Filling steel tubes with epoxy granite- machine before or after?

[catahoula]

I like the brazed end caps idea. I have an oxy-acetylene torch and a little bit of know-how left over from building a few bicycle frames some years back, and I think I have both brass and silver solder left over, too.

The context for this machine is a fixed gantry layout, so it sounds like a capped tube filled with loose sand might be the best DIY way to go?

I called my steel supplier and they have 8" x 8" x 1/2" tube in regular stock. So altogether this would provide mass, stiffness, no free edges, bolted assembly with the gantry supports.

From a machine performance standpoint is there any difference between through bolted and tapped fasteners? I was originally planning on through-bolting the gantry supports to the gantry tube, but with capping the ends, a tapped hole is more realistic. Unless I only brazed the cap across half of the tube diameter, and used some kind of bolted plate to cover the rest of it.

[pippin88]

Fixed gantry = don't worry about the mass increasing with damping measures.

Sand in fact works best when it is on the top surface of the tube. Hard to achieve in real life.

I will try to find the reference, but I believe EG is better than sand.

[peteeng]

Hi Brian - If you have brazing equipment I recommend you braze the entire frame. This means no frame metal melting so no large shrinkage, the red heat stress relieves on the go and it will be stronger then required. Tobin bronze is a good material to use. I'm an ex bike frame builder as well (steel, aluminium and titanium) and those methods are very appropriate for machine builds.

Through bolting may not be reliable. Long bolts stretch more and the side walls are springy and may come lose over time; and as you torque up, the section shape may change. Two methods, drill and bolt to 1/2" section very good. Or make a straight backing bar with the threaded holes that match clearance holes in the section. This method is good for thin sections... Get tracks as close as possible to edges Cheers Peter S

If you are into experimentation re damping. Get a long 25mm dia (1") thin steel tube that can be bolted to a heavy stiff bench or beam. Say 1.5m long plus. The idea is to use something that will wobble well. If you can, 3 side by side all the better. One leave empty, one fill with sand, one EG or PU foam. Then deflect (same distance each) and let go. The one that stops wobbling first is the best damped. Remember to post the results here, we need real evidence , a video is worth a million words.

Re "capping " ends. The free edges just need support so you could use an external ring or cap with hole ie a flange or return is alls that needed to give it stiffness in two directions. They can be angle iron glued on (or soldered) , doesn't take much. Plus do you have any drawings yet of this machine? Maybe helpful...

[peteeng]

Hi Pippin - Start a thread we want to know more about mills... I highly recommend brazing/soldering over welding especially since laser cut parts can be made self jigging via slots and tabs etc. The assemblies don't move much at all when silver soldered and I'm about to do some with soft solder so will be even less heat (none that I can determine) Peter

[catahoula]

Ah, great to run into another bike nerd! There's actually not much brazing or welding to be done on this as whole, by design. I'll probably still weld the gantry support towers because they are small and can be put into an annealing oven without too much difficulty.

I think I didn't describe the through-bolting well: it'd just be to the bottom surface of the gantry tube, so no chance for distortion of the main tube. I'm thinking of just capping the upper half of the open end on each side, so I can still access the support tower bolts easily but it closes the upper free edge. The bottom free edge is supported by the towers, anyways.

Attached is a very rough beginning of a machine sketch.

As far as vibration testing, I'll see what I have around the shop. Do you think a small tube like 5/16 x 0.035" (8 x 0.9 mm) would be appropriate for the test? I buy a lot of that in cromoly steel for a different purpose.

[peteeng]

Hi Brian - If the gantry tube is big enough to get a hand in then just use nuts. Yes any size tube will work. Make it long so it has a good wobble. I'm reading up on transient analysis so will try to get some models running to look at damping. Why not braze the "towers" directly to the gantry? Peter

https://www.pedalroom.com/bike/technicomps-28794 one of my alloy frames from 1991 aluminium. Its been repainted not original stickers. I had extrusions done to my own spec and we heated treated in house.

https://carbon-works.com.au/about/interests/ Ti frame from 1992... stopped making frames in 1992 got into boats

I don't think the towers will need stress relief as designed just make the plates thick and re-machine flat after welding if welded to take the bow out. But if brazed I'm sure they will not bow.

[pippin88]

Hi Pippin - Start a thread we want to know more about mills... I highly recommend brazing/soldering over welding especially since laser cut parts can be made self jigging via slots and tabs etc. The assemblies don't move much at all when silver soldered and I'm about to do some with soft solder so will be even less heat (none that I can determine) Peter

Will do once I have some CAD done.

