Cycling UK Forum

Oh no!
Frame stiffness is the hoariest of old chestnuts, we debate it all the time on here.
I'm no engineer, but I have been "designing" Spa's frames for a while now, I have some willing volunteers to test ride them, and I'm beginning to formulate some ideas of my own (shock horror)

With apologies to all the followers of Jan Heine's "planing" theory, I like my bikes pretty stiff laterally at the bottom bracket, so that when I pedal the bike goes forward, rather than the bottom bracket flexing to the side. Trouble is, a bike frame is almost 2-dimensional, there is not much lateral bracing that you can do.
Similarly, I think "vertical compliance" is just about a no-go, a bike frame is fairly well braced and triangulated vertically....some flex occurs, but not much.
So, apart from taking up stamp-collecting, what can we do?
My thought processes begin with designing frames in Titanium....bear with me on this! I think a really major difference between steel and titanium frames is that the chainstays and seatstays are virtually the same size in both materials....yes, thats right, that is a difference, because titanium is less stiff than the same size chunk of steel, and about as stiff as a similar weight chunk of steel. However, because there is a restricted amount of space between the chainrings and the rear tyre, chainstays tend to be about the same size in either material, that is 28 x 18mm oval, give or take a millimeter or three, with a wall thickness of 0.9mm for titanium and butted steel tubing being maybe 0.8/0.6mm wall. ....not a staggering difference. Seatstays tend to be pretty similar dimensions in steel or titanium too, I suspect because thats what we are used to looking at.
So, we should expect titanium chainstays to be less stiff than steel, and give less lateral bracing to the bottom bracket. (The frame tubes can be made bigger diameter in titanium, to gain some stiffness.) I said before, I'm no engineer, but this is my shaky understanding of the stiffness of tubes....if you double the diameter of a round tube keeping the same wall thickness, the "doubled up" tube will be something like eight times as stiff, not twice as stiff. Stiffness goes with the cube of the diameter....roughly.
Which brings me to a thought.....I (now) think that denting in the chainstays to clear the tyre and/or chainwheel is a pretty silly thing to do, because it reduces the effective size of the tube and makes it more flexible laterally, when what I want is lateral stiffness in that area.
I can't put a number on how much more or less stiff, but I can give an example...
We built some titanium prototypes, with the frame tubes chosen to more or less match the sort of stiffness you would expect in a fairly lightly-built "audax" steel frame, with the difference that the chainstays were bent so they angled between the tyre and the chainring, in preference to flattening the chainstays. Also, just to make it more complex, we also ovalised the bottom bit of the down tube so that it is wider than it is tall, and should be stiffer laterally than vertically. These bikes, although not "racy" in general design are laterally stiff and accelerate really quickly, the difference you get by changing the profile of 3 tubes is really quite staggering.....they are laterally stiffer than "similar" steel bikes with the chainstays conventionally dented for clearance, but even with forks for disc brakes they are pretty comfortable (on 32mm tyres)
.......picture.....



thats a 32mm tyre, so there is lots of clearance, and the chainstays are 440mm long, not exactly a conventional racing set up.

They tell me that tapered carbon steerers are the future, so a bigger head tube is wished upon us, in order to fit a 1 1/2" bottom bearing. But even this cloud has a silver lining, in that it means I can ovalise the top tube where it joins the head tube so its wider than it is tall, giving lateral stiffness, and on big frames with a big downtube the front end of the downtube can be left round instead of ovalising it vertically just so it can be welded to the head tube. All this should give greater lateral stiffness against shimmy, and at the same time not give rise to unwanted vertical stiffness.

So thats my current thoughts, really.....to try to get the right degree of stiffness, in the right plane, by optimising the sizes and profiles of the tubes, within the constraints of the materials available and the necessary physical shape of the bike.
I'm thinking of calling the bike "Damascus"......
I'm sure anybody with an "Engineering" turn of mind can drive a coach and horses through my approximations....

