Newbie new build

[Desertboy]

You're not going to do it on your budget sorry to tell you this, it's depressing I know and I thought I could build one on a budget and I got a lot of stuff for almost nothing including the frame and it still cost £1500

You need a vfd for the spindles they won't work without it.

I'd say neale's right £2500 is the zone I had to work real hard and had a lot of luck to get mine to £1500 some things are not worth wasting money on like those motors and drivers you linked to other things you can scrimp on get a $5 BOB for now, make your PSU, beg a cheap table to put it on. But don't cheap ass the drivers, the frame, the slides, the motors or the ballscrews as these can;t be upgraded easily without the first ones having little resale value.

Also 10mm pitch ballscrews, don't make my mistake

[Heliox]

Hey Heliox

Welcome to the forum

I think your budget might need to stretch abit more if your looking to do aluminium aswell, there’s plenty of people who can refer you to contacts to get parts cheaper an deals, also don’t order your electronics until your design in in place as things will change and you don’t want to be stuck with parts you’ve brought that aren’t no good,

go onto the guys channel he has got a build log to give you a better understanding of his design, and I believe them Nema motors your looking at are too small nM wise, I think they would suit a 3D printer more, the standard seems to be 3nM or more from what I’ve read.

With regards to the spindle water cooled is the best option, again people can give you contacts to sort that for you with a vfd to go with it

Sam, Thanks alot for that input, I'm a good listener, and what you said makes sense.

Those kits look attractive, but are underpowered .............The electronics is not the right place to start.

Although you have talked about the materials you would like to cut, you haven't mentioned cutting area which is a major design driver. And although everyone would like to cut aluminium, that's a big step in rigidity and accuracy over and above what's needed for wood.

You need to go down that route with your eyes open. I'm not saying don't, just be aware of the compromises you are making and decide realistically if that's for you.

Ask questions and listen to the answers - there's a lot of experience on this site from people happy to help.

Just as a stake in the ground, thinking about budget - my roughly 1500x750 cutting area machine so far cost around £2500
But good luck - it's great fun building a machine like this! Just be realistic

You're not going to do it on your budget sorry to tell you this, it still cost £1500

You need a vfd for the spindles they won't work without it.

I'd say neale's right £2500 is the zone I had to work real hard and had a lot of luck to get mine to £1500 some things are not worth wasting money on like those motors and drivers you linked to other things you can scrimp on get a $5 BOB for now, make your PSU, beg a cheap table to put it on. But don't cheap ass the drivers, the frame, the slides, the motors or the ballscrews as these can;t be upgraded easily without the first ones having little resale value.

Also 10mm pitch ballscrews, don't make my mistake

Guys already Iv'e made some some design decisions, I am only going to cut card, I won't bother with any electronics, card will be 2mm thick, work size 300mm x 300mm.............aye, Iv'e already got the kit for that, a good stanley knife and my right hand I'll spend the other 499 quid on Ibuprofen and beers to aid the repetitive strain syndrome I'll have lol.

o.k Jokes aside, I get it, and I don't believe in reinventing the wheel after receiving good advice so thanks all.

o.k here goes.

my main desire/requirement right now is to cut carbon or GRP plates, and likely never to be above 3mm thick. (though I assume if it were 30mm thick I would just need more tool passes, not a stronger machine) So realistically from what you guys say, if I get sidetracked from that objective to wanting to cut Aluminium and wood etc then I am looking at bigger costs.

(I do want to keep cost down as this is a startup venture with one specific material to cut to my designs).

1. work size will be approx 1000 x 600mm, I am open to change on this for cost effectiveness, the components will not be wider than 250mm square so having a 1000 x 600mm CNC I reckon I can produce 4-6 at a time. but if it were more cost effective to build 2 machines that cut 2 at a time I would do that, not that I intend to, but for example If I only could cut one at a time, then I only need a machine that will cut 250mm square 3mm plunge into carbon fibre plate, so I am open to changing the size depending on cost, if I have to run the machine twice to produce 3 parts, 4 parts,,5,6,7 etc so be it.

