[StrawberryBoi]

pete, there seems to be some disconnect. Your own words were comparing bolted to a bonded connection and I assumed you meant adhesive vs bolting after reading lead up posts. My mistake. Some careful clarity when a word has multiple meanings in a thread is always helpful to avoid these kinds of things.

[peteeng]

Hi Strawb - unfortunately plain english and technical language sometimes have different intent. I do try to be careful with these things. Peter

[peteeng]

Evening all - I looked around for some light fibreglass cloth to use as an interleave in the aluminium laminate. Came to the conclusion to use glass veil as its cheaper then cloth. Plus I had to find a company that did mail order as I generally use queensland suppliers and they are now difficult to got to...

https://trojanfibreglass.com.au/product/surface-tissue/
https://trojanfibreglass.com.au/prod...loth-685-mm-2/
https://trojanfibreglass.com.au/prod...cloth-1000-mm/

Start designing again.... Peter

[peteeng]

Evening all - Thinking through the lifting gantry concept. Here's a big one at NASA. Its a printer but it shows if it was subtractive it would not work unless it had a very very long tool. Since I want to make moulds which can be deep I think the lifting gantry is out and I'm back to the Mazak type arrangement. Peter

https://www.dezeen.com/2021/08/10/bi...re%20exploring

[peteeng]

Hi All - Milli is starting to permeate my gray matter (and grey hair) again. The aim was to make cold cast parts but that has not worked out. Another was to take advantage of a big mill at one of the machinists I use and another was to develop a new machinable E70 material and that didn't work. So in short we are back using aluminium, maybe laminated then finish machined in the big mill. There are three pathways to do this:
1) billet machine
2) bolted plate construction
3) fabricated heavy extrusions and finish machine - this has not been on the menu before

Billet machining makes a lot of expensive swarf to get to my 50kg per part rule. Bolted plates have various issues and edge bolting is expensive due to the double handling required to do edges. The machinist does a lot of aluminium welding and does regular solution heat treating to T6 of these assemblies. But for machines I don't think that's necessary although it would fulfil the same function as a thermal stress relief.. many of the large parts when run through the form finder have triangulated internal shapes which a plate build can't achieve.

The steel trussed mill is really interesting and this brings me to an aluminium version. Recently I had to design a large dolly for moving a large machine and I did it in heavy aluminium sections welded and machined and it worked out quite good weight and cost wise. So my thoughts mashed around the idea of taking advantage of the laser cut sections (tube and plate) then welding, then finish machining. Also using laminated aluminium wherever possible for vibration mitigation... I think this is a valid path. It uses the aluminium with minimum swarf production, keeps the manufacture within two companies and the welding and machining is done by one company which means they understand what the outcome target is... I have to check what the max thickness they think they can weld easily...

So philosophically the Mazak VTC 200 config is the sort of thing to aim at. Its pyramidal in structure and can be long. My things tend to be long. Since my aim is to make aluminium moulds and mill composites I can neglect the cutting steel requirement briefly and spec a high speed spindle.... so starting at the spindle I can work up a minimum size aluminium truss mill to see what happens...

Also I have news that the chinese company with chopped carbon is sending me a sample but it would be 150g. I need 500g to make Plank3 so have asked for 1000g.... see what happens...

A new Milli chapter begins - Peter

[handlewanker]

Hi Pete.......have you considered making a frame........a complete machine........ out of clay so that you can hand mold it to get the desired shape then firing it in a kiln to give it a brick like hardness........when it cools you would have a frame as tough and hard as a fired house brick and that will not deflect or move no matter how you impact it..........fitting linear rails would be by gluing them on not bolting etc........at the same time the fired one piece outcome without any metal etc would not be affected by ambient heat like a composite metal frame would.
Ian.

[pippin88]

Evening all - Thinking through the lifting gantry concept. Here's a big one at NASA. Its a printer but it shows if it was subtractive it would not work unless it had a very very long tool. Since I want to make moulds which can be deep I think the lifting gantry is out and I'm back to the Mazak type arrangement. Peter

https://www.dezeen.com/2021/08/10/bi...re%20exploring

Peter, I don't think this is true.

