Small desktop gen-purpose CNC mill on a shoestring budget

[GreenLead]

Having been bitten by the CNC bug ever since this, I'm interested in having a go at putting together a small general-purpose CNC mill.

Basically, I'm here to seek help on finding the general direction to getting started.

The mill is to have 200-300mm of X-Y travel, and 100-150mm Z travel, with 1/20000" resolution; the spindle is to be mounted on a moving gantry. I'm leaning towards building the frame out of extruded aluminium sections.

The machine is intended for cutting small aluminium or epoxy molds for injection molding of plastics, plus other things like milling gears and such like.

As I am a university student I will be working on quite a tight budget.

I have experimented with EMC2 and the Allegro UCN5804B Unipolar Stepper ICs, although the result was less than satisfactory (I may need a chopper driver)

Here are some issues I have so far:

1) As EMC2 seems to be the only free control software around, I've decided to go with it. The sim-axis setup looks sweet, but now I'm wondering how I would convert a 3D design from PTC's Pro/Desktop to the required gcode.

2) Referring to a CNC retrofit of a TAIG2108 for mold-cutting, my colleague told me he used NEMA23 276oz-in steppers, and the mill had 1/2"-20 ballscrews. What sort of controller (and motors) should I go for that will meet the above specs and not break the bank in the process?

Any help/advice/pointers would be gladly appreciated.

[Harryman]

What's your budget? No point in anyone suggesting stuff you can't buy.

[shadrag]

Well, "CNC Mill" and "shoestring budget" can not be put in the same sentence. Also, you specified 1/20000" which is 0.00005" resolution which is not something you are going to find within 1000 miles of "inexpensive".

[scrapman]

I don't think you're likely to find a machine of the desired resolution on a university student's income. Not meaning to rain on your parade, but most of your mid-range ($60K-$100K)industrial machines only have .0002 unrepeatability, and it's quite a leap to go from university student to toolmaker/die-sinker. Unless you have a solid background in machining and machine theory, you're setting yourself up for a big disappointment.

[cadmonkey]

I have to agree with the previous posters - you're asking for a LOT for very little. A stepper machine won't do that kind of resolution without extreme gear reduction which limits speed hut that wouldn't matter for the size you're looking at. As for G-Code - that's something you're best to contact the manufacturer about. EMC can handle any standard G-Code thrown at it, and has some features in it that some other packages don't - I hand code most of my own work but I don't do, or should say I haven't done any relief milling. To even come close to the accuracy you want, regardless of stepper vs servo you're looking at HIGH DOLLAR ground screws and my bet is you'd need HIGH DOLLAR glass scales for position feedback. Even at short lengths they won't be cheap. EMC can handle all of that, but the mechanics are going to be pricey.

If you do manage to prove us wrong - do tell and brag to the heavens! (and patent anything that isn't already covered and then start manufacturing these cheap high precision units and selling them to payback your school loans).

[Stepper Monkey]

If you are bitten by the CNC bug, starting out by trying to make a mill from scratch is just not the way to go. Learn to be a machinist first. Without a CNC mill already at hand to make many of the parts required, a good manual lathe, and a few other more mundane goodies of a well stocked shop like a bandsaw and the like, it is going to be a pretty lost cause anyway (at least making one that performs better and/or is cheaper than just buying one).
Even more importantly, without experience and a background actually doing machining - not just for the actual manufacture of the machine but the design parameters you need to specify - designing a machine is not going to work. It is not at all trivial, even for engineers.
Even if you were to make a machine appear out of thin air right now to your imagined design, it still likely wouldn't do what you wanted it to as you likely don't have enough knowledge base now to even know what to specify in the design.

From what I see from the link you posted, nothing even close to what you specified may even be necessary though. Even a basic used Sherline or Taig setup will handily manage that sort of work without issue, either cutting the actual pieces or the molds for making them, and those mills are at the very bottom range for as cheap and light as CNC gets and still function adequately.
Try something like one of those and get your feet wet. Then determine what you need. That may be all you ever need, maybe not. When I needed more accuracy I eventually built my own machine using a beater stock Taig that held an order of magnitude less tolerance than that machine I built from it. If you really want to tackle building a machine you'll need to collect a lot of other tools that cost more than a small mill package anyway. Get your feet wet, learn, and once you collect the whole set maybe you'll be ready for a build. Or maybe you will find you won't need to and can be happy with modding the one you have. In any case, building first is not the answer.

