[ardenum]

when filling steel fabrications, you need to account for thermal expansion. most of these high end machine bases will have active cooling and live in precision climate controlled buildings, so they never change more than 1-2 degrees in operation. if you dont have that, you could end up with a machine that warps itself.

That's certainly problematic, I wonder if you used metal powder instead of granite dust in EG, would it bring the expansion closer to steel.

[catahoula]

That's certainly problematic, I wonder if you used metal powder instead of granite dust in EG, would it bring the expansion closer to steel.

Right, have been thinking about using steel shot in the epoxy matrix, I think one would still get improved damping due to a solid structure and the myriad boundaries between epoxy and the shot, though probably not quite as good as sand/granite fill.

[catahoula]

Right, have been thinking about using steel shot in the epoxy matrix, I think one would still get improved damping due to a solid structure and the myriad boundaries between epoxy and the shot, though probably not quite as good as sand/granite fill.

But also quite expensive

[ihavenofish]

i think steel shot costs 2-3x more. other than that, i expect it would be fine.

i dont know if it resolves the expansion issue either. the composite matrix will behave differently than the plain steel.

i think the primary way to deal with expansion is to have thermal symmetry. for example, you dont want a thick steel plate back with an equal amount of "granite". a sandwich with steel on each side would counter warping.

[Mud_]

Hi Mud - Like this? But allowing the gantry to lift. Peter

I mean one face of the gantry beam is used for attachment of both the vertical bearing blocks and the rails for motion along the gantry length. This way you only need a single flat surface, and you aren't worried about parallelism of the gantry section. Pretty much just reverse your columns on #2 or #4. It should move the CoG usefully too.

[catahoula]

ihavenofish- Yes, that's a good point about symmetry in sections. One mistake I think I made in my build is having some 45 degree sections in the machine base, which may contribute to uneven thermal expansion.

[peteeng]

Hi All- A) mild steel and concrete have very similiar thermal expansion coefficients. They are both about 12e-6 m/m/C so makes a good match B) epoxy is expensive and provides no real stiffness (epoxy E=3GPa whereas concrete starts at 25 and could go to 60GPa) concrete is cheap for mass. I have been an epoxy advocate but looks like I have been swayed to the dark concrete side!!

In my part of the world concrete CSA is about $2/kg and epoxy is about $18/kg C) I've looked at steel shot and epoxy and it does not provide the same stiffness as CSA or UHPC... So HP concrete it is for me... C) for a one off making a steel form and filling it with concrete makes sense but if you want to do 2 or 3 or more then making a mould is considered. There will be a break point on quantity and complexity....

for those people with FE capability they can explore thermal expansion in their models. Most solvers do thermal analysis....Peter

[ardenum]

.
The metal inserts that people use are mild steel? I thought it was tool steel.

epoxy is about $18/kg

$18/kg is a low viscosity resin or just the cheapest? Because I was about to switch to CSA too in my build due to large volume.
But if it's only 20 it's something I can swallow, right now my thing weights 1.1 ton, so setting an upper limit on epoxy content 20% it would cost me max. 4000

[peteeng]

Hi Ard - good mild steel is all that's needed. For threads a good engagement length is needed. Can use "better" steel if you have it available or free machining steel....Peter

[ihavenofish]

concrete warps, and more importantly shrinks by a LOT when cast. it will continue to change shape for weeks or months after casting. You cant make a precision part with it. im assuming you mean portland type cement with a stone aggregate here. epoxy granite is also "concrete" so you need to be clear with your terms.

highly compacted granite with minimal epoxy is used because it effectively does not change dimensions when cast, as it heats up during curing and cools down.

[ardenum]

concrete warps, and more importantly shrinks by a LOT when cast. it will continue to change shape for weeks or months after casting. You cant make a precision part with it. im assuming you mean portland type cement with a stone aggregate here. epoxy granite is also "concrete" so you need to be clear with your terms.

highly compacted granite with minimal epoxy is used because it effectively does not change dimensions when cast, as it heats up during curing and cools down.

