View Full Version : Desktop size build to learn on - PCB routing / small MDF work.

25-02-2014, 08:42 PM
I'm looking towards a bit of a long term project to knock up my own CNC machine as I'm fascinated by the things and it would be lovely to do away with the chemical based PCB creations I'm currently knocking up.

Initially I would like a bit of guidance on what to research as I'm not entirely sure what general design is best for learning the build / CNC basis and which lends itself to working with PCB's as the accuracy / ability to work small would be pretty important.

I've already got a couple of parts which I got cheap, namely a double internal ball, ballnut running on a 25mm diameter, 10mm pitch, ballscrew and a couple 16mm supported rails/bearings. I'm not sure on the full specs as such but for the 50 I paid it seems a very hardy bit of kit so potentially I have 2 axis's already part way there.

I'm leaning towards a moving gantry design mainly due to space as this is unfortunately a bedroom build for now but I'm very new and the more I read, the more I'm not sure what design would best suit my needs.

Do we have anyone using there setup for PCB work? Advise on designs or just good reading to give me a place to start?

26-02-2014, 08:55 AM
Many years ago I used a resist pen to draw PCB's, then came photoresist but that was always out of my reach, now we are onto cnc cutting, next it will be conductive ink printing PCB's, which probably already exists. The ball screws you have are quite meaty for the size of machine you propose but that's better than too small. If you use the forum search feature you will find quite a few references to PCB making.

26-02-2014, 09:28 AM
The ball screws you have are quite meaty for the size of machine you propose but that's better than too small.

Not really and at this small size machine it would be hard to find a ball screw too small but like in this case very easy to find one too big.!!

Problem with this large screw is the power required to accelerate it to full speed then de-accelerate will be high so the your creating your self more expense due to needing larger motors, drives etc while at same time costing machine performance due to lower acceleration times compared to the correct size screw and motor.
Yes it will work but at a cost that isn't always cheap in the long run.

Cutting PCB's or Aluminium the general machine design hisn't really any differant other than strength and spindle type with maybe some dedicated fixture for work holding very thin matarial. But the general principle of building one is just the same really so all you really need to do is research any well built machine.

The areas you need to focus on for PCB milling will be Accuracy of axis alignment in regards to bearing rails being on same plane and the resolution of the ball screws provide is sufficent and with minimal backlash as the tolerneces you'll mill to will probably be small.
Ridgidity of the frame and in particular the Z axis will be important as any flex will transfer directly to the cutter, the cutting forces will be generally low so strength hisn't much of a concern but often ridgidity and strength go hand in hand.

Spindle selection will be important for PCB work as you'll need very low runout and high speed for the tiny cutters you'll be using.

Good luck and golden rule is don't buy anything without asking first.!!

26-02-2014, 09:55 AM
Thanks guys, I'll have a bit of a search around and see what I can dig up.

I'm not totally gutted as at 50 am sure I can sell on the ballscrews I have at a profit anyway, should I never find a use for them :)

I'm thinking to aim for a working area around a4 sized, very likely the first build will be MDF as I don't have much in the way of metal tools.

What sort of size ballscrew will be most suitable? Is it as simple as the lower mass ballscrew/nut, the smaller motor or faster acceleration/deceleration can be achieved, for any given motor?

26-02-2014, 10:48 AM
I'm thinking to aim for a working area around a4 sized, very likely the first build will be MDF as I don't have much in the way of metal tools.

BIG BIG mistake don't go near MDF. Just on way out now so can't get into it now but just look around and you'll see why. I'll drop you a PM with my number as I can probably help with making something more substancial.

26-02-2014, 02:24 PM
I have an MDF router. Not a big mistake if you want something to "learn" with. Big mistake if you actually want to do some serious cutting, though! And in the kind of size you are looking at, MDF thick enough to have some kind of strength(*) will just make the whole thing too clumsy.
(*) Strength - until it gets damp/humid/tired/too hot/too cold/there's an R in the month...

26-02-2014, 03:17 PM
Well main reason for MDF was lack of metalwork tools and experience, and lots of woodworking tools to hand. Don't have much in the way of a workshop tho - I'll take a photo later as you'll all laugh at it :)

More then happy to look into a metal frame if this is the best way forward, some ballpark cost ideas would be handy as I'm guessing using metal will bump price up. I was looking to maybe start with a dremel or my Trend T4 router as the spindle (especially the Trend if I move towards wood CNC work) as the spindles seem to be a hefty part of the overall costs.

