calibration for the design and initial leveling

[RonVais]

Hey guys
I must say thanks to all the people here helping and sharing so much information!

Im currently designing moving gantry CNC mill, the design will be a poor mans version of this:
29134

The Design :
The main use for the machine will be plywood/mdf, with some occasional aluminium milling .

The table will be made of 40*40 box section profile with 1.5mm wall, partialy welded and partialy bolted (so i could move it later)
For the frame - ontop of the table I plan to use 100*50 RHS steel profile with 3.2mm wall - (two beams under the linear rails) in between them i will be welding 40*40 box section supports for the spoil board. also the two beams with bolts to the table.

for the rails i plan to use 20mm HGR linear bearing

I tryed to use the calculator that"CNC router" made in the pined topic in this forum, as far i understand the results the tool deflection should be less then 2.5micrometers

http://www.mycncuk.com/threads/2023-Some-thoughts-on-cnc-machine-design-principles

For the gantry i plan to use 300mm 50*100 RHS on witch the catridges are mounted and those pieces will be welded to the Y axis beam made of 2 beams length of 1500mm and 100*50 RHS

and here the deflection (bend and twist) should be less then 50 micrometer

so as far as i understand the calculations give 4 time the precision i try to achive


The questions:
1. To get the X axis flat and coplanar between the two beams i have seen lots of methods but as i sayed this is a poor mans build - i dont have access to precision grinding machines and using a dial - i dont have reference flat surface... so what other methods can i use?

2. The X Frame will be welded to get more stiffness - although the steel can get twisted - how can i plan my design to compencate for that later?

Edit:
i have added some screenshots:
29135
29136
the red arrows shows where the frame will be welded and the steel dimentions, i intend to use 2 types of profiles: 50*100mm with wall 3.2mm and 40*40 with wall 1.5mm

[mekanik]

I am a bit reluctant to comment on your design, if you had read any of the build logs you should be aware that Rolled H section is a bad choice. Did you read AVOR ?
better option is to fabricate a beam from 50x50. or get your hand in your pocket and buy some heavy duty ally profile. Dean uses this on some of his big builds,steel is cheaper.
Draw your creation in CAD and you might get better feedback.
Good Luck

[Kitwn]

One option for leveling is to use low viscosity, slow setting epoxy (such as West Systems 105 resin and 209 slow hardener) on top of the beams. Search the build logs and you'll find several designs that discuss how best to do this.

[RonVais]

i think i wasnt that clear about my design - so i sketched something to explain it alittle better (in the meantime im working on somthing more detailed)

any way i will not be using no i beams or H sections but regular rectangular hollow section steel tubes dimentions of 100*50*3.2 and 40*40*1.5

for the epoxy method i have seen some builders use that method - but are there any other methods?

[JAZZCNC]

i think i wasnt that clear about my design - so i sketched something to explain it alittle better (in the meantime im working on somthing more detailed)

any way i will not be using no i beams or H sections but regular rectangular hollow section steel tubes dimentions of 100*50*3.2 and 40*40*1.5

for the epoxy method i have seen some builders use that method - but are there any other methods?

Think I've told you this in another thread but those tube thickness's are not a good choice, esp the 1.5mm.

Regards methods of leveling then if you don't have any way to check the rail plane then you'll be better using the epoxy method as it gives the best accuracy for a large machine without needing a reference surface.

However, If this is your first machine build then I suggest you exercise caution and research much more than you have already. Many people think it can't be anymore difficult to build a large machine as it would be a small machine but in practise this isn't the case. Every little error is magnified and if you haven't learnt the gotcha's on a small machine then they will bite you hard on a larger machine, many first time builders fail when they try to build large.

[routercnc]

I assume by 50 you actually mean 5 mm (wall thickness)?
The calculator is correct for ideal conditions, perfect joints, and does not include vibration. So it is just a guide and great for comparing relative performance but real deflection will be a bit higher due to factors not included in the calculations.

[Neale]

I was the guy who wrote up the "AVOR" design as mentioned earlier - which has proven to do a decent job for me even if it looks like a disaster area.

