Milli a new composite mill kit

[peteeng]

Hi Guys and Gals - I looked at the bearings and there's a large cost up if I go to 30 or 35mm rails. 25mm seems to be reasonable even less then 20mm in some cases. So 25mm it is.... I was hoping to find a stack height match but that's out as well. I could look at the 12mm screws and bearings but that sounds a bit small.

Have to do some numbers on buckling and vibration.... seems were rebalancing the design Peter

20x900mm $177 set 25x900mm $172 set 30x900 $360 set I've just placed my first order with BST for my router build will be interesting in timing to get here. Xmas is looming and various things slow down about now...

[peteeng]

Hi all - While my computer was thinking about a mesh I did a quick calc on the 12mm vs 16mm ballscrew. I intended using 5mm pitch, this gives 345kgf thrust, a 10mm pitch will give half that at 173kgf which is probably plenty. A 1000mm long 12mm screw whirls at 1867rpm so steppers won't get to that, servos can. A 1000mm 12mm screw buckles at 403kgf and usually a SF=2 is used for bump loads etc. so a 203kgf allowed for buckling means the 10P 12mm screw is OK as its 173kgf but the 5mm pitch is not OK.

Looking at 16mm buckling, it buckles at 563kgf so will do all cases.... 16mm stay with, I think said Yoda.... Peter

My overall screw lengths at the mo are X985mm Y715mm Z660mm. The distance to the nut is maybe 200mm shorter so maybe do to the nut to be a little more accurate.

checking nut distances X746 Y321 Z445 so using 750mm instead of 1000mm gives an allowable buckling load of 361kgf which is OK with P5 ballscrew.. Maybe worth a look... just finished meshing some 1M elements too big have to remesh...

[spumco]

Peter,

If the rail assemblies are a limiting variable due to deflection, there are other options besides up-sizing. Cylindrical roller rails tend to have higher load ratings than ball rails. And depending on where the flexing is taking place (in the truck/trolley or the rail itself), a wide-type rail might be more favorable.

Neither of these types are easily available as off-brand knock-offs, but if you're planning to use name-brand components all the major manufacturers offer some variation of these two types.

-Ralph

[peteeng]

Hi Ralph - I have not yet decided the exact issue. The rails are 200GPa steel siting on 30GPa FG. The local bed maybe the issue not the rail. It won't hurt upsizing and it seems the cost is near equal. The current model is 100% bonded so once bolted every 60mm it will be less stiff locally especially in lift. The column is 270x240 which weighs 116kg/m. A 200x200x14 steel SHS is 81kg/m and of equal stiffness. So I have to tread carefully to not go the steel route (as I didn't want to do steel) but steel would give it a better bed stiffness. I'd fill it with expanding concrete.... or 400kg/m3 PU foam... Its not about the load rating its about the local stiffness... I've attached an image of the rail squirming under the X axis load. You can see the foundation has lifted up. This maybe the issue not the rail itself. More analysis needed with bolts I think but that's at a detail level in the future. For Now I want to down size it a bit, go to 25mm rails/cars use 16mm screws and get a solution for the motor mounts. As modelled this machine should be about the same rigidity as yours @ 10N/um. All the axes are over 20N/um and I use an 50% efficiency factor. I wanted to keep the component weight to <50kgs so I can manage the build by myself without a crane... another factor in the design... Peter

The current FG column weighs 78kg but could be hollow.... I'm currently using BST house brand with medium preload. Will see how they go... obviously quality brand name bearings can be better but cost more as well....

[peteeng]

Hi All- I downloaded a Hiwin 25mm car and some track. I simplified them so they are easier in the FE world. Here's a snapshot of 25 and 20 side by side on Milli to guage the difference,,, Peter

[peteeng]

Morning All - News - US Govt investigates concrete CNCs.... perhaps they should read CNCZONE more.... Peter

https://www.mmsonline.com/articles/i...ol-moonshot(2)

[peteeng]

Morning All - This morning I've been looking at the column more closely. If I freeze the bearings at the arm I get (all in N/um) X87 Y225 Z174 Then freeze the bearings at the column X26 Y137 Z53 then hold on the base feet X20 Y46 Z32. X load puts the column into torsion so seems I have to up the torsional stiffness. Using geometry I go bigger, but would seem the 30GPa material is the hurdle here. Al would be 70GPa and steel 200GPa.... which relates to the shear modulus E=2G(1+v) its the G (shear mod) that resists the torsion... I'll start on No10 and see how it floats. Peter

[maxspongebob]

So, lets build you a 3D printer to print the shell/molds for Milli.