For deep pocketing, all designs depend on using long tools / holders (gauge length). The spindle nose / head is always a limiting factor and will get in the way regardless of column / Z axis design.

C frame is not really different from lifting gantry in spindle clearance issues.

How much the actual spindle sticks out down below surrounding frame is a factor.
But all spindles are a lot bigger than the actual tool.
E.g. 10mm tool. Spindle nose is 100mm. You are going to need a long tool or a long gauge length holder to do deep pocketing in almost all circumstances.

(Yes, there are a few workpieces where frame design may be more of a factor. If you are machining large parabolas maybe...).

[peteeng]

Hi Ian -Thanks for chirping in. Milli is a production machine so I need to be able to build one per week say. Hand making clay objects then firing etc is not a precision production process. I did some quick looking for the elastic modulus of clay and its no better then CSA concrete (the numbers I found where around the 20GPa mark plus they were compression as that's what buildings are interested in and I think it will suffer the poor tension/flex properties) so may as well make a mould and cast CSA. Clay brick CTE is about 7um/m/C and concrete is about 12 & alum is about 26 so it still expands and contracts.... Thanks for thinking about stuff for me. Peter

Oh Ian - don't all hobbyists have a fully air conditioned workshop with a cleaner on fridays? Plus an AC material store for normalising stock?

Pippin - I keep thinking about this and on Mondays I think its Ok and by Thursday I think the lifting gantry is no good. I notice that on very large machines they still have a Z axis and the lift is called the W axis. I agree the lifting gantry is a C frame mirrored and the spindle put on the side vs the end. I think the issue in my head is that at a certain width the gantry design is required but under that width its not. I'm thinking 250x500mm working envelope at the moment and at 250mm I don't see the advantage in a gantry. Gantry or column I'll see how the alum truss goes. Can't tell until its modelled and stressed..... Peter

Machines are designed for a job so perhaps this thread has been an exploration of geometry limits vs configuration with no job to do. If I had to machine 500x500x500 the gantry is in or its a very big C frame. The biggest bits I "want" to machine are composite gantries and they are 1400-1500mm long so a moving column C seems to be what I "need" vs a general machine,. But a 1500mm machine is a big thing....

I have started modelling using a Dia100mm 3.5kW G-Penny metal spindle will publish when it weighs a bit more.

[peteeng]

Morning All - The Grail is out there somewhere. The pain begins again. So I have built a block model. It currently moves X400 Y600 and Z400. Will massage this a bit to make it a bit smaller. Say X250xY600. Then start tuning parts. The machine base is 75mm thick and weighs 128kg in Al or CSA so it has to come down. It will become a truss. I look at the column and arm they just says cast me in CSA with aluminium inserts and post machine. So maybe a hybrid approach is the go... Peter

[StrawberryBoi]

Morning All - The Grail is out there somewhere. The pain begins again. So I have built a block model. It currently moves X400 Y600 and Z400. Will massage this a bit to make it a bit smaller. Say X250xY600. Then start tuning parts. The machine base is 75mm thick and weighs 128kg in Al or CSA so it has to come down. It will become a truss. I look at the column and arm they just says cast me in CSA with aluminium inserts and post machine. So maybe a hybrid approach is the go... Peter

Are you thinking to truss just the orange part or the entire base?

[peteeng]

Hi Strawb - Initially the lot. Then it comes down to weight/stiffness. Heres where I'm at. I've done some FE and its running at about 10N/um target is 20N/um plus... thickening some bits...Better get back to some commercial stuff...Peter

[peteeng]

Evening all - So the journey begins. I built the model in blocks, and ran the FE in stages with each block being aluminium. All the rest of the machine except the rails (steel) is "rigid". This is an infinitely stiff material property. This allows a quick understanding of which bit is the most compliant. Arm 157N/um Column 68N/um Saddle 84N/um and machine base 24N/um (all in X Axis). So the base is letting the team down. This was the conclusion last time I got to this point and spent a lot of time looking at different base restraints and materials and configs. If all parts are aluminium with steel rails the machine stiffness is 14N/um which is not bad but 20+ will be better...

