[peteeng]

Hi Klaas - Good to hear from you. Rigimill moves everything so I expect its a design thing vs a real thing. I'm setting up comparison models now... Millis file is 21Gb some weeding may have to be done....Peter

[peteeng]

Morning all and Sundry swarfers - I've built a comparison model of the gantry vs the moving column. I made all the parts the same (as much as possible) so that the config is the delta. I used E35 material for the parts and steel for the rails.. I also made one gantry with dual rails each side to see the delta of that. The dual rails lifted the stiffness about 2N/um in the X dirn. Maybe not enough to justify the extra cost.

So Gantry style X10 Y15 and Z25 N/um and Moving column X14 Y13.5 and Z23 N/um.. Seems the MC version is the go. If the table is narrow I expect this to be the case. At some "width" the gantry would take over. The gantry style has a lot of wasted space at its rear and requires two drives. The moving column is less deep suiting a bench mount and long objects can go transverse and overhang. Seems the moving column is the go for now. I do have to decide on the table size, that drives all other decisions....

The model is a trade study so has no cars. Its apples to apples in as many aspects as possible. Its a close call really but the MC is my pick... I use the dual rail to gain some real estate at the bottom of the gantry but prior work has not shown it to be a big delta to one. Better to go up a size in rail vs two rails maybe...

So some time into the MC design.... Peter

[peteeng]

Hi All - I should have mentioned these models have an envelope of 250x500 with Z=450mm. There is a 100mm perimeter around the envelope for fixturing. An interesting thing, I had to make the carriage a little thicker to get the tool to reach forward and back correctly so the leverage is more then on prior models. I can fiddle with that now and get the motors, screws etc in to check everything fits.... Peter

[clmenz]

Peter
Possibly really stupid way for alternative geometry would be to make the moving carriage into a taller movingv hoop gantry and suspend the fixed "work table" inside the hoop. The hoop is suspended above and moved on rails either side and mounted on a subtable along with with a central single screw moving the hoop gantry. Z becomes maybe 50 mm higher, but the actual Z travel is still 450mm rails. The subtable is the real table for the whole mill.

Among the bads, more weight on newer moving hoop gantry but stiffer although Z stiffnes going down cause higher and bearings more stacked apart.
I'm sorry, I have zero skill in fusion 360 or the like to either redraw or simulate.
Again, thanks for all the exprrtise

Sent from my SM-N986U using Tapatalk

[peteeng]

Hi Clemz - Interesting idea but to achieve the target stiffness the base needs to be 125-150mm thick. Your suspended table would need to be that thick plus the supports would need to be massive. This config would also make the loadpath much longer (the longer the loadpath the less stiff it is). Plus the columns would need to be higher to acheive the same Z travel. Columns are cantilevers and they are wobbly so need to be minimum height. In summary of the whole Milli design process I have been limited in stiffness because of the "benchtop" and 50kg per part design rule. Plus now the backstep from E70 material to E35. But I may be able to get that back, more physical testing needed. A 2.5cm thick steel table is not thick enough for this size machine. Typically 40-50mm would be used. Plunging loads of mills can be very high 250kgf plus and this will deflect the suspended table. Thanks for participating I never know when I good idea may pop up... Peter

[peteeng]

Morning All - I was going to start detailing the next version but decided to use the trade study for development. Lets get the big picture right so to speak. So Trade-No1 is the original, Trade-No3 is the next one (Trade-No2 is a the dual rail version)

Trade-No1 X10 Y10.5 Z22 N/um baseline model
Trade-No3 X10.6 Y10 and Z22 I simplified the base to a flat chunk and lightened the column. I have a stonemason nearby and maybe the base can be a chunk of granite. No real change so keep lightening the column. Its holey version weighs 122kg so want it about 75kg?

The way I use the loadpost is always a worry. It models say a vice in the centre of the table. If I constrain the base the stiffness jumps 4x but that's not how the machine will be used in reality. Its static stiffness should be tool to job. Its current footprint is 850mm deep 700mm wide and 750mm high. Overall model mass 325kg. Peter
EDIT
So I moved the column into inventor and ran it thru the shape generator in torsion. It took 40% off the weight and basically made it hollow. So I shall shell the column in various thicknesses until it drops stiffness.

So I shelled the column to 20mm thick and the weight is 49kg yippee. The stiffness is X9.5 Y9.5 & Z20.1N/um so as expected dropped a little but not bad. So refine the column a little bit more and move onto the base....

[ardenum]

if you dont want to make cutouts in the gantry, you're gonna have to make the steel insert+rail quite thick to allow for ballscrew assembly. That hole in the middle? Great for cables... Also if you dont want alot of cables to stick on top you can use this https://www.igus.eu/info/3d-energychains-non-robots. Config is spindle echain mounted on top of the bridge(moving with the bridge horizontally) and then goes all the way down to the edge of the gantry and then up to the hole.

