View Full Version : NEW MEMBER: Total noob about to embark on a build.

01-04-2012, 07:36 PM
Hi all.

I'm an absolute beginner with CNC, but have been that impressed and blown away by some of the home builds out there that I feel I must have a go at it. Of course, I'd like to set my sights high, but even if all I manage to cut out is a circle or a square then I'll be very happy..! My engineering experience is very limited but, in my favour, I'm reasonably good with my hands and like to do things 'properly'.

I've got a whole heap of questions (somehow answers seem to lead to more of 'em...) but I guess I have to start somewhere so here's my proposed concept (feel free to tear it apart..!):

The build will be based around 50mm square section aluminium extrusions (3mm wall).

I mostly intend to cut wood.

I'd like the cutting area to be around 450x550mm with a minimum plunge of around 60mm.

I'm fairly aware that I may well end up 'over-engineering' this, but would rather end up with a machine that's sturdy, even if it costs me a bit more. I'm budgeting on around 1-1.5k for the parts (extrusions/motors/spindle/rails/lead-screws/support bearings etc.) +electronics +software on top of that.

I'm more interested in repeatability and precision than outright speed.

Motion will be via 12mm lead screws (3mm pitch) ACME nuts (possibly pillow block ACME nuts for the Y & Z) all driven directly by steppers.

The gantry will ride on blocks/rails.

You can see what I'm thinking in the following clip:


Both the gantry and the Z-box will be made of chunky aluminium u-channel with side plates held in place by the tension from the nuts holding the guide rails in place. Again, here's a clip to illustrate what I mean:


I think it'll work, but you may well think otherwise! The reason I took this particular approach is that I can be reasonably sure that the aluminium channel will be sturdy and manufactured to a fairly tight tolerance. I also know that I can very accurately cut and mill the side plates for both channels on a friends ShopBot. This approach also saves me having to drill and tap into narrow edges!

If possible, I'd like the gantry to be driven by a single lead screw, but this would mean that the screw sits quite far down from the gantry with the ACME nut attaching to some sort of bar which in turn is bolted to the back of the gantry. Is this madness? Will the ACME nut working this far away from the load introduce prohibitive torque at the base of the gantry/blocks?

Where do I start? I guess a bench and frame would be a good place, but after that? I'm thinking that I buy the lead-screws, ACME nuts, rails and support bearings and then begin to design in greater detail... Would that be an acceptable approach?

The electronics I'll most likely leave until a later stage. I'll be using standard (easily available) stepper motors, I'd like to be able to use a bit/probe to zero the machine and the control software I'll go for will more than likely be Mach 3. Although I'm fairly good with computers, this part's a bit of a mystery to me... More research needed...

Finally, if you're in the Manchester/Stockport area and you wouldn't mind talking me through and showing me your build in exchange for a fine bottle of wine, then I'd welcome the chance to do so.

Thanks for sticking with me this far, let me know your thoughts.

Once I start I'll be sure to keep a diary.


02-04-2012, 10:36 AM
Hello and Welcome,

First let me explain that with what I'm about to say I'm not being negative or in anyway pulling your thinking apart. I will be just passing on my experience which will hopefully help guide you in a better direction regards some points.!!

First if the Aluminium is just 3mm box section then It's a bit on the thin side so you may get flex and resonance from vibration, I would strongly suggest you up the wall thickness to 6.35mm(1/4") or change material to steel which is much cheaper than thick walled aluminium and handles the vibrations better with 3mm wall.
The frame has drawn will also need some bracing on the uprights to stop sway. Don't under estimate the forces of the Y Axis moving across the Gantry they can be very high and impart much force on the frame, esp when changing direction at higher feeds. Same applies to forward and back motion of the gantry along the X axis.
This also another reason why thin walled Ali is less than ideal, Ali doesn't like to flex and bend so a thin wall will crack over time if not ridged.

Don't use Acme thread.!! With 1500 you can afford ballscrews, much much better. I would also very very much urge you to think about twin screws on the long X axis or change the design to one which suits a single screw better.?
The way you have it draw with the gantry sat directly on the bearings is the best way regards strength but not if you have long dangly sides with cross brace for the central screw and I pretty much guarantee you it will Rack or twist on the outer edges when under cutting load.!!
If you want the best machine possible regards accuracy and repeatabilty then it's a NO BRAINER you must go with twin screws.

