[peteeng]

Hi Sus - By reinforced do you mean steel rods? ie steel reinforcement? The only documented UHPC/Grout/concrete at >E70 is a german one using ALOX as an aggregate. I have made a 65% alox coupon with 35% epoxy by volume and the test flexure modulus was around 20GPa (my plywood is better then that) so I don't see how it helps getting concrete to E80. I also spoke to the professor who developed that material and he was helpful but not helpful enough to replicate the result. I think all concrete compression tests are high biased...

So to steel reinforcement if that's what you are talking about. Steel reo is used in concrete to provide a stiff loadpath within the structure when the concrete cracks ie it bridges the cracks. Plus the steel is very strong and ductile so it does the heavy lifting in terms of strength. Its placed on the outer surfaces to maximise its geometric inertia. If you did not have a strong element in the concrete its global strength would be very small and its natural cracking would mean the structure falls down really quick (by the way roman concrete is different to modern concrete in that it self heals when it cracks, interesting stuff if you look it up)... Now in terms of a machine design strain and loads travel along the stiffest path so if you have continuous steel in the cast then that's where the loads travel and it bypasses the concrete. Steel E200 concrete E40. Machine elements tend to be shear dominant so loads travel in the entire section unless there is a stiff highway which the steel is. Being such a small volume component of the casting the steel does not give you much global stiffness gain. You can model it if you like to figure out the benefit or not.

I read an article on a mill made by NASA and it was a 3D printed plastic mould with steel reo in it and std concrete poured in. Looked just like my foundations for my shed... sans plastic, I don't think the team did much homework on that one to figure out the state of the art in concrete mills. There are a few threads and on other sites where people have used std concrete from the hardware and it works out for them. Just depends on how high a bar you set yourself.

So use a commercial grout that does not crack, is double shrinkage compensated (Class C) and preferably is 100% CSA cement. Peter

Damping - any composite material is damper then a homogenous material. Its to do with the internal hysteresis of the material as its constituents rub against each other as it strains (moves around) Dynamic stiffness of a machine is complex and you need to get the static stiffness right to side step a lot of the vibration issues. Don't over think it... settle on a build material and get to it...

[jackjr-123]

I've ordered some Durigid stones to try making this elusive E80 concrete. I'll make a few samples with the exact Nanodur recipe and test the stiffness.

[peteeng]

Hi Jack - How are you going to test? Peter

[jackjr-123]

Good question
I didn't think much about it, but I was thinking a beam supported on both ends, the load in the middle and a DTI measuring the deflection.
How would you do it?

[peteeng]

Hi Jack - Just as you described. I'd use a manual calc to back calculate the modulus vs a FE model. I think I've described this process elsewhere here. Peter

[peteeng]

Hi Jack - Your in USA? Some time ago I tried to buy Durigid from Germany and they said they did not supply outside of CE so would not ship to Australia. Did you have trouble ordering this? Peter

[jackjr-123]

No I'm in Europe. I can't find any settings to change my location...
If you want some Durigid I could send you some.

[peteeng]

Hi Jack - Yes I have tried to change my location as well and its impossible. No durigid is ALOX and I get 3mm sand blasting grit here easily. I have 20kg of it. Peter

[G-Spot]

Hi Sus - By reinforced do you mean steel rods? ie steel reinforcement? The only documented UHPC/Grout/concrete at >E70 is a german one using ALOX as an aggregate. I have made a 65% alox coupon with 35% epoxy by volume and the test flexure modulus was around 20GPa (my plywood is better then that) so I don't see how it helps getting concrete to E80. I also spoke to the professor who developed that material and he was helpful but not helpful enough to replicate the result. I think all concrete compression tests are high biased...

You have to pre-stress the steel in your concrete otherwise you lose a lot of stiffness.
Reinforced concrete cant hold a candle to prestressed when it comes to stiffness.
I have seen a DIY(owned a CNC business) CNC build where they made the back and bottom of the C design using prestressed columns.

Thick steel caps at the end of your beams.
Drill multiple holes for rebar
thread ends of rebar
fit to machine and torque up close to the maximum for the size of rebar
pour epoxy

Only apply torque just before you fill with epoxy. Otherwise you might distort your structure.
Release torque bolts after 7 days and retorque.

Now your lump of epoxy to flex must overcome the tension stored in the rebar.

You have to use a metal that will grip the epoxy not smooth/threaded bar. Just use a high tensile rebar designed for that very purpose.

"Prestressed concrete has important advantages over reinforced concrete. First, the entire section is effective in resisting the applied moment, whereas only the section above the neutral axis is fully effective in reinforced concrete. This leads to greatly reduced deflections under service conditions."


https://www.sciencedirect.com/topics...essed-concrete

[peteeng]

Hi - G spot - Prestressing or post stressing does not improve stiffness. It does improve the composite beams strength. Concrete is very poor in tension so the pre/post stress moves more material into compression, so it takes more load to get the concrete to break.

