You’ll get too much flex at the tube junctions, so the welds will fail. But go ahead and prove us all wrong!

Might be worth trying a curved tube down from the top of the seat tube to the down tube, probably look horrid though. You’d get some flex in the seat tube, but it would still be triangulated.

If you want some decent FEA doing (i.e. not the SolidWorks solver) I have an MSC Patran (technically Nx) license that’s good for linear stuff. If you sent me the model, constrains and materials of each I could perhaps have a bash at it.

I’d be looking at some sort of a flexure linkage. No friction, no worries about mud, and you control the flex by the size of the hinge.
Plus compliant mechanisms are cool!

I’d look at a trunion type interface between the stays and the seat-tube.

So a bracket to create a vertical slide on the seat-tube, with stub shaft’s either side to create a 90deg pivot through the stays.

Like some others have said the IGUS range (i’m sure I pointed oliverracing towards these at SSUK) would be good bearings for this, a Drylin for the vertical and there top hats for the stays.

Might be an idea, put a split sleeve slide surface where your original brass sleeve was – something like 17-4 stainless or hard anodised aluminium could be good for this.

I can’t imagine any of these ideas working out first go, without a number of prototypes and potentially a few failures. And adding more weight/complexity than its worth.

Maybe try something a bit more like the Volagi Viaje?

Volagi Viaje

Maximus, very kind FEA offer, but still very early days. I’ll bear it in mind if I decide to try and persue it further, however as many have pointed out it’s unlikely to be something which works on first attempt. As I’m not planning on manufacturing, marketing and selling these, multiple prototypes for a single final frame could end up being a rather expensive and time-consuming process 😀

If I do forge ahead I’ll probably go the boring route and making a variation to the WTF, which I notice also has a similar chainstay to the Santos for tyre clearance.

Cheers, Rich

Maybe you’re overcomplicating things with all these linkages and bearings why not simply have a bracing tube between the downtube or head tube and and seat tube, a bit like these old kona frames:

Use that to constrain the seat tube, and then rather than have a conventional top tube, use your double curved arrangement without any, connection to the seat tube as a means of making your soft tail work…

One extra tube Vs elastomers/bushes/linkages etc…

Turbo,

Id be more than happy to. I am also about to do an FEA validation on a space frame chassis, and a bike frame isn’t that far off (and an awful lot simpler). I will be doing some for various things I’m attempting so it wouldn’t be any bother at all. A lot easier to do multiple FEA than to build multiple prototypes.

I N R A T S but

urboferret – Member

Just because I can, here it is in slightly jerky motion, although in reality the seat tube would be moving backwards and forwards slightly at the same time

It’s be flexing back and forth like billy-oh IMO, way more than the top tubes will move up and down.

V creative!

Better yet, look at the way this old schwinn works:

A

Essentially it’s a conventional front triangle with the seat stays looping past the top/seat tube junction and intersecting with the downtube just behind the headtube, just do away with any connection between the stays and seat tube.

Seat stay length is critical with belt drive.
Best to work out the gearing you want first, then pick a chain stay length to suit the limited range of pulleys and belts.

Conventional 454mm seat stays on my Qoroz 29er with a 50t pulley.

Chainstay, not seatstay length, but yes.
Heavily indented rear stays add to flexibility, and that can’t be good for belt retention.

It’s been a long day. I got it right one time out of three. 😕

I’m not sure exactly how to interpret this, but I think it means the frame deflects around 4.7mm horizontally and 0.7mm vertically under a 26kg load.
http://www.qoroz.com/docs/msg_bikes-stiffness_test.pdf
It would be interesting to see how that compares to other “laterally stiff, vertically compliant” frames.

Actually a bit more poking around and it turns out that schwinn did precisely what I described with their “cantilever” cruiser frames.

Good old Sheldon says:

it was what Schwinn called a “cantilever” frame, where the seat stays pass by the seat cluster and continue on in a graceful curve to join the bottom of the head tube.

1950’s engineering…

cookeaa – Member
Actually a bit more poking around and it turns out that schwinn did precisely what I described with their “cantilever” cruiser frames.

Good old Sheldon says:

it was what Schwinn called a “cantilever” frame, where the seat stays pass by the seat cluster and continue on in a graceful curve to join the bottom of the head tube.
1950’s engineering…

WTF am I looking at O.o

MidlandTrailquestsGraham – Member
It’s been a long day. I got it right one time out of three.
I’m not sure exactly how to interpret this, but I think it means the frame deflects around 4.7mm horizontally and 0.7mm vertically under a 26kg load.
http://www.qoroz.com/docs/msg_bikes-stiffness_test.pdf
It would be interesting to see how that compares to other “laterally stiff, vertically compliant” frames.
POSTED 8 HOURS AGO # REPORT-POST

Yes, but if you zoom in on the pics of the frame in the test, it doesn’t have any chainstay indentation.
Your frame has internal and external indentation shown on the drawing.

Good point, I’d never spotted that.
The frame in the test is a Qoroz Expedition Won, which is a 26er.

