One ring to rule them all?

[Jonathan]

Using HIWIN HG25 linear rails http://cnc4you.co.uk/resources/Hiwin%20General.pdf
basic dynamic load rating (lowest rating)
26kn / 2600kg. At a separation distance of 300mm and 1m Z that would be 780kg or 7800nm (but there is 2 of them so this times two really). Clearly the linear guides are adequate.

Adequate in terms of load rating, yes, but adequate in terms of stiffness ... you need to work that one out. There is stiffness data available from THK if you dig around.

[JAZZCNC]

Completely do-able but just beware if you dream Big then be prepared to Fail Big.!! . . . . The Chances of failure are very high.

I also Agree plane vanilla 5mm box section without any special treatment will be Big fat Fail.!! . . . . It's not just deflection you need to think about. Think about resonance and quality of cut etc. Box section without being filled will ring and large section box with only 5mm walls at over 2mtr will chime like Holy Trinity bells.
I've got a large router with gantry 2mtr wide with 10mm thick walls and that rings like a bell, doesn't deflect much and for router the resonance from it is more than acceptable but for milling Steels, even mild steels then forget it. It would sing like an. .

The point being you'll seriously need to Re-think your material choice and look hard at design methods if this is to come close to working at an accepatable level. Milling steels takes you another level of building compared to routers and cutting exotic steels takes you several levels higher.!!!

But like I say still do-able but with massive chance of failure.!! . . . . Good luck thou.

Edit: I'll revise that to Massive chance of dissapointment.!! . . . . Not to try would be a massive failure...

[Rufe0]

RE:routercnc
The problem with a machine where you move the work around is you loose working area. Although that machine looks great it only has a working area of 500x500mm.

I was thinking I would have a 5axis head attached to a normal xyz gantry machine like this http://www.5-axis.com/.

I'm thinking of designing the gantry like this, cross section

The Z axis would probably run down the middle.

Anyone know how to work out the deflection? I'm thinking If you had a load of 100kg at 1m thats equal to 333kg with a separation of 300mm. Then calculate the bending deflection of one beam with 333kg, then multiply that by 3 to get the deflection at 1m.

To be clear I was thinking the frame would be 100x100x5 and the gantry 200x100x8 but only because those where readily available sizes. If I have to special order some bigger then so be it.

[Lee Roberts]

If I have to special order some bigger then so be it.

I like your enthusiasm!

I did have an idea though, what about in addition to the Z lowering the spindle another axis could raise the bed?

Damn it, I had that idea myself months ago but kept it to myself for "the big reveal" lol, reading your original post in full, this project has the potential to be a big deal if you come up with something and pull it off.

I wish you all the best of everything and hope you can pull it off, you sound determined and that's a good start!

.Me

[JAZZCNC]

To be clear I was thinking the frame would be 100x100x5 and the gantry 200x100x8 but only because those where readily available sizes. If I have to special order some bigger then so be it.

To be honest I would avoid any box section for stiff areas of the machine like Gantry, Z axis etc and look more towards plate and weldment setup then have the structure machined flat and true. This is the only way you can control strength and resonance properly.

This can and will get expensive fast.!! . . . After all There is a reason why the Big boys toys cost so much money.!!

It's very easy to get carried away at this design phase and create a monster CNC but it's all different thing when the fat lady sings and reality hits you in the balls.! The reality of costs alone sting like a B@~~rd for Strong/heavy high quality machine and thats without getting Silly with 5 axis heads etc so to quote Ali-G try to "Keep it Real" otherwise your doomed to dissapointment from the start.

[routercnc]

Hi Rufe0

These are my thoughts for a machine like that -

1. Double gantry, as per your drawing, and the machine I linked to, and MechMate:



and Jonathans machine a while back-
http://www.mycncuk.com/threads/6484-...strong-machine


2. Box Z, as per the link machine and Jonathans using 4 rails on the Z

3. Twin ballscrews on the X,Y and Z axis to avoid racking and allow the bearings to be close to maximise travel. Z twin ballscrew is to help with lifting the weight of the 1m axis, plus keep the loading central to the spindle.

4. Hiwin rails all round, 20mm or bigger.

5. Height adjustable bed with vertical rails at each corner -
Manually
or hydraulic using 1 ton bottle jack in the centre
or ballscrews
Which ever option you must be able to 'lock it down' before cutting - like you would lock a quill or lathe saddle once in position. Maybe thick angle brackets around the edge to nip up and hold it fast.

The alternative to height adjustable bed is to use bedplate mounting blocks (scroll mid/bottom):
http://cnzbzg.en.alibaba.com/product...eers_cube.html


6. Spindle(s) capable of high speed medium torque [alum, wood], plus very low speed high torque [steel]. Maybe via a single spindle linked to a separate cartridge spindle with pulley drive to give different ratios.

7. Automatic Tool Changer (air powered), optionally with ATC carousel.


In terms of your calculations the bending stiffness calc is straightforward (load one beam then half the deflection for 2) but that only accounts for the Z axis weight and the plunge stiffness. For the sections you are talking about (200x100x8) etc this will no problem for simple bending. However, where you will come unstuck is having enough torsional stiffness when cutting in X (gantry moving forward or back), especially at 1m extension. I don't think hand calcs would be possible on a double beam in torsion so you need an FEA program if you want to work it out. I notice Fusion360 has FEA built in and this is 'free' so worth a look. This will give you deflections once you have learned how to use the CAD part of the software.

[Rufe0]

Hmmm well I've tried but I just cant get to grips with Fusion360. I tried one called LISA which was ok drawing wise but the actual FEA never worked.

As I see it the gantry and Z are like a T in cross section, which means its a simple bending calculation. I'm getting around 0.025mm deflection at 1 meter with 200x100x8 seperated 400mm center to center.

Think I'm going to have to scale down my plans though. I will have to find a cheaper alternative to what i'm looking at the moment, maybe just wait for stuff to come up on ebay. Currently I'm pricing it from Zapp Automation for ballscrew related stuff and cnc4you.co.uk for the rails/carriages.
£110 Ballscrew support fixed end
£36 Ballscrew support floating end
£88 Ballscrew
£132 Ballnut
£37 Hiwin 25 Carriage
£37 Hiwin 25 Carriage
£108 Hiwin 25 Rail
£40 Motor
£10 DIY motor driver
Thats about £600 per screw. I need 6 screws for the basic XYZ table thats £3600. +2 for the lathe tailstock £4800. +4 for the moveable bed £7200.

So now I have to decide whether it would be best to get an old bridgeport for around £2000.

[JAZZCNC]

Head East my friend.!! . . . .China is the place for ballscrews. Contact Fred Lee at BST automation on Ali express and get a quote. You'll be very pleasantly surprised. If you want higher spec ballscrews then just Ask because he sells higher spec than class 7 advertised.

He Also sells Hi-win rails much cheapper than you'll buy in UK.

Whats the £10 DIY Motor drive.? . . . . If your thinking to build your own then unless your really really really good with electronics and have full knowledge of whats required then I'd strongly urge you to forget it and buy some decent drives with the Money you'll save from Fred Lee.!!

[Rufe0]

Bloody internets been playing up for the last couple days but yeah I had a look and thats great, this will change everything. Thanks

[Rufe0]

Yeah I dunno about the controllers, I'm using an arduino and it seems promising so far, i've just got a rotary encoder in the mail today so will be testing that out soon as.