[MarkLiu]

Take a look at the STEPPERONLINE t6 series servo motors. The reviews and prices seem to be pretty good.

[Ashley-techprolaser]

Just little suggestions,AC servo may be better for cnc,and Leadshine servo motor drivers are more affordable.

[suspension]

I've done several specimen samples at home to understand the most effective mixture (grout + chips (max 15mm)) for stiffness/cost. Apart from chips, I also tested with sand, gravel dust/powder and coarse gravel.


Based on a simple deflection test, the pure grout showed the lowest deflection followed by (almost identical) 50:50 mixture with chips. 70% chips sample seems to have many voids. course gravel sample broke while tightening to the vice during test. 50:50 with sand showed worst performance.


This made me chose the 50:50 mixture with chips (max 15mm).


Next step is to send the standard sized specimens to the lab for the tests for compressive strength, flexural strength, EM, flexural modulus and dry shrinkage. While reading about flexural modulus test here, I started to wonder why cannot I do this at home since I have a high resolution dial gauge, a way to create a known force (just use a known weight), etc. To understand the accuracy I can use a sample from a know material and compare. What are your thought on this? Of cause, the motive here is to reduce the cost.


In addition, I have attached the setup I had for the mixture test. I was wondering if I can calculate a rough figure for the flexural modulus based on this setup as well.


Thanks
Sus

[peteeng]

Hi Sus - You need to make a long cantilever or beam for a flexure test. What you have drawn is a short beam which is shear deflection dominant. If you have a steel or aluminium solid bar at least 10x its thickness or more (prefer 20x). So if its 20mm thick it needs to be 200mm long prefer much longer... So if you deflect the steel or al bar with a known load and the cast concrete bar is the same geometry then you can proportion the deflection to the modulus. Since we know steel=200GPa and Al=70GPa Say we have an AL bar 20x100x500mm long and a cast bar the same size. If we place 10kg on the Al bar and it deflects 20mm then we set up the concrete bar so its cantilever is the same length (apples for apples) and it deflects 40mm then the modulus is half the AL: ie 70/2=35GPa... That would be all you need to know... Peter

a 15mm chip means you need say a 30mm deep coupon, so the chip is not dominant in the section, deeper is better but then you need a big steel bar or al bar with lots of load. if you make a cantilever of known size and place three prefer 4 loads on it and measure the deflections, I can calculate the modulus for you.. Or if you have FEA you can model it and change the model E until it agrees with the actual deflection include gravity if you only use one load. But you need to use 3or4 to figure out an average. Plus turn the bar over and repeat in the opposite direction to check the bar does not have a bias.. Peter

what grout did you get?

[suspension]

Hi Sus - You need to make a long cantilever or beam for a flexure test. What you have drawn is a short beam which is shear deflection dominant. If you have a steel or aluminium solid bar at least 10x its thickness or more (prefer 20x). So if its 20mm thick it needs to be 200mm long prefer much longer... So if you deflect the steel or al bar with a known load and the cast concrete bar is the same geometry then you can proportion the deflection to the modulus. Since we know steel=200GPa and Al=70GPa Say we have an AL bar 20x100x500mm long and a cast bar the same size. If we place 10kg on the Al bar and it deflects 20mm then we set up the concrete bar so its cantilever is the same length (apples for apples) and it deflects 40mm then the modulus is half the AL: ie 70/2=35GPa... That would be all you need to know... Peter

a 15mm chip means you need say a 30mm deep coupon, so the chip is not dominant in the section, deeper is better but then you need a big steel bar or al bar with lots of load. if you make a cantilever of known size and place three prefer 4 loads on it and measure the deflections, I can calculate the modulus for you.. Or if you have FEA you can model it and change the model E until it agrees with the actual deflection include gravity if you only use one load. But you need to use 3or4 to figure out an average. Plus turn the bar over and repeat in the opposite direction to check the bar does not have a bias.. Peter

what grout did you get?

Thanks a lot Pete, so the test I have done to check various mixes may not be very correct as with the dimensions i used, its the shear deflection that is dominant.
Assuming a 40mm thick specimen (two times the max chip size + some margin) the minimum length should be 800mm (20X factor). Loading a similar sized Al or steel would not be very practical. So I guess what I can do is make a cantilever and put 4 different loads and measure the deflection? I also thought about the FEA path you suggested, but it is good if you can also help me calculate it.

