Bicycle Mechanics - Steel vs alu scientific result

Hi, this subject has been talked over and over and over in this forum. Don't want to start a debate, just curious about measurable, scientific approach to the frame material subject. Is there a test showing how much a steel frame flexes, compared to an aluminium one? mm travel per N of force, etc, on two similarly made frames, one alu, one steel.

Your question is moot because the ideal frame design for each material varies.

Further, the parameters are affected by other choices such as method of construction (welded, lugged and brazed, fillet brazed, bonded) and specific tubing choices eg normal cromo or 853, or hyrdoformed ally versus butted and ovalised.

In short, there's no point considering identical frames made of different materials, because at least one frame will be making poor use of its material.

The closest you can get to a fair comparison is the output of a high-end builder equally exerienced and talented with each metal.

And, as anyone who's looked into the subject can tell you, the frame's design is far more important than what it's made from, in the same way that buid quality is the most important component of a wheel.

this is a good place to start reading:

http://sheldonbrown.com/rinard/EFBe/frame_fatigue_test.htm

Sheldon does a reasonable job with translation and describing the issues.

here's a more recent version of the results (in English)

http://www.efbe.de/testergebnisse/fulltest/enindex.php?typ=1&sort=15

the actual testing methods are here:

http://www.efbe.de/pruefservicenew/enindex.php

also, it is Germany, so I can't imagine anywhere else doing more rigorous tests (as far as I know there are no specific minimal requirements in the US).

also, it's quite a reasonable cost, €2k for a full frame test.

Yes there is. It's called maths. You'll need to know the Young's modulus ("stiffness") of the two materials you want to compare, the geometry of the frame and the dimensions of all the tubes, as well as the point you intend to apply the force and the direction you intend to apply it in (i.e. is it flex under pedalling, or under the weight of the rider that you're wondering about?).

Since a lot of frames these days are made of oddly-shaped tubing, it's likely to be really quite complicated to work all of this out, but it's a fairly simple principle. It'll give you the deflection in the frame for a given force.

Yes there is. It's called maths. You'll need to know the Young's modulus ("stiffness") of the two materials you want to compare, the geometry of the frame and the dimensions of all the tubes, as well as the point you intend to apply the force and the direction you intend to apply it in (i.e. is it flex under pedalling, or under the weight of the rider that you're wondering about?).

Since a lot of frames these days are made of oddly-shaped tubing, it's likely to be really quite complicated to work all of this out, but it's a fairly simple principle. It'll give you the deflection in the frame for a given force.

actually there's something called empirical testing and that what the links I provided did. the inherent attributes of the material are irrelevant when they're used in a bicycle frame ... the entire frame must be tested ... as per above.

the inherent attributes of the material are irrelevant when they're used in a bicycle frame

Oh dear....

Oh dear....

of course, as the points are failure to to be elsewhere, as my links clearly state. most failures are within a few cm of a modification to the tube (i.e. a weld, a boss, a hole) etc...

for all practical purposes, Young's modulus doesn't matter, the modifications to the frame material matter much more. that is, of course, unless you just want to intellectually wank it to arguing which material has "more desirable" properties ... which is a useless metric as it pertains to a manufactured/assembled frame.

if you can't see that, then feel free to intellectually masturbate.

of course, as the points are failure to to be elsewhere, as my links clearly state. most failures are within a few cm of a modification to the tube (i.e. a weld, a boss, a hole) etc...

for all practical purposes, Young's modulus doesn't matter, the modifications to the frame material matter much more.



I took a 20 year old MTB for a ride. Salvaged old friend's bike. Intend to use it in the winter. It is a steel frame. It felt so much more comfortable on bumps, than my aluminim treking bike. I was sure it was down to a lot fatter tyres, but it got me thinking. How good a frame can be in absorbing shocks, making ride comfortbale? I heard lots of talk about steel being more comfy, then some people swear in carbon fiber, but my personal feeling is that it's 99% down to fit and tyre thickness/pressure.

This test, however, shows that carbon fiber and steel are less stiff. Still not sure how much it influences ride comfort, but it must make some difference.

It's also comforting to know that alu bikes don't have to break apart. Not sooner than steel bikes.

I'm not sure about alu rusting? I thought it doesn't rust besides surface oxidation. Here it says: "...some aluminum alloys typically used in bicycle frames are comparatively susceptible to so-called "grain boundary corrosion", which can operate very briskly in the material and so cause a break..."