The current base is 75mm thick of aluminium which is a monster base I feel for a benchtop machine. It weighs 193kg so needs stiffening and lightweighting.... Hopefully the truss can do this.... I'll be making it out of 152x75x6mm thick RHS sections. Peter

[StrawberryBoi]

Hmmm it almost seems like you might want to make a laminated composite of sheet steel and carbon fiber with an epoxy granite or even foam filler.

The specific stiffness of carbon fiber is about the highest you can get, though Aluminum is probably the most attainable metal out of high specific stiffness.

You could combine the two, use a pre-preg carbon fiber with sheet aluminum or steel in layers and clamp it together during reflow. Or use fiber mesh and epoxy to fill and bond.

Leave a thicker layer of metal on the top and bottom layers to machine flat, maybe?

[peteeng]

Hi Strawb - CF is too expensive for this product. Earlier in this thread I looked at various sandwich type constructions with no joy. Specific stiffness of the fibre is not much help once you make it into a laminate. If you make a CF laminate with general properties it will be less dense then aluminium (1500kg/m3) but not as stiff about 55-60GPa. If you make a bending optimised laminate it will be 70-80GPa in one direction and 10GPa in the other (using std modulus fibre) aluminium wins against CF unless you need the lighter weight or exceptional strength (I build CF and glass laminates with flexural strength of 900MPa plus). Both of these are not needed in a non moving machine base. Another issue with sandwich construction for a machine base is that over time the oil will penetrate the interface and slowly break it down. Then there is torsional stiffness. Sandwiches do not do well in torsion. Although CF has good inplane properties its shear stiffness is poor around 8-10GPa whereas aluminium is 27GPa. The machine base is really a candidate for steel as it is stiff, cheap and torsional stiff. But I don't like steel and it has to be painted (that's why I use stainless steel on my production routers and maybe thats a good solution here) ... and I have to ship it around the planet... If you look at commercial VMC machine bases they are massive and this work points out why. So compromises have to be made in this area for this class of machine.... So maybe make the base as best as possible (2 parts 50kg each) then tune the rest of the parts to it to see what happens.

In terms of the analysis I have an ongoing issue with how to restrain the model. The articles on machine stiffness do not discuss how they restrained the machine. Did they load the machine from the tool to the table? Did they load the machine from the tool to something bolted to the floor external to the machine (ie the loadpath is quite different) Was the machine bolted to the floor like HAAS recommend? if so we suddenly have 5x more stiffness then how I model it. Many machine builders don't want their machine bolted to the ground... so many rabbit holes to cover.

Last night I did the first truss base and it came in at 4N/um so dismal result. rethinking that at the moment...Peter

[StrawberryBoi]

My first thought for modeling would be to restrain it at the tool holder end (infinitely stiff tool), unless you are including gravity. Then I would constrain just a back edge of the base and load the tool holder and the part or part mounting equal and opposite with your theoretical load. I would use a separating constraint on the bottom so it doesn't "fall", but isn't bolted so no support to stop lifting. This of course assumes an infinitely stiff and flat table surface.

Compromises almost everywhere. You could constrain by separating or sliding friction constraints on just the "feet" or small areas where it stands on a non-level but perfectly stiff table or floor.

[peteeng]

Hi Strawb thks for the input - The idea of simulation is to be as real as possible. 1) There is no restraint at the tool holder in practice and you say place a load on the retrained tool? that's no good. The rigid tool is a good idea. 2) there is no restraint at the base back edge unless the machine base is screwed down there. Most machines are free standing.
3) you can use gravity and have the machine sit on feet with friction to a surface but that's overkill as the machine is so stiff gravity does not affect it much. Next run I shall include gravity to quantify that statement.

