New build with recycled parts on a ridiculous budget 120cm*60cm work area

[Nickhofen]

You might like that.
It is a diy handheld MPG to linuxcnc a guy from Denmark post at a cnc group on Facebook. The group is the DIY CNC and the member, Allan Madsen, said that today he is going to provide a link, to where he will post the files, to make it work in linuxcnc.

[Desertboy]

Excellent work by Allan will have to check it out, I know people have also got PS3 and Xbox controllers working in Linuxcnc for controlling the machines ;)

Having some latency issues with the I7 not geting the performance I should, it could be a few things from what I read up.

But the main candidates seem to be Nvidia GFX card (Need to work out how to turn it off in Linux so I can use the on board, machine is dual boot with windows want the Nvidia card for Fusion.) Possibly SMI, disabling cores and a couple of other things.

Toying with swapping the Y axis nema mounts so it's belt driven 1:2 motor:ballscrew as the 1604's have a critical speed over 2000rpm so I'm hoping I can drive the nema's at 1000rpm and get 8m/min out the screws and drive the 2010 at 1000rpm direct drive so I get 10m/min on the gantry. Resolution will be halved doing this but then the resolution of the 2010 ballscrew is still less.

Using 2010 ballscrew will give positional accuracy of ~0.05mm, slightly better with the 1604 geared 1:2 (Which essentially becomes an 8mm pitch for accuracy purposes) the microswitches I'm using (Omron) have a repeatability between 30-50 microns almost the same as the cheapo inductive sensors most of us use which works out to 0.05mm. So my repeatability on any part is limited to ths it might be better of course but in theory it should never be worse. (Ignoring wear and tear)

I have some 40mm inductive sensors I recovered that run on 48v I got from a large printer I scrapped I suspect these have high accuracy over larger distances, when I googled them they were over £100 a sensor the main reason I kept them lol.

I might have to strengthen the gantry with angle steel ground flat both sides and cnc the holes to bolt in into the T slot so the machine is rigid enough to actually take advantage of the speeds and can my PC generate enough pulses for that speed. I can see a step generator in my near future lol. I'm glad I have a friend who works for a big cnc firm as a foreman and who's job for the first 10 years was grinding ;) He also told me they been binning Hiwin's so he going to see if he can get me a set before they go in the skip, he laughed at my 15mm Hiwin's lol.

Thinking when I upgrade or build new if I don't go servos then going nema 24

https://www.ebay.co.uk/itm/Nema-24-S...item1c54c443d4

Good price for 8 wire motor, 4nm

Had a few crazy ideas for upgrading including 2* 2m travel gantries ;)

Still not got dual homing sorted but I did get an engineers square to straighten the gantry (It was well out, which is the technical and scientific term) and was happy to see when I straightened one side and checked the other side it was perfect. I had the machine switched off and rotated one ballscrew by hand.

Recessing the extrusion into the plate really helps in squaring the gantry you can feel it click into place when you've got it right.

[Desertboy]

I have a 1204 ballscrew, it's a ground THK C3 with matching bearings 450mm long.

I paid £150 for it a few years ago from South Korea, I was wondering what it's good for? I was wondering if I could use it for a Z axis? It would have another advantage that it will mount easier with 15mm Hiwin's but I worry it's not strong enough.

Still floating ideas
https://forum.linuxcnc.org/38-genera...ared-work-area

This one's a doozy ;) I think linuxcnc is probably capable of driving something like this especially now everything is joints but CAM software could be a bigger issue.



I just cloned the gantry not a proper design yet just floating ideas

[Davek0974]

Is it a twin-head or a duplicator ?

One is hard, other is easy ;)

[Desertboy]

Is it a twin-head or a duplicator ?

One is hard, other is easy ;)

I have a few ideas how to implement it, what switches I need god knows lol. I would like multiple modes.

1. Clone 2 identical machines ran from a linuxcnc controller, 8 AM882's.
2. Individual 2 sepearate machines how I would run it mostly, eyeing up the ddscv controller for this mode, 1 for my 2nd machine and linuxcnc can run the first machine.
3. If possible 1 linuxcnc controller utilising all 8 axis for a shared area this I see as maybe not possible with my level of skill and cam solutions.

How I switch all this I don't know but it can't be that hard.

I'm sure an inductive sensor can be used on one the gantres as an emergency stop so they don't crash into each other if the programming is haywire.

[Desertboy]

Had a quote back for 2020 ballscrew, 2m long with fittings 3 of them just under £500 delivered to UK, (Will be £100-£150 in import tax on top). I should probably just build a new router square with 1.8m*1.8m (Going to need a strong gantry) with steppers for my 2nd router I can get some Nema 34's for £5 each off a mate who has a dolab full of them lol for Y and drop a 4nm nema 23 from cnc4you for the X. This is a lot more doable in the short term than dual gantry with servos all round lol. It's Chinese New year so I can't order for a month, well I can but nothing will be posted for a month. So have a month to think about thing before rushing in lol. 1.8m travel is what I'm after having it on both axis would be truly amazing. I've been doing a lot of test cutting now on my machine for the things I want to build and I know I need at least 10m/min cutting speed to stand a chance of making money on it. What I want to build uses 2 types of wood and I actually couldn't fit it on a 1.8m*1.8m bed to cut it in 1 but if I keep the original machine as well then I could run them side by side and make the item in 1 run giving me a fighting chance lol.

I would love to build a dual gantry machine but having spoken to some people on the linuxcnc forums it seems the software just isn't there to take real advantage of it. So in the end I'd be building essentially 2 cnc machines in one frame I could get it to clone easy enough (Parallel port switch) but even that would have dubios advantages as you loose homing on 1 of the machines. After all if I have 2 controllers it's not hard to run the same code on both and then have homing and forget the cloning. I can still see an advantage to have a dual gantry with a shared area if 1 is R&P powered and the other ballscrew and you only work 1 gantry at a time. Would be easy to use a limit switch so if the gantry not in use is not parked then power to the other gantry is off. 1 gantry for speed and 1 for aluminium. They could then share electronics at least if I can find the right switches after all I only need to switch 4 wires for each stepper which would be 2 sets for this setup (X&Z). It very easy in linuxcnc to have multiple profiles for the machine so wouldn't be hard to configure. This way sharing the rail, frame, electronics & controller you're saving a lot of money and floor space to get essentially 2 machines.

Maybe build No. 3 will be dual gantry ;)