The comment about reduced deflections maybe because if you change the beam shape in that process eg curve the beam upward (crowning) then the live load and gravity straighten the beam. If the beam is made flat, then gravity and live load bend the beam downward (sagging) But this does not improve its stiffness just it shape.

You could look at it like this. If you get a spring and stretch it, its a certain stiffness. Then if you apply more load its still the same stiffness same with the reo and concrete. Peter

[suspension]

Hi Pete

I talked to a local company specialized in creating new concretes for different purposes. I told them about the CSA cement that you proposed for this work but according to them, this cement variant is not available in here. They however have PPC cement (which use fly ash) which they recommended for this application (this will be mixed with various other chemicals to get the final mix). I wanted to check with you also about your opinion on this.

Thanks
Sus

[hanermo]

One of the most successful, rich, engineering company founders made videos on youtube, showing how strong or rigid parts are.

2 steel parts bonded by rubber bands are as strong as if they were welded, as long as they do not separate.
Dan Geldorf or something like that, iirc .. more or less.

It´s "unnatural" how machine tools are rigid.
Likewise, how rigid they need to be.
Typical max loading for lathes and machine tools is 1-2% of breaking strength.

A medium 12" lathe needs a strength of about 50 metric tons, to have adequate rigidity.
Yes, I know, it´s ridiculous, but all good lathes have followed this formula for 100 years.

[peteeng]

Hi Sus - "concrete" makers all over make grout that is double compensated. Its called class C in USA. It will have a plasticiser to minimise water addition (drying shrinkage) and use fillers to minimise the curing shrinkage. Can't really comment unless they have a data sheet. Sika have companies all over the world making grout. Can you get Sika products? Peter

[peteeng]

Morning Sus - I have gone off the idea of making grout/concrete parts 1) The material is not stiff enough in small machines that I am interested in 2) Moulds have to be made or you make steel shells. 3) Its difficult to post machine it unless you have access to stone masonry type machines 4) I still have to take it to a machinist to have the inserted lands machined 5) if I use ceramic aggregates to improve its stiffness it becomes nearly impossible to machine & these have not improved the stiffness in my tests. I would consider it if someone wanted a production machine like 1 per week, but then metal is viable at my machine sizes. If the machines are very big then cold casting makes sense.

The only material that I have been able to make that has a modulus greater than aluminium reliably is carbon fibre. I 've made E80 material in compression, tension and flexure in std modulus CF. Its costly and fiddly but a great solution.

So I come back to metals Aluminium, steel or cast iron the traditional material. For me aluminium works as I can machine it on my routers. But in production the cheapest path so far has been steel laser cut parts. If you need really thick parts they can be laminated. You may say what about damping? People harp on about damping but stiffness trumps damping. I have researched dynamic stiffness of machines enough now to know that no material is damp enough to solve the vibration issues in traditional CNC subtractive machines. There's been heaps of research and effort in this area and there are no material answers. The only answers lie in input shaping, active shaping and tap testing at present. These identify the vibration modes of the machine and you just stay away from those operational areas. Using tap testing, productivity improves 10x. So using the stiffest cheapest material is the way forward, then tap test your machine and your on your way. This means steel will be the obvious solution. Either billet machined, fabricated or self-assembled (bolted). If fabricated ideally use laser welding that may sidestep the stress relieve process.

Moving fwd commercial machines will 3D print their parts in metal (this is already occurring), use shaping in their control systems (which is occurring) and tap testing will become a std workshop practice (which is occurring)... unless additive overtakes subtractive in economics. Which in 5-10 years that will be the case... Then those fancy spiderweb designs that AI produce become easy to make and since they use 10X less material become viable to print... and printing is limitless in size... so hot casting will stop... when I'm 80 and in my dotage machines will look entirely different to how they look now...

So for me the conclusion is that laminated aluminium works for me its as damp as cast iron or composites and I can machine them on my routers. Once I get two more routers made I design a fully laminated aluminium mill.... Peter

[suspension]

Hi Sus - "concrete" makers all over make grout that is double compensated. Its called class C in USA. It will have a plasticiser to minimise water addition (drying shrinkage) and use fillers to minimise the curing shrinkage. Can't really comment unless they have a data sheet. Sika have companies all over the world making grout. Can you get Sika products? Peter

Hi Pete,
I can get Sika, but it is 4 times expensive than the one I am evaluating now. In addition, the local company (country office of Sika) is not that helpful. The grout which is available in here has no EM specified. My guess is that, if I did similar tests (EM/Shrinkage), I will get similar results to the one I am evaluating. My issue with current grout is Shrinkage. Although I thought the dial indicator was fluctuating for temperature differences, there is a average shrinkage like .5um per day it seems. This may be small, but the way I see, given the size of my column and base, this shrinkage will move steel plates too much after a while and the accuracy will be gone. I honestly think this will be the case for Durcrete as well. And I think the current grout I am using is class A.

