I keep coming back to moving table long-Z 5-axis as my next project, but I doubt myself, because so few people have gone in that direction.
Hi All - To discuss machine configurations a bit. 1) There is no ideal config. Each has a purpose and all machine manufacturers make several types of configs for different applications. If there was an ideal config we would have figured it out by now. We have been making CNC and manual machines for a very long time and evolution has not narrowed the configs it has broadened them. 2) The only "recent" thing that has changed mills and lathes is linear guides. Guides have opened up possibilities that did not exist with gibs and ways. Many different configs are possible with guides. Applications & production have driven development of VMC's and HMC.s in various ways 3) future machines will be influenced by 3D printed parts (lighter stiffer damper in steel and titanium) and composites (lighter stiffer damper, more stable over time).
I was convinced that a lifting gantry was to be my next build but this study has pushed me back to the moving gantry design. So Millis brief is a small benchtop or stand CNC Mill. With maximum stiffness which is not specified at the moment but all decisions have been made via the stiffest studied solution. Aiming at 20N/um which is quite a high target for a small mill. We shall see how that goes as the project develops... Peter
Last edited by peteeng; 10-28-2020 at 03:49 PM.
I keep coming back to moving table long-Z 5-axis as my next project, but I doubt myself, because so few people have gone in that direction.
Hi Zorbit - Nothing ventured nothing gained. Mechanically I think the 5 axis is not an issue. Its the CAM side I think is the hurdle. So find the CAM. 5axis is on my list too. Happy to collaborate. I see the 5 axis as a technical challenge but you must have something in mind for it to do? A good purpose is the best driver to get a machine built! What envelope do you need? Peter
Last edited by peteeng; 10-28-2020 at 03:49 PM.
I already have access to a 5 axis Morbidelli router and Openmind Hypermill CAM, but the Z clearance of the Morbidelli is not enough for my needs, also a second machine would allow me to have jobs on the machine for much longer.
If you don't have access to full-fat simultaneous 5-axis CAM then an indexable system is still useful and cheaper.
I'd like a 3 axis envelope of about 1500 x 1500 x 750 ( Z ), full 5 axis would be less. The Morbidelli has a nutating or cardanic head which allows excellent 5 axis machining of envelope edges, but it's difficult to build, so I'd go for something more conventional.
The machine will be for making foam plugs ( undersized ) , then skinning with something more solid and finish machining - preferably without removing the job from the machine until it's finished.
I like this machine, but prefer a moving table system:
Last edited by Zorbit; 10-28-2020 at 06:02 PM.
Hi Zorbit - The 750mm Z is the trick but its only for foam so does not need to be uber stiff. So you have 4m floor space for the moving table? Peter
Hi Zorbit - unlimited space that's a bit of fresh air. That solves so many problems... I think its a laminated timber and aluminium beast then. What CAM would you use? stepper or servo? If money short I suppose stepper.... As long as sparse brains are used well thats enough...Peter
I already have a 1kw and 2kw Mitsubishi AC servo motors with matching drives, fitted to 32mm x 10mm lead ballscrews and blocks, 25 and 30mm profile rails ( all bought used ), a spare copy of UCCNC and UC300ETH-5LPT ( I used these when I replaced the original control on my Camtech Z7 and really like them ). I have various steppers that could be used for Z B C, or some brushed DC servo motors around 200w.
I don't have any harmonic drives, but I do have some very good quality reduction gearboxes with very little backlash. Unfortunately the reduction ratio is pretty high at 535 to 1 ( they're two stage, gear then worm, but it might be possible to remove a stage ). They're very robust units, easily capable of supporting a 5hp motor, the output shafts run on 75mm tapered roller bearings. I got 4 of these, plus 3 ready made profiled rail Z axis setups from a scrap yard that had just done a factory clearance of a PCB manufacturing plant. The guy in the yard let me load up a pallet with all the bits I wanted and charged me by weight, £20 .
He even warned me that the black granite surface plate I'd chosen wasn't metal.
Zorbit - Your more than half way there! Have you started a CAD model? Lucky about the granite tip , you may have tried to weld it I'm working through the Milli BST parts for final checking hope to get order out today....Peter
Hi All - Have sent PO to BST so now onto the sheetmetal parts checking.
Last edited by peteeng; 10-28-2020 at 10:25 PM.
Hi Everyone -I have been reviewing spindle motor options:
DHT-86M 0.75kW servo 3000rpm 2.4-7Nm max 3.1kg dia14 shaft dia80mm mount VSD seems to be ideal. maybe use a 1:2 pully 6000rpm
HLA inductance 0.75kW 3000rpm 2.5-7.5Nm 12kg!! dia19mm shaft 120mm dia mount VSD direct drib-ve or 1:2 pulley to give 6000rpm
N34 stepper 12Nm holding 6Nm at 750rpm Dia14 shaft 5kg frame 86mm sq VSD, drooping torque curve is the worry 1:4 pully? 0-3000rpm?