I took a 20 year old MTB for a ride. Salvaged old friend's bike. Intend to use it in the winter. It is a steel frame. It felt so much more comfortable on bumps, than my aluminim treking bike. I was sure it was down to a lot fatter tyres, but it got me thinking. How good a frame can be in absorbing shocks, making ride comfortbale? I heard lots of talk about steel being more comfy, then some people swear in carbon fiber, but my personal feeling is that it's 99% down to fit and tyre thickness/pressure.

This test, however, shows that carbon fiber and steel are less stiff. Still not sure how much it influences ride comfort, but it must make some difference.

It's also comforting to know that alu bikes don't have to break apart. Not sooner than steel bikes.

I'm not sure about alu rusting? I thought it doesn't rust besides surface oxidation. Here it says: "...some aluminum alloys typically used in bicycle frames are comparatively susceptible to so-called "grain boundary corrosion", which can operate very briskly in the material and so cause a break..."

I agree with your assessments. Most of the forces transferred from road to rider have more to do with the non-frame components of a bicycle.

For example, if one compares the vibrations transmitted from the handlebars to the riders hands/arms via the hand grips ... you've have two different intensities of transmission via two similarly shaped grips:

GP5 BioKork: http://www.ergon-bike.com/us/en/product/gp5-biokork

GP5 Standard: http://www.ergon-bike.com/us/en/product/gp5

same with the tires, wheels, pedals, seat posts, handlebars and hand grips. although I haven't seen an empirical test, i would bet the bank that the items are responsible for a greater percentage of total variability in ground-rider vibration transmission that the frame itself.

i do find the German testing quite interesting. not only the testing methods, but also the consumer's desire to see the test results, and the average German's compulsion for owner the "highest quality" product at a given price point.

i do find the German testing quite interesting. not only the testing methods, but also the consumer's desire to see the test results, and the average German's compulsion for owner the "highest quality" product at a given price point.

:))) Yes. Typically German. Funny and admirable at the same time. :)

Those test results are interesting. I'm just not sure they're easily applicable accross the board to all models made of those materials regardless of price point. I checked a couple of those frames - they were $1,500 mtb bikes. Since the averge consumer in NA typically buys a $400 to $600 hybrid - it might be of more practical value to see tests done on items in that price range.

Yes there is. It's called maths. You'll need to know the Young's modulus...You can't use "maths."

A frame is too complex to model. It's not like predicting how far a tensile test specimen will stretch. You could say we don't yet have the maths.

A computer could break it down into a bunch of simplifications and compute a result using finite element analysis. So could a person but it would take forever.

Sorry old chum, couldn't resist.

An inexpensive steel bike can be fairly stiff.

A bike made from 28mm tubing with 0.6mm thick walls can feel quite flexy especially if you exceed the rated weight by 70lbs.

A good aluminum frame feels stiff but it's in the direction you want. The steering seems to respond quicker and even standing up and hammering produces no feeling of flex. However, if you have a carbon fork and stays, it can ride better than the one made from the 0.6 mm tubing by far. I hate that they're going away from carbon stays.

My 2 cents.

actually there's something called empirical testing and that what the links I provided did. the inherent attributes of the material are irrelevant when they're used in a bicycle frame ... the entire frame must be tested ... as per above.

Curious, so presumably I could make a frame out of leaves or something, if I did a really good job of it?

Curious, so presumably I could make a frame out of leaves or something, if I did a really good job of it?

Don't know about leaves, but certainly bamboo.

essentially one can make a bicycle frame out of anything solid with reasonable properties ... several companies make frames out of "plastics."

although common sense for most, it should be stated, that the individual characteristics of the original materials pales in importance to how those materials are "joined/connected" to make a complete frame.

and, leaves would be fine for a bicycle frame as long as they were joined/held together by some type of resin.

plastic bike.

285046

how do you think the characteristics of those plastics compare to steel/aluminum alloys? probably not so well, eh? it's about the design/connection

full story:

http://www.gizmag.com/frii-recycled-plastic-bike/19337/

it would be great to have two of those plastic bikes, snap them apart and have a picnic bench/table for a brunch next the river, then snap them back into bikes and ride home.

that's more important to me than Young's modulus.

oh, and, the table better have two champagne bottle holders.

