[MikeGM]

Hi Peter (and others)

Given your interest in material properties, I thought you might be interested to see the following papers:

https://www.fhwa.dot.gov/publication.../13060/003.cfm - If you scroll down to the 'modulus of elasticity' subheading, this study subjected untreated and heat-treated samples of a UHPC mix to modulus testing in both compression and direct tension tests. The samples tested in direct tension achieved higher elastic modulus than their compression counterparts; with heat treated samples achieving 51.9Gpa in direct tension and 50Gpa in compression respectively. This trend of slightly higher elastic modulus in tension is in agreement with an earlier paper I posted which analysed a standard concrete. This may be a significant distinction between UHPC and epoxy granites/polymer concretes.

https://link.springer.com/article/10...70-021-07118-y - This paper investigates: a natural granite; a UHPC and two types of epoxy granite for creep behavior with a specific focus on precision machinery structures. I wonder whether the elastic modulus in tension could explain this papers finding of lower creep deformation in the UHPC relative to the epoxy granite samples.

Just thought I'd share what I've come across.

Mike

[peteeng]

Hi Mike - Thanks for chirping in. E=50GPa is a common number for UHPC and its commercially available in europe/OZ and USA. I expected the tensile and compressive figures to be the same as the material is by nature nearly isotropic. Just getting the data is the tough bit. The first paper is very good thanks, haven't seen that one before. The poisons ratio is low but not too bad and would fall into the realm of isotropic analysis. But we would then need the shear modulus to confirm isotropic equations are valid. Another number to chase. I can buy CSA grout that gets 40-42GPa out of the bag. The concrete chemist says it will get to 50 if I add basalt aggregate to it. The trick is to get the right basalt. Since the aggregate suppliers can't tell you what the modulus of the agg is, its a can of worms. My main holy grail is the machining side of the part. If I cast in concrete or EG I have to use metal inserts. My aim is a material that I can machine or a machinist is happy to machine. The search goes on.

Sorry second link does not work. Would be good to fix this...All plastics; and epoxy is a plastic, are long chain highly entangled materials. They can and do creep and relax given the correct conditions. ie they become un-entangled or oriented along the strain line. The strain levels in machine parts are very low and I'd expect creep and relaxation to be zero in a correctly cured epoxy granite casting. I was involved in a long term bolting test in a fibreglass structure and did not detect change in the preload over a period of over a year in a highly loaded part. Around 50-100MPa continuous which is 50 to 100x more than a machine part would experience. I have various chunks of epoxy and composites around here that are over 20 years old that have not changed shape but they have been unstressed. Other parts out of other resins have changed shape weeks after casting...

Creep is a phenomena where the material elongates under stress in which it is allowed to elongate eg if you hang a weight from a tree on a plastic rope it can stretch then the molecules orient and slid and allow the rope to elongate over time as the molecules disentangle further and further. It is a condition of constant stress with increasing strain (stretch) over time. So the EG machine part is unlikely to creep.

A long gantry under gravity would have constant stress this may creeeeeep. eg timber beams sag with a similar mechanism over time. but if the gantry is big and correctly sized I cant see the strain being high enough to creep it.

Relaxation is the change of internal stress (usually a decrease) due to constant strain (ie no change in length). Plastics are viscoelastic and this is the viscous side of the material. It means the molecules are slightly sliding or rotating but not disentangling. This slight sliding allows the local strain to reduce so the global stress reduces and equilibrates then the mechanism stops. This is similar to the yield point in metals. At yield the grains slide over each other, stress reduces until the grains lock up again then the stress goes up more sliding etc until breakage. Its also similar to hysteresis in cast irons.