I have used excem 80GP which is available in Sri Lanka.
https://www.chryso.lk/p/8921/chryso-excem-gp-80

[badhabit]

[...] So if you deflect the steel or al bar with a known load and the cast concrete bar is the same geometry then you can proportion the deflection to the modulus. Since we know steel=200GPa and Al=70GPa Say we have an AL bar 20x100x500mm long and a cast bar the same size. If we place 10kg on the Al bar and it deflects 20mm then we set up the concrete bar so its cantilever is the same length (apples for apples) and it deflects 40mm then the modulus is half the AL: ie 70/2=35GPa... That would be all you need to know...[...]

This is the way :-)

I've been meaning to do the exact same thing with my EG(R&G resin with Silimix282 in 11% by-weight mix)...just haven gotten around to do it yet... i would be very interesting if we somehow could do the "same'ish" test, then we could do a proper "apples to apples" test, to use Peters words...

/Thomas

[peteeng]

Hi Sus - The excem 80GP's data sheet says >47GPa so why test? use 50GPa and you will be very close. With this sort of material you are looking for comparative results not absolute results. A "thin" poorly distributed test sample will be different to your large real part. So the FE is used as a guide to make bigger decisions vs detailed decisions in this sort of case.

The thing to check if you use FEA is that some linear elements do not include shear deflection. So you have to use lots of elements to capture the shear deflection in short beams or you have to use a second order element that includes shear deflection. So read up in your FE system manual or help and check that shear deflection is included in the simple element or change the meshing to an element that includes shear or do a test with a small beam and then check or calibrate the FE against the short beam.... What FE system do you use? Peter

This problem also translates to your and others machine models. Lots of these thick short structures in machines are "short beams" vs "slender" beams. If you use a simple mesh element it will miss the shear deflection and appear stiffer than it actually will be. So read up on your elements before you hang your hat... Peter


Short Beam Bending: Roark's Formulas For Stress and Strain (physicsforums.com)

https://en.wikipedia.org/wiki/Timosh...st_beam_theory

Why Euler-Bernoulli beam theory does not work in this case? (physicsforums.com)

[suspension]

Hi Sus - The excem 80GP's data sheet says >47GPa so why test? use 50GPa and you will be very close. With this sort of material you are looking for comparative results not absolute results. A "thin" poorly distributed test sample will be different to your large real part. So the FE is used as a guide to make bigger decisions vs detailed decisions in this sort of case.

The thing to check if you use FEA is that some linear elements do not include shear deflection. So you have to use lots of elements to capture the shear deflection in short beams or you have to use a second order element that includes shear deflection. So read up in your FE system manual or help and check that shear deflection is included in the simple element or change the meshing to an element that includes shear or do a test with a small beam and then check or calibrate the FE against the short beam.... What FE system do you use? Peter

This problem also translates to your and others machine models. Lots of these thick short structures in machines are "short beams" vs "slender" beams. If you use a simple mesh element it will miss the shear deflection and appear stiffer than it actually will be. So read up on your elements before you hang your hat... Peter


Short Beam Bending: Roark's Formulas For Stress and Strain (physicsforums.com)

https://en.wikipedia.org/wiki/Timosh...st_beam_theory

Why Euler-Bernoulli beam theory does not work in this case? (physicsforums.com)

The main reason I want to test is because the value given in the data sheet is for grout only mixture while I am using 50:50 grout and chips. In addition as mentioned by you some time back what is most relevant in this application is flexural modulus than young a modulus. The lab test include a flexural modulus test.


As per your last comment the dimensions used in the machine are very different from a long thin sample that will be used for the flexural modulus calculation. If the deflection in these dimensions are mostly influenced by other factors that is not captured by flexural modulus would doing this test give me any significant insight? Perhaps a better test is to make a sample that has scaled down dimensions as the pieces used in the machine then apply scaled down forces and measure the deflection and extrapolate it to machine size?