Brazing is of great interest. I was happy to go down stress relief route, but if it can be avoided that would be excellent!

In fact I suspect I can use brazing in some awkward areas that may have been difficult to weld. I only have an arc (stick) welder and that makes getting into small spots problematic.

I was already planning on tabs and slots anyway for alignment.

[catahoula]

The first frame I brazed I did fillets with brass, and given my lack of experience got it rather hot and experienced some warping. The second frame I did with lugs and silver solder and it went back into the jig exactly as positioned before joining. So nice to work at those low temperatures. The only reason I don't braze the towers to the gantry is because the machining is easier to do with the parts separate and I want to minimize any chance for residual stress or misalignment. Plus, drilling holes is easy enough and I would be surprised if the bolted structure isn't rigid enough. Though that's based on feeling not calculation.

Interesting to see you moved from bikes to boats. I grew up on boats and have been putting off that inclination for years now but the older I get the stronger the pull is. Currently I make bicycle saddles with carbon/fiberglass hybrid shells, 2-part urethane molded foam, and formed cromoly tubular steel rails. With the CNC I hope to build carbon rail molds as well as more accurate 2 sided shell molds.

[peteeng]

Hi Brian - Then you should be making composite machine parts!! If you can publish the step file of your gantry I'm happy to run a modal solver on it. If you put the bolts in I can include their torque up and friction. If you do a welded one I can compare. You can publish STEP here as a zip file... Peter

Before I set up the bike factory I was working for a spar maker (very late 80's) and we played with CF a lot. Then I set up to build CF frames (early 90's) but that was too hard at the time so did aluminium (that was hard to at the time due to heat treatment, so I developed my own in-house HT and temper) Once I'd made a few 1000 frames and for various reasons, I moved onto CF masts then hull analysis for a couple of naval architects. More $$$ in big boats then bikes...

[peteeng]

Hi Pippin et al - Here's a good read. Haven't digested it yet...Peter

here's the resin, will be better then epoxy as its cheaper and softer. Even has "vibration damping" in its description. its about car tyre hardness
https://www.freemansupply.com/produc...t-polyurethane

[pippin88]

Peter, thanks for the paper and product link.

4" casting thickness on the RenCast 6442 may be a limit.

A good paper is attached:
Influence of the addition of styrenebutadiene rubber on the dynamic properties of polymer concrete for machine tool applications

For a metal skinned structure with epoxy granite (aka polymer concrete) fill, the strength of fill does not really matter, we just want damping really.
Addition of a viscoelastic (e.g. rubber) appears to be beneficial.

For moving elements, the question is the lowest weight (inertia) material or composite to improve the damping.
Rubber granules.
Resin ?RenCast 6442 ?epoxy (want something that does not shrink)
What else?

Or is polyurethane foam any use? Still can't find data.

From my reading it appears that part of the benefit of EG / polymer concrete is the composite nature with different materials.

[pippin88]

Characteristics of Polymer Concrete from Pumice Stone and Rubber Thread Waste with Polyurethane as Natural Bonding

Characteristics of Polymer Concrete from Pumice Stone and Rubber Thread Waste with Polyurethane as Natural Bonding :: Science Publishing Group

[peteeng]

Hi Pippin - 1) If the structure is all EG then its damp as the load strains the structure. EG being multimaterial multi interface is naturally damp If you fill a steel tube with nearly anything the steel tubes rigidity is orders above the fill so the viscoelastic effects are minimal as there is hardly any strain transferred to the fill. So its the mass gain that's being helpful. And in the case of thin steel it supports the sheet so stops panting (or oil canning). If mass gain is the effect then its cheaper to and more effective to use thicker steel then nearly any viscous fill. 2) The article on aluminium foam filled tubes is enlightening due to its discovery that a uncoupled fill is better then a coupled fill. As the stiff outer tube deflects the interface creates friction which damps much better then the coupled fill. So if you fill with epoxy spray wax or line with plastic so you establish a differential movement between the two structures. This is the same effect as leaf springs use to damp trailers etc. Coupled PU foam creates an asymmetric structure which damps panting. How good it is we'll have to do some actual tests.

