of replacing brake pads (high wear rate = more frequent replacement) might offset the cost of the less frequent rim replacement you might get out of the ceramics, standard rims are a better deal.

If you are a racer and race in the wet, they are a pretty good thing of course. To solve the economic problem try to find a sponsor who will buy them (and brake pads) for you.

4. Welding - supposed to be stronger than pins, and creates a smooth side wall so that brakes won't grab on the gap.

Welding a rim is tricky, and it will make a stronger joint if it's done properly. It will make a weaker joint if it's not.

Joint strength is really not a simple issue, and it is not an important one in most cases. Pinned joints do not normally fail in the field. A rim is normally loaded by the tension in the spokes and by terrain and rider induced loads. During these the joint is rarely stressed near it's load carrying capacity. If the rim fails laterally (tacos) the joint can fail, though it often doesn't. If it does, it does so well into the overall failure of the wheel. In other words, the strength of the joint is normally only an issue when the rim is already bent to a point where it is doomed anyway. The joint will not be the cause of the failure; it is not the first thing to go. It is a consequence of a lateral load that causes the rim to fail by bending. This is true for most pinned and welded joints. I am sure there are exceptions, but they are rare on high end parts.

The advantage of the pinned joint is that a failure rarely ends up as a separation of the joint if it ever fails, whereas a poorly designed or executed welded joint can separate completely. In either case this type of failure would be very rare. Most welded joints are backed up in the internal cavities as well. Extra metal spans the joint to make up for the strength reduction in the weld HAZ and to keep the weld together. This is effective. There are a few other tricks to increase the strength of a welded joint that the production wizards know about, and they're not telling. It is a very subtle and complex process.

Welded rims can be slightly better when the a rim is subjected to an impact directly on the joint. By impact, I mean the result of a direct hit on a rock, the sort of event that would lead to a pinch flat and ding in the rim. The connection across the welded joint between the brake walls, a connection that can't be made on a pinned rim (Mavic tried spot welding the brake walls at the tips on the M231 pinned rims that they ceramic coated. They did this so these walls would not move from the aligned position as the wheel was built - this is the only exception I am aware of) makes the welded joint stronger. It is not a huge advantage for welded rims though. The probability of hitting the rims in this spot exactly is not very high, and pinned rims perform identically to the welded rims in this load case away from the joint. I have only seen one instance of a pinned rim that took a hit at the joint in a way that could not be repaired. I think that the probability of this event is small.

We were reluctant to weld rims before machining was possible. The braking surfaces of older welded rims without machined braking surfaces used to be horrible around the joint. The old RM20 Arayas were like this. The rims were welded, and then the welds were ground down until the braking surface was flat and smooth - in principal anyway. This never happened in practice. The grinding operation removed too much material from the braking surface and the low spots affected braking noticeably. Many were okay, and if they were decoratively anodized, a poorly finished weld wore in quickly. Hard anodized RM20s were a real mess though. This is the primary reason we stayed away from welding our rims back then.

The real advantages of welding and machining a rim become obvious in an analysis of the manufacturing problems and efficient production of rims with good specifications and tolerances. It is easier to make rims to very high alignment standards (flat and round rims) when they are welded and machined than it is when they are pinned. The cost and technical requirements of the welding and machining operations are high, but once mastered and stable in production, the rims are flatter and rounder on the average.This makes the wheels they are built in to stronger and easier to build and maintain.

There are complexities of manufacturing pinned rims that are subtle, and rarely discussed outside the doors of the engineering department of a rim manufacturing plant. A pinned rim is joined by pressing the rolled hoop of extrusion together at the joint with pins aligning the inner cavities of the extrusion. The alignment of this press fit depends on many things including sawing tolerances, extrusion tolerances, pressing process variables, and cleanliness. Sawing tolerances are a big deal here.

The extrusion is rolled into a helical coil, three rims worth, before it is cut into individual rims. The tolerances of the surfaces that will but together eventually are very tricky to manage, and determine the strength and stability of the wheel assembly. If the rim is pressed together and the joint surfaces do not align perfectly or are not completely flat, the rim will have a kink in it, at the joint, after it is pressed together. You can see this when you lay it on a flat surface.

Everyone who makes pinned rims has to deal with this and there are klinkers out there from every brand, the big M included. The defect is fatal to the wheel assembly too. It is not possible to true this defect out during the wheel build. It is difficult or impossible to adjust spoke tension to make a rim like this straight and strong since the defect is confined to a small length of rim. There are not enough spokes around the joint to straighten it. If the rim can be made true enough to get by, the spoke tension around the rim is so uneven that the wheel's stability (its tendency to come out of true when loads are applied) and lateral strength (its taco tendency) are very poor.