[peteeng]

Hi Max- a large scale printer has been on my list for soooo long.... but There's only 20 hrs in the working day... I'm off to put the last pieces of YaGs base together. Then's its wait wait wait for the Shanghai parts... Peter

[peteeng]

Hi All - Seems one answer is carbon fibre. If I build the current solid column out of CF it would weigh 61kg call it 60kg for the math. If I build at std ratios 1kg of CF plus 500g of resin gives me a 52% by volume laminate. This gives me a 200x0.5=100GPa UD stiffness which in a quad say laminate will be 55-60Gpa with good shear stiffness. This would be double the FG laminate stiffness but not quite as good as aluminium.... so I shall put an enquiry in for how much 60*0.666=40kgs of CF. cheers Peter

[maxspongebob]

How big do you want it?
I just finished building one for myself. See it here https://grabcad.com/library/retnirp-d3-the-3d-printer-1
All it needs is to be scaled up.

[peteeng]

Very very nice Max - Did the H drive work out OK? and did you do the 3D printer sdrawkcab on purpose? Cheers Peter

[maxspongebob]

Very very nice Max - Did the H drive work out OK? and did you do the 3D printer sdrawkcab on purpose? Cheers Peter

LOL. I looked at your reply on my phone while I was at dinner. Initially I didn't understand your question so I put it away until back on my computer. Now that I look at it, I am laughing because I didn't know what sdrawkcab was. Then it hit me. It's backwards.

I think you agree that every machine needs a name. When I was building my printer I couldn't think of a name, so I just reversed 3D printer and came up with Retnirp D3.
Also, yes, the H drive works great. However, to be honest I still can't figure out the real advantage of using a CoreXY design over a Cartesian.

[peteeng]

Hi Max- Yes every project has a name, some people don't get that but it has to be described somehow. Everyone has a name. I've looked at the H drive a few times and thought why? Onward to Milli No10... Peter

[peteeng]

Hi all U CNC people out there - I've put the 25mm rails on Milli and adjusted the table size and envelope. Here's the current stats....

Table and envelope same size X300 Y700 and Z450mm tool reaches entire table surface
Bottom spindle nose 50mm from bed top nose 500mm from bed. Table will be slotted aluminium anodised
FE model weighs 475kg about to put Milli on a weight loss program...
Footprint 1060mm wide and 875mm deep so it may still end up on a bench.

Current stiffness numbers are in N/um (drum roll please) X24 Y25 and Z77 so about 12 in reality so happy there at the Mo. I now want to sort the motor mounts and a few other operational things. This model I placed on a floor and gave the contact friction 0.35. I made a block on the bed and pushed the tool one way and the block the other way. I also turned on gravity so the friction at the feet would be correct. Milli "slipped" 0.0002mm on one of the feet. This has to occur to balance the tiny out of balance analytical forces that would occur in the solver. So I'm happy with that.

Using the table/friction approach gives the same machine stiffness in the X and Y as if I just hold the feet. The Z however is not as stiff at 40N/um. Pushing down on the table with 100kg makes a diff I suppose including the upward tool load... Peter

[peteeng]

granite cast mills

[peteeng]

Update on "cast" granite. Watched a few more vids. They are all cut from solid granite using water jets. cheers Peter

[peteeng]

Evening All - I have started adding motor mounts but I can see the next level of development will take quite a bit of time (tensioners, pulleys, covers maybe a pan) I need to finish my router YaG build (or Brevis-HD) so will put time into that and then come back here. I'm happy with the current level of development, its got good DNA. Some printed motor covers and various odds and sods will round it out. Got to get YaG built to build Millis parts with!! Peter

[peteeng]

Yo! and good morning - I've been trying to find a benchmark machine for Milli. Would be good to have a commercial comparison. Looking through Millis predecessors files ( P2P and HS-Mill) I'm also tempted to redesign in plate aluminium as an alternative now I have config and spindles etc figured out. If anyone knows of a similar commercial machine I'd appreciate a lead on these. I'll have to model a collet/tool and give it more table height. I also have to consider the arm length for the different spindles and the case of using a high speed spindle. The arm sub plate may have to be quite thick to accommodate that variation. Details details details... Peter

[peteeng]

Morning all - I have a quote for the CF. Its $81AUD/kg and Glass is $5.80AUD/kg. So 40*81=$3240AUD for the column alone plus the epoxy cost @ $16AUD/kg. That's commercially out. I pay $7AUD/kg for aluminium and that would be better $/GPa. I have a source in France for recycled short fibre CF that's really cheap but costs a motza to get it here.... so more development needed. Peter

So the column material would cost:
40*81=$3240AUD
20*16=$320AUD ie = 3560AUD too much...

In glass it would be 1kg glass to 420g resin at a resin ratio of 30% by weight. so a 60/1.5*1.8= 72kg in glass column would cost.

72*0.7= 50kg glass and 22kg resin

50*6=$300AUD
22*16=$352AUD

total $652AUD that's acceptable ... but CF will be 70GPa and FG will be 30GPa