[peteeng]

Hi JS - I too have struggled for a couple of years doing FE and building virtual mills trying to get reasonable stiffness. Its a mission. But I have come to the conclusion to use laminated steel for the construction. The parts would be laser cut and stacked together, dowelled or bolted and bonded with epoxy, then finish machined where needed. This gives the opportunity to make parts the exact size you want, no welding, laser cut parts are economical and the lamination makes it damp. I have made a few parts like this for my router and they worked well. I could also make it in aluminium on my router but steel is 3x stiffer so is very attractive....I've played with epoxy granite, UHPC, carbon fibre and glass fibre but keep coming back to steel is its the stiffest & cheapest material. Keep chasing the grail... Peter

[Hezz]

Hi JS - I too have struggled for a couple of years doing FE and building virtual mills trying to get reasonable stiffness. Its a mission. But I have come to the conclusion to use laminated steel for the construction. The parts would be laser cut and stacked together, dowelled or bolted and bonded with epoxy, then finish machined where needed. This gives the opportunity to make parts the exact size you want, no welding, laser cut parts are economical and the lamination makes it damp. I have made a few parts like this for my router and they worked well. I could also make it in aluminium on my router but steel is 3x stiffer so is very attractive....I've played with epoxy granite, UHPC, carbon fibre and glass fibre but keep coming back to steel is its the stiffest & cheapest material. Keep chasing the grail... Peter

Has anyone tried a tubular steel space frame with heavy rebar inside and filled with EG in the gap. Would be very strong and stiff and pretty good damping properties

[peteeng]

Hi Hezz - yes that's been done - if you look at the epoxy granite thread and many other threads. Everything has been done. But if you use a tubular steel frame, why fill it with rebar and EG? if the EG is not in the load path (which it is not) then you do not gain stiffness or dampness just mass. EG is nowhere near as stiff as steel and strain goes down the stiffest path. If you add rebar inside the tube you do not gain stiffness again as its only the outer skin that is doing the flexural work. All of these things have been done but no before and after tests have been done to correctly establish if it works. My guess is it does not. If the machine is designed in steel and it is designed stiff enough then it will perform correctly. Hollow section steel tubes are great for global bending stiffness but have poor local stiffness. They are also noisy as they are hollow but this acoustic noise does not mean they are affecting the cut. If you build a tubular steel spaceframe which is highly triangulated then there is no reason for it to be tubular. As the members are in tension or compression so they are only as stiff as their area. So a 75x75x5 SHS is 1400mm2. This is the same tensile/compressive stiffness as a 37x37mm solid bar and same weight. If you do build a highly triangulated 3D truss structure then the members could be quite small as its using the triangular geometry for stiffness not the geometric inertia of the section. If you look at 3D printed optimised structures you will see they are spider web type thin trusses very light and very stiff.

If you choose to use steel hollows then make them thicker to add mass and add local stiffness. EG is way more expensive than steel by volume and if it is only adding mass then its better to add steel thickness... steel 7800kg/m3 and EG is say 2500kg/m3 so you only need 1/3 the volume of steel to get the same mass at maybe 1/10 the cost.

Mass is an interesting earth bound concept. If you took a mill to space it would work just as well but it would have zero mass. So you have to be careful with the mass is great argument....Peter

[Hezz]

Hi Hezz - yes that's been done - if you look at the epoxy granite thread and many other threads. Everything has been done. But if you use a tubular steel frame, why fill it with rebar and EG? if the EG is not in the load path (which it is not) then you do not gain stiffness or dampness just mass. EG is nowhere near as stiff as steel and strain goes down the stiffest path. If you add rebar inside the tube you do not gain stiffness again as its only the outer skin that is doing the flexural work. All of these things have been done but no before and after tests have been done to correctly establish if it works. My guess is it does not. If the machine is designed in steel and it is designed stiff enough then it will perform correctly. Hollow section steel tubes are great for global bending stiffness but have poor local stiffness. They are also noisy as they are hollow but this acoustic noise does not mean they are affecting the cut. If you build a tubular steel spaceframe which is highly triangulated then there is no reason for it to be tubular. As the members are in tension or compression so they are only as stiff as their area. So a 75x75x5 SHS is 1400mm2. This is the same tensile/compressive stiffness as a 37x37mm solid bar and same weight. If you do build a highly triangulated 3D truss structure then the members could be quite small as its using the triangular geometry for stiffness not the geometric inertia of the section. If you look at 3D printed optimised structures you will see they are spider web type thin trusses very light and very stiff.

