Milli a new composite mill kit

[ardenum]

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

Hey Peter,

Looking really solid, almost x2 jump in stiffness than previous model. I guess there's a reason that rigimill has 5 rails on the column. I'm interested to see how much damping will the hollow column provide...

[peteeng]

Hi Ard/Jon - The question of damping has always been a tough one due to the complexity of analysis and the conflicting data available. I ran the modal analysis on No23. The first mode is the column rocking at 203htz then it torques at 354htz then there's quite a few global modes. Since you mention the hollow the first local mode is the column walls panting at 800htz and 870htz. I'm thinking of making the column a bit asymmetric fwd and aft to counter this and provide a bit more stiffness. I have started to play with the base to reduce its mass but not successful with that yet. This is partially because I restrain it via the post which is a small object on a large base. I have to be careful not to optimise too hard here as in reality the "restraint" is a moving target depending on the part fixturing. Then there is the difference between dynamic stiffness and static stiffness and we'll only know that once its built! The static stiffness approach is sort of wrong in terms of what's happening when the machine is cutting as the tool and workpiece are intimately connected. In the static stiffness approach the tool and workpiece move apart which can't happen in practice... But its an easy and understood approach to getting a stiff machine... I will run a damping solver on this once the base is better resolved...Peter would be good to get videos up here. You need to pester admin to allow MP4 vids

[peteeng]

Morning All - maybe work towards something like this ... Peter

[peteeng]

Hi All - "Cement All" looks good. Available in Oz as well.. 38GPa in compression but add high stiffness granite or blue metal will get to 40+ Really low shrinkage (100% CSA cement) Peter

[peteeng]

Afternoon All you peps out there! Mucked about with a stand. If I "fix" the stand to the earth the stiffness jumps up to over 60N/um (with 35GPa material so if I get higher then E35 its a bonus) and this is very encouraging. I need more geometry in the base, flexing a flat object in torsion is inefficient. So if I design a suitable stand integrated with the machine I can minimise the base weight and use the stand for the extra stiffness. The stand can be made by anyone on the planet... no need to ship it somewhere. So onward to the next version incorporating the trade model insights... Peter

[peteeng]

Evening all you Peoples - I added the stand and found that as the loadpath is now really long (tool to stand feet) it did not really step fwd as expected. If I screw the base to the floor, yes massive advantage but I don't want to wear my knees out working at the table. I suppose I could cast a concrete block 700mm high to sit it on? Any how here's the guff. So I think this Trade model is at its furthest development. Need to move onto something that is close to production quality now...No27 stats are X 40 Y 45 Z 36N/um.. The stand is 202kg mainly 100x100x8mm steel tube... Peter

zip file is mp4 file of X axis jigging

[peteeng]

Evening All - I have been thinking about the issue of damping analysis (see attached it states there are 11 common damping parameters so its a bit complicated The viscous damping ratio and the log decrement would be the most common . Since there is a lot of conflicting info about damping rates and different FE systems handle damping differently I think that a sensitivity approach would be useful vs an absolute analysis. For instance in Simsolid it applies damping at a structural level vs a material level. In Strand7 the damping is applied at a material level which is more accurate as each material has individual damping factors. Setting up transient dynamic models' in Strand is very time consuming vs in Simsolid its very fast. So in SS I could run say three models one would have 2% damping next 5% then 10%. If each model showed one particle part to be troublesome then I'd attend to that part... I'll have to make a simple model and test the theory. I picked up the materials for the PlankNo2 today. I'll rip the melamine sheet tomorrow and start setting up the cavity... Peter

The attached article states to use these figures with caution due to the many variables involved.
viscous damping (zeta)
metal structures <1%
metals structures with joints 3-7%
auto shock absorbers 30%

https://journals.sagepub.com/doi/ful...87814019847009

[Zorbit]

I know you appreciate "out of the box" ideas, so here's one for you.

In the rifle shooting field we use moveable weights to damp the vibrations of rifle barrels with particular ammunition. It's very much a trial-and-error process, difficult to calculate, much like the vibration of a particular tool at a particular speed and feed would be difficult.

So, for your mill, how about a couple of extra linear guides on the machine that allow movement of a weight until it reaches a sweet spot - this could possible be an automated process, either with sensors reading vibration and a circuit adjusting the damper or with pre-programmed positions you use with a particular tool.

Perhaps rails just on the prototype, if you discover sweet spots you could add some threaded fixings for optional damper weights.

This could all be rubbish though, it just popped into my head. Too much caffeine perhaps.