If you were to take two bike of the same material, brand and model but in different sizes (eg S and L), then a deflection test would give different results. It almost always true that bigger frames have more deflection for any given force, compared to smaller frames. This is the inverse of the requirement, where big frames are ridden by stronger, heavier riders.
You could build big frames with identical deflection or even better, with less (ie a stiffer frame made with fatter tubes) but it is expensive because you need to hold more stock in tube types. I know that Principia used to build like this.

Is there a test showing how much a steel frame flexes, compared to an aluminium one? mm travel per N of force, etc, on two similarly made frames, one alu, one steel.

Steel and aluminum frames can be designed to have nearly identical stiffness, but since steel is much denser than aluminum, the steel frame will tend to be heavier and have thinner tubes. But your query presupposes that "stiffness" should be the over-riding design criterion in a bicycle frame. Is it really?

OP, see what you have done?

This is the internet, where people talk **** about stuff that they have no clue about.

it would be great to have two of those plastic bikes, snap them apart and have a picnic bench/table for a brunch next the river, then snap them back into bikes and ride home.

that's more important to me than Young's modulus.

oh, and, the table better have two champagne bottle holders.I'm depressed now.

I am sure you would experience a substantial savings on your train fare if you took enough picnics.

:(

I'm depressed now.

I am sure you would experience a substantial savings on your train fare if you took enough picnics.

:(

I live right next to the river, no train fare required :)

However there's some nice vineyards upstream and downstream that make great spots for picnics :D

It's impossible to generalize, because while the materials have different properties, the manufacturers can alter dimensions such as tube diameter and wall thickness to compensate and produce a frame with the ride characteristics of their choice.

It's not like there are ideal characteristics where more is good and most is best, except maybe for weight, where least may be best. Good frame design is a balance of characteristics, some of which conflict, to give the best overall ride. This can be done effectively with any material.

plastic bike.

285046

how do you think the characteristics of those plastics compare to steel/aluminum alloys? probably not so well, eh? it's about the design/connection

full story:

http://www.gizmag.com/frii-recycled-plastic-bike/19337/

Here we go, another design student dreaming something up while being almost entirely clueless about the subject:"There are no external brakes... Peleg says that he sees such matters being handled in a similar way to the way BMX hub braking operates - when a rider pedals backwards, the bike comes to a halt."

Coaster brake anyone? And when did you see that lately on any BMX intended for serious use?

And then there's this:"The uncomfortable-looking saddle would be made in different sizes to suit different riders and, like much of the design, would snap into place."

So, assuming the saddle also includes the same-coloured bracket that attaches it to the frame, all in one piece - how many different saddles would a seller be forced to stock in order to offer a decent level of adjustability? 10? 20?

For anyone interested, do a net search for "itera plastic bike". Apart from starting from recycled plastic, that bike has already been made. Admittedly, what really killed that one was probably poor business decisions, but it wasn't a very good bike even at the best of times.

You can't use "maths."

A frame is too complex to model. It's not like predicting how far a tensile test specimen will stretch. You could say we don't yet have the maths.

A computer could break it down into a bunch of simplifications and compute a result using finite element analysis. So could a person but it would take forever.

Sorry old chum, couldn't resist.FEA works, however the problem is deciding what forces to input. It's too hard to model. Hence bikes with metal tubes are mostly built by rule of thumb, developed over time. Shaped tubes are modeled with FEA, but tested by humans, then modified further. Same with carbon bikes, but they're easier. They're modeled with FEA, built, ridden, and then material is added or subtracted to that same frame until the rider/engineer decides that's right. Then they mold another frame with that layup and ride it again.

If it really Matters a Lot,
.. maybe you can do a post graduate degree in engineering, materials analysis,
and write a Masters Thesis on this ?

maybe someone at MIT or Cal Tech has done that, IDK.

then a link to that Publication can be found..

Here we go, another design student dreaming something up while being almost entirely clueless about the subject:"There are no external brakes... Peleg says that he sees such matters being handled in a similar way to the way BMX hub braking operates - when a rider pedals backwards, the bike comes to a halt."

Coaster brake anyone? And when did you see that lately on any BMX intended for serious use?

Rear pegs, still good for people transport if you don't have a rack.


And then there's this:"The uncomfortable-looking saddle would be made in different sizes to suit different riders and, like much of the design, would snap into place."