Thanks
Sus

[peteeng]

Hi Sus - This is an interesting discussion. Scaling down may not be a good thing. The aggregate does not scale relative to the part size. Unless you scale that down too. If the 15mm agg becomes large compared to the part size then the "material homogenisation" used for modelling may not be accurate. There are all sorts of things that won't scale. Poisson's ratio, shear modulus etc etc. Can of worms unless you intend do a PHd on it... So as I said this issue of modelling is relative not absolute. Use 50GPa isotropic and compare designs in the knowledge that your looking for major differences and insights into the design not absolute answers. But a real test is good if possible. Reality sucks. I was really disappointed with my alumina beam tests... But that's a great outcome as I would have assumed it was much better if I didn't do the test... So are you adding agg to reduce cost or to improve modulus? Peter

[peteeng]

Hi Sus et al - I made small cantilever in F360 50x50 200mm long and placed a 5000N load on its end. In F360 the alum version deflected 0.40mm the steel deflected 0.13mm. I transferred the model to simsolid and rerun and the Al deflected 0.38mm. I checked and the value of E for AL is different in each database. Correcting for the difference the result is the same. I then manually calculated the deflection using the usual slender beam equation and the answer is 0.38mm so the beam is slender or the FE is missing the shear deflection. I'll have to check that out in F360 and SS to understand its algorithm. But I think you could tune an FE model to the real beam and get a good figure. So build the best beam you can and load it up!! Peter

[suspension]

Hi Sus et al - I made small cantilever in F360 50x50 200mm long and placed a 5000N load on its end. In F360 the alum version deflected 0.40mm the steel deflected 0.13mm. I transferred the model to simsolid and rerun and the Al deflected 0.38mm. I checked and the value of E for AL is different in each database. Correcting for the difference the result is the same. I then manually calculated the deflection using the usual slender beam equation and the answer is 0.38mm so the beam is slender or the FE is missing the shear deflection. I'll have to check that out in F360 and SS to understand its algorithm. But I think you could tune an FE model to the real beam and get a good figure. So build the best beam you can and load it up!! Peter

Hi Pete,

sorry I am bit lost now Are you suggesting following?

1. Do some cantilever tests with a beam of size say 40 x 40 x 800 mm and measure the average deflections for different loads.
2. Setup FEM software so it takes in to account shear deflections. Models the same beam in CAD and do some FEA simulations until you get the same deflection as the real test for each load. To do this I will have to keep everything constant except some modulus parameters? Which parameter should I alter in this case? If it is more than one, I am sure there will be multiple parameter combinations giving same deflections so I wont be able to find a value for the modulus I am after?

Thanks
Sus

[peteeng]

Hi Sus - If I'm manually calculating a modulus I require 3 or 4 load points so I can plot a graph slope with good correlation. But by using FE a single load is acceptable if your careful eg include gravity in your model unless your on the moon? You don't have to include gravity if you only measure the deflection of the single load. A 40x40x800mm cantilever will be good. Model it and determine a suitable load that does not break the beam so you know what sort of deflections you will get. What FE are you using? Then turn it over and repeat to check it is not biased. Use 2 loads and the second load should be twice the first so the deflection should be twice the first. If not then recheck things like how you attach it to the bench as the bench may have bent... Once you have a consistent deflection/load number you model the beam and vary E until it matches the measured deflection. In fusion you also have to be careful as with isotropic materials you have to add the shear modulus. In isotropic materials (which we are assuming the concrete agg mix is) the shear modulus is related to the youngs modulus by

G= E/(2x(1+u)) u=poissons ratio say 0.3 so for aluminium G=70000/(2.6)=26900MPa commonly accepted as 27000MPa... Good Luck Peter

this process is called calibration you are calibrating the theoretical object to the real object....

to cover some basics - In an isotropic material the tensile, compressive and flexural moduli are the same. The load history does not change the end deflection and the rules of superposition hold. So you should be able to repeat the process and get the same result. If its not isotropic the material can relax or stiffen during a load cycle and if extra loads are applied during a load cycle the total deflection can be different. If this occurred our machines accuracy would be in trouble. In reality we know that machines behave in a linear fashion and load history and viscoelastic affects do not happen. ie the material is isotropic....

[peteeng]

Hi Sus - et al - Reality check - SUCCESS!! So I have some 80x6 flat plate extrusion and I decided to reality check the figures and see if a one off measurement is close enough. I clamped the plate over the bench edge at 1000mm and let gravity droop the bar. I measured the defection at 15.77mm then went inside and did the calcs and the FE model. Probably a fluke but got the same answer. So by calc or FE you will get thew same result and can back calculate the modulus from the beam test. My reference beam is 3mm thick that's why the reading is 18mm.

80x6x1000mm measured deflection = 15.77mm

Fusion 360 FE deflection 15.81mm
Beam calc deflection 15.77mm with E=70GPa

Close enough... Peter

[suspension]

That's nice. I will do something like this for a known material and then move to the unknown mix. Thanks a lot Pete.