He's casting with CSA, apart from that only UHPC is an option.

[ihavenofish]

csa shrinks and warps too. uhpc is portland and csa mixed. portland and ca is another similar cement - mix, chuck it in a hole, and by the time youd turned around, its hard.

i use CA (calcium aluminate) for art and pizza ovens. i had entertained it for machine castings briefly. shrinkage, warping, cracking, and, it doesnt actually bond to anything very well, so any steel part is only held mechanically. its not "glued" in.

[catahoula]

So regarding thermal expansion of EG, I ran some numbers based on my current build, by separating the components into their respective proportions of a given tube length. It actually turns out that the epoxy, at CTE of 55 unit/unit/degree, overwhelms the CTE of the aggregate, even at low ratios. This is grossly oversimplified: assuming the aggregate/matrix blend has no air, is not taking into account the relative moduli of the different materials and how much the epoxy will deform restricted by the aggregate or steel tube, but seems like a good place to start.


Using my machine with a 34" (860mm) long gantry tube, let's compare linear thermal expansion at 80/20 Granite/Epoxy. CTE values from Engineeringtoolbox.com, averaged

Epoxy CTE 0.000055 - Length 34" x 0.2 ratio = 6.8" portion. Temp change 20 degrees F (50-70 degrees, winter shop vs summer shop). = 6.8075" final length
Granite CTE 0.0000082 - Length 34" x 0.8 ratio = 27.2" portion. Temp change 20 degrees F. = 27.2044
Combined = 34.0119"

Steel tube CTE 0.0000115 - Length 34" x 1.0 = 34" portion. Temp change 20 degrees F = 34.0078"

Difference = 0.004" / 0.102mm



If we go with 90/10 ratio (hard to achieve in real life, but for filling tubes, lack of voids and surface quality isn't such an issue):

Epoxy - Length 34" x 0.1 = 3.4" portion. Temp change 20 degrees F. = 3.4037"
Granite - Length 34" x 0.9 = 30.6" portion. Temp change 20 degrees F = 30.605

Steel: as above, 34.0078"

Difference = 0.0009" / 0.023 mm



From this rudimentary analysis, it seems like reducing the epoxy content and CTE of aggregate as much as possible is desirable

Edit to add: I also did not account for ratio by weight vs volume

[peteeng]

Hi IFish -
1) In my world epoxy is not concrete. Epoxy and minerals or epoxy and fibres if correctly cast and post cured is very stable., That is why machine tool beds are made from this
2) There are a few companies around the world making machine tool bases from UHPC if processed properly it does not shrink or crack and is very stable
3) CS is a cement additive, CSA cement is not portland cement and is very stable, grows slightly and if processed and post cured properly is exceptional stable.

Like metals there are many types of concrete with many types of properties. You can't can't tar all of them with the same brush. There are metals that are unsuitable for machine tools as well. So I am happy to get some CSA and some UHPC and do some tests and go from there. It is clear that steel fabrication is limited, casting iron is limited and plate builds are limited. I want to investigate other options and CSA or UHPC offer solutions not available using other methods. Just have to learn about the medium... Peter

Hi IFish - epoxy granite is sometimes called polymer cement. I think this is because the people involved are from the concrete side of the technology and wanted a word that had some familiarity in it vs mineral epoxy, EG etc etc... In regard to "mechanically" bonded all adhesives except ones that melt the surfaces are mechanical bonds. Epoxy does not chemically attach itself to surfaces, it penetrates the surface interstices and locks itself to the surface. Same with other adhesives. This is called wetting the surface. If your surface is very smooth or has a very low surface energy or has a moisture barrier on it, adhesives can't flow into the surface valleys & holes.