The way I have been doing it is to restrain a rigid post placed central on the table. This mimics a vice say on the table being held. Since the machine is "weightless" (no gravity) the post restraint is the only restraint needed in a linear analysis. I then make the spindle and tool rigid and apply 0.001mm (1um) deflection to the tool end in X, Y and Z. The FE calculates the load required to move the tool 0.001mm so that then becomes the machine stiffness in each axis. Target is 20N/um in the thought that the FE model will be about 50% efficient once all the bolts and connections are really taken inti account. So I want a machine in reality to be stiffer then 10N/um... I have been doing this from the beginning to keepthe proces consistant. Sometimes I restrain the base to see what happens. I have also made floor frames and included those occasionally but they make the load path longer and don't help much... keep thinking Peter

[peteeng]

Morning All & Sundry - Misty morning here in lock down... I revisited the Skyfire model kindly supplied by Ralph some time ago. I'm not sure if Ralph has rebuilt his machine and formed an opinion on it after filling the casting with goop?? When restrained by the base bolt holes skyfires static stiffness is X17 Y24 Z45N/um. Its a cast iron machine so I have used E=114GPa for its modulus. When restrained by a patch on the table like a vice its X drops to 10N/um. The base weighs 106kg and its not a sophisticated shape. The load is applied 320mm above the table. Millis load is applied 425mm above the table so the moment is a bit more. But this means the target stiffness figures are at least right.... I had a play with stacking the 175 extrusions side by side and 200mm of solid CSA with poor results. So I'll rethink the base from first principles steel may have to creep in. Looking back there have been designs with 20+ all round. Maybe time to review the past work. Peter

since I had the skyfire model up and gravity has been mentioned I added a gravity case. it slumped 20um under gravity. This may affect its dynamic stiffness like a whip on a ballscrew. With a light cut if there is an up force (as mills us up cutters) then gravity will want to slump the machine this will create a vibration. maybe not just a thought.... Peter

[StrawberryBoi]

Hi Strawb thks for the input - The idea of simulation is to be as real as possible. 1) There is no restraint at the tool holder in practice and you say place a load on the retrained tool? that's no good. The rigid tool is a good idea. 2) there is no restraint at the base back edge unless the machine base is screwed down there. Most machines are free standing.
3) you can use gravity and have the machine sit on feet with friction to a surface but that's overkill as the machine is so stiff gravity does not affect it much. Next run I shall include gravity to quantify that statement.

The way I have been doing it is to restrain a rigid post placed central on the table. This mimics a vice say on the table being held. Since the machine is "weightless" (no gravity) the post restraint is the only restraint needed in a linear analysis. I then make the spindle and tool rigid and apply 0.001mm (1um) deflection to the tool end in X, Y and Z. The FE calculates the load required to move the tool 0.001mm so that then becomes the machine stiffness in each axis. Target is 20N/um in the thought that the FE model will be about 50% efficient once all the bolts and connections are really taken inti account. So I want a machine in reality to be stiffer then 10N/um... I have been doing this from the beginning to keepthe proces consistant. Sometimes I restrain the base to see what happens. I have also made floor frames and included those occasionally but they make the load path longer and don't help much... keep thinking Peter

Hi pete, I can see how what I felt I communicated and what it looked like was not a good matchup.

1) I provided and incomplete thought with an implied force application strategy and then in the same sentence expressed a grammatically separated thought if not constraining this way, so I understand the confusion.

2) I see an area of confusion here between, I think, several of us and your approach, will address in a moment.

3) It would definitely be good to quantify it, though the truth is it wouldn't matter even if there is microns of movement.