However, one of the main things I would change if I were doing it again (and I don't have plans to do that anytime soon) is to go up from 3mm wall thickness steel box to 5mm. When cutting harder materials especially there is a certain amount of resonance/vibration that you can feel as a buzz through the whole machine. May not be needed for strenght but just simple mass. There's a reason why commercial milling machines are made of cast iron - yes, rigidity is one reason but simple vibration damping is another.

I can only think that 1.5mm is going to ring like a bell!

[RonVais]

Think I've told you this in another thread but those tube thickness's are not a good choice, esp the 1.5mm.

Regards methods of leveling then if you don't have any way to check the rail plane then you'll be better using the epoxy method as it gives the best accuracy for a large machine without needing a reference surface.

However, If this is your first machine build then I suggest you exercise caution and research much more than you have already. Many people think it can't be anymore difficult to build a large machine as it would be a small machine but in practise this isn't the case. Every little error is magnified and if you haven't learnt the gotcha's on a small machine then they will bite you hard on a larger machine, many first time builders fail when they try to build large.

Hey JAZZCNC
Thanks for your replays it realy helps me yo get in the right direction with the build
As i wrote in the original post i want to do a simulation of that and realy understand the problem (may be i could dampen some spots and it will be enough?) But as you can see im still learning how to use fusion360 and havent yet made a normal model to simulate

As for deflections - as i wrote i tryed to use the calculator i mentioned above and in the ideal world the total deflation should be ok as far as i understand the results of the calculator



I assume by 50 you actually mean 5 mm (wall thickness)?
The calculator is correct for ideal conditions, perfect joints, and does not include vibration. So it is just a guide and great for comparing relative performance but real deflection will be a bit higher due to factors not included in the calculations.

Hey routercnc
I have been following you build and some posts you wrote and it taught me a lot !
Acutaly i meant 100mm hight and 50 width of the rectangular profile with wall of 3.2
That would be the largest stock profile that the stores around here hold


I was the guy who wrote up the "AVOR" design as mentioned earlier - which has proven to do a decent job for me even if it looks like a disaster area.

However, one of the main things I would change if I were doing it again (and I don't have plans to do that anytime soon) is to go up from 3mm wall thickness steel box to 5mm. When cutting harder materials especially there is a certain amount of resonance/vibration that you can feel as a buzz through the whole machine. May not be needed for strenght but just simple mass. There's a reason why commercial milling machines are made of cast iron - yes, rigidity is one reason but simple vibration damping is another.

I can only think that 1.5mm is going to ring like a bell!

Hey Neale
Thank for the replay !
As i wrote to CNCJAZZ i want to try and simulate everything and may be dampen the effect with
As far as i understand it should be a factor of beams length - or may be i could use soft connections at certain points

[JAZZCNC]

Hey JAZZCNC
Thanks for your replays it realy helps me yo get in the right direction with the build
As i wrote in the original post i want to do a simulation of that and realy understand the problem (may be i could dampen some spots and it will be enough?) But as you can see im still learning how to use fusion360 and havent yet made a normal model to simulate

As for deflections - as i wrote i tryed to use the calculator i mentioned above and in the ideal world the total deflation should be ok as far as i understand the results of the calculator

Ok then, I'll revise my last post and add this to it.!

If your building for the first time using only Simulation data and choose to ignore experienced advice then your SCREWED!!

You can't possibly simulate all the variables involved in building a machine even a small machine, ask anyone on here who's tried and gone to the trouble to simulate deflections, etc, they never work out as the model predicts. Multiple this x10 for a large machine then add a little more for safety and you might get some idea.

Then understand that getting anywhere near the predictions of the model requires every detail to be exactly like the model, just a small percentage change on any one will affect the result. Then, the finished result will be MASSIVELY influenced by the builder and His/Her Skills, mechanical experience, and ability with specialized measuring equipment and machinery.

Sorry to be so blunt in this next bit but I say this to save you BIG disappointment and Expense.!

You have NO experience, NO Machinery, NO Specialised measuring equipment and because of a lack of these things I'll assume limited skills in using any of them and seemingly choosing to ignore advice from experienced builders.!! . . . . Do I need to say more.!

Read my Tag line, it's very true.!