As an example, this type of joint flaw was a frequent problem with our BCX 1,2 and 3 rims, which are demonstrably strong enough for most riders, but which were made with bad joint tolerances on occasion. The kinked rims were built into wheels by a machine which could often figure out how to get the wheel true, but the resulting spoke tensions were uneven, and lateral strength and stability were compromised, especially on an 8 speed rear wheel. Wheels came out of true in the field. We studied the problem a fair amount and worked with Weinmann to correct the problems. The lesson: When tight tolerances are not held in the sawing and joining process pinned rims will not be flat across the joint, and these rims will not build into strong wheels - no matter what. The degradation of the wheel's strength is proportional to the "kink angle' at the joint. It's a tricky measurement to make in QA, immediately after the sawing operation but before the rim is joined, the point in the rim manufacturing operation where it really matters, because the tolerances on the surface are so tight.

A pinned rim can be made properly, and they have been for a long time. Mavic does it. Sun Metal does it, and so does Weinmann. It's not easy, and they each have their problems when they are not firing on all cylinders. We bond the pinned joints on our Red and Blue label rims to align the brake surfaces and rim section reliably and to make this step less sensitive to sawing tolerances. The process is expensive, but it is designed to assure us that there is no problems with misalignment or kink at the joint. The brake walls are held in alignment, parallel across the joint, in a heated fixture while the adhesive cures. This eliminates the effects of saw tolerances and many of the potential problems from extrusion tolerances on the straightness of the joint.

The bottom line for wheel builders: This is not a flaw that's easy to spot once the wheel is built. You have to measure the tension of the spokes all the way around the rim to see the problem and you will have to replace the rim if it is. When the rim finally gives up or the rider gets tired of truing the thing, a new, flat rim is the only way to solve the problem. It is rare that a rider will put the same rim back on, and the new rim works better. The old rim gets blamed, which it deserves, and the new one gets credit, which it may deserve if it does actually work. But the problem with the first rim is a manufacturing problem, not a design problem. This maybe a trivial distinction to the rider who has to go through the ordeal, but it matters to rim designers and manufacturers who have to identify, understand and solve the actual problem.

Check to see that a rim is flat and round before you build it up. You don't need a surface plate. Lay it on a trued wheel, brake surface to brake surface. Make sure there are no "kinks". Look at the surfaces where they are in contact, near the joint. If the new rim rocks on this point, or has a gap there (check it from both sides by flipping it over), take it back and get a flat one. Do the same inspection of the rim's OD. It only takes a minute and it can save a lot of work.

The bottom line for riders and shop mechs: A wheel that will not stay true may be due to an irreparable rim flaw. If the rim works okay on some bikes and not on others, this is very likely. It may be easier for everybody to rebuild the wheel than it is to fight it out with a kinked rim.

Back to welding and machining rims. It's obvious that a welded and machined rim can be made with brake surfaces that are very flat at the joint. In production, these tolerances are less finicky and more stabile using the combination of welding and machining. There are only a few ways they can go wrong, and these can be engineered hard to be reliable. The critical tolerances in the welding and machining operations are not as difficult to manage - they are not dependent on saw blade wear, chips in the fixtures, burrs on pins, etc. Then the machining operation makes sure the surfaces are flat all the way around the rim. If you want to make a lot of rims, and you want the specifications held high and tight, welding and machining is the way to go.

And the welded and machined rims have a huge marketing advantage over a pinned rim, since the alignment and flatness of the welded joint is much easier to understand than the more complex pinned joint. It's new looking, an easier story to tell, and the hypemeisters generally dig this.

Bottom line: Welded and machined rims are here to stay. The processes do not add to the cost of rims if the volume of rims made is high enough to justify the tooling investments. They do not offer substantial performance benefits over a well manufactured pinned rim. They do make the tight tolerances that are required to build consistently strong, lightweight rims easier to achieve, and this is the basic reason for them to be adopted as the manufacturing means of choice.

5. Heat Treated - ?

All rims are made of heat treated aluminum. No rim is heat treated as a second operation (as far as I know).

Most rims are made of 6061 or a similar alloy. The differences between the 6000 series alloys are small, though not always trivial. Many alloys in this family are not as strong as 6061 for various reasons. These are often easier to extrude and heat treat. Some alloys are potentially stronger than 6061, but are very sensitive to heat treatment process variables and can end up weaker due to the limitations of "press quenching" aluminum to heat treat it.