Mills
Sieg 500-1000W brushless 100-3000rpm VSD
grizzly 750w 50-2250rpm VSD
taig 187W 1000-11000rpm 4 speed pulleys
H&F BF-16V 500W VSD 100-3000rpm with 2 speed gearbox so its a 1500rpm motor
Hmmm keep looking - Peter
Hi All - This morning I rebuilt the 20mm car and rail for the next FE round. The prior models were the manufacturers models and these have a lot of extra fillets and details that are not needed and just take uptime sorting and dealing with. When the FE system searches all the connections it also generates 10x more connections then needed. Now it finds 4 which represents the ball tracks. Very good. Peter
Hi All - Here's the "usual" connections at a car. The mill model has over 1500 connections to check. The std cars have 17 in the car then theres the lubricators and nipples to remove as well. Next models will be much easier.. Hiwin call their parts Lump 1, Lump 2 etc thats funny...Peter
While I had the model up I thought I'd check the car stiffness. So I looked at Hiwins car stiffness data and at 5000N up it deflects 15-20um with the high preload car. I usually model the cars as aluminium so ran the model. The car deflected 15um which is good. The transverse did about the same so the car stiffness is close enough for modelling.
Last edited by peteeng; 10-29-2020 at 07:58 PM.
Mactec54
Hi Mactec - Yes a DMM servo drive will be needed. By VSD I mean an electronic speed control of some description vs pulleys or gearboxes.... I've been looking at a 750W axial flux motor looks encouraging for the basic model. But have to buy 20 sets as its an OEM motor. Anyone know anything about axial flux motors? and yes Mactec it is a pool pump and a solar car motor but its torque curve and utility are good....Peter
Hi All - Been looking at the small Syil as a benchmark. 1.1kW 100-5000rpm 275X 155Y 230Z Peter
https://www.pinterest.com.au/pin/638596422140633305/
Huge spindle overhang and a saddle.....
Hi Klaas - I assume you are talking about the Syil in the prior post? If so yes the saddle function is replaced in the moving column with the column moving in the X dirn. The table moves in the Y dirn. The conventional directions are Xdirn+ to the right as you stand at the machine, Ydirn+ away from you as you stand at the machine and Zdirn+ up.... Spindle overhang (which is the "Arm" or overarm is governed by what Y travel is required. At some point the arm becomes very long and then you have to go to double columns or gantries to get the stiffness. Its rare to find a column VMC with a Y over 400mm (if there's one out please tell me) so that would seem to be the commercial machine limit... The Syil has 155mm so I don't think its very long, so its a case of making the arm stiff enough... In Millis case its about 300mm tall and 200mm wide solid. So its arm is very stiff. Peter
[QUOTE=peteeng;2411050 Its rare to find a column VMC with a Y over 400mm (if there's one out please tell me) so that would seem to be the commercial machine limit... [/QUOTE]
I presume you're referring to 'commercial' hobby-sized VMC's and not actual commercial machines. If not, there are a ton of single-column VMC's with significantly more than 400mm/15" of Y travel. Internet search for Fadal 6030 - and that's an old one.
For a newer VMC - Haas makes a VF-12. 3810mm x 813mm travels.
The only significant difference I can see (other than scale) is that a typical commercial C-frame VMC's saddle is the full width of the X-axis so there's no table overhang. Columns are massive - twice the width of the head (or more) and as wide as the base at the mointing points. Bases are massive, saddles are very wide on the Y-axis interface, etc.
But nothing really different other than full support of table at travel limits like your 2nd pic above.
Oh, and putting a CAT40 (or 50!) face mill in the spindle is a back-wrecker. If the head isn't down low, you're climbing in the machine to load the spindle.
Hi Spumco - Thanks for the machine references. I looked up the Haas VF-12 and that's alot of metal! It uses 3 rails under the table and the table is 7" thick. Fadal are releasing a bigger one as well.
You mention scale and that's one of the issues with small machines. They just don't have the geometry available to be as stiff as a full size machine. Materials don't help as steel is the top modulus of low cost material at 200GPa. I have used silicon carbide for the wear strip on race car air dam leading edges. But it becomes nearly unmachinable once cast. So we are stuck with steel, cast iron, carbon fibre, aluminium, carbon fibre, mineral castings as materials (in descending order of stiffness). So geometry is our friend. One day we may be able to use graphene that has a E=2 TPa or 2000GPa that will solve some of these things.
https://meehanitemetal.com/wp-conten...09.07.2013.pdf
Maybe I missed it during one of your revisions, but if stiffness is your goal what was wrong with an up/down gantry & table? Fixed & triangulated columns with dual Z drives (at each end of gantry), plus a moving table in Y. Same thing basically as the Vulcan.
What was the compromise? Looks easiest to reinforce/stiffen through geometry than any of the other variations.