[peteeng]

Hi Zorbit - some boring bars work like that. When I played competition tennis there were small weights that you clipped into the strings. You moved them around until you felt the ping go... especially with aluminium rachets at the time, then CF became reasonable...and they were quite damp...



I'm not concerned about vibration too much. It will be damp because its composite or concrete and it will be very stiff. I try to take advantage of analysis if the tools are here so its good to explore such things if you have the tools. Should be able to predict the vibs with the FE... Plus it allows me to develop an understanding of the software before someone asks me to do a commercial transient dynamic analysis. This happened a little while ago with a vibrating very large router...so I was prepped for the task...

This evening I thought I'd look at casting the bench vs steel fabbing it. So I drew up a CSA bench that weighed the same as the steel to keep it fair (220kg) . Its stats are X 21 Y43 Z38. The unbraced X direction is very poor other dirns good. Wouldn't like to cast <50mm thick.. So I could shift some material around but I'll leave it at that for now. Time to start again with production in mind... Peter

[peteeng]

Hi Pippin - Have a look at Soraluce DWS system - Peter

[peteeng]

Late evening all - Bought the mould bits yesterday and cut them up today. Now need to wax it and drill some holes and set it up for Plank No2... Maybe saturday. Also a 5 axis one box controller arrived today from china. Have to set that up and test it. Its for YaG. 36V 3.5A 5 axis (extra axis for future rotary) $87USD looks good. and the documentation is good and its in English bonus....Peter

[ardenum]

a different shot for rigimill

[peteeng]

Evening All & Sundry - I have found an ER20 power head and have started hacking the bits down to its scale. Its much more the type of thing my customers would graduate to from a router I think. Anyone can still put whatever they like on it and a pro user would prefer a BT30. Once I get the bits rescaled I'll decide on the working size... This series is to be called Milli-Pede. I have thoughts about billet machining the bits in aluminium. The machinist I use has a big Mazak 5 axis machine and it would be so easy to bolt down a billet and go for it. That would solve all the alignment issues as everything would be machined absolutely. He has tonnes of anodising done as well so I could have the parts nicely finished in red or purple or green... He also does a lot of 7075 stuff but I'd stick with 6061-T6 the workhorse alloy. Its a back step from composite but maybe an answer if plankNo2 is disappointing. I also have someone working on an asian source of chopped carbon fibre so that maybe another answer. If I billet the carriage I can do an integrated spindle Hmmmm... More ZZZZ's needed. Or maybe an extra coffee like Pippin has occassionally... Peter

[ardenum]

I've seen many videos on these milling heads but I'm wondering if its actually any better than having a regular motorized router spindle from hiteco/teknomotor...any data on that?

[ardenum]

A different composite:
https://docplayer.net/storage/104/16.../165236944.pdf
https://pdf.sciencedirectassets.com/...qb&type=client

[peteeng]

Hi Ard - Can't open the first article. The second article is one from an Indian Uni. I understand Universities needs for students to do stuff but they continue to publish stuff that says "we confirmed the FEA predicted the outcome" or to that effect. The FEA developers have heaps of post grad doctorates people churning away at code and checking stuff so of course its correct!! The Unis need to do some new work and use the tools to move forward. The days of having to prove simulation is correct every step of the way have long gone... They need to move onto optimisation and AI solutions would save me years of time.. Peter

Hiteco solutions are well out of the budget for the sort of customers I envisage....

[RubeG]

Hi Peter, great progress so far and I'm really enjoting your update, cheers!

[peteeng]

Morning All - I started sorting the PlankNo2 mould this morning. One of the many things to do at present. Last time I used pre primed timber and the paint failed (maybe the solvent in the wax) and created a mess. So this time I'm using aluminium and that will be good.

plank1 - using the example 80x6mm aluminium flat bar to check set out
plank2 - long shot. In the background is YaG's controller board. I test it this week when the PS shows up. Its a 5 axis board, want to run a rotary on the machine. YaG will make the moulds for Milli...

plank3 - checking the resin and vac port tubes won't get squished too much when closed. I think I'll run two resin lines and a single vac line...

Onward - Peter

[peteeng]

Afternoon All - A carbon fibre diversion. Today I went out to cut the paddock and I put my muffs on and they snapped. I was going to take them back but then they were over a year old so... I haven't used any carbon in a while so I set the break up and gave it a stabilising wrap. Tomorrow I'll give it the structural wrap.... Peter

[peteeng]

Morning All - Here's stage one done of the Muffs - Makes me think I should cost out Milli in CF properly. Peter