I find that all an interesting read. I thought the idea of tapered head tubes was purely to create stiffness by allowing the larger downtube to interface directly to the headtube. This is how you describe it but it read as a secondary effect forced upon us by the trend for carbon steerers. Anyway the question i want to ask is pictured in your image, that is do you feel the chainstay bridge has an important part to play in adding stiffness? I ask because prior to buying a condor frame i had been on the edge of buying a titanium road frame. I noticed it did not have a chainstay bridge, this put me off a little as titanium having more flex, it concerned me that it may exaggerate that flex. A website had some cfd simulations with bridge and without and they show a change in stiffness at the bb. From your experience do you feel it helps, aside from its mudguard fixing function? The frame i viewed was from a mainline manufacturer, in a medium 56cm

The quadrilateral formed by the bottom bracket shell, the 2 chainstays and the chainstay bridge is a very stiff bit of frame, i think the significant chainstay flex occurs between the chainstay bridge and the wheel....on the prototype pictured, there is more than enough mudguard clearance, in the production bikes we will be moving the chainstay bridge 10mm towards the axle, giving 10mm less "unsupported" chainstay to flex.

So if we remove the bridge altogether as the in the frame i viewed it would make bb flex worse? On your frame alls well but it shocked me to see it missing on a titanium frame. My condor deda tubed frame does not have one but i feel the tensile strength and tube diameters\profile then play their part. It is very stiff. As you say it forms a quadrilateral at the bb enhancing stiffness.

A little earlier today I was reading Jobst Brandt opining on how chainstay bridges add no stiffness and are only there to prevent the wheel getting jammed back in horizontal dropout times. I think his view was that the pair of stays attaching to the bb were sufficiently stiff already and the bridge was not helping much. One of Jobst's frames had lateral stiffeners brazed to the side of the seat tube, supplementing the lug. I have wondered previously if stiffeners extending from the seat tube and down tube to the bb shell would be of any value. Or ovalising the ST and DT so they attach further across the BB. Essentially any way of increasing their contribution to reducing bb flex. I suppose given your view that lateral stiffness is more important ovalising the DT would make more sense.

old_windbag wrote:So if we remove the bridge altogether as the in the frame i viewed it would make bb flex worse? On your frame alls well but it shocked me to see it missing on a titanium frame. My condor deda tubed frame does not have one but i feel the tensile strength and tube diameters\profile then play their part. It is very stiff. As you say it forms a quadrilateral at the bb enhancing stiffness.

I have seen a few frames with short chainstays where the chainstay bridge was omitted....if the stays are very short there is really little room for a bridge.
On a frame with longer stays I think omitting the bridge is a mistake.

I would have thought the chainstays were loaded mainly in tension and compression* (like members of a space frame) and that the lateral rigidity of the rear would therefore not be much affected by denting the stays.

Certainly some of the lateral stiffness comes from the ‘rear triangle’ actually being two rear triangles spaced wider at the hub than at the bottom bracket (i.e. tapered). This forces the chainstays to change length to let the back wheel move laterally. I suppose the OLN distance is 135 mm here, which would be automatically slightly stiffer than 130 mm.

Not sure if you were serious about the Damascus name, but I think that would be a great name for a bicycle of this type! Spa Damascus has a nice ring to it. Mind you, the road to Damascus is pretty war-torn these days – hope that wouldn’t complicate things.



* Well, mainly tension with a rider sitting on the frame.

alexnharvey wrote:A little earlier today I was reading Jobst Brandt opining on how chainstay bridges add no stiffness and are only there to prevent the wheel getting jammed back in horizontal dropout times. I think his view was that the pair of stays attaching to the bb were sufficiently stiff already and the bridge was not helping much. One of Jobst's frames had lateral stiffeners brazed to the side of the seat tube, supplementing the lug. I have wondered previously if stiffeners extending from the seat tube and down tube to the bb shell would be of any value.