So because I would like to batch produce these components, if I could do 20 in a day I'd be happy. and that will determine machine size.
But it seems a waste to limit my machine to just 250mm square 3mm carbon fibre plates if with a little extra grunt and money I could build something capable of more.

2. forget cutting aluminium (I'll have to build or buy another machine in the future.....shame that

3. water cooled spindle,,,,,,,,,,whats a VFD?........variable frequency drive ? i.e speed controller?

4. decent nema motors (even if I oversize them to able to use on another machine).......how do you calculate the nM required, what is the relationship between the force and table size? I guess the longer and wider the ball screw and hardness of materials being cut to travel speed determines the motor force needed? but how do you decide or calculate it,,,,yes I could copy another design for that but is there some rule of thumb?

5. What's a $5 BOB?

6. I already got a Table, it is an old school classroom table, 1" steel legs/frame and a formica type top bolted to it, I can attach my machine to that.

This forum needs a glossary for us noobs ;-)

Thanks for the input guys.

[Desertboy]

Sam, Thanks alot for that input, I'm a good listener, and what you said makes sense.
3. water cooled spindle,,,,,,,,,,whats a VFD?........variable frequency drive ? i.e speed controller?

4. decent nema motors (even if I oversize them to able to use on another machine).......how do you calculate the nM required, what is the relationship between the force and table size? I guess the longer and wider the ball screw and hardness of materials being cut to travel speed determines the motor force needed? but how do you decide or calculate it,,,,yes I could copy another design for that but is there some rule of thumb?

5. What's a $5 BOB?

6. I already got a Table, it is an old school classroom table, 1" steel legs/frame and a formica type top bolted to it, I can attach my machine to that.

This forum needs a glossary for us noobs ;-)

Thanks for the input guys.

Can you weld? Or do you know someone who can and will for free. This will save you a lot of money.

You will need slides
https://www.ebay.co.uk/itm/15-1500mm...oAAOSwyi9aJP-x
This is the cheapest I've found them and I have this exact set very happy with them, you will not find them cheaper than this new.

You will need a drive as well, ballscrew is your best choice here Ebay is your friend but expect £250 and maybe some import tax.

3. Variable frequency drive it's a inverter that drives the spindle. Spindles are not running 240v single phase they are 3 phase and need the VFD to drive them you cannot connect a spindle directly to either single phase of 3 phase otherwise you get the magic smoke. Water cooled spindle are much quieter if you need a water pump pay postage I'll send you one for free I have a box of them. £200 for spindle with VFD.

4. Everyone seems to go for 3nm Nema 23's ~£30 a motor, I have 2nm on my machine £20 a motor. You definitely don't want less than 2nm.

5. Breakout board to connect to parallel port on a PC, this means you need a desktop PC to drive the machine but it's the cheapest possible way (Yes even cheaper than arduino)
https://www.ebay.co.uk/itm/Mach3-CNC...AAAOSwhqhaRkJb

Buy 2

Think about how your going to control the machine you can buy Mach 3 £200 or use linuxcnc free I chose linuxcnc ;) You need to generate Gcode for the machine most of us are using fusion 360 for this but there's many ways of doing it depending on the job.

I understand the need to cheap ass it I really do I would never have built my machine if I knew what it would cost lol but I'm so glad I did. The best thing you can do first is learn! Read people diaries ask questions. If the job you have for it is profitable you can probably find someone on here to make them for you and still make money giving you a cnc fund ;)

Don't get put off by the cost because in the end the cost will seem nothing to amount of work you have to put in lol, there's a reason we all take 3-9 months to build our machines the first time it's a lot of work when you include what you have to learn.

But once you've done it you're in a special club for sadists that only members can ever understand lol.

It's also highly addictive ;) Most of us on this site would rather get 2 3m lengths of 25mm Hiwin's with carriages for Xmas than a new Iphone, which is the way it should be to ;

Now I built one I can say you build must start in cad! I thought I could build one without cad I was so wrong lol I was being idle. I could have worked it out on paper but damn it would have be sadistic and expensive with mistakes.

