Milli a new composite mill kit

[mactec54]

They are all pretty poor. they have radial bearings and most are good for 2000-3000rpm - a few say 6k but i wouldnt.
They are not what you buy when you go to the effort of building a nice machine frame. they are what you buy to bolt to a sheet a plywood to stick a grinding wheel on.

How do you figure that the one he is going to buy has radial Bearings the Bearing type is listed and a 7007 ( 2 ) is a 25 degree angular contact bearing, and can run at 13,000 RPM, there are people here on the Zone that are using these spindles and they work well for what they are some also have a motor drive setup as well for not much more money

[joeavaerage]

Hi

Not saying that they are wrong just in a few years printing will replace most of that and CI will die out from machine tools as well. Composites will replace them, printed or RTM or cold cast so may as well get on the bandwagon now. I could make Carbon Fibre parts cheaper then CI parts maybe if the volume was there.

I disagree, foundries are certainly fewer than yesteryear, but for certain products casting just cannot be beaten for price/performance.

Printed metal parts are mega expensive by comparison, that is dollars/kg.

Composites for all their versatility are still quite expensive and unless fibre reinforced low stiffness, about 1/3 that of steel and about 1/2 of grey cast iron.

When it comes right down to it there is a solution that has the required properties and you could achieve that result in a number of different ways, casting is one of them.
You seem to be suggesting that because casting is an old process that it is going to be obsolete.......and my opinion on that notion is 'nonsense'.

May I suggest you look around for a foundry in your area. If my experience is anything to go by you'll find superbly experienced tradesman, whose knowledge and confidence
will impress.

Craig

[peteeng]

Hi Craig - I don't disagree about foundries and tradesman etc etc. I deal with foundries for steel and aluminium on and off for marine and mining equipment. I think there are other ways to skin the cat so that's what part of this project is about. Printing is expensive now but will become the norm soon...We don't make valve radios anymore and soon we won't make IC engines either just a matter of time I feel for CI to phase out in may areas. I suppose I'll have to check back in 10 years to see how those thoughts go... Peter

[ardenum]

Some links for you mate.

bt30 for belt drive : https://www.aliexpress.com/item/4000307172516.html
blue bt30 : https://www.aliexpress.com/item/1005001640465000.html
direct drive bt30 tapping spindle : https://www.aliexpress.com/item/4000452497782.html

[peteeng]

Evening All - Thanks for the spindle feedback will be handy. Some of those power heads look very good.

I have been thinking about the column and how to remove it. I have also been thinking that the 20N/um maybe a high target for a small machine. So I went back and looked at Ralphs mill and here's the Mikinis/Skyfire stats. I restrained the mills at their feet to remove the issue of how to hold the table. X 94 Y 46 and Z 41 in N/um. The Mikini model is 341kg in weight and is cast iron. Ralph is busy stiffening this mill so its understood that it maybe not stiff enough for some. So to remove the column I went back to a fixed column, lifting gantry (to remove the saddle) and moving table. I turned the table 90deg from last effort so the mill would run along the bench. This also minimises the gantry width.

It has the same table size as the Mikini but the mill can reach the entire 600x300 table. This is Mill-LG No3 and it weighs 399kg. Its main parts are E70 material (aluminium or high modulus CSA). Its stats are X35 Y17 Z 28 in N/um which is considerably stiffer then the Skyfire. Both Mills have a nose height of 430mm so its nearly apples for apples in the loading... I have been playing with the bridge and its probably not needed but it does help... next step is to make the gantry a bit deeper to help the X deflection... see Milli-LG No4

I think this configuration has more scope. Its close to my original thoughts but the moving and lifting column was way too complex to be successful as a small mill. I would like to get the table drive out of the muck however!! As always onward and upward... Peter

[peteeng]

Hi All- Opps I quoted the actual deflection in um vs the static stiffness in the last post. Converting the Mikinis deflection to N/um is X 94um is 11N/um as I applied 1000N, Y is 22 and Z is 24N/um. Milli-LG No3 is X 1000/35= 28N/um and Y is 58N/um and Z is 35N/um. So the X Axis is still over 20N/um so a little work needs to be done... Peter

[ardenum]

A $500 2.2kw teknomotor solid shaft motor for example running 12000rpm

You mean one of their rectangular motors? https://www.teknomotor.com/eng/prodo...ular_motors-2/

Some of their distributors post torque charts but I can't find any directly on their website. Do you have personal experience with these motors?