Brandt's seat tube...http://forum.ctc.org.uk/viewtopic.php?f=5&t=104401&p=992831&hilit=brandt#p992831
Brandt's book on wheels is inspirational....its a lesson in how to build wheels, why to build them like that, how to write a text book, how to analyse a complex subject, and so forth. It feels disloyal, but I have to say that I don't think everything Brandt wrote subsequently quite measured up.
Working with titanium, you can ovalise the down tube, which is more elegant than adding bits, and possibly also more effective. You can also ovalise the bottom of the seat tube, as long as you don't give the front mech. an oval bit of tube to mount onto.

Samuel D wrote:I would have thought the chainstays were loaded mainly in tension and compression* (like members of a space frame) and that the lateral rigidity of the rear would therefore not be much affected by denting the stays.

Certainly some of the lateral stiffness comes from the ‘rear triangle’ actually being two rear triangles spaced wider at the hub than at the bottom bracket (i.e. tapered). This forces the chainstays to change length to let the back wheel move laterally. I suppose the OLN distance is 135 mm here, which would be automatically slightly stiffer than 130 mm........................

* Well, mainly tension with a rider sitting on the frame.

This stuff makes my head spin. I can tell you that the prototypes I posted a picture of are noticeably stiff, and noticeably quick. Several people whose opinion I respect have ridden them. I know it sounds like marketing hyperbole....all I did was change the profile of 3 tubes.

colin - are you familiar with columbus max steel tubing? as i read it your original post outlines ideas similar to that re. tube profiles.

mig wrote:colin - are you familiar with columbus max steel tubing? as i read it your original post outlines ideas similar to that re. tube profiles.

Sadly no. It would be nice to be given a big pot of money and be told to "Go away and learn....." Not really going to happen, I think.

Colin, I don’t for a moment think it’s marketing hyperbole or that you’re mistaken about the stiffness you observed.

I’m curious how much stiffness comes from the space-frame-like properties of the rear triangle and how much comes from the tubes resisting bending. My intuition may be wildly off!

Another thought: a stiffer main (front) triangle may hold the bottom bracket squarer under pedalling forces, thus making the rear laterally stiffer because it’s sitting on a stiffer foundation.

I once saw a finite-element analysis of a bicycle frame under pedalling forces and all I can remember is that (a) the whole system was remarkably interconnected, and (b) it made my head spin.

I volunteer to test ride the latest bike

531colin wrote:

mig wrote:colin - are you familiar with columbus max steel tubing? as i read it your original post outlines ideas similar to that re. tube profiles.

Sadly no. It would be nice to be given a big pot of money and be told to "Go away and learn....." Not really going to happen, I think.

http://www.ciclicorsa.com/en/negozio/serie-tubazioni-max/

i have a frame with race clearances built from this tubing. the effects of acceleration you mention in your own frame are evident in my own.

if it is not clear from the pictures and information in the link then the frame's main tubes are ovalised along the length with each end having the oval in a different plane. hence the top tube is ovalised 'horizontally' where it meets the seat tube but 'vertically' as it meets the head tube. similarly with the down tube. the chain stays are ovalised 'vertically' as seen in the photo. the tubes are oversized too (apart from the seat stays.)

my frame is nearly 20 years old from back when i raced regularly. i have some photographs of it somewhere if you are interested. i think there is also a website dedicated to frames built in this tubing with goods shots of how it 'works.'

Samuel D wrote:CUT
I’m curious how much stiffness comes from the space-frame-like properties of the rear triangle and how much comes from the tubes resisting bending. My intuition may be wildly off!

Another thought: a stiffer main (front) triangle may hold the bottom bracket squarer under pedalling forces, thus making the rear laterally stiffer because it’s sitting on a stiffer foundation.
CUT

When pedaling hard my notsostiff titanium bike I can clearly see the BB moving sideways under the power of my 63y.o. legs but the rear tyre does not move at all in the rear fork.
That's why I think that stiffness of the rear triangle plays no significant role in the picture but much more the lateral stiffnes of the complete frame from head tube to rear axle.
Local stiffer parts may make a small difference (F.I. ovaling down tube to BB) but still the whole frame can bend.
BTW I don't bother the least about stiffness of frames, it plays only a tiny role during sprints and I'm not even sure it's a negative role...