[peteeng]

Hi Ard - I usually use catenary cables. For some reason I don't like cable chains. There is a producer of curly cable local and I was looking at using those this time around. But that's a long way off... Milli-Trade has kicked off well I think. very hopeful this is the one...Peter main concern is 250x500 too small?? maybe 300x500?

I use 20 core high flex cable and run everything through that.... flex 2 is a retail stand for a local distillery. was a complex little job..

[peteeng]

Hi All - Updated the column to an open backed ribbed thing to see how it goes. Its 46kg so well within weight budget. And it didn't drop off too much stiffness. X9.1 Y9.7 and Z 18.6 N/um . I am using a forced deflection of 0.001mm at the tool end and the FE calculates the load required to move it 0.001mm.. So now to do the same with the base and then regroup... Its called No5 to bring it inline with the CAD file numbering...Peter

[peteeng]

Evening All - I sheeted in the back of the column and noticed the load post was not connected quite correctly so fixed that as well. The column now weighs 54kg happy with that. The new stats are X 21 Y 15 and Z 19N/um that's much betterer... So I went back to Trade No1A and fixed the post connection and it was X 23 Y 16 and Z22 N/um for solid parts. The target model stiffness was 20N/um so we are nearly there. Next is to refine the base as I think this is the main culprit at the Mo... Once thats dome then I'm into a detailed model.. Peter

I do like the "hollow" feet on the column means I can bolt the bearings directly to it. Will look at the bearing size to figure out column base...

[peteeng]

Evening all - I roughly diagrided (if that's a word) the base and the stiffness dropped severely. I'll have to be careful in this area. Milli-Trade No7 X8 Y6 and Z13 N/um... I'll try an isogrid next... Peter

[peteeng]

Morning All - I looked at the history of the Milli-Trade stiffness and the base is the difficult area. Other prior designs had the same issue. I looked at aluminium sandwich with CSA foam core and some gridded bottoms. Will need to spend some time on getting the base right. Maybe into shape generator for a clue.... 132mm thick of aluminium seems a bit overdone. Peter

[peteeng]

Another great morning in Dum Dum Oz - I updated the column and the gantry a little. But the main thing is that I made the base as a laminate of 25mm thick plates. This meant I could start at 150mm thick and delete each plate and rerun to get an idea of how the thickness affected the stiffness, I've been thinking about it being 200mm thick but hollow with ribs (geometry wins). The open backed column seems to be doing well at 45kg. I can optimise it a little better later... Maybe the base top and bottom can be aluminium. This gains stiffness and I can have the top machined with holes etc... More head bubbles needed.. Peter

[peteeng]

Hello All - I've done a couple of variants of the webbing but the hollow design is much better then the webbing. I'm thinking of a CSA hollow column with fibreglass bottom or legs. Plus now looking at the model I think a central drive for the column may work as well... that makes it easier. So I shall space out the cars to the max and see how that goes.... time to add the drives to get the landings and stack heights right... Then back to the base. Peter

[peteeng]

Evening all - I added the simplified car and rail and corrected some of the stack heights to agree with the bearing and nut height. I added a third rail to the column and this jumped up the stiffness. So I added one to the base as well.... Looking very good at 26/31/41 N/um
Now to hollow the parts . Next Tuesday I pick up some materials to build the new alox plank mould....Peter

[peteeng]

Evening all - Milli-Trade 19 has been the stiffest one yet. 29/36/34 N/um and this mainly uses E35 material. The Base is 175mm deep and the top 25mm and bottom 25mm are aluminium, the centre is E35 material as a sandwich... The 25mm flanged bearings have made the machine gain size, so I'll drop back to 20mm I think and also find an ER20 power head. I need to make it alot smaller... But the configuration is solid even with the E35 material. Next week I'll cast and test some more Alox... So No19 has an envelope size of 250x500xZ450 Peter

[hanermo]

Still an excellent thread.

I really want to see what the final design turns out to be like, and the actual practical mounted first kinematics from it.

Aka how stiff is it, measured at the linear carriage, under some typical loads like 0, and 10 kg mass.
And how repeatable.
Real world.

[peteeng]

Me too Hamermo - Peter

[peteeng]

Morning All - I did some structural loop testing. If I hold the gantry bearings the static stiffness is 201N/um. Then hold the column bearings its 59N/um then hold at the load post its 28N/um. This is for the X axis loading. This means the column is the most compliant loop at 201-59- 142N/um delta... So work on the column more. The base loop delta is 59-28= 31N/um. Peter

[peteeng]

Afternoon All - The Trade study version has been useful to allow quick development of the concept. Now to use square cars to remove some real estate and find a smaller power head. Version No23 stats are X32 Y39 Z39 N/um and the column weighs 66kg. Next round I will chew away at the base. The gantry's only 20kg so the major parts are starting to look reasonable... The hollow column will be cast in UHPC (CSA) using styrene cores to create the voids.... The side buttresses seem to have contributed to the Y axis being stiffer then the solid baseline model interesting...Peter