What is CNC - check out!!!

 

Why CNC?

Actually the CNC project came up as a consequence of the cockpit project and my decision to try to make "everything" my self. Put $700 in a CNC rather than in a MCP (Mode Control Panel).

In spite of having a lot of different tools, I first of all missed something to make precise cutout of panels - so I started to surf the net to try to see how others succeeded in the area and found several cockpit builders talking about CNC and a few actually owing one too. It took me 5 minutes to find www.cnczone.com and that's the place to be if you consider to build a CNC-machine. So many clever and helpful people out there (just as in our community) and after a few weeks of investigation I started to collect / buying some parts - both abroad and local. It took me about 6 months to complete the CNC to the current level. It's 100% worth the sacrifice of time and money. I believe you will get a very fast payback - just look at the price level of simparts out there. Ok - you have to spend more time on the project, but making drawings, build/produce, integrate and finally the product finished and working - that's one of the major part of the whole deal!!!

Per-Erik - March 2, 2008

---

Materials & components

So, how much money? I will say you can build a decent CNC unit for approx. $700 (NOK 4000). That implies that most are materials from Home Depot shops ( in Norway - e.g. Maxbo, ByggMakker, Clas Ohlson, BilTema) or from the junkyard/dump. With a CNC budget of approx. $700, one have to be creative.
Underneath you see a table showing a kind of a budget and where to find (in Oslo area) the stuff. It will vary from place to place and you can be lucky and find some parts which makes the project cheaper and "stronger". For instance in some old printers you can/will find steppers, strong pips/tubes & slide bearings etc. New / unused rack rails for servers/computers could be an alternative to pips. Look at cnczone and you will see/learn how different techniques ends up with almost the same result except with various precision, speed and strength. These elements are the main things to decide on up front and will tell you approx. how much money you have to spend. What are you going to route/mill - wood, plastics, aluminum, steel etc. and at what precision. In my case I believe I could mill aluminum with a low cutting speed and run several passes through the material, but mainly it will be wood/MDF and acrylic/lexan.

Parts	Supplier	Approx. price (NOK - kroner)
Table	Junkyard / dump	0
4 x Square steel rod 25x25x1000mm, wall thickness: 2mm	Norsk stål or equal and their dealers/customers.	250
Ball bearings (20 x 22mm, 6 x 32mm)	Clas Ohlson / Biltema	480
Misc. 5, 6, 8 and 12mm all threads	Clas Ohlson / Biltema	320
Misc steel fittings/mounting brackets, nuts, bolts etc.	Clas Ohlson / Biltema	200
Nuts for the thread	From the kitchen - plate for bread cutting. I don't know what kind of plastic, but it works. Wife not to happy until I replaced the plate.	0
MDF- Very rigid and easy to play with.	Junkyard / dump	0
2 x 300oz step motor and control electronics	HobbyCNC	1600
2 x 100oz step motor	Old Printronix matrix printer/plotter	0
Electronic components	Elfa	600
Router	Christmas present, but Dremel is OK	0
Router bits, end mils etc	Crown Norway or Clas Ohlson etc.	350
		3800
Software	Supplier	Approx. price (NOK - kroner)
Mach3 - control software	Mach support	800
Sheetcam - CAM software	Sheetcam	850
Corel Draw, AutoCad, TurboCad - something which can produce DXF, HPGL etc. - Corel already in house. There are some simple drawing tools for free out there.	Corel	0

Back

Design & build

A CNC at this level is not rocket science. The clue is to select and find the pieces which are good enough for the task to be done. I'm not looking for 0,05 mm precision in a steel object and then I can skip high class ball bearing/screw and linear rails. My design is straight forward - based on square steel pipes and 22mm skateboard ball-bearings

 

 

Pictures equals 1000 words - click and take a closer look. You will see an ordinary office table (1500mm x 500mm) with steel legs. At both long sides there are two steel rods - one at top and one underneath. These four pipes makes up the X-axis. The Y-axis is traveling on these pipes and consists of two verticals made of 25mm MDF. Each vertical carries 8 x 22mm ball bearings - two on top, two on bottom (8mm bolt through the MDF) and four on the side ( two and two bearings on a 8mm steel rod) - all clamping around the two X-axis square pipes like in the Corel drawing.

 

 

 

 

 

I made a few serious drawings - this is one of them, and a lot of sketches not "suitable" for the web. The most tricky part was to generally do precision drilling and routing of the slots in the vertical MDF for the 8mm steel rod. But with the correct tools, anything is possible. This phase is the absolutely most important part of the building. You have to be precise!!!. It must be 90 degrees, square and straight!!!! - ALL OF IT!. If not, you will not be satisfied with end-result.

The Y-axis (gantry) was original based on the same principle as X-axis, but I came across two very solid rails with unknown history. I got those for free from a local metal supplier. It's a aluminum profile each carrying two drill-rod "super-duper" steel. I tried to cut one of them but had to give up - not a scratch in the surface. These two rails are firmly attached to the verticals by some Home Depot 3mm steel brackets and 6mm bolts and nuts - 20 all together. The distance between the rails are locked by two predrilled flat steel plates (3mm).

 

Then we have the last - Z-axis - running the router up and down. Z is traveling along the gantry. Together with the Y-axis rails there came two steel plates each having four V-profile ball bearings - two of the four are "concentric" - I think that means the bolt in the bearing is off center. These bearings can be turned by an umbrako tool and in that way tightening (preloads) the contact with the rails - a very smart type of bearing. Make it very easy to get a solid attachment. I'm using two normal V-bearings at the top rail and two concentric bearings at the bottom rail. Ref. picture to the left.

 

The Z-"housing" is story for it self. Due to the small size and weight constrains I was looking for a suitable aluminum profile, but couldn't find anything without involving a lot of money and many hours "workout" - so why not choose MDF. This is by far the best, affordable and rigid material to form and play with. Easy to find around (dumped furniture) even if you have to buy (Maxbo). Z-axis consists of mainly two pieces - one stationary (attached with 6 x 6mm bolts to gantry) and the other the moving part. Ref. pictures which tells the whole story. The construction is fairly simple. Both pieces are made of 12mm MDF. A computer-rack rail is taken apart and attached to each side of both - all firmly bolted together with 5mm stuff. The moving part is somewhat preloaded (tightened) to have a precise and smooth travel. The spindle mount is easy - just a hole matching the spindle-nut and at the other end a MDF-bracket firmly locking the motor in place.

So, what's next. We have to get some movement - how? All axis have a "allthread" rod attached to a stepper and it's the same principle for all axis. Just to illustrate I have made a simple drawing of X-mechanics underneath the table.

Pictures from underneath the table - X-axis with the nut in place

The Y and Z are very similarly to X concerning stepper attachment, bearings, bracket etc. but Y have a different nut arrangement. I would like to try another setup, now with two 12mm steel nuts as in the pictures of gantry down under. These two nuts are pushed from each other to make contact with left and right thread wall (understand?). Check cnczone for topics concerning backlash.

An attempt to draw the Y-nut / attachment .

I will soon replace the four 6mm bolts attaching the nut arrangement to the traveling steel plates with two blocks of MDF or aluminum the get a more rigid structure.

This should be all about mechