3D printing a set of hubs (or what ever I think of)

Have you seen how Gokiso hubs work? http://cyclingtips.com/2014/02/gokiso-climber-hubs-review/

right off the bat, these posts are getting long. im trying to answer everyones questions, but its pretty tough. if i havent answered a question, please PM me and ill be happy to clear anything up.

Have you seen how Gokiso hubs work? cyclingtips.com/2014/02/gokiso-cl­imber-hubs-review/

That seems like a lot of effort to reduce bearing wear, which isnt really an issue. plus the "suspension" theyve added is surely nulled by the tyre? Not the first hub with suspension like that Ive seen though, and really, a 3D printer would be great at making it, but to make it effective is really hard.

Anyway, im sending a few jobs through the printer. If you want to see what it looks like, check out a video here:

https://www.youtube.com/watch?v=p7WFVw6JOKg&feature=youtu.be

You can see a chris cross pattern on one of the rings in the top right. This is because the direction the laser is sweeping across the material changes, which in turn gives it a different reflective surface. Sorry about the shoddy video quality though, my phone was struggling to focus on it through the glass.

You should really do a convergence plot of multiple tests to make sure the model is working and tending to an ultimate stress value

I am doing convergence analysis, but not in the area you expect, instead im looking at impact. If a shed load of weight is applied to the hub over 0.01 seconds, as opposed to 0.1 seconds, how do we think it will react? well we can find out:

0.01 second loading:

0.1 second loading:

We can see the high loading rate slightly increases the stress in the heavily loaded spoke. But this is pretty meaningless because i have got no idea what the loading rate is when you go over a bump. The tyre will absorb alot of it. How much? no idea.

This is so cool. I played around with FEA tools with fusion 360, as it not only shows the stress areas but the plastic deflections on the item. Might be worth a look.

Im using Abaqus CAE for this analysis. It can essentially show me anything I want it to. And if it cant, I can always use a script to plot the analysis elsewhere. Heres the plastic strain for the model around a spoke hole. as you can see were talking about small values, which are most likely associated with the course mesh and loading conditions, so are greater than what abaqus actually thinks.

Im going to update the geometry based on these results and resubmit it.

Ideally, everything should be red. That way we will know that optimum weight has been achieved, however, im going to leave quite alot of weight on around the spoke holes, just to be sure (although, the chances are this hub will still explode :P).

Ive also shaved alot of weight off the hub. Heres the latest iteration, we can see weight savings going on in the cups

The other side of the cups is the interseting part though, these shapes ive cut out of it can be CNCd, but i really wouldnt want to try it, and it would be expensive. For the printer however, this is easy stuff.

The hub flange has also has a treatment:

Interesting stuff. Why does the laser jump around so much while printing? Is it to allow the material to cool? It looks random but I assume there's some logic to the pattern?

This thread pushes my nerd buttons. A++

Theres all kinds of reasons it would want to jump about. Residual stress is a big one, if you had a ring and started sintering along it, so the laser takes a circular path, the residual stress is huge. By breaking it up into sections it can reduce the residual stress.

If you had two parts really close together (I print parts with a 0.3mm gap), you need leaked thermal energy to dissipate between printing each side, otherwise material can bridge the gap.

The laser direction also alters the material properties, its strongest in the direction of the laser sweeping (i think), this generic pattern is an attempt to make it uniform (but you can do very clever stuff to make it strongest in the direction you want).

How solid (in terms of the percentage) are the parts that your DMLS machine produces?
My only worry (great thread btw, I love nerdy DIY stuff like this), despite all the FEA and safety factors you've shown us, is that because laser sintered metal is never going to be 100% solid as a material, this, combined with the extent of your weight reduction, could make the end result a bit more brittle and fragile than you might be expecting.

I remember being taught that sintered materials (was titanium we discussed) can have a higher strength than machined materials. That is all I remember mind.

is that because laser sintered metal is never going to be 100% solid as a material

I didn't know there were degrees of solidness ;)

Do you mean the sintered material will contain more impurities / air pockets / other whatnot would not be present in a forged billet?

How solid (in terms of the percentage) are the parts that your DMLS machine produces?

Its probably in excess of 98% solid. In order to avoid any issue with this, im not making any load baring part too thin. Also, im applying a factor of safety of 2.