Would be greatful of any assistance with it mind, really don't know what I'll use it for long term but I imagine once I get to grips with a bit of the design software and realise what can be done I would find more and more things to experiment with :)

26-02-2014, 03:53 PM
When I started building my machine, it was with the intention that I could find out what this kind of machine could actually do (nothing beats trying it for yourself, no matter how much you read beforehand) and so that the most expensive bits would be reusable for a Mk2 if that's the way I wanted to go. To give you an idea of the scale of my machine, it has about 1200x600 cutting area, and maybe 150 Z travel. It's essentially the JGRO design (lots of info about this with a bit of googling) but extended slightly. Basic frame is a couple of sheets of MDF - 30? Guides are mild steel tubes and skate bearings - 60? More nuts and bolts than I would have expected, plus bits of angle aluminium and other small components - 50? Leadscrews are stainless studding with home-made acetal anti-backlash nuts - which are one component that actually works quite well! The nuts, that is - the leadscrew material is much too small diameter and severely limits rapid travel. But it was very cheap compared with ballscrews or even trapezoidal thread leadscrews. So, that was the cheap bit, and has to be regarded as throwaway as a better machine just wouldn't use any of these bits. Not even the nifty adjustable tube holders I made on my 3D printer!
The expensive bits are the drive electronics and motors - but these are all reusable for the Mk2. When I bought mine you could still get analogue drivers which were a bit cheaper than the digital drivers now supplied (although the digital drivers are, by all accounts, rather better than the analogue ones that I have). The nearest equivalent to what I have is the 3-axis high-voltage kit from Zapp which is about 500. I built the electronics into an old PC case, reusing its power supply, and I use it with an old PC running LinuxCNC (which is free!). Oh, there's also the spindle - 2.2KW water-cooled, which is probably the most popular option even if hopelessly over-powered for my machine, and mine cost around 350 from CNC4You (UK supplier) although there are plenty cheaper on eBay.

So, round figures - maybe 1100? Of which about 80% is going to be reused in a much better machine (I'll be going for welded steel). For a smaller machine, you could use lower-spec drives and spindle (yes, maybe a router but I wasn't too happy with any of the ones I looked at - and the water-cooled spindles are so much quieter in use, will run all day without any issues, etc). For PC milling, engraving and similar, not heavy cutting, an aluminium extrusion-based machine might make sense, but I've no idea of what this might cost.

KB Aluminium
26-02-2014, 07:33 PM
If you do decide to base the build on aluminium please drop me a pm to see if I can help.


26-02-2014, 10:18 PM
Thanks for taking the time to write those details up Neale. Afraid my budget certainly isn't spanning towards the 1000+ mark right now but as a steady project over a year or so I might be able to assign those kind of funds.

I'll have a look into the design you mentioned to see what its like. My plan was to create a mdf machine which 'works' to learn and make mistakes on so if I get to the point where I'm more serious I could justify the extra spend on something more solid.

For last 12 months I've been ripping out random printer bits with a view to reusing bits but its quickly seeming like that's a no go lol

26-02-2014, 10:50 PM
Remember that if you chase up JGRO information it's usually from a US site (such as CNCzone) and there are a number of things that the US tends to do differently. Mainly, this seems to be avoiding anything that is sourced from China in favour of good ole US of A manufacture! So, they tend to use threaded rod on smaller machines and rack-and-pinion on larger to avoid ballscrews; Gecko drivers rather than the Leadshine or similar; various guide bar systems such as the skate bearing style of thing in place of profiled rails. They tend to use routers quite a lot as well in place of the ubiquitous Chinese spindles. People seem to have built the JGRO design with little more than a cordless drill, jigsaw, and the kitchen table, however. You might also look at the Joe's 2006 design which is a much-improved version of the JGRO but really needs a CNC router to make the components. It's all a bit Catch-22 here!
I'm almost certainly in a very small minority (might even be a minority of one) on this forum by actually building and using a machine in MDF. There are very good reasons for avoiding MDF! However, you could probably halve the cost of the electronics by choosing lower-spec components, as mine are definitely OTT for the machine I have, and that might make it an accessible "starter" machine. I'm sure that you would learn much more than going out and buying one of the Chinese aluminium-framed machines, even if they are probably more capable.
Just as an indication, I've profiled and engraved an oval 400x300 hardwood plaque for a presentation, engraved reasonably fine detail in a small jewellery box lid, profiled 30mm thick curved sapele rails for a new hatch for my boat, and done 3D carving, all on my machine. I'm currently playing with an idea taken from the 'net for a cable chain cut from interlocking pieces of MDF, to be cut on the router. I haven't tried making PCBs on it yet, but I think I probably could, and shall probably have a go at it some time. Toy? Absolutely, compared with the machines that most people here have built, but not too expensive, not too difficult to build, and great fun into the bargain. Needs a biggish desktop, though - my machine just about covers a 6'x3' bench.