I have been looking for a source of sonic modulus testing and found one in europe but they wanted a really exorbitant amount of $$$ to do the testing. Since I do a lot of composite work this would be an easy way to establish the materials modulus even on existing parts. May have to get some ultrasonic cells and a scope to do this myself.... So the search for the Grail continues. My current conclusion is laminated metal trumps all of these solutions. 200GPa for steel or 70GPa for aluminium machinists are happy I'm happy... I have 2 machines to sell then (one just sold yippee) I can get onto Milli proper at last... Peter

[peteeng]

Hi Mike - I did manage to open the second link from my email... I don't think they successfully measured material creep. The rig is complex and over 66 days (1600hrs) the rig would introduce factors due to the many part contacts involved. The thermal cycles need to be removed from the data or test somehow etc etc. What on face value evident is, that the millions of years old granite, maybe is a bit more stable than the recently made materials. To summarise I expect that correctly cured UHPC and EG will perform satisfactorily as machine parts over a machine's life. Peter

[ardenum2]

Hi Mike - I did manage to open the second link from my email... I don't think they successfully measured material creep. The rig is complex and over 66 days (1600hrs) the rig would introduce factors due to the many part contacts involved. The thermal cycles need to be removed from the data or test somehow etc etc. What on face value evident is, that the millions of years old granite, maybe is a bit more stable than the recently made materials. To summarise I expect that correctly cured UHPC and EG will perform satisfactorily as machine parts over a machine's life. Peter

cooking a concrete makes it much stronger, think I've seen 2x stats(uhpc) after proper cooking, well the romans did it too.

[peteeng]

Hi Ard - Its not about strength - its about is material stiffness and long term dimensional stability. Peter

[peteeng]

Evening all and sundry - Been working on the Tetrium material again. The test lab is happy to run some more modulus tests and I wanted to test the ALOX. But the test machine can only grab up to 4-5mm thick samples and this would not be thick enough to characterise the alox as the grain is 2-3mm spheroidal so would only have 2 grains across the coupon. So I thought I'd park this but then I thought about doing a flexural test in which we can test very thick bits easily. So did some FE modelling to check the sample length was not too short (if too short shear effects come into play and the calculated stiffness will be wrong) and that checked out so first order linear math's work fine sop the test will be fine. So talked to the technician and he's happy to run flexural tests to determine modulus. In many ways this is easier then doing it in tension... So tomorrow I discuss how many tests he's happy to do then I can split that up into three or 4 materials. Very good. Peter

extra - I did the initial FE in Simsolid then decided to do it in Fusion since that's the future system for me. Was straight forward and answer matched hand calc as it should.... simsolid solves this sort of thing instantly was unusual for me to wait a few minutes to get an answer,....

[ardenum2]

Evening all and sundry - Been working on the Tetrium material again. The test lab is happy to run some more modulus tests and I wanted to test the ALOX. But the test machine can only grab up to 4-5mm thick samples and this would not be thick enough to characterise the alox as the grain is 2-3mm spheroidal so would only have 2 grains across the coupon. So I thought I'd park this but then I thought about doing a flexural test in which we can test very thick bits easily. So did some FE modelling to check the sample length was not too short (if too short shear effects come into play and the calculated stiffness will be wrong) and that checked out so first order linear math's work fine sop the test will be fine. So talked to the technician and he's happy to run flexural tests to determine modulus. In many ways this is easier then doing it in tension... So tomorrow I discuss how many tests he's happy to do then I can split that up into three or 4 materials. Very good. Peter

extra - I did the initial FE in Simsolid then decided to do it in Fusion since that's the future system for me. Was straight forward and answer matched hand calc as it should.... simsolid solves this sort of thing instantly was unusual for me to wait a few minutes to get an answer,....

I don't understand what the point of testing 2-3mm big grain is? I remember your bowl test a few months back, it looked like someone got shot from a shotgun and removed the pellets

Fusion FE uses meshes? you coming from simsolid, the pain will be excruciating...don't even think about running an assembly, not unless you have a second pc to use while this cooks for days

[peteeng]

Hi Ard - Yes its granular like shot pellets. The aim is to be able to pour a dry material into a mould, then pour thin epoxy in until filled or use a sealed mould and use vacuum casting. The size of the grain does not matter its the solid fraction that matters, then the efficiency of the bulk strain transfer. The test coupons will be 40mmx10mmx300mm long. ALOX should be the stiffest (300GPa) easily available material at reasonable price. I have played with Silicon Carbide (400GPa) many moons ago but that was in a wear prevention exercise... but SiC is $$$

If strain transfer is perfect the ALOX E=300x0.6=180GPa which is unlikely. If efficiency =0.5 then ALOX E=90GPa which would be excellent... and attractive, test will answer that question.