If you choose to use steel hollows then make them thicker to add mass and add local stiffness. EG is way more expensive than steel by volume and if it is only adding mass then its better to add steel thickness... steel 7800kg/m3 and EG is say 2500kg/m3 so you only need 1/3 the volume of steel to get the same mass at maybe 1/10 the cost.

Mass is an interesting earth bound concept. If you took a mill to space it would work just as well but it would have zero mass. So you have to be careful with the mass is great argument....Peter

Thanks for the info, I have been out of this DIY CNC headspace for a while. My interest is in getting a pretty stiff and well damped machine in a reasonable light weight for a portable home CNC medium size. Something that can be moved around and does not weight more than 500-800 pounds. I read a research paper once in high end CNC gantry design. They seem to use a constrained layer type of system with two hollow steel members. They can be round or rectangular. One inside the other and a low hysteresis thin material between them. Gantry systems have particular problems due to the spans and stiffness requirements. With a standard type mill I think pure mass and stiffness are better options. I want to build a clone or quazi-clone of the Daytron M8, which I think is just about the most useful medium size mill that I have seen. It's more like a heavy duty router that can do steel and alloys. It's pretty much a steel/EG hybrid machine.

[peteeng]

Hi Hezz - To duplicate a Datron you need a really solid super stiff gantry. Constrained layer damping is difficult to get right when you do it in telescoping tubes, but easier to do in layered laminates of metal. Worrying about damping is a second order issue. The first thing to get right is the machine rigidity, if thats right then damping is nearly a non event. There are heaps of fabricated tubular steel routers and mills that work just fine and they have no damping strategy, they are simply uber stiff. Designing to a weight limit is tricky, the way forward if you have Fusion 360 is to design block geometry then use the generative design function, so F360 designs/generates the minimum weight same stiffness structure... It does provide insight into what the structure should look like. Datron uses high speed machining techniques that allow its light structure to have good MRR's. Modern machines are getting stiffer and a bit lighter. The usual hiccup is that if you fabricate a steel frame it needs to then be stress relieved so when its finish machined it does not move. If you can have your frame thermally SR (or vibrationally SR) then you can build a very very good machine with no thought about damping. Peter

You have read the Bambach thesis and all machine designers should have this in their back pocket. But he caved into a design that was not ideal for some reason. If you follow his logic through the document his conclusion does not line up with his body of work. As he was building a machine for a commercial customer and that being the patron I think realities kicked in near the end... and then you have to pick your poison.

[Hezz]

Hi Hezz - To duplicate a Datron you need a really solid super stiff gantry. Constrained layer damping is difficult to get right when you do it in telescoping tubes, but easier to do in layered laminates of metal. Worrying about damping is a second order issue. The first thing to get right is the machine rigidity, if thats right then damping is nearly a non event. There are heaps of fabricated tubular steel routers and mills that work just fine and they have no damping strategy, they are simply uber stiff. Designing to a weight limit is tricky, the way forward if you have Fusion 360 is to design block geometry then use the generative design function, so F360 designs/generates the minimum weight same stiffness structure... It does provide insight into what the structure should look like. Datron uses high speed machining techniques that allow its light structure to have good MRR's. Modern machines are getting stiffer and a bit lighter. The usual hiccup is that if you fabricate a steel frame it needs to then be stress relieved so when its finish machined it does not move. If you can have your frame thermally SR (or vibrationally SR) then you can build a very very good machine with no thought about damping. Peter

You have read the Bambach thesis and all machine designers should have this in their back pocket. But he caved into a design that was not ideal for some reason. If you follow his logic through the document his conclusion does not line up with his body of work. As he was building a machine for a commercial customer and that being the patron I think realities kicked in near the end... and then you have to pick your poison.

Thanks Peteeng, I have access to Solidworks and Inventor as I am a designer. I have designed a few portable CNC hobby designs and other machines but not yet built one. I will look at Bambach and see what his ideas are. I am not hard core on the weight. But something that is light enough to move around on heavy duty castors. I have some engineering college training in statics and machine design and mechanisms but have been running and programming a CNC router for the last four years. Some of the design skills are rusty.

My idea to make the EG plinth like the Daytron has but more light and stiff would be to have steel torsion box embedded inside of the EG casting with lost of holes to bind them together and make it less thick. maybe four to six inches instead of 12 inches thick. Also, lower gantry rails would be mounted upside down on the bottom of the EG plinth and steel subframe so they are out of the way of dripping coolant and chips.

