[mactec54]

I'm really interested in your response. So if I cast in the anchors, the anchors aren't enough to straighten the rail? The anchors have a thread inside them to screw the rail bolts in.

in other words, I can't straighten the rails prior to casting them into the CSA/EG ? They wont stay straight after curing?

No

What ever you mount the rails to is not going to give you perfect alignment after casting, machining is always needed after casting, any precision machine built like in the video is machined after casting, they are not showing you the whole process in that video

If it was as easy as this to mount linear rails everyone would be doing it like this

[peteeng]

Evening all you G33.1 people out there. - I have added the motion parts, cut the gutters and placed mounts in approx position to get an idea of what happens next. I've had to adjust widths of parts and will have to adjust the envelope again. I'd like to reflex the motors and use a belt, will have to think about that. So nearing the full picture.... Peter

[peteeng]

So closing thoughts for the night before ZZZZZ - I will upsize the rails from 20 to 25mm so the motor mounts can go under the bearing plates. This will stop them from sticking out... I will also use flanged bearings as I prefer these. The current stats for Milli no19 are in N/um X 46 Y 77 Z 36. These seem to be big especially the X number so I'll closely look at the model.... Peter

ahh I see the problem, many of the parts are aluminium vs CSA I'll fix that. So X 40 Y 63 and Z 28 not bad but I'll check through the model again before some ZZZZ Peter

[peteeng]

Hi All- I've started looking at tucking the motors under the saddle and column. Looks like 20&25mm cars will do the trick with a spacer on the nut mount.. This also brings me to reviewing using subplates for the rails. My original idea was to have aluminium subplates machined with all the rail holes, registers for the rails and holes for bearings etc etc in them. Then these would be bonded to the CSA bulk parts. The saddle will be monolithic 6061-T6 aluminium. The sub plate on the base has become very big. The one on the front of the column is not in the model yet. So do I make inserts and have them machined or make parts and bond them on? We are very close to that decision point... Now its Milli No20 perhaps I start with a series 2? Milli2-No1?? Peter

I've been asked to help with the design of a 16m x 8m wide concrete printer (for building panels) next year. That's going to be fun!! I'm not worried about the printing aspect but the 600mm dia diamond saw is going to need some serious stiffness to cut proper... The printer uses a 30sec cure concrete.

I found a Mill/Lathe or Lathe/Mill that looks interesting.

[klaas123]

A Shoptask 1720 gold, nice.

[peteeng]

Hi all - Soon it will be jiggle bells time, so will try to get as much done as possible this year. I've started sorting the drives and have found I have to cut considerable amounts of material out of the structural parts if I put them down the center of the rails. The Y axis I'm thinking of a side drive vs a central drive to cure that....

I've decided to embed aluminium blocks under the rails then machine them after casting (orange bits) . This will solve the caustic cement issue. But these thoughts change daily. Get the drives sorted and that will provide the path...I'd like the base to be less then 100kg , its 125kg at the moment. Peter

[peteeng]

Hi all and sundry- have decided that the side drive for the column is the go. It means no stiffness chomped out of the column and saddle and makes assemble easier. Now to look at flanged bearings or not and maybe a rear drive for the saddle for the same reasons. Then its a bit of a rebuild.... Peter

20mm flanged bearings are 63mm wide and square are 44mm so there will be a gain of width of about 19mm for each part Hmmmm

Bearings from BST arrived a couple of days ago. Packed well and assembled easily. Can start to get serious with YaG now.

[peteeng]

Evening all - I think I'll mount the drive for the X on the rear. Will tidy it up and won't have gutters aka "stiffness reducers" in the base. It may also be the time to look at using two rails on the saddle vs three...Peter

[ardenum]

You know, you could always make the saddle thicker, you're casting it so there's no issues with geometry for that.

[peteeng]

Hi Ard - Good thought but I also need to reduce weight. Millis growing again and I need to control its size and weight as the intent is a benchtop machine. So stiffness first, weight second then size. Peter

[peteeng]

Hi all and Sundry - I think Milli has come of age. She's version 21 and I think I have settled its basic morphology. Now to let it rest a little, while I concentrate on my router build so I can make Millis moulds. Still have to sort many things especially in the spindle area. But I think Millis DNA has been set. So now to maximise each parts stiffness, cut the size back a little, add some flair to the design.

I did investigate the rear drive but it just added length to the bed so canned that side trip. I'm away next week, so this week is frantic with getting 2020 rounded up. Shall pick up some CSA for casting tests in my travels and get serious again next year. YaG may even be able to make the aluminium bits, that would be a bonus! Have a great week out there, its nearly Jingle Bells time. Peter

Just run the FE and the latest stats are in N/um. target >20N/um on all axis.
X 12 Y 15 and Z 24 making the base not solid and using the central web has reduced its stiffness. Will review soon. In the image you can see it flex in the middle. I have also started supporting it via the bed rather then its feet, this reduces its apparent stiffness as well. If support is via feet its stiffer then in practice. By supporting at the bed this means the load loop is correct. Peter

[jaguar36]

Do you have a limit on the height of the machine? If not, just make the base casting thicker, but with larger voids. That will increase your bending stiffness without increasing your weight.