[Neale]

1.That's a good size. Big enough to be useful, but not so big that you start to give significant construction issues. It's not really worth going much smaller unless you have to because of lack of space. The finished machine will end up around 300mm bigger than the cutting area in each dimension (give or take a bit).

2. By keeping the size down, you can get reasonable strength again without going to anything too ambitious by way of construction. You might end up with something that can cut aluminium with care.

3. As said, water cooling means that the spindle is quieter, and can run for hours without problem. So can air-cooling, but you need to be a bit more careful about dust management so that it doesn't clog. You are right about VFD. Generally these can be controlled manually via the front panel or connected to your control box for software control of start/stop/speed.

4. There are some spreadsheets on this site which will allow you to calculate the exact motor spec required. However, to be pragmatic about it, the answer for this general class of machine is NEMA23, 3Nm or 4Nm. Might as well go for the more powerful as there is barely any cost difference. Don't go for cheap eBay motors as these are generally high-inductance and for various reasons this is a bad thing!

5. BOB is break-out board, and functionally it is a kind of junction box that takes the connections from your PC parallel port and makes it easy to connect the stepper motor drivers and things like emergency stop buttons and limit switches (for end-of-travel detection). It does a little more than that (like provide isolation between PC and drivers for protection of one from t'other) and might also provide the signals you need to control the VFD but if you think of it as a junction box, you won't be far wrong. There are really cheap BOBs that will do the job. However, they assume that you are using a parallel port on your PC and a lot of modern PCs don't have these although you can add them via an add-on interface card. That brings up questions of Mach3/Windows, or LinuxCNC. That's another whole story but not really one that needs too much thought at this stage. There are other alternatives which complicate the story as well - we'll leave that one for now!

6. You will be astonished at how fast and how violently a decent machine can shake itself around. Even if you are not doing fancy 3D carving (and you will be able to) even things like boring holes involve frequent and rapid direction changes. That table will hold the machine while static but it's going to need quite a bit of bracing to handle dynamic loads.

Keep asking. Don't get upset if you get your leg pulled from time to time. Goes with the turf! But you'll get plenty of useful info as well.

[Clive S]

What Neale has said.
Plus:-

You could go with supported rails but if you do you will wish you had gone of proper profile rails, although 15mm rails will do and are plenty strong enough the 20mm are much easier to mount and are not much different in price.

For a machine that size I would go with ally profile 90x45 The holes in the ends of them can be tapped M12 use two pieces for the gantry arranged in an L config and three for the main frame with ally end plates 20mm thick. Check out Joe Harris on here and see his vids all built with simple hand tools.

Start a build log to keep all the questions in one place

[Heliox]

Can you weld? Or do you know someone who can and will for free. This will save you a lot of money.

You will need slides
https://www.ebay.co.uk/itm/15-1500mm...oAAOSwyi9aJP-x
This is the cheapest I've found them and I have this exact set very happy with them, you will not find them cheaper than this new.

You will need a drive as well, ballscrew is your best choice here Ebay is your friend but expect £250 and maybe some import tax.

3. Variable frequency drive it's a inverter that drives the spindle. Spindles are not running 240v single phase they are 3 phase and need the VFD to drive them you cannot connect a spindle directly to either single phase of 3 phase otherwise you get the magic smoke. Water cooled spindle are much quieter if you need a water pump pay postage I'll send you one for free I have a box of them. £200 for spindle with VFD.

4. Everyone seems to go for 3nm Nema 23's ~£30 a motor, I have 2nm on my machine £20 a motor. You definitely don't want less than 2nm.

5. Breakout board to connect to parallel port on a PC, this means you need a desktop PC to drive the machine but it's the cheapest possible way (Yes even cheaper than arduino)
https://www.ebay.co.uk/itm/Mach3-CNC...AAAOSwhqhaRkJb

Buy 2

Think about how your going to control the machine you can buy Mach 3 £200 or use linuxcnc free I chose linuxcnc ;) You need to generate Gcode for the machine most of us are using fusion 360 for this but there's many ways of doing it depending on the job.