[pippin88]

Hi Hezz - yes that's been done - if you look at the epoxy granite thread and many other threads. Everything has been done. But if you use a tubular steel frame, why fill it with rebar and EG? if the EG is not in the load path (which it is not) then you do not gain stiffness or dampness just mass. EG is nowhere near as stiff as steel and strain goes down the stiffest path. If you add rebar inside the tube you do not gain stiffness again as its only the outer skin that is doing the flexural work. All of these things have been done but no before and after tests have been done to correctly establish if it works. My guess is it does not. If the machine is designed in steel and it is designed stiff enough then it will perform correctly. Hollow section steel tubes are great for global bending stiffness but have poor local stiffness. They are also noisy as they are hollow but this acoustic noise does not mean they are affecting the cut. If you build a tubular steel spaceframe which is highly triangulated then there is no reason for it to be tubular. As the members are in tension or compression so they are only as stiff as their area. So a 75x75x5 SHS is 1400mm2. This is the same tensile/compressive stiffness as a 37x37mm solid bar and same weight. If you do build a highly triangulated 3D truss structure then the members could be quite small as its using the triangular geometry for stiffness not the geometric inertia of the section. If you look at 3D printed optimised structures you will see they are spider web type thin trusses very light and very stiff.

If you choose to use steel hollows then make them thicker to add mass and add local stiffness. EG is way more expensive than steel by volume and if it is only adding mass then its better to add steel thickness... steel 7800kg/m3 and EG is say 2500kg/m3 so you only need 1/3 the volume of steel to get the same mass at maybe 1/10 the cost.

Mass is an interesting earth bound concept. If you took a mill to space it would work just as well but it would have zero mass. So you have to be careful with the mass is great argument....Peter

Peter,

Filling tubular steel does improve damping (maybe not as much as if it's only EG without steel). There are papers that have tested and proven this.

I would like to see testing of filling steel sections with things like urethane and rubbers

[peteeng]

Hi Pippin - Please direct these papers to me sounds interesting. Often these tests us cantilevered samples that in no way represent how the member is used in a machine. I feel its only adding mass which is then a bit harder to excite. There are many videos of hitting open and filled tubes with hammers to "prove" they are damper. This does not prove they are damper in a machine application unless you are hitting things with hammers. Machines are under very very low strains and those sort of tests are high strain tests. Peter

Hi Hezz - In my Milli thread I spend a year or so working through EG, UHPC, carbon fibre, glass fibre and metals and combos. I feel EG is a dead end for hobbyists. Too expensive and the modulus achieved is too low for the cost. UHPC is cheapest per volume but again is low modulus around 40GPa. Steel at 200GPa and cheap is very attractive and if you laminate it its damper then cast iron. Laser cut steel is a cheap and accurate process, then finish machine and your done. The inside of the steel layers can be tailored to an optimum shape.

Embedding a steel box into EG is in my view the wrong way to go. But get busy and make models to prove out a few things for yourself and get quotes for epoxy that will scare you. I buy epoxy at $22/kg or $22/litre AUD (3.5GPa stiffness and I just had a quote for some aluminium 25mm plate at $40AUD per litre and its 70GPa and easy to machine. If you get the sand for free then the epoxy is not too bad but then you will only get 30-35GPa out of the material... So you need a lot of it to get to the same place as steel for a given geometry... Keep at it. If your machine vibrates its simply not stiff enough. Peter

Filling tubes with stuff is clearly the wrong approach. All aerospace structures are spider web type trusses. Fusion360 has generative design that will figure out the shape something should be at half the weight and same stiffness as the bounding box solution. Bambach did good work but that was over 20 years ago and we have moved on! Hobby builders are a bit obsessed with SHS and RHS and welding and fettling and damping, we have to move on way past that... and Hezz look up input shaping that the 3D printing industry uses. This is an approach that cancels structural vibrations using electronic damping. This is a much better strategy then material damping as it is a direct accurate method vs a splatter gun hope for the best approach via material damping. material damping will never be good enough for high performance low weight structures. If material damping was good enough we would not have shock absorbers on our car wheels we'd just have a slab of rubber. Rant ended....

[ardenum2]

to have steel torsion box embedded inside of the EG casting

When it comes to cold casting EG/UHPC you do not embed any structures inside it, only inserts and anchors and finally if you have a series of holes that need to be more precisely positioned to one another, a plate with anchors, For rails you can put a long square steel piece with anchors or use segmented steel blocks instead, but you could mount the rails directly on the cast and bolt to the threaded inserts underneath. That's it. Mounting surfaces need to be sanded/lapped or ground.