[suspension]

Hi

I am in the process of finalizing the Z bracket which carries the 3.7KW spindle motor (40Kg) + BT30 spindle + its bracket + pneumatic stuff. I hope the full weight would be close to 100Kg (except the bracket ).
Two options are of course possible, use steel plates welded together or work with a foundry to make it using cast iron. Steel plate based process is much simpler but definitely require stress reliving. (based on my limited experience/knowledge). I was wondering if cast iron based method would also require stress reliving. I created designs for both options and cast iron bracket weighs around 100Kg and has wall thickness of 30mm in most places.

Thanks
Sus

[peteeng]

Hi Sus - There are cast irons that auto stress relieve in the mould prior to release. Talk to the caster about options. Peter

[joeavaerage]

Hi,
one of the reasons that grey cast iron is so favored for casting is its modest shrinkage, about 1%, whereas steel (which requires a hard mold) is about 2% and SG iron at about 2.6%.
Shrinkage is the reason for residual stress in the first place, ergo the less shrinkage the less residual stress. If all parts of the casting cooled at the same rate there would be NO residual stress,
and while that is not practical, if you design the sections to be all the same or least least similar you will achieve next best thing to uniform cooling.

I am of the opinion that a small to moderate part, say 5kg, designed with equal or near equal sections will have little residual stress. If the part required no machining then it would most likely never move at all.
Its when you 'peel' the skin off a casting that the residual stress shows up.

My advice is cast it and be done. I wouldn't worry about stress relief, it will be fine as is. If you are concerned find someone local with a pottery kiln and have them take it up to 650C for an hour or
so before a slow cooling. Pottery kilns regularly have to do 1100C, so 650C they could do in a canter.

Even if you paid for it, I priced stress relief for 20kg welded steel and the quote was $134NZD plus GST (15% local tax) or $4.60USD/kg (inclusive of tax). A 5kg part is hardly going to break the bank.

Be aware that I am a cast iron aficionado...its expensive but the results are stunningly good, stiffness and damping.

Craig

[suspension]

Hi Craig,

The thing is in my case the bracket is roughly 106Kg and size is like 350 X 350 x 350 mm. Also the foundry I will be using (in fact any foundry) use recycled iron and these usually come from vehicle engine blocks. I am not sure what material is being used to manufacture engine blocks (but I was told it is cast iron) and the shrinkage % may vary. It is not practical to find a foundry which use standard materials (except a government owned facility which quoted 3 times higher costs). Given this, do you still think it is safe to go without stress reliving? I have also attached the design which also indicate where machining should be done.

Also, after seeing many DIY heat treatment builds in youtube I feel that building one to accommodate this part can be done at a reasonable cost. What are your though on that? In case I decide to build one, it can of be used to stress relive steel plate welded design as well. In this case what would be the better choice if we forget about the cost?


See one build here:
https://www.youtube.com/watch?v=0ZLMmbYKK9c


Hi Pete,
The foundry said they will keep it in the sand until it cools down - but not sure how long, probably 12 hours max. My assumption is this is not enough for a proper cool down.

Thanks
Sus

[peteeng]

Hi Sus - Engine blocks are generally made from Grey Cast iron grade 20 or 25. This is stress relieved after casting to remove the cooling stresses in complex casts caused by the variable cool down and heat zones from the pour. Around 600C for 1hr per inch thick is the the usual soak time. In your case leave a generous machining allowance and have it rough machined. If it looks stable go the rest of the way. If its not stable then find a TSR source. In your case for a single part, I'd fabricate in steel and finish machine it. Steel is twice as stiff as CI and if you leave a generous machining allowance and the welder does a sensible job I'd expect it to clean up quite well. Being small and symmetric I'd expect it to work out OK. Peter

[peteeng]

Hi Sus - Engine blocks are generally made from Grey Cast iron grade 20 or 25. This is stress relieved after casting to remove the cooling stresses in complex casts caused by the variable cool down and heat zones from the pour. Around 600C for 1hr per inch thick is the the usual soak time. In your case leave a generous machining allowance and have it rough machined. If it looks stable go the rest of the way. If its not stable then find a TSR source. In your case for a single part, I'd fabricate in steel and finish machine it. Steel is twice as stiff as CI and if you leave a generous machining allowance and the welder does a sensible job I'd expect it to clean up quite well. Being small and symmetric I'd expect it to work out OK. Peter