Home Made Bikes

Author Message
Posted: May 16, 2018 at 14:07 Quote
Thought I would have a play about while I was on the work station:

This is why you should not weld up a leaf gusset at the end, this gusset is fully attached to the tube around the whole perimeter and the stress at the tip is higher than any other point around the head tube. If you coupled that with a butted tube its a horror.


This is the effect of side plate gussets in the exact same loading conditions. (crappy lazy plate gussets that meet the head tube badly, if you think these gussets are too crappy and lazy then I will refund your consultation fee)Joker

Easy to see how these improve the loading.


Posted: May 17, 2018 at 3:28 Quote
l-plates wrote:
How much heat went into those tubes when the frame was put together? Looks like it may have been fillet brazed? (at least on this small laptop screen).

Is it possible you softened the tube in the heat affected zone of the joint? Always a potential problem with low alloy steels like 4130 and why I prefer the maraging steels as they age harden and do not need heat treatment.
4130 does not need heat treatment but too much heat into it when joining can weaken it, a big issue for those building roll cages and chassis out of it as it's not so easy to heat treat a whole car. lol
Yep the whole bike is fillet brased, keeping the heat low compared to tig. This frame is 3 years old now and riden alot and had some hard hitting cases.. so im pretty happy with that. Getting started on next frame so looking at different alloy tubes and keen to add some gussets. I see Henry James doesnt have a lot of tubes in stock, any other suggestions?

Posted: May 17, 2018 at 6:33 Quote
I'm in the UK so get my tubes via the Bike Academy as they are only a few miles from me or from either Aerocom or smaller local steel stockist as most can get T45 tubing in.

T45 is pretty ubiquitous over here which helps.

Posted: May 17, 2018 at 18:04 Quote
l-plates wrote:
Thought I would have a play about while I was on the work station:

This is why you should not weld up a leaf gusset at the end, this gusset is fully attached to the tube around the whole perimeter and the stress at the tip is higher than any other point around the head tube. If you coupled that with a butted tube its a horror.



This is the effect of side plate gussets in the exact same loading conditions. (crappy lazy plate gussets that meet the head tube badly, if you think these gussets are too crappy and lazy then I will refund your consultation fee)Joker

Easy to see how these improve the loading.



You should increase the mesh density of the weld/braze fillets. Standard practice would put at least 3 elements across any feature like that (you should be able to refine the mesh in that specific area, rather then the whole model - which will SIGNIFICANTLY increase solve times). Then re-run the results and compare.

Also, sometimes you can run into "singularities" in an FEA analysis that don't necessarily reflect reality. I recommend doing a mesh sensitivity analysis - refine the mesh in that area and see if the stress converges or goes to infinity.

Note: I'm not refuting or reinforcing anyone's gusset/reinforcement theory, just pitching in some FEA advice.

Posted: May 19, 2018 at 0:46 Quote
klopp wrote:
l-plates wrote:
Thought I would have a play about while I was on the work station:

This is why you should not weld up a leaf gusset at the end, this gusset is fully attached to the tube around the whole perimeter and the stress at the tip is higher than any other point around the head tube. If you coupled that with a butted tube its a horror.



This is the effect of side plate gussets in the exact same loading conditions. (crappy lazy plate gussets that meet the head tube badly, if you think these gussets are too crappy and lazy then I will refund your consultation fee)Joker

Easy to see how these improve the loading.



You should increase the mesh density of the weld/braze fillets. Standard practice would put at least 3 elements across any feature like that (you should be able to refine the mesh in that specific area, rather then the whole model - which will SIGNIFICANTLY increase solve times). Then re-run the results and compare.

Also, sometimes you can run into "singularities" in an FEA analysis that don't necessarily reflect reality. I recommend doing a mesh sensitivity analysis - refine the mesh in that area and see if the stress converges or goes to infinity.

Note: I'm not refuting or reinforcing anyone's gusset/reinforcement theory, just pitching in some FEA advice.
Yeah I agree, stresses on a single knot doesn't mean much. And modeling a welded structure just like that is a huge simplification. Can't quite remember what's the "right" way to do it as I'm more of a plastic guy lately...

Posted: May 22, 2018 at 8:46 Quote
Fully agree on both the mesh density and the single points, normally I would use the mesh control to increase the localised density of the elements in any areas of interest but didn't see any point for what I was trying to illustrate (i.e. joint config > pure weld area).

In this case I know that bringing the mesh density down from the ~3mm its currently set at to ~0.5mm needed to get suitable accuracy maintains the same trends but brings the stress up quite a bit, it does however increase solve times exponentially so just doesn't make sense when only doing a couple of throwaway runs to prove the point about joint configurations.

RE: welded structures, they are a tough challenge for FEA as there is so much un-quantifiable variability thrown in by the welding process, we were always taught to work off the base geometry and apply concentration factors post FEA. I.e. design as though one solid piece and downrate to account for welding.

Would be curious to hear any thoughts as I know there has been a lot of work into more advanced methods of simulating welded structures in the years since I left uni. (although as far as I am aware a lot of the papers published on the subject come back to either simplify or invest in a super computer).