Aluminum Oxide | Al2O3 Material Properties (accuratus.com)

Silicon Carbide SiC Material Properties (accuratus.com)

Sand or Silica which a lot of people use in machine parts is only ~70GPa Fused Silica | SiO2 Material Properties (accuratus.com) same as E-Glass fibres

Basalt and granite are only marginally stiffer than sand. There are some granites same stiffness as cast iron but hard to come by. But you've followed the thread so know a lot of this stuff. But currently I'm convinced laminated metal is the go. This test is being done to round out the data I have.... Peter

Re - Fusion yes F360 uses meshes and I'll have to wait for answers. But currently someone else pays for SimSolid and one day that will cease so need to take advantage of SS as best as possible and have a backup when that stops. I have Strand7 which is good but hopeless with contacts. So soon when I semi-retire and won't have SS I'll have to look at other FE systems or use hand calcs... or seat of the pants stuff... So Milli is looking like laser cut steel, zinc plate the parts, then laminate together, then finish machine. Use generative design or adaptive design in Fusion to establish internal geometry with maximum stiffness. Or make aluminium parts myself on router then laminate etc. Al won't be as stiff as steel... but I can make AL until I get Milli running then I can make steel parts. The Grail is out there...

[pippin88]

https://youtu.be/0yFKaCn9soo

DMG rising gantry

[peteeng]

Hi pippin - I do like Mori vids. The raising gantry (RG) is great for 5 sided work. I intend to do deep moulds and this is the issue. With the RG you need massive tools see the thickness and length of some of them! But maybe the same with conventional Z. I'm finishing up Frankie then I'm into the next design. A mill maybe the next thing... or at least a router of the same config to sort out the issues. I need a long router/mill to make gantry moulds or gantries.. Peter

On another note has anyone made parametric spaceframes in Fusion? I want to compare spaceframe to prismatic gantry construction... Working on 60 deg planes is a pain!! and wireframing then piping is ditto a pain.... My dots don't line up yet... Peter

[pippin88]

Hi pippin - I do like Mori vids. The raising gantry (RG) is great for 5 sided work. I intend to do deep moulds and this is the issue. With the RG you need massive tools see the thickness and length of some of them! But maybe the same with conventional Z. I'm finishing up Frankie then I'm into the next design. A mill maybe the next thing... or at least a router of the same config to sort out the issues. I need a long router/mill to make gantry moulds or gantries.. Peter

On another note has anyone made parametric spaceframes in Fusion? I want to compare spaceframe to prismatic gantry construction... Working on 60 deg planes is a pain!! and wireframing then piping is ditto a pain.... My dots don't line up yet... Peter

Ultimately the spindle nose must clear the work. Only if doing quite large and open moulds would there be a difference.

[peteeng]

Hi Pippin - Yes I think your right. I think the high wall is stiffer though... Peter

[peteeng]

Nope the RG maybe better - have to make a model to determine that. But if the gantry is the same so thats out of the equation and the saddle is stiffer then the conventional design as it does not move up&down, is short (only one set of bearings to stack) and stiff. Then the double column is stiffer then the single long Z axis... Hmmmm Peter

[peteeng]

Evening All - has anyone joined square rail? BST can only supply to 2m and I want 3m. They can match grind lengths so has anyone done this? Peter

[ardenum2]

Evening All - has anyone joined square rail? BST can only supply to 2m and I want 3m. They can match grind lengths so has anyone done this? Peter

This guy has: (TUTORIAL) Butt-joining profile linear rail with limited tools

[peteeng]

Hi All - Vibration is a big topic at the moment so...

https://www.mmsonline.com/articles/w...ics-in-milling Peter

[peteeng]

more good vib info - peter

https://www.mscdirect.com/betterMRO/...proach-chatter

[ardenum2]

rails usually come in 4m so that's weird, personally I'd just find a different vendor than attempt to buttjoin 2 rails. I'm eyeing these guys as of now : https://tnkbearing.en.alibaba.com/pr...ear_guide.html

They have rexroth stuff too, and they sell rails for the carriages as well, the prices are without customs and tax though, because its alibaba not aliexpress

[ardenum2]

from hiwins manual:

[peteeng]

Hi Ard - Its not the suppliers its the couriers to get it here. They have length limits. Peter