[sansbury]

Stepper Monkey has some very sound advice. I will throw out a slightly different idea, which is to start by building a small (~120mm X/Y/Z) router around a Dremel-type rotary tool. This will allow you to learn all about steppers, drivers, CAM, etc. without spending too much money. I started this way spending less than $300 total for absolutely everything.

The routers can be built from hardware store parts with ordinary home shop tools and a lot of the parts and all of the knowhow can grow with you into a larger machine. Also you should see if you can get access to a machine shop through your school. If you can, that opens a lot of doors in terms of being able to fabricate parts as well as hooking up with people who can help you succeed.

[tmarks11]

for the type of thing you want to make, I would second the recommendation of a cnc router. Take a look at the stuff built by widgitmaster (he may have one of his mini-routers available as an almost ready-to-go option). Doesn't look like you really need much room in the XYZ axis, and his stuff would be perfect.

My NEW Mini CNC Router Design!

Also look at some of the other build logs in the cnc router forum.

[GreenLead]

Hey guys, thanks for the response(s).

I was aiming for something with equivalent specs to this, although obviously I must've gone ahead of myself.

My budget is around NZD$600-700 (USD$450-550). Obviously, not even enough to even get a Sherline or a TAIG, or even one of these.

Had a look a widgetmaster's setup - something along the lines of what I might need, but his units are pretty pricey at USD$1050, and I'll have to get the steppers, controller and dremel myself.

It's very difficult trying to do just about anything in New Zealand, considering the lack of resources.

Yeah, I'll look again at the Sherlines and stuff. Thanks guys - will keep you posted.

[bones]

Hi,
I agree with most of the guys. Good luck with builting one for that price.
You will be lucky to but he motors and drivers with that budget.
I CNC'ed a Hafco AL-50G Lathe and it cost me about $1200 AUD for the parts.
And that was getting alot of them cheap.
I also did a small mill. $300 for the mill $120 for a new compound table, ballscrews and you've hit your budget already.
And if I didn't have a mate that had a mill, saw etc there was no way I could of built them.

Personally I wouldn't muck around with electronics, buy a board already done! And electronics is my trade!
They are cheap and work straight away.
Xylotex board and their motor package will get you started.
http://www.xylotex.com/3AxSysKit.htm
$372USD shipped to you. That's your motors and drivers done.

As for resource in NZ, you don't need them.
I hardly buy anything from Australia. Too bloody expensive.
It's cheaper to buy it from the US and pay $50 shipping than it is to buy it here.

A router machine is the go, as the guys said. Use a small 30K rpm router.
They are cheap. Use acme threads (oops I swore) as they are cheap and you can can replace he nuts cheaply when they wear.
But ballscrew are heaps better if you can get them cheap.

Built it strong. re-enforce stuff where ever you can.
Rigitity is the key for a good machine.

Best way it to buy stuff bit by bit.

And make it pay for itself. Do some wooden carvings and go sell them.
Wooden signs for the front of peoples house's.
eg. DIDYOUBRINGYOUROWNGROGALONG

You can sell then to family members and mates.
Then buy a proper mill with maybe an ballscrew kit to suit.

Easy money.

Cheers
Rodney

[darkith]

Hey guys, thanks for the response(s).

I was aiming for something with equivalent specs to this, although obviously I must've gone ahead of myself.

Heh, the Taig's resolution of .0005" is 1/2000, not 1/20,000 as you originally spec'ced. That's a good part of the confusion cleared up.

D.

[sansbury]

Look at the Mcwire design on Instructables:

http://www.instructables.com/id/Easy...lling-Machine/

This is a good basic outline for a cheap but minimally functional machine. If you haven't done this before and don't have a big pile of scrap and parts you will spend twice as much as you plan so with this design I'd give you a good chance of staying at or under $500USD