Since you have a router for wood I suggest you make MDF moulds and go all EG or composite. You can make optimum shape parts. Using sheet metal you will be restricted to the geometry you can create...I think that is a far better approach then using metal and trying to mitigate its properties. Step ahead of the rest, break new ground (but its not new really, top machine companies have been doing this for decades just haven't told us much)... happy to help I'll be going down this path soonest....If you are using fusion 360 it will give you optimum shapes automatically.... and you certainly won't be able to do these in metal. Peter

[pippin88]

Damping and mechanical properties of composite composed of polyurethane matrix and preplaced aggregates

https://sci-hub.tw/10.1016/j.conbuildmat.2017.03.233

[peteeng]

Thks Pippin - good info. But they don't measure stiffness. All rubber like materials are very damp (visco-elastic) . But they are not stiff. So using this sort of flexible PU/filler as a structural material is no good as its not stiff. If you put it into a steel structure as I have said before the load path is through the steel (the stiffest path) not the fill therefore the dampness of the fill is moot. Make moulded EG or epoxy fiberglass parts, your well ahead.... Peter

Plus all of the research uses large deflections to enable measurement of damping ratios. If we design a machine correctly the deflections hence strains are microscopic so the effects we want are not in the large deflection arena as we need to think in terms of micro frictional effects and the structures loadpath. Filling things with EG/foam/granules does not fulfill these micro level requirements. The only thing it does do is add mass and maybe support thin sections a bit better. but then they should not be thin anyway in this application....

Pippin pick a single material (or a single composite epoxy fibreglass, epoxy CF, epoxy sand ) and work with it. Trying to use mixed media is unhelpful. Do the best you can in the chosen material, get it working then solve any particular issues that come up. KISS Simples

[peteeng]

Hi Pippin and others - I reread the conbuildmat doc and found the material modulus. They called it beam modulus. The PU with FG is as stiff as published EG figures and stiffer with the CF fibres. This is as expected. The PU composite has very good damping loss factors and I'd expect if epoxy was used then it would be similiar. Not quite as good, as the PU has extremely low modulus compared to std epoxies. PU <10MPa epoxy ~3000MPa. One of the PU's had a E= 1380MPa and this produced a stiffer composite as expected but not by much.. To use PU as a matrix has a big hurdle, its gel time is only a few minutes. I've done large PU castings and even with a complex open mould you can be up against the gel time to distribute the stuff. So I shall talk to my friendly epoxy chemist and see what plasticisers we can use to make the epoxy a bit stretchier. I do know of one but I'd have to buy it wholesale in a large qty and its costly. So for a one off thing it would be prohibitive I think as you only add 3-5% by weight to change the elongation from 6% to 25%. Its a shame they didn't publish fibre ratios so I could do some analysis on the mix. But they used a vacuum bag to consolidate the laminate and going by the modulus they achieved a good fibre ratio. One comment was that the laminate construction (mat, 45/45, 90/0) did not change the damping ratio. This makes sense as damping is all about the interface not the fibre direction. This points at making good structurally optimized fibre directions and use a stretchy resin. The CF panels achieved the highest stiffness and highest damping so there's the take away I think. So Pippin make FG or CF epoxy parts and you will have a superior dampness machine. I can supply Tetrium as well as this is along the same lines. You cut your MDF moulds I'll vacuum cast the part. The Uni that does my testing wants about $1000 to do a damping test on Tetrium.... one day the right project will fund this test... Peter

[peteeng]

Hi Pippin et al - If your in Australia here's a source of basalt and carbon short fibres. Make a part mould, fill it up with epoxy and short fibres (prefer consolidate with vac bag) or infuse to fill.

Basalt is about 90GPa whereas E glass is 70GPa stiffness. Same density as E glass...

https://beyondmaterials.com.au/produ...-carbon-fibre/ Cheers Peter S

[catahoula]

Peter- Appreciate the offer of running calcs on the gantry. I have a fairly good idea of the overall budget and layout for the whole machine now, but need to work on some other things for a bit. I'm feeling fairly confident that the gantry is comfortably overbuilt with the 200x200x12.7 tube, and am more concerned now about z-axis and gantry support deflection. Maybe in a few weeks time when I have a more complete CAD model of the whole machine you could give some input on those aspects?

Re: composite machine parts- you have to start with a square and flat surface to end up with one! As this will be the first piece of machining or CNC equipment in my shop, I don't have that capability until it's done. So for this time I think I will continue with granite and machined steel, then this machine can build molds or templates for others.

[peteeng]

Hi Brian - Yes so many rabbit holes to get out of and chickens and eggs to chase. Peter

[pippin88]

Structural Damping Enhancement Via Non-Obstructive Particle Damping Technique

https://sci-hub.tw/10.1115/1.2930221