[peteeng]

Hi Jag - No limit on height so yes the thicker the base the better. I am trying to limit base to 100kgs, so its a juggle. Has been varied from 80mm to 120mm thick. Last night I started playing with thicker and even an infinite stiffness base. Trying to understand which bits are important now. May set the base up in inventor and do a generative design. I have also cut the Y from 300 to 250mm but left X at 600mm. Z currently 425mm...Peter

[peteeng]

Morning All - This mornings stuff mucked about with different thickness bases. I want to keep this base a simple cast so looked at aircrete 120mm thick. I looked at parts in various materials plus making them rigid trying to see a trend. The trend is that steel wins on the machined parts, which is obvious I suppose. So steel zinc plated is the go. The chart attached shows some results of this study. The numbers are the deflection from a 1000N tool load. The target deflection is <50um to get to 20N/um. The hurdle is still the X axis torsion. May make the arm wider. Will get the flanged bearing model running then rebuild. I suppose the interesting thing is that if I make all the major parts rigid the deflection is 18/25/47 which means the small parts are contributing a lot to the deflection. Maybe its sound to go to 25mm rails just for the stiffness gain.... If I made everything in aluminium I'd get there but I don't have a foundry here... this also points to carbon fibre as its the same stiffness as Al. But $$$. Maybe we are at the Holy Grail point. Peter

I'm also thinking about making the base as two pieces. This will make assembly easier. Many machines with moving columns have the table part and base part seperate....

[jaguar36]

Looks like you're getting killed by the narrow cross sections needed to attach the head and column to the linear bearings. Rather than necking down to add those attach pads, can you bump out and envelop those pads?

You could consider embedding some carbon fiber strips in the castings. If you put them at the edges you could get a pretty substantial benefit with a small amount of material, and pre-cured carbon strips are pretty cheap. Course it would still be alot cheaper to just embed some steel. Gotta trade the cost vs weight.

[peteeng]

Hi Jag - I have looked at the attachment areas and they seem to transfer stiffness quite well. Its their thickness that has to be looked at. If thin they warp which is no good but at a certain thickness they work. If I add flanges vs the recesses the width or depth blows out. A small amount of carbon does not do much. Carbon fibre is 200GPa then its only at 50-55% by volume so its only 100- 110Gpa vs steel at 200Gpa for same volume. The CSA material will be steel fibre reinforced. If blue metal is used in the CSA it gets to 40GPa, if I add a lot of steel fibre it should get better then 40. Once I get some CSA (2021) and do some tests I'll know the correct modulus for the CSA /steel composite. I'm hoping for better then 40GPa.... Being a small machine I can't make things as big a I'd like. It grows and shrinks as I refine each machine part. At the end of the day it will be as big as it can be and as stiff as it is. Its already much stiffer then small benchtop machines as far as I can see. My initial concept was to use materials that could be machined so Tetrium-S was the big candidate or infused FG. But CSA came along and its 26GPa (2100kg/m3) by itself and only $2/kg which is a winner vs epoxy. So now I have to look at having inserts. Design is a juggle... I'm off to start setting up the motion bits on YaG will need moulds soon!! Peter

[peteeng]

Hi all - I booleaned Milli together and imported into Inventor to do an optimisation but the mesh failed. My guess is too complex because of lots of small features. So I tried a simplified version and it bonked to. So I'll have to make some block geometry to try that. I'll have to start thinking about paint. Gully is good with that. Metallic blue with pearlesant or yellow hammertone? Maybe a poll will sort that. Peter

I did get a optimisation done but its just bending not torsion. So I'll add a block arm and redo....

[spumco]

Metallic blue with pearlesant or yellow hammertone?

I vote for yellowish off-white. Try to duplicate that dried coolant patina...

[peteeng]

Hi Spumco - the one with the faint rust streaks through it? Trying to make the machine look 15 years old vs new... Peter

[peteeng]

Hi All - I tried to gain some insight into the base stiffness issue and it became self evident. Use the thickest base and stiffest material. I thought there may have been an overlap with geometry and material. But in hindsight no.... Peter

There is a slight overlap with Al at 100mm thick and aircrete at 150mm thick. But 150mm is too thick I feel and
100mm of aluminium sounds silly! I added steel to see if maybe a thin one in steel will work but geometry wins...

aircrete 20GPa CSA 40GPa and Aluminium 70GPa