If where you live, they do heat treating for reasonable prices, your best option is steel fabrication. Can't beat it. If they have insane prices, you could risk it and do vibratory stress relief, but it can be tricky without proper equipment. You'd have to do a lot of trial and error to find the correct frequency of your fabrication, the vibration motor must reach close to the resonant frequency of the part. It's a high frequency, low amplitude process. It will be tricky, I don't think it's been done in the diy community yet.

If there's anyone who could walk you through the process it's Pete, who did this on very large fabrications in his work, so there's hope...

[peteeng]

Hi - This is the company that supplied the VSR equipment to the company I do FE for. If you do a modal analysis on your structure that will tell you what freqs you need. TSR is simpler if you have oven access but 20m or 30m mining vehicles are a bit big to put in ovens....It may not be low amplitude, we have had things crack during the process anything that can waggle needs to be braced.......Peter

https://www.vsrtechnology.net/

[ardenum2]

Hi - This is the company that supplied the VSR equipment to the company I do FE for. If you do a modal analysis on your structure that will tell you what freqs you need. TSR is simpler if you have oven access but 20m or 30m mining vehicles are a bit big to put in ovens....It may not be low amplitude, we have had things crack during the process anything that can waggle needs to be braced.......Peter

https://www.vsrtechnology.net/

do you mount the vibration motor only in one place? do you have to determine the correct place to mount it? or does the process need to be repeated for all the weldments/joints? what's the strategy?

[peteeng]

Hi Ard - Usually a single motor is used. Plus accelerometers are attached to record a signature. The part is suspended or put on cushions (old tyres) so it can freely move. Yes once the modes are known the motor or cushions can be moved to allow better movement. The motor starts slow and winds up to top speed (0-8000rpm). As the structure goes through its vibration modes it stress relieves and vibration intensity decreases. At resonant freq's the operator spends time at that freq and it will drop slightly as the SR occurs. The SR is achieved by taking advantage of the Bauschinger effect that occurs in metals. Under alternating strains the compressive strength drops allowing the metal to stretch more in that direction, then on the next cycle this repeats etc. This reduces the local strain state. See attached study that outlines the process...Peter

Bauschinger is not disccussed in VSR literature and is not a widely known mechanism. I came across it when I was operating an aluminium tube draw bench and found I could stretch the metal further if I turned the tube around each pull. I then researched this...

https://www.vsrtechnology.net/wp-con...l-10-27-08.pdf

https://en.wikipedia.org/wiki/Bauschinger_effect

Vibratory stress relief - Wikipedia

[ardenum2]

Hi Ard - Usually a single motor is used. Plus accelerometers are attached to record a signature. The part is suspended or put on cushions (old tyres) so it can freely move. Yes once the modes are known the motor or cushions can be moved to allow better movement. The motor starts slow and winds up to top speed (0-8000rpm). As the structure goes through its vibration modes it stress relieves and vibration intensity decreases. At resonant freq's the operator spends time at that freq and it will drop slightly as the SR occurs. The SR is achieved by taking advantage of the Bauschinger effect that occurs in metals. Under alternating strains the compressive strength drops allowing the metal to stretch more in that direction, then on the next cycle this repeats etc. This reduces the local strain state. See attached study that outlines the process...Peter

Bauschinger is not disccussed in VSR literature and is not a widely known mechanism. I came across it when I was operating an aluminium tube draw bench and found I could stretch the metal further if I turned the tube around each pull. I then researched this...

https://www.vsrtechnology.net/wp-con...l-10-27-08.pdf

https://en.wikipedia.org/wiki/Bauschinger_effect

Vibratory stress relief - Wikipedia

thanks Pete, good study

[peteeng]

Hi _ I modelled a spaceframe truss and did a comparison to a square tube. I have to work on the truss it is missing some elements. But the truss and the SHS are 1200mm long and support 100kgf. The shs wins as I have members missing. Trying to build on 60deg planes and parametric is a bit tricky. They weight the same 5kg. I'll work on the truss model. Plus the ends and load introduction points have to have the correct shear distribution... But you can see that the geometry is much bigger in the russ and this will win out of the triangulation is correct.... The shs is 47x47x3mm the truss is 100mm deep and consists of 10mm diameter members. Once I get the parameters working it can be optimised easier... Peter

[pippin88]

Hi Pippin - Please direct these papers to me sounds interesting. Often these tests us cantilevered samples that in no way represent how the member is used in a machine. I feel its only adding mass which is then a bit harder to excite. There are many videos of hitting open and filled tubes with hammers to "prove" they are damper. This does not prove they are damper in a machine application unless you are hitting things with hammers. Machines are under very very low strains and those sort of tests are high strain tests. Peter

Hi Hezz - In my Milli thread I spend a year or so working through EG, UHPC, carbon fibre, glass fibre and metals and combos. I feel EG is a dead end for hobbyists. Too expensive and the modulus achieved is too low for the cost. UHPC is cheapest per volume but again is low modulus around 40GPa. Steel at 200GPa and cheap is very attractive and if you laminate it its damper then cast iron. Laser cut steel is a cheap and accurate process, then finish machine and your done. The inside of the steel layers can be tailored to an optimum shape.