[peteeng]

Morning All - I have been looking at laser alignment again an can get construction lasers for <$400AUD that are accurate to 2mm over 30m. But they are useless for a machine. The latest quote is about $40,000AUD for a full blown tool. A bit out of my range. Peter

[joeavaerage]

Hi,
that laser relies on a collimated beam, its not for measurement.

For measurement or interferometry, you require a coherent laser. Bang for your buck the old HP Zeeman stabilised helium/neon lasers
are still available and have effective coherence lengths of 15 plus m.

The Zeeman stabilising results in a dual mode beam with the modes only 3.25MHz apart and orthogonally polarised. This allows for accurate differential
measurements which rival iodine cell stabilised single mode lasers of ten of thousands of dollars whereas an old HP is $1k.

Craig

[peteeng]

Hi Craig - Collimated or not its sold for machine alignment and measurement, I'll look around for the Zeeman, thanks for the lead. My collimator is used for straining veges Peter

[mactec54]

Morning All - I have been looking at laser alignment again an can get construction lasers for <$400AUD that are accurate to 2mm over 30m. But they are useless for a machine. The latest quote is about $40,000AUD for a full blown tool. A bit out of my range. Peter

Like the level

[peteeng]

Morning All & Sundry - So Milli-LG No3 has hit the marks. But its a bit big. So I shall redesign using 25mm rails and the ER16 power head. Come back to a 250x500mm table size and see what it looks like. The 20mm rails are squirming a bit and the Mikini has 25mm rails on the Z axis so they must have figured that out as well.

Looking at the materials I divide them into low modulus and high modulus with aluminium being the middle.

epoxy E=3.5GPa 1.05 specific gravity
concrete 28GPa 2.3
concrete + aggregate 40GPa 2.8
advanced fibreglass 30GPa 1.8
aluminium 70GPa 2.7
E80 UHPC 80GPa 2.8
general CF 60GPa 1.5
standard CF optimised 80GPa 1.5
intermediate modulus CF 105GPa and 1.6
cast iron class 35 114GPa 7.1
titanium 110GPa 4.4
steel 200GPa 7.8
beryllium 303GPa 1.85

estimated materials
CSA AlOx 80GPa plus
Epoxy Granite 25-40GPa


One of my aims was to use a material that I could near nett cast and finish machine. The concrete and alox materials can't do that. Then I have to use inserts which is possible but fiddly. The next bet was laminated metal. Laminating aluminium with fibreglass interlayers produces stiff damp blocks that can be machined. Milli-LG is currently monolithic aluminium with thoughts to laminate it. Look up GLARE material. Steel fabrications with CSA fill is the next thought but that sounds like a compromise but sometimes simple is best.... Beryllium is a new interest will try to get some Be powder from the USA. Be has same modulus as AlOx and can be machined. Probably have to sell the house to get a kg here... Peter

Today I'm off to get more resin so I can finish the workshop floor. Then I can get serious in there...

[peteeng]

Morning All - Today I looked at the geometry and the rails and changed a couple of things. I now feel like I have control of the stiffness. This config allows adding or subtracting stiffness much clearer and easier. The bridge is not critical but I'm sure it helps. Currently the general material is aluminium but think of it as E70 material. This could be aluminium or CSA-ALOX. The Beryllium company engineer got back in touch but no $$$ yet. No4 stats are X 22 Y 26 and Z 66 N/um and this is to the base of the post so is quite realistic. Its weight is 399kg and its footprint is 600mmx1200mm. I'm particularly happy with the footprint for a 300x600mm table. I'd like to get the X stiffness up a bit then shrink everything down to 50kg parts.... my only concern at the mo is how to bolt the cars to the gantry that has not yet been considered. Feeling like I have a firm direction now... Peter

[peteeng]

Hi All & Sundry - I have updated Milli to No5. This mainly concerns the tool carrier. I have to have a place for the spindle motor and in another gantry thread we worked thru so many options and a trend came thru. By placing the rails on the diagonal the "action line" is made closer to the shear centre (look it up its the point in the section that it wants to twist around, in a car its called the roll centre) this minimises the secondary moment created if you mount the rails on the front. It also puts the bearings further a part so their loading is a bit smaller. I also compared the static stiffness of restraining the machine on its base and on the table post that I use.