Ok, now to explain better how I'm thinking of the issue of simulating any mill. Even if gravity is pulling down a heavier mill like you mentioned after the quoted post to the tune of 20µm, this is a constant sag so will be trammed out. Basically, meh. So, ignoring gravity as it's only important for the purpose of avoiding stress failures the machine should never get within a mile of, you only have one goal in the FEA constraints: Simulating the stiffness/deflection of the machine between the cutting tool tip and the part. Given that the part and the tool are variable areas of variance, it is absolutely something I agree with to make the tool infinitely stiff. If you have a known or common tool stick out, then you can apply the side load to the tool end and get the best representation of the real moment load from that side loading. Likewise, you would ideally put an infinitely stiff part in the vice and do the same. Given that these stickouts are likely minor and can even just be incorporated as a moment loading manually, you can also eliminate the part or the tool and directly load the holders. For example, if you didn't want to constrain any part of the machine at all to allow full deformation studies of the base, etc, you can fix either the vice OR the tool holder, load the unfixed one of those two, and see the machine deformation/stiffness between the fixed and unfixed ends. This is, again, all you care about.

You can alternately NOT fix either of the load terminus you care about and fix an area of no concern, such as an edge that won't deform or a bolt hole, even a bolt hole edge. FEA just requires as you already know (just writing this out for those who don't) a fixed reference. This is an issue if you have no non-critical features in the load scenario, but that is just not the case. Pick a non-critical feature for the cutting load on a part that is substantially stiffer than the general stiffness and fix that. Then do the fixed deflection or fixed load testing. Do be sure that when you are using fixed deflection that it is relative to the other load terminus, ie that the deflection is tool holder (or infinitely rigid tool tip) relative to the vice (or infinitely rigid part cutting point).


I would like to briefly address a concern. Your input and your work is incredibly valuable to me and I think many others. You are though I feel, quite quick to interpret something as confrontational when I have almost never read such confrontation myself in the posts themselves. This isn't just in response to me, but in dozens of other threads I've silently just read. Please consider this: We are not your enemy, we are not seeking to show you to be wrong. You have plenty of experience in your 16 years (yes, quoting you there , yes, I know you meant 61, no, nobody cares it's just a typo we all know what you meant) and you undoubtedly know that everyone learning and sharing and making mistakes and discoveries together is a positive thing and not meant to tear anyone down. We can all learn, no matter our age or experience. You don't stop learning until you die, you stop learning and you are dead even if your heart is beating.

Assume that if you think something I or anyone has said is supremely incorrect or unhelpful, that the issues lie somewhere between you and that person. It could be you, it could be them, it is almost certainly both. I'll continue to pop my fractions of a dollar in despite the 'tude', but I think you might find the forum overall a much better place if it were, as one might say, adjusted.

[peteeng]

Hi Strawb - I am not a confrontational person and I write in a blunt technical style that does come across a bit exaggerated to some people. Any forum is not a good place for deep technical explanations or polished literature. You are not dealing with the person but a series of discrete written reactions to various stuff. Its up to the reader to filter the info they need at the time. Can't please everyone all the time & I don't try to... I take all input as good input. Thanks for your contributions. Overall the forum has provided me with information that can't be found other then by experience and time. I share my time and experience back...that's all I expect....Peter

Re - stiffness of machines
1) The "static" stiffness of a machine allows machines to be compared. There is no agreed method to do this so all of the articles and various sources that quote the machine stiffness have to be taken lightly. I feel its important enough that it should be quoted in the machine specs just like the horsepower of the spindle is, once a std method is published. I have communicated with a few machine makers and they have either declined to comment on this or have said they do not know the static stiffness of their machines. The current range of mill stiffnesses that I can find is 1N/um to 650N/um
2) The dynamic stiffness ie the interaction of the tool and part in the cutting process is another issue and is not represented just by the static stiffness. The dynamic stiffness is a combination of static stiffness, machine damping, fixture stiffness, tool holder and many other issues. The dyn stiffness is where the rubber meets the road and although you can have a good statically stiff machine it may be poor dynamically. FE is not yet good enough to capture all of that.... I look fwd to seeing some of your modelling, and as I've said elsewhere, welcome to the shark tank.

[ardenum]

Since I want to make moulds which can be deep I think the lifting gantry is out and I'm back to the Mazak type arrangement.

What do you mean? You can only go as deep as your tool is no matter what type of machine you have?

and wasn't this a much stiffer design? 30+ on all 3 axes?