Going by the pictures and specification here, Van Nicholas appear to use fully tubular chain stays on the belt drive 29er Zion and presumably that’s passed the same test.
They’re shorter than mine too, at 437mm.

If you are making it in Ti
I would just use a horizontal blade attachment to the seat tube which joins the twin seat stays further forward.

That way it is still constrained fore and aft, and side to side, but with vertical compliance 😉

If you wanted some degree of tuning to this, then make it bolt on, so you can try different thickness and or length blades.
You could make these In Ti or carbon to get just the flex you are looking for.
You could make the front attachment points clamp around the tubes, which would also allow you to vary the position / length of blade.

Dom – I like that idea, and reckon that might be a goer 🙂

Cheers, Rich

I would just use a horizontal blade attachment to the seat tube which joins the twin seat stays further forward.

That would be a flexure then!
A single flexure results in a rotation, and you’ll need something to limit the movement to keep it from over-stressing the flexure. Two seperated by a few mm give a nearly parallel movement which may be the way to do it. Easier to work out the stress and deflection if you use metal, something like copper beryllium
is often used, as it has good stress/deflection properties.

What about an idea similar to a Tonic Fabrication flex ring:

http://tonicfab.com/news/?tag=crusher-bike

Two seperated by a few mm give a nearly parallel movement which may be the way to do it.

The moving end still tries to move in a circle, it now just does it while staying parallel to the other end, thus putting a big stress in at both ends, it’s really no better. At any rate, in this case the movement will be dictated by the much larger and stiffer frame tubes, any flexure plates will just be forced to do what they have to to keep up.

To the OP, the upwards curving twin tubes look nice but aren’t they really curved the wrong way to make the rear end flex much vertically? Imagine, if you took the seat stays away, the rear wheel would be cantilevered on the chainstays and over small deflections move like it was on a swingarm, pivoted somewhere near the BB. So the axle would be moving up and forwards in an arc. That arc is pretty much in direct line with your curving tubes, so effectively you are putting them into compression and transferring a huge load to the top of the headtube, thus stressing both connections to the headtube very heavily. If the tubes arched downward (yes, like a Jones), you would be flexing them like leaf springs and quite probably get on better. Your current design will mainly soften the ride by allowing the seat tube to flex backwards (until it bends or snaps at the weld).

I’d be tempted to do the twin tubes arched downward and weld a thin horizontal flat plate between them in a couple of places to allow them to work together as one beam sideways in order to give lateral stiffness then run a couple of small diameter tubes from the high on the seatube to around the headtube/downtube junction to stop the seat tube flexing back. add in a flat plate section behind the BB in the chainstays to take away the vertical stiffness there is need be. It probably wouldn’t look as nice, and it’d probably be heavy but you might get it to flex at the rear wheel if the stays were slim enough.

richmars – Member

I would just use a horizontal blade attachment to the seat tube which joins the twin seat stays further forward.

That would be a flexure then!
A single flexure results in a rotation, and you’ll need something to limit the movement to keep it from over-stressing the flexure. Two seperated by a few mm give a nearly parallel movement which may be the way to do it. Easier to work out the stress and deflection if you use metal, something like copper beryllium is often used, as it has good stress/deflection properties.

How much movement do you think its going to see ?

Its a Ti Frame, a Ti blade flexure with sufficient length will be more than adequate to accommodate the vertical movement.
The trick will be having a big enough fillet radius where it joins the seat tube so it doesn’t concentrate the stress at that joint.
you want the blade to deflect where its thin in the middle rather than try and break at a welded join to the seat tube.

I`ll try and scheme something up if I get a minute.

Rich, what version of SW are you on ?

Rich, what version of SW are you on ?

2014

The moving end still tries to move in a circle, it now just does it while staying parallel to the other end, thus putting a big stress in at both ends, it’s really no better.

It depends what you’re trying to do. If you need a (nearly) parallel motion, that’s very stiff in other direction, the two (or more likely, four) flexure design is pretty good. Typically you can get a few mm’s of movement in the required direction but only a few 10’s of microns in the unwanted direction.

Was the SW version question directed at Rich, me, Turboferret, or Rich, richmars?

I’m currently on 2013 but have a subscription and m machine has been telling me that 2015 SP0 X64 has been available for download for a while now, just not got around to doing so.

Fully agree that the simple arc isn’t necessarily the best shape for the stays, but it does look quite simple and elegant IMHO 🙂 Perhaps that isn’t necessarily the best design criteria though when experimenting with weird combinations of forces and potential failures!

Certainly more design iterations to go…

Cheers, Rich

Sorry it was aimed at you Rich 8) (Turboferret)

Ok I’ll save it as a step file when I`ve schemed it if you like as I’m on SW2014

Send me your email address too if you want me to send you a file over ?

Ha, too many rich’s in this thread!
(Also waiting for 2015, but we normally wait for the first SP before installing)

A step file would be great. Email coming your way.

Interesting to hear about waiting for the 1st SP before installing, I’ll hold off on 2015 for the moment.

Cheers, Rich