All rims are press quenched as the first heat treatment step. As the aluminum is extruded, it leaves the extrusion press at a high temperature. It is cooled quickly from this temperature, using water spray or air cooling. Then the material solution treated material is aged in a pizza oven (or the industrial equivalent). This step is called precipitation hardening. If the material is quenched quickly as it leaves the press, it is called T6 material when it's done. If it is cooled more slowly, it is called T5 (T5 is only used for a few alloys. They are not as strong as 6061 in most cases and are only used to simplify production rather than optimizing mechanical properties - 6005 is often supplied in the T5 condition). Press quenching is not an ideal way to heat treat aluminum, but it's all you can do since the rims are too delicate to heat treat after they are rolled and joined. (We tried). It is a nightmare of distortion when it comes back, though it's strong as hell.

The problem with press quenching is that the exit temperature of the extrusion out of the press cannot be controlled in an ideal way. The range of exit temperatures is tool large. The ideal mechanical properties of high strength aluminum is achieved with very close controls of quench rates and quench temperatures in the heat treatment process (optimal T6 practice). Claims to have a T6 aluminum don't mean much. Strength differences on the order of 10 to 15% can be seen with small variations in press practice on 6061 and related alloys.

Bottom line: Heat treated is a trivial claim since everything is heat treated. The details of the process determine the strength of the rim, and these are often too complex or subtle to discuss in marketing literature.


6. Double eyelets

Double eyelets are a good way to reinforce the spoke bed area of a very light box section rim. They distribute the load from the spokes to both horizontal walls of the rim, rather than leaving the lower wall to resist it by itself. But they are expensive to install and they are losing ground to other in off road rims. Multi-cavity cross sections, where there are inner walls to do the job of the double eyelets, and more sophisticated box sections where the walls that support the spoke bed of the rim are configured to resist spoke loads better are taking over. Examples of the former are our Red and Blue rims, and new Mustang units, Ritchey Rock, M231 Mavics, WTB powerbeams and some of the newer Sun Metal shapes. Humbly speaking, our's is the better design of course.

A Mustang rim cross section. The rim is supplied with a single eyelet, and the inner walls support the spoke bed by distributing the spoke load between the upper and lower walls of the rim.

The figure at the right shows an improved box section with a single eyelet. Note that the walls that connect the spoke bed to the outer braking walls are rotated around closer to a parallel orientation to the spoke axis. They are not parallel to the spokes, but as they incline in that direction, the load on them changes from a (primarily) bending load to a mix of bending and tensile loads. Tensile stresses are more desirable in this case.

The box section of 217 Mavics is a good example of the improved box section designs. These are better at resisting spoke pull than the older box designs (like wider, less sophisticated design used in old Ritcheys, Suns, and Arayas - all abandoned now) because the lower wall is angled to align better with the load. But they are not as strong overall as the multi-cavity shapes when they are made with single eyelets. The thin inner walls and spoke beds are prone to fatigue failures.

I oversimplified the discussion above a little. The advantages of reinforcement in this area, like we design into our multi-cavity rims and like what is achieved with double eyelets, is not simply to resist spoke pull through. There are increases in lateral strength in the wheel assembly when spoke tension is resisted and not allowed to elastically deform the rim in the vicinity of the eyelet as the wheel is loaded. If the spoke bed is not rigid, and the spoke nipple seat area is pulls in toward the center of the rim when the spoke tension increases under load, the rim can deflect to the side slightly further. It is about the same as a lighter gauge spoke. The onset of a tacoed wheel is determined by both the yield strength of the rim and by the lateral deflection of the rim. There is a critical point in the lateral deflection, a second minimum potential energy state if you want to get fancy about it, that the wheel wants to be in once the rim is pushed a certain distance laterally, and a rigid rim gets to this point at a high lateral load - it's stronger because it's stiffer, and this is not a common characteristic of mechanical structures.

The spoke bracing angle affects the lateral stiffness and strength of a wheel too. This is why wheels that are dished more are weaker laterally. This is why an 8 speed wheel is weaker laterally than a 7 speed wheel, and why an asymmetrical rim cross section (this is an informal product announcement by the way - you heard it first here!) can be made to increase lateral strength.

Which of these rim features are actually functional, and which are marketing gimmicks?



Don't believe the hype. I think you can safely believe most of what I wrote above, I'm not selling (with one or two tongue in cheek exceptions). There is nothing there that requires hand waving to defend. Check it out.