[peteeng]

Hi Catahula - Your ratios are a bit optimistic. I think you will acheive 70% volume ratio tops with granite so 30% epoxy. Most people use a "dry" mix and end up with quite a bit of air in the mix. 90% granite/mineral by volume is practically impossible. Some people do the math based on spherical particles with a graded mix but in reality 90% solid is not possible. So yes epoxies CoE does affect the total temperature stability if this is something you are trying to account for. Means concrete is more attractive as its CoE is closer to steel. Peter

[peteeng]

Hi All, back to Mill-G,
My initial plan was to use a material that could be machined so I developed Tetrium-S and Tetrium- A. But concrete requires wet machines with diamond tooling to cut and polish surfaces with. So I think I'm going to make parts in a two step process.

Step -1 vacuum cast the surfaces that require machining in Tetrium-S Maybe 5mm thick or as needed
Step - 2 back fill with concrete then post cure whole part
Step - 3 set up part in mill or router and finish machine surfaces and holes/threads...

I'm happy with this approach which means I can do all of this myself which was another rule in the project Cheers Peter

One of my commercial projects is designing and manufacturing composite car panels for a low volume vehicle being built here. We have looked at so many mould materials and processes its becoming a mission. I now think that using a fibreglass shell with integrated liquid heating tubes then backfilling with CSA or UHPC is the way forward. I think that the shell maybe unnecessary and I could use the concrete surface directly for the mould with appropriate sealers and top coat then polish.... we shall see

[catahoula]

Hi Catahula - Your ratios are a bit optimistic. I think you will acheive 70% volume ratio tops with granite so 30% epoxy. Most people use a "dry" mix and end up with quite a bit of air in the mix. 90% granite/mineral by volume is practically impossible. Some people do the math based on spherical particles with a graded mix but in reality 90% solid is not possible. So yes epoxies CoE does affect the total temperature stability if this is something you are trying to account for. Means concrete is more attractive as its CoE is closer to steel. Peter

Yep, my only point there is that the epoxy makes a surprising contribution to CTE. The ratio benchmark came from here: https://hackaday.com/2019/06/01/expe...epoxy-granite/
Where 80/20 is a relatively wet mix. I don't find the video to be a rigorous study as it seems to be only focused on the aesthetics, but it does seem to indicate that 80/20 is workable, at least with a small and consistently sized aggregate like sand.

[peteeng]

Hi Catahula - In the EG thread a lot of this was covered but I'll redo here as the concrete adds a new layer. In the EG I usually neglect the stiffness of the epoxy as its 3GPa in a stiff material. The two stiffest materials I can get easily are steel 200Gpa and Alum Oxide 300GPa. Concrete by itself is about 26Gpa. So if you have 50% by volume CSA and 50% steel you get (0.5x200*.5) + (0.5*26) = 63Gpa which is the upper limit for Epudur , see epudur data attached.. ie if you used a mineral of E=200GPa plus you would expect to get the same value or more eg AL2O3. The extra 0.5 in the equation is the efficiency for a strain transfer in spheres. If it was aligned long fibres like a stack of directional glass fibres the strain efficiency would be 1.0.

I have always aimed at the same modulus as aluminium and the Vf=35%+ of epoxy doesn't help. In concrete the 26-30GPa helps heaps.... The trick is to couple the two mediums. Epoxy definitely couples to most things, concrete however may not stick to carborundum or alumina so the result may not be to the calculation. This is called translation. In carbon and glass UD laminates this rule of mixtures translates very well from experience. Have to make a test coupon to check actual modulus. I do this by buckling a strip and back calculating E from the buckling load.... Peter

[peteeng]

Hi Cat - I expect 80/20 to be a weight ratio not a volume ratio. Convert the 80/20 to volumes... Peter

[peteeng]

Hi All - So now I start designing a concrete machine and I have to pick a design modulus for the material and I haven't decided on the material yet!! The figure lies between 40 and 80GPa. Prefer highest possible but have to be able to achieve it. I can buy mixes from overseas that allow E=80GPa but that's a bit hard logistically into the future. So I'll pick a middle value of 60GPa and were away. Peter

so if I use steel at a Vf=50% rule of mixtures predicts E=63GPa and density of 4965kg/m3 if I use AL2O3 I get E=88GPa and density = 3015kg/m3 I'll get some alumina next year and test the theory... Peter