Embedding a steel box into EG is in my view the wrong way to go. But get busy and make models to prove out a few things for yourself and get quotes for epoxy that will scare you. I buy epoxy at $22/kg or $22/litre AUD (3.5GPa stiffness and I just had a quote for some aluminium 25mm plate at $40AUD per litre and its 70GPa and easy to machine. If you get the sand for free then the epoxy is not too bad but then you will only get 30-35GPa out of the material... So you need a lot of it to get to the same place as steel for a given geometry... Keep at it. If your machine vibrates its simply not stiff enough. Peter

Filling tubes with stuff is clearly the wrong approach. All aerospace structures are spider web type trusses. Fusion360 has generative design that will figure out the shape something should be at half the weight and same stiffness as the bounding box solution. Bambach did good work but that was over 20 years ago and we have moved on! Hobby builders are a bit obsessed with SHS and RHS and welding and fettling and damping, we have to move on way past that... and Hezz look up input shaping that the 3D printing industry uses. This is an approach that cancels structural vibrations using electronic damping. This is a much better strategy then material damping as it is a direct accurate method vs a splatter gun hope for the best approach via material damping. material damping will never be good enough for high performance low weight structures. If material damping was good enough we would not have shock absorbers on our car wheels we'd just have a slab of rubber. Rant ended....

Some papers attached for interest

[peteeng]

Hi Pippin - Thanks for the docs. I have read them and have some comments. I may do this in my Milli thread so this does not clog up this thread. In short - these strategies do improve the apparent dampness of the structure but only by very small margins. The dampness of a system can be expressed by zeta which is the damping ratio of the system. Critical damping ie a system in which a vibration stops in one cycle has a Zeta=1. Most good motion systems have Z=0.4-0.6 say. the steel filled tube achieved Z=0.003 about so you can see that its not really damp in terms of a good motion system damping. [steel Z=0.001 EG fill Z=0.002 combo somewhere in the middle from one of the docs] Its often expressed as a % and most structures have 2-3% natural damping (connections, friction, bearings, air contact etc) so we have to jump along way to 0.4 say to get a "good" damp machine. That's why electronic damping (input shaping or active damping) is the go as materials will simply not get us there. The materials damping discussion is all about splitting hairs or rearranging deck chairs we need to leap forward bigtime like the 3D printer people have. Peter

.

[nateman_doo]

I remember this thread from over a decade ago. I feel like the problem with engineers, is they keep engineering. They never stop and commit to make the project, so 12 years later you can have a way better design and still no chips on the floor or parts cut. Vs a layman can design one nowhere near as rigid and just take lighter cuts and get stuff done. I am sure my machine if it went through the computer analysis would look like spaghetti under stress. Sometimes you just have to take the plunge. Glad you at least bought a benchtop mill. That will help you build the main one you are designing and perhaps make you more anxious to finish.

[ardenum2]

Forgot to mention, but there's also brazing, which I think doesn't require any heat treatment(?)

Could be a viable method for you

[Chmfhm]

hi peteeng

i am still in phasing of cad world for my diy i am taking time on cad work.

i want to know what you mean by laminating steel or aluminium. do you want to say laminating with alu aur steel with cf >>>??

what thickness of steel or aluminium plate is at least good for laminating.

also it will be good if you refer any video or any study and instructions material how to do

thanks again

[peteeng]

Hi CHM - Laser profiling of steel or aluminium is an economic process. So you design any part as a stack of layers, then you have them profiled by the laser cutter. Then you use epoxy or other suitable adhesive to bond them all together. Then if needed you finish machine the part. The stack can be held together with bolts if you make provision of these on the design while you bond the part, they can stay in or even be bonded in. The thickness is up to you and the part you are designing. Often thick laser cut parts are similar price to thin parts as handling costs are similar and the laser if powerful will cut at similar speed. So use thickest plate that is economical and fits your geometry. If I used steel; "skins" I would use FG middle not CF. CF is too expensive to use it for a filler. Peter