Post restraint X 23 Y 31 Z 76
Base restraint X 60 Y 104 Z 145 these are big numbers!!

A little work on the tool carrier and then I can start the next level which is adding the motion parts at last!! Patched the workshop floor this morning now can get serious with the second coat and its done Yeh! Better get serious with some commercial work now days nearly half done. Peter

[ardenum]

I also placed the rails in the same config on my build, the only difference being I used a 50x300 plate instead of a square one. The rail on top will also better support the heavy spindle motor.

[peteeng]

Evening all & sundry - I was looking at No5 and was about to cut the gantry down to save some weight. Then I thought it looks ideal/easy to set up a twin rail on the columns. I have looked at this before and the results have been inconclusive. Intuition says two has to be stiffer then one per side. But alas its come out no delta again so I think that approach has to be skuppered. Its commonly done on very large machine but I think this is to make assembly easier.

No5 single rail each side X 22 Y 31 and Z 73 N/um
No6 twin rails X 22 Y 28 and Z 83 N/um does help in Z but I was hoping for a leap of stiffness in X... Its interesting that Y dropped. I ran non linear and got the same so seems odd. But I think this is to do with the tool plate which I have to add webs to...

Next time 25mm rails and some more organic shapes... Peter

[peteeng]

Good Morning All and those that assign themselves as "other" - I have looked at Milli No5 closely and removed the bridge. I have made the table steel and that's about it. Time to redesign around 25mm rails. The tool plate needs stiffening and that's about it. Stats are X 22 Y 40 and Z 74 N/um.... I've had word back from the Beryllium people and a recycled grade of Be is $450USD per kg so will need a special project to use that material. So for now its ALOX.

Be E303 D1.85
ALOX E300 D3.95 currently concerned with stiffness more than weight and The E is the same.

Plus I have to sign a declaration that I won't use the material in missiles or nuclear warhead technology. Don't seem to remember doing that at the sandblasters.

[peteeng]

Hi All - I couldn't get it out of my head that 4 rails is same stiffness as 2 rails. So I modelled the gantry by itself with 2 and 4 cars and looked at the all the models closer. Turned out the major deflection was in the Tool Plate so fixed that and then we could see the delta with the 4 vs 2 cars.

Milli-LG No7

4 cars on columns - X 24 Y 29 and Z 89 N/um
2 cars X 20 Y 28 and Z 68 N/um so maybe that's worthwhile, we shall see what the 25mm cars do. Peter

[ardenum]

You can always just screw a 20mm plate on top of the columns to connect them, no need to do anything more fancy then that.

I really do hope you stay with this build, you were over so many configs, Imagine if all that work went into polishing just this setup, I can't wait to see what you end up with. Exciting times ahead of us.

[peteeng]

Hi Ard - I was thinking of a stout aluminium channel, easy to bolt up and quite stiff. But happy to cast a good bridge if needed. Currently seems no need for the bridge but I think it will help in the vibration area, especially when the gantry is down and those tuning forks are sticking up in the heavens. Every time I go through this process I learn something so its not all in vain. The build is the tough part, reality sucks and time travels about 1000x less fast in the real world vs CAD/FEA world. Like I've had a bag of CSA for 2 weeks now but haven't made the moulds, been painting the workshop floor, so many things to do.... then there's my day job...

The bridge is connected at the back so the columns can be adjusted sideways vs in the "middle". I could attach them at the top, maybe next round the hurdy gurdy ... should do a modal FE I suppose. Peter

post script a 20mm "plate" will probably vibrate so it needs to be stiffer then a "plate" plus damp that's why the interest in composites....