The best thing you can do first is learn! Read people diaries ask questions. If the job you have for it is profitable you can probably find someone on here to make them for you and still make money giving you a cnc fund ;)

But once you've done it you're in a special club for sadists that only members can ever understand lol.

It's also highly addictive ;) Most of us on this site would rather get 2 3m lengths of 25mm Hiwin's with carriages for Xmas than a new Iphone, which is the way it should be to ;

Now I built one I can say you build must start in cad! I thought I could build one without cad I was so wrong lol I was being idle. I could have worked it out on paper but damn it would have be sadistic and expensive with mistakes.

1.That's a good size. Big enough to be useful, but not so big that you start to give significant construction issues.

3. As said, water cooling means that the spindle is quieter, and can run for hours without problem. So can air-cooling, but you need to be a bit more careful about dust management so that it doesn't clog.

4. to be pragmatic about it, NEMA23, 3Nm or 4Nm. Might as well go for the more powerful.

6. You will be astonished at how fast and how violently a decent machine can shake itself around. That table will hold the machine while static but it's going to need quite a bit of bracing to handle dynamic loads.

Keep asking. Don't get upset if you get your leg pulled from time to time. Goes with the turf! But you'll get plenty of useful info as well.

What Neale has said.
Plus:-

You could go with supported rails but if you do you will wish you had gone of proper profile rails, although 15mm rails will do and are plenty strong enough the 20mm are much easier to mount and are not much different in price.

For a machine that size I would go with ally profile 90x45 The holes in the ends of them can be tapped M12 use two pieces for the gantry arranged in an L config and three for the main frame with ally end plates 20mm thick. Check out Joe Harris on here and see his vids all built with simple hand tools.

Start a build log to keep all the questions in one place

Guys what can I say, thanks I'm off to a flying start already with all those suggestions, many thanks.

re: preferring Hiwins to a new Iphone,,,lol,,,I totally get that angle ;-)

You've thrown me with the "can you weld" question? I can as it happens (I used to be a coded welding inspector offshore, among other things) but can't think what you'd be welding on a CNC machine, unless your talking about welding some of the Ali (or Steel) sheets/brackets/side plates? And to be honest I'd probably get someone else to do it as I haven't in a long time, other than a bit of mig and stick on various bit's in the workshop. (oh and I don't have roof on my workshop at the moment).

Software: I have been using Sketchup for a number of years, so no issues with a bit of CAD, I was about to download Fusion 360 a few days ago, haven't got round to it yet, but I did see you can use it free as a small business or hobbyist/student so that's good.
I also downloaded a few weeks ago a plugin for sketchup that produces G code 'SketchuCam'
I got that here http://phlatforum.com/xenforo/

I do have a question on designing the CNC machine in Cad though, I can't see the benefit in designing in CAD, what I mean is am I missing something? because to design a CNC m/c in a CAD program you have to know the accurate Dimensions of each part, or box profile and then Draw each into the program. Which means unless your provided accurate drawings of each individual part before buying it, then you have to have the part in your hand and take a vernier to it, which means you have probably already bought the part? I can see how while your building it it might be helpful to make a drawing of it so you can add parts or modifications, but I can't see how your accurately going to draw it before building it? That's not me being awkward, I just don't get it? can you elaborate how you found it helpful?

Re: The controller/PC,,,,,,it just so happens I scrounged a windows XP (or vista) desktop tower off my father in law about 3 yrs ago, and funny enough I wanted it for the parallel port as I have an old laptop with a parallel port that I use to send files to a large A0 A1 printer plotter I bought off eBay a few years ago, the laptop screen is dodgy sometimes so I got the tower for that as backup,,,it's sat in the box just in case but I took it out a few days ago and powered it all up,,with the intention of possibly using it as a dedicated CNC computer, but I also bought a 3d Printer last week that I thought I might use it dedicated for that too. (Tronxy x5s, not arrived yet).

BOB's,,,,,,,,,,familiar with the concept of them, and why opto isolation etc.