So, assuming the saddle also includes the same-coloured bracket that attaches it to the frame, all in one piece - how many different saddles would a seller be forced to stock in order to offer a decent level of adjustability? 10? 20?

For anyone interested, do a net search for "itera plastic bike". Apart from starting from recycled plastic, that bike has already been made. Admittedly, what really killed that one was probably poor business decisions, but it wasn't a very good bike even at the best of times.

Crotch mold, 3D scan, 3D print ... seems pretty easy.

You can't use "maths."

A frame is too complex to model. It's not like predicting how far a tensile test specimen will stretch. You could say we don't yet have the maths.

A computer could break it down into a bunch of simplifications and compute a result using finite element analysis. So could a person but it would take forever.

Sorry old chum, couldn't resist.

just curious, but what do you think the FEA program is doing, voodoo? it's all math. if you do it by hand; it's math. if you model it in a computer and hit run; the computer is still doing math.

just curious, but what do you think the FEA program is doing, voodoo? it's all math. if you do it by hand; it's math. if you model it in a computer and hit run; the computer is still doing math.Actually, it models highly complex geometry as a series of simplifications and then solves thousands and thousands of interrelated equations. No way a person could do all that. Not enough time, no chalkboard or paper big enough to keep track of all the equations.

FEA works, however the problem is deciding what forces to input. It's too hard to model. Hence bikes with metal tubes are mostly built by rule of thumb, developed over time...I think they can do pretty good with FEA. They can't model the real world but then again if something really wierd and extreme happens to your bike, you're likely not to hold it against the manufacturer. You probably can't sue.

Steel frames have been developed with tradition born of long experience, but it's a working method. Plus, people expect and accept tradition in a steel frame.

Here we go, another design student dreaming something up while being almost entirely clueless about the subject:"There are no external brakes... Peleg says that he sees such matters being handled in a similar way to the way BMX hub braking operates - when a rider pedals backwards, the bike comes to a halt."

Coaster brake anyone? And when did you see that lately on any BMX intended for serious use?

And then there's this:"The uncomfortable-looking saddle would be made in different sizes to suit different riders and, like much of the design, would snap into place."

So, assuming the saddle also includes the same-coloured bracket that attaches it to the frame, all in one piece - how many different saddles would a seller be forced to stock in order to offer a decent level of adjustability? 10? 20?

For anyone interested, do a net search for "itera plastic bike". Apart from starting from recycled plastic, that bike has already been made. Admittedly, what really killed that one was probably poor business decisions, but it wasn't a very good bike even at the best of times.

Yup - had one of those come in the shop last year. Still completely functional - outlasted the company! :lol: It was unusual unough that I took pictures - later in the week I'll see about uploading something.

Actually, it models highly complex geometry as a series of simplifications and then solves thousands and thousands of interrelated equations. No way a person could do all that. Not enough time, no chalkboard or paper big enough to keep track of all the equations.

Solving a load of interrelated equations isn't maths?

OK, it's an approximation, but it's still maths.

Actually, it models highly complex geometry as a series of simplifications and then solves thousands and thousands of interrelated equations. No way a person could do all that. Not enough time, no chalkboard or paper big enough to keep track of all the equations.
Correct but isn't that why we have computers? There are a huge number of mathematical techniques that are completely impractical to do manually but take only seconds with a computer. FEA is one of those techniques.

Back in the '60's I recall a camera lens manufacturer designing a very complex zoom lens using a laborious but very accurate "ray tracing" technique that they estimated would have taken 20,000 people, 20,000 years to do manually but took only a weekend using the computers of the time. These days it probably would take less than a minute.

Solving a load of interrelated equations isn't maths?

OK, it's an approximation, but it's still maths.We don't have the math to model complex shapes precisely.

FEA programs simulate a complex shape as an interconnected lattice of many, many simple members.

If you want it to be a reasonable approximation you often have to make the mesh so fine that it would not be humanly possible to solve all the equations.

So we can't solve it with math. We can approximate it with simplifications, but there is so much math it would take forever.

We don't have the math to model complex shapes precisely.

FEA programs simulate a complex shape as an interconnected lattice of many, many simple members.

If you want it to be a reasonable approximation you often have to make the mesh so fine that it would not be humanly possible to solve all the equations.

So we can't solve it with math. We can approximate it with simplifications, but there is so much math it would take forever.