I remember being taught that sintered materials (was titanium we discussed) can have a higher strength than machined materials. That is all I remember mind.

So, alot of people are asking about the material properties so I decided to actually look it up! Elastic modulus - 200GPa, Yield stress - 470MPa, Tensile strength - 570MPa. If we compare that to stainless 316, Elastic modulus 193GPa, Yield stress - 205MPa, Tensile strength - 515MPa.

The Elastic modulus is roughly on point, but the tensile strength is increased if printed, and the yield strength is HUGE if its 3D printed. Unfortunately, this isnt the whole story. 3D printed parts are not isotropic (if you take a single point in a part, the material properties are dependent on the direction to which the point is stressed). Im guessing they are cherry picking the best material properties and going with it. This is the exact same story for carbon fiber. Carbon fiber is incredibly strong in the direction of the fibers, but in any other direction, its shite. To get around this you have to make a carbon fiber weave or wrap a ribbon in multi directions, but then all of a sudden your strength/weight ratio has dropped significantly (carbon fiber rant over, steel is real).

Now what do these anisotropic material properties mean for a hub? the material properties are best in the plane the printer prints in. If the hub is sat on its end, then the flange will be in this plane. Since the front hub predominantly has to deal with radial stress, im laughing :)

Finally, when i looked up the spec of the printer, the layer thickness is 20microns, and that laser moves at 7m/s, and the machine weighs in excess of 3000kg!

Yep, precisely that.
Air in the build chamber is removed/replaced with an inert gas before the laser starts its work (steel uses nitrogen, titanium requires argon), but while a forged billet part is 100% 'solid' (or so you'd hope), a 3d-printed part is not.

How about fatigue resistance for the sintered material? I would expect that to be the dominant failure mode, and surface finish around the spoke holes will be critical.
http://pardo.net/bike/pic/fail-009/000.html

Interesting - why does this happen? And is it possible to make a printer that overcomes this? Or are there post-printing techniques that can be used?

I read that SpaceX is printing their rockets engines.

Interesting - why does this happen? And is it possible to make a printer that overcomes this? Or are there post-printing techniques that can be used?

SLS printers work by melting a powder. unfortunately theres always gaps between the powder particles (like balls in a ball pit). some of the gaps disappear (surface tension of the liquid steel pulls it in to the gaps i think?).

considering this, and the surface finish that Screwball brought up, check out the manufacturers example finish, no gaps, pretty darn smooth:

Personally, ive never achieved something like this, so i guess they are heat treating the parts afterwards. when ever i print something it looks like like very fine sandpaper. heat treating isnt off the cards, but i dont actually know anyone at work who does it.

This is mainly excellent, with some sprinkles of superb

So, i wanted to know what affect the mesh has on the simulation, so i just amped the mesh resolution up (5.5 million nodes). we get a much nicer picture of whats going on, problem is, the file size of the output data base is 30Gb, so i wont be doing this often.

Heres a few pics of the simulation results:

We can see the areas that need more material added, and the areas which can have some cut away, but as a whole, i think its approaching its final geometry. so its time to buy some carbon fiber :)

small update, i was wondering through the office and actually found a piece of carbon fiber tubing that was the exact correct diameter for the axle. happy days.

i was wondering

What were you wondering?

:p

i was wandering through the office and actually found a piece of carbon fiber tubing

Sounds like a good office

I cut the carbon tubes today for the axle/hub body. cutting carbon is a pain because it can produce nasty particles, as a result im using a large slow spinning liquid cooled cutting disc:

with a target length of 67mm, i was pretty happy with the result:

Ive been having issues getting my shady printing done, but hopefully i can take a different route to get it all sorted though :)

High speed diamond coated blade and plenty of coolant is best! Basically anything that doesn't have teeth which will chew up the ends of the fibres. But that looks ok to me.

I took some inspiration from your thread, downloaded Fusion and had a go at designing my own hub. Bad idea, as now I need to buy a lathe and a mill to make it. I haven't yet got around to simulations, I think that'll take a bit more learning.

1 Attachment

Good work! you will have issue milling it though because the disc brake mount might obscure some one the flange cut outs. you would have to laser sinter this (not a problem, just google a company in the UK and get a quote).

Has this been resolved yet?

How about now?

HTBRY?