Re: the table and how violent the machines can be, I get that Neale and agree with you, many moons ago I worked installing multitool CNC machines in Cosworth in Northampton and I also worked for Sony in Bridgend and Festo pnuematics, so I totally get where your coming from re: the forces, In fact anyone who has had a run-off with a decent hand held plunge router knows when you bolt something like that down and start cutting things there's alot to go wrong and fly about and all round alot of harnessed energy. (I'd thought about bolting the table down, adding a decent slab of something on top to give it a bit more mass and some form of diagonal bracing and beefing up the legs).

Dust management,,,,,mmmm yes that has been in the back of my mind as an issue, not for clogging a spindle but just the sheer pain in the rear factor of capturing it all, I have an axminster dust extractor for a Jet table saw, but extracting carbon fibre dust is going to be a different ball game, I'm contemplating a glass fibre wet tray into the design so the parts are submerged in water during the cutting, something similar to a wet tile cutting machine, has anyone seen anything like that on a CNC on this forum?

I see you all recommend the linear profile rails opposed to the round supported rails, why is this? are the profile rails less susceptible to play? or wear? less vibration? what's the negative issue with the round supported rails?

Thanks again guys,,,off to do a load of reading on here.

[Desertboy]

Cad models aren't so bad, you can download dxf's from KJN aluminium if using aluminium profile, Hiwin cad models are available from their website make sure you get the right model (It's the carriages you have to watch mostly) I never cad the ballscrews out I just did the maths using the datasheets but you could cad them if needed the thread is irrelevant. I certainly never picked up a pair of verneirs during my build lol

The information you need on the ballscrews are on the datasheets for BK12/BF12 and dsg16 if using 16mm ballscrews.

Further to what clive said about 15mm being harder to mount you need spacers see below.



also greasing 15mm rails is also a PITA.

If you have trouble finding cad models grabcad is good but I always grab the models directly from manufacturer or reseller if I can as accuracy is important.

Lots of people makes their frames from welded steel box section and then use epoxy resin to level the box section where you fit the linear slides. Steel is a lot cheaper than aluminium extrusion, it's also popular to mix steel everything with an aluminium extrusion gantry which is my plan for my next build.

There's £600+ in aluminium and fittings in my router if I'd paid retail for it and I have 120cm * 70cm travel. You could certainly do something in steel a lot cheaper than that and of course it's stronger.


Accuracy with a steel a frame is nowhere as important as you think depending on how you fit the hiwin's the most important thing is they are level and parallel to each other. I'm going to weld mine together myself but have the steel cut by someone else (Who will be super accurate and square) to make things easier then you need a level working space and so decent magnetic clamps should be fine.

This is a quick mockup I did in fusion 360 for my next router, see how I'm planning to mix steel with extrusion. The cost of those 2 pieces of 9090 extrusion is £240 alone! My design is a nightmare because I want the Hiwin's sideways on so the frame does have to be very accurate and epoxy levelling for the Hiwin's will be difficult as I'll have to turn the router 90° to level each side. Mount the Hiwin's on top and you can don't need the frame to nearly as accurate.



That's not a ballscrew it's a cylinder I extruded with BK fitting's I downloaded the important bit is the DSG ballnut housing as this is your mounting point.

Prices add up quickly I've spent over £80 just on T nuts to fit the Hiwin's to the extrusion and M4, M5, M6, M8 & M12 bolts. If you use corners it becomes even more expensive.

[Heliox]

Ah ha,,,,,welding now it makes sense.

I see now, yes for sure if your including a frame there would be alot of welding, and I see that shimming out the rails with epoxy etc will give you the alignment accuracy where needed.

Lot of interesting points in your post, I'll digest that some more later.

Out of interest why are you choosing to mount the rails on the side rather than on top? is there an advantage or is it to do with working width?

Also to reduce materials; at a cursory glance if you wanted side mount rails I wonder if you would gain strength and reduce flexing movements by putting those rails on the inside of the frame?
(Though no doubt you'd probably lose some working width).