OK, we can't solve it precisely with maths, the point is, we can come up with a decent approximation using maths.

http://2.bp.blogspot.com/-r7vSo4tM_pI/TgYdcof_OnI/AAAAAAAAD0Y/bBnSxAhJog8/s1600/cats%2B-%2Bmasturbation.jpg

I think it's more pertinent to ask whether there's a test that gives a real world result which translates into whether a bike has a good ride or how much the flex affects steering or power transfer.

They have tested a few frames for millimeters of flex under laboratory conditions but even if every frame came with a table of flex values it would still mean very little.

I don't think there is such a real world test save for subjective analysis by bike authors.

The original jerkoff's question doesn't have much point.

The original jerkoff's question doesn't have much point.

The OP asked a simple and IMO legitimate question. He specifically said he wasn't looking to start a debate. I don't think it calls for name calling, especially from someone who seems only interested in prolonging the debate that wasn't called for in the first place.

Of course you'll want to have the last word, and as far as I care, you're welcome to it.

The OP asked a simple and IMO legitimate question. He specifically said he wasn't looking to start a debate. I don't think it calls for name calling, especially from someone who seems only interested in prolonging the debate that wasn't called for in the first place.

Of course you'll want to have the last word, and as far as I care, you're welcome to it.I can expand on his point.

I can point out that those who live in glass houses shouldn't throw stones.

So there.

Rear pegs, still good for people transport if you don't have a rack.

I have no idea how that is meant to tie in to what I wrote.


Crotch mold, 3D scan, 3D print ... seems pretty easy.

Oh, sure.
In a world where it's real easy to find shops who won't have tension meters, won't use torque wrenches, won't have loaner programs for finding the right stem length or a suitable saddle - and you think 3d-printers and associated hardware is reasonable to expect?

Now, I'm all for a way of turning recycled plastic into something more useful, but inventing a new, but less adjustable and less compatible way to attach a saddle seems rather pointless. He might as well have gone for a brand new wheel size.

:rolleyes:

http://en.wikipedia.org/wiki/Finite_element_method

If every maker tested their frames for how much a standardized load makes the front and rear axle move, and how much a torsional load like standing on one pedal makes the bottom bracket deflect, and how much the bars deflect from a standardized load, then you could compare frames, but only if the test was standard.

If every maker tested their frames for how much a standardized load makes the front and rear axle move, and how much a torsional load like standing on one pedal makes the bottom bracket deflect, and how much the bars deflect from a standardized load, then you could compare frames, but only if the test was standard.

All that might be somewhat helpful, but it'll never happen. In any case you'd still end up with apples and oranges comparisons. Ie, one has a stiffer BB/seattube, the other less head twist in hard cornering. Given all the variables involved, the only meaningful test for frames or bikes remains a real world road test. There are just too many variables, not the least of which is the engine/payload.

In a world where it's real easy to find shops who won't have tension meters, won't use torque wrenches, won't have loaner programs for finding the right stem length or a suitable saddle - and you think 3d-printers and associated hardware is reasonable to expect?


I'm not sure where you live, but bike mechanics make 30-50€/hr and we use the tools to ensure that everything is within proper specs. There's also no drop-in service here, per-arranged appointments are mandatory.

I have theorized that it is impossible to bring up the topic of
frame materials on Bike Forum without starting a rather acrimonious
and often ill informed debate.

As another point of experimental data, this thread brings me that
much closer to scientifically publishable results. Thanks to OP and
all participants...............kudos

I'm not sure where you live, but bike mechanics make 30-50€/hr and we use the tools to ensure
that everything is within proper specs. There's also no drop-in service here, per-arranged appointments are mandatory.

I know it's hard to believe, but there are many differences between
Germany and the USofA. Prevailing wage scales for bike mechanics
is only one of them.

All that might be somewhat helpful, but it'll never happen. In any case you'd still end up with apples and oranges comparisons. Ie, one has a stiffer BB/seattube, the other less head twist in hard cornering. Given all the variables involved, the only meaningful test for frames or bikes remains a real world road test. There are just too many variables, not the least of which is the engine/payload.

Another reason it would never happen is manufacturers would have to get sharper. Anyone could see that a larger frame or a wider bar was flexier than a small one...and they'd demand the manufacturer use a different tubing for each size....even if sometimes it's not an important difference.