Edit - I might run some finer ones on the gussets once I have finished the design for my new frame, would be curious to see precisely how much difference they make rather than the fairly obvious trends.

SleepingAwake - Curious to see how your 3d printed carbon molds come out, I have been considering doing similar myself. What I can say is that depending on your infill settings PLA can be surprisingly strong, I use it to prototype and fit check race car parts every so often, just did some roll center spacers and the PLA at 20% infill was strong enough to sit the car onto the ground and check the geometry.
They do however give little warning of failure, so if your mold starts to crackle then its about to explode, the PLA just seems to shatter unless it is printed at very high density.

Posted: May 23, 2018 at 6:19 Quote
l-plates wrote:
SleepingAwake - Curious to see how your 3d printed carbon molds come out, I have been considering doing similar myself. What I can say is that depending on your infill settings PLA can be surprisingly strong, I use it to prototype and fit check race car parts every so often, just did some roll center spacers and the PLA at 20% infill was strong enough to sit the car onto the ground and check the geometry.
They do however give little warning of failure, so if your mold starts to crackle then its about to explode, the PLA just seems to shatter unless it is printed at very high density.

I made quite a few moulds this way. With a printed plug which is then used to make the negative mould. Works a treat and as PLA is quite stiff and moulds usually can be (initially) cured at ambient temperature it works a charm.

The one mould in my gallery is different as I am planning to use it for bladder inflation moulding with prepregs. So high stress at elevated temperatures. I actually never seen a composite mould for a BIM process and I'm really still not sure if it is going to work or not. The mould is currently in the oven for post curing. I'll post my conclusions in here, no matter the outcome.

cheers, Reto

Posted: May 23, 2018 at 11:46 Quote
Just checked your gallery SleepingAwake - that must have taken quite a while to 3D print!!

Posted: May 23, 2018 at 13:59 Quote
Although not a mountain bike, here is my latest frame;
Carbon road frame.

Posted: May 23, 2018 at 14:10 Quote
octanejake wrote:
Although not a mountain bike, here is my latest frame;
Carbon road frame.

Something is very very weird about your handlebar set up

Posted: May 23, 2018 at 18:21 Quote
SleepingAwake wrote:
l-plates wrote:
SleepingAwake - Curious to see how your 3d printed carbon molds come out, I have been considering doing similar myself. What I can say is that depending on your infill settings PLA can be surprisingly strong, I use it to prototype and fit check race car parts every so often, just did some roll center spacers and the PLA at 20% infill was strong enough to sit the car onto the ground and check the geometry.
They do however give little warning of failure, so if your mold starts to crackle then its about to explode, the PLA just seems to shatter unless it is printed at very high density.

I made quite a few moulds this way. With a printed plug which is then used to make the negative mould. Works a treat and as PLA is quite stiff and moulds usually can be (initially) cured at ambient temperature it works a charm.

The one mould in my gallery is different as I am planning to use it for bladder inflation moulding with prepregs. So high stress at elevated temperatures. I actually never seen a composite mould for a BIM process and I'm really still not sure if it is going to work or not. The mould is currently in the oven for post curing. I'll post my conclusions in here, no matter the outcome.

cheers, Reto

I have seen BIM moulds made using tooling resin and carbon fiber with metal and wood backing, which you could do with a 3d printed plug. I hav enot seen 3d printed BIM moulds and would be interested in how your project goes.

Posted: May 24, 2018 at 10:50 Quote
"Also, sometimes you can run into "singularities" in an FEA analysis that don't necessarily reflect reality."

Effective notch

Posted: Jun 1, 2018 at 8:05 Quote
Wondering if any of you builders have tried low bbs on dj frames? Currently rising a frame I built with 30mm of drop on 26 in wheels and 100mm fork. Wanting to try 50 next. But nobody makes anything like that. I use this mainly for bmx and pump tracks. No street or anything like that.

Posted: Jun 1, 2018 at 8:22 Quote
I'm going to book a place on the bicycle academies 7 day frame building course later this year, has anyone here done one of their courses? I have precisely 0 experience but think it'd be a really cool thing to take up.
I'm also going to learn a bit of TIG from a friend.

I've been lurking here for years now, interested to find out how people got into building frames, and if it took them anywhere. Absolutely don't want to get ahead of myself but I'd love to have it as a side project that covers its own costs.
Salute

Posted: Jun 2, 2018 at 9:38 Quote
Tsoxbhk wrote:
Wondering if any of you builders have tried low bbs on dj frames? Currently rising a frame I built with 30mm of drop on 26 in wheels and 100mm fork. Wanting to try 50 next. But nobody makes anything like that. I use this mainly for bmx and pump tracks. No street or anything like that.

it'll jump like a bathtub

i've made 26" hardtails with up to 50mm drop and they make nice DH sleds (very stable), but are an absolute struggle to pop off lips. i find them hard to pump because your feet are too low relative to your arse above the seat/you're in a quite upright position. hard to explain that one, i just hated it when i tried it on a pump track.

i'd say just give it a go with a cheap/free tubeset to see what its like for yourself...dont go spending heaps on the frame set but Wink


 
Copyright © 2000 - 2019. Pinkbike.com. All rights reserved.
dv65 0.018501
Mobile Version of Website