On the other hand, this engineering company which produce ad-mixture will do testing for both EM/Shrinkage (for free) which will give me confidence. I will also do the testing my self. Overall this mixture will have similar costs to the current grout I am evaluating which is reasonable.

The did not give me lots of details of the PPC cement apart from the fact that it is made by adding fly ash.

Thanks
Sus

[suspension]

Morning Sus - I have gone off the idea of making grout/concrete parts 1) The material is not stiff enough in small machines that I am interested in 2) Moulds have to be made or you make steel shells. 3) Its difficult to post machine it unless you have access to stone masonry type machines 4) I still have to take it to a machinist to have the inserted lands machined 5) if I use ceramic aggregates to improve its stiffness it becomes nearly impossible to machine & these have not improved the stiffness in my tests. I would consider it if someone wanted a production machine like 1 per week, but then metal is viable at my machine sizes. If the machines are very big then cold casting makes sense.

The only material that I have been able to make that has a modulus greater than aluminium reliably is carbon fibre. I 've made E80 material in compression, tension and flexure in std modulus CF. Its costly and fiddly but a great solution.

So I come back to metals Aluminium, steel or cast iron the traditional material. For me aluminium works as I can machine it on my routers. But in production the cheapest path so far has been steel laser cut parts. If you need really thick parts they can be laminated. You may say what about damping? People harp on about damping but stiffness trumps damping. I have researched dynamic stiffness of machines enough now to know that no material is damp enough to solve the vibration issues in traditional CNC subtractive machines. There's been heaps of research and effort in this area and there are no material answers. The only answers lie in input shaping, active shaping and tap testing at present. These identify the vibration modes of the machine and you just stay away from those operational areas. Using tap testing, productivity improves 10x. So using the stiffest cheapest material is the way forward, then tap test your machine and your on your way. This means steel will be the obvious solution. Either billet machined, fabricated or self-assembled (bolted). If fabricated ideally use laser welding that may sidestep the stress relieve process.

Moving fwd commercial machines will 3D print their parts in metal (this is already occurring), use shaping in their control systems (which is occurring) and tap testing will become a std workshop practice (which is occurring)... unless additive overtakes subtractive in economics. Which in 5-10 years that will be the case... Then those fancy spiderweb designs that AI produce become easy to make and since they use 10X less material become viable to print... and printing is limitless in size... so hot casting will stop... when I'm 80 and in my dotage machines will look entirely different to how they look now...

So for me the conclusion is that laminated aluminium works for me its as damp as cast iron or composites and I can machine them on my routers. Once I get two more routers made I design a fully laminated aluminium mill.... Peter

Hi Pete

This answers many questions I had in mind!

I have actually evaluated many of these options. In summary, Al is too expensive. Cast iron path is difficult due to lack of big foundries which can pour close to 1 ton. The bigger foundries do not take me seriously.
I dumped steel plate base path because I could not get the stiffness I require with simple, non optimized frames economically. In addition, there are no stress reliving facilities that can accommodate a huge pieces like this. The only path for SR would be to build one and then after the work, sell it. I have not considered bolt based assembly path yet. Although I have read your posts and simulation results on bolting, I am still wondering how a bolted joint can give same stiffness/strength to a fully penetrated welded joint. With welding, you are basically combining two pieces together completely while with bolts we are holding them together at discrete points. In addition, I was concerned about damping as well. Apart from sub optimal surface finishes, the noise it will create will be an issue as I will be working with this machine in the mid night. Although I will be sound proofing the room, the noise still could be a problem. The way I see, even if you pour a concrete or Epoxy based mix to cavities, the damping issues will not be fully fixed?

The only thing that is left is Epoxy Granite. I believe the only issue it can have is lower EM. At-least there will not be any cracking or shrinkage.

Thanks
Sus

[suspension]

Hi Pete

I contacted Local sika again and while 3350 is super expensive, this one is 1/3 of it.

I know the modulus is not that high, but I can test it and compensate for it by larger frames. Do you see any other issues in this data sheet?

Thanks
Sus

[peteeng]

Hi Sus - No data sheet?

Strength is not in the equation. Obviously a 100% pen weld is stronger than a bolted assembly. The strain levels in a cnc machine are tiny. There is not enough strain in a CNC bolted connection to allow it to slip. Friction is huge and a prestressed connection due to the preload behaves just like a 90% efficient welded connection in terms of an elastic connection... . and that's in its near slip limit condition. So if you say compare that to al or cast iron you could say the material is 200x0.9=180 MPa vs al at 70 and CI at 100 so your still ahead The connection is not as stiff but it does not slip. Your IC car engine is a self assembled part. It has many many bolts holding it together and it runs for a million kms reliably (or it should) None of the joints slip otherwise the engine would not work or it would leak. The Eiffel tower and other rivetted structures have stood for 100 years and they don't slip. High strength bolts C8.8 provide more pretension then hot rivets do. So C10.9 or C12.9 are even better.