The Z axis is THE most important piece of a CNC machine and unfortunately you have the worst possible combinations in your design.!
It use's unsupported round rail with acme thread all placed under tension. In time this tension will release so the Z axis will change shape and accuracy.
The use of round rail dictates the bearings must go onto the front plate. This then means the distance or extension from the cutter to the support of the first bearing is always the same and a constant length. Often this extension is quite long so by the time you add a spindle you have a very nice spring board so when cutting allows the cutting tool to do a "triple back salto" on the work piece giving poor inaccurate work.
If you want good quality finish then again I VERY STRONGLY URGE you to use supported rails with the rail on the front plate and bearings on the rear plate. Just have a look around the forum you will Jonathan and I often explain how and why with some example pics from me.!

The same applies to the gantry design you have a very bad combination of tension and unsupported rails.!! Drop the unsupported rails and tension isn't needed or required. You don't want tension in a CNC machine because at some point it will release then things change shape.!!
Don't be afraid of the weight of the gantry as well so trying to keep it light as possible and losing strength. Mass really helps with a CNC machine and if used with correct lead or pitch screws and motors/drive/PSU then will easily handle the weight giving a far superior finish and quality of cut.

This brings me nicely onto rails.! You say square block and by this I think you mean profiled.? If so then with only 1500 you will struggle unless buying second hand rails.? New profiled rails will eat nearly half the budget at the size your thinking to build.
For the type and size of machine your wanting then you'll be far better off using ballscrews all round and twin screws on the X axis. Too enable you to do this with 1500 then you'll most likely have to use supported round rail on all Axis because it's much cheaper. With 1500 this is very easily possible if you buy them as a package from China.
I use Chai at linearmotionbearings on Ebay and he can provide all you need in one go. Complete with ballscrew ends machined to suit which ever type of end support bearing blocks you choose which he provides as well and also make building that much easier. He also sell's cable chain etc so everything can be got in one place at very good price. You won't even buy the machined ballscrews in this country for less then he will probably provide the whole lot and that won't include the end support bearings.!!

Hope this helps and like I say I'm not "tearing apart" just passing on what I see. Look forward to seeing you progress.

Oh and the thinking to buy electrics, motors etc later on is a wise decsion. .:tup:

02-04-2012, 12:14 PM
You don't want tension in a CNC machine because at some point it will release then things change shape.!!

Misleading, you don't want tension in a CNC Router. On a CNC mill or lathe it's quite the opposite - tensioning the ballscrews is necessary if you want the best possible accuracy, however you can't do this on a CNC Router as in general the frames are nowhere near strong enough. Applying tension to the screw would just flex the frame.

+1 to everything else you said.

I'd add in the cross pieces at the front and rear of the frame. I know it's nice to have one of those missing for better access, but it will severely compromise the rigidity of the frame.

You've stated 50mm aluminium extrusion - if you replace that with 50*50*3mm steel box section and weld it the frame will be strong and cheaper. Don't be put off by welding if you've not done it before. Welding my frame was the first welding I had done and it turned out fine...

02-04-2012, 12:24 PM
Misleading, you don't want tension in a CNC Router. On a CNC mill or lathe it's quite the opposite - tensioning the ballscrews is necessary if you want the best possible accuracy, however you can't do this on a CNC Router as in general the frames are nowhere near strong enough. Applying tension to the screw would just flex the frame.

Don't start Jonathan he's building a router not a Lathe or milling machine.!!:rolleyes:

Also tensioning the ballscrew is a completley differant thing to tensioning the frame. The frame on a CNC mill or Lathe would be built in such a manner to accomadate this ballscrew tensioning but not impart tension else where on the frame so it does still apply.!!:tup:

02-04-2012, 08:24 PM
Hi Wal,

You might want to look at Florin's machine which he posted a while back. I wouldn't copy the Z axis with the unsupported rails and long Z plate overhang, but there are some other good features.


03-04-2012, 01:00 AM
Hello all.

Firstly, thanks for all of your advice - some seriously helpful pointers there!

Routercnc - you're right - Florin's machine is indeed very interesting. I've been to check out the heavy duty alu-extrusions that he's used and I'm thinking that this could be what I end up using - I take Jazz's point about opting for steel in terms of strength and weight, but this extrusion appears to be up to the job and (not least) will require a great deal less work to get this project off the ground - don't get me wrong, I don't mind putting in the work, but with my very limited fabrication experience, coupled with a lack of serious tools I could end up floundering before I've even started!

I'll have another look at everything I've been mulling over and try to put together some new sketches based on all of your advice.

Thank you again - it's all a little clearer now. No doubt a whole heap of silly questions coming soon!



21-12-2012, 01:48 AM
Moved to: http://www.mycncuk.com/forums/router-build-logs/5501-5x3-ideas-build.html#post39701