As far as I can tell at the moment Fusions bolted connections are good compared to simsolid bolted FE. So I believe you can rely on Fusions bolt answers.

if your worried about slippage then use a couple of dowels and it can't go anywhere...

Al is not expensive if you use the correct amount of it. If you are happy to have steel parts machined, laser cut or machined then they should be the stiffest cheapest solution. I think the noise issue is misplaced. I do not see why a steel machine would be any more noisy then any other material. The noise is from the cutting tool and if the machine is compliant its noisy as the tool is jumping around. A stiff machine makes a low throbbing hum...

The material cost component is small compared to the total cost of the machine especially if the material is stiff and minimised. That's why I pursue the generative design process... Keep at it.. We all come to different conclusions as our circumstances and rational are different... Peter

compare costs by $/m3 and see how that goes? and $$/stiffness and see what wins...

OK missed the link to Sika 214. Its expansion compensated. Use as little water as possible, see the expansion ratio is tiny when mixed "stiff". It has some calcium sulfate in it and its modified Portland cement. So its fine. The data sheets do not quote modulus so I expect its around 30-35GPa. Add basalt and it will lift to 40GPa...

[peteeng]

Hi Sus et al - So called optimised geometry is independent of material modulus (material stiffness) so you can design a machine totally disregarding material till the end. There are two streams of optimum a) highly triangulated spider web type structures typically what AI produces ie 3d space frames or trusses b) maximum envelope "skins' or monocoques with internal secondary structures such as webs or ribs.

Typically in machine design b) is the usual. We design large thick slabs of material at the "hard" points and then join the dots between them. I'm sure machines are much heavier then they need to be at the required stiffness because designers are conventional and to an extent slow to grasp the idea that a lightweight machine could be as good and cheaper to produce. In my 40 years designing stuff I have seen racing boats, cars and planes go from steel to aluminium to carbon fibre and get lighter and lighter and stiffer and stiffer... Peter

[suspension]

Hi Sus - No data sheet?

Strength is not in the equation. Obviously a 100% pen weld is stronger than a bolted assembly. The strain levels in a cnc machine are tiny. There is not enough strain in a CNC bolted connection to allow it to slip. Friction is huge and a prestressed connection due to the preload behaves just like a 90% efficient welded connection in terms of an elastic connection... . and that's in its near slip limit condition. So if you say compare that to al or cast iron you could say the material is 200x0.9=180 MPa vs al at 70 and CI at 100 so your still ahead The connection is not as stiff but it does not slip. Your IC car engine is a self assembled part. It has many many bolts holding it together and it runs for a million kms reliably (or it should) None of the joints slip otherwise the engine would not work or it would leak. The Eiffel tower and other rivetted structures have stood for 100 years and they don't slip. High strength bolts C8.8 provide more pretension then hot rivets do. So C10.9 or C12.9 are even better.

As far as I can tell at the moment Fusions bolted connections are good compared to simsolid bolted FE. So I believe you can rely on Fusions bolt answers.

if your worried about slippage then use a couple of dowels and it can't go anywhere...

Al is not expensive if you use the correct amount of it. If you are happy to have steel parts machined, laser cut or machined then they should be the stiffest cheapest solution. I think the noise issue is misplaced. I do not see why a steel machine would be any more noisy then any other material. The noise is from the cutting tool and if the machine is compliant its noisy as the tool is jumping around. A stiff machine makes a low throbbing hum...

The material cost component is small compared to the total cost of the machine especially if the material is stiff and minimised. That's why I pursue the generative design process... Keep at it.. We all come to different conclusions as our circumstances and rational are different... Peter

compare costs by $/m3 and see how that goes? and $$/stiffness and see what wins...

OK missed the link to Sika 214. Its expansion compensated. Use as little water as possible, see the expansion ratio is tiny when mixed "stiff". It has some calcium sulfate in it and its modified Portland cement. So its fine. The data sheets do not quote modulus so I expect its around 30-35GPa. Add basalt and it will lift to 40GPa...

Thanks Pete. I am getting a sample of Sika 214 LK tonight. New testing will take two more months, so during that time I can model a new steel plate based frame and see how it goes.
I saw this paper (The only one I could find that had some real implementation details) which has details of a design using steel plates. No testing has been done to check the stiffness, etc however. I am thinking of may be following a similar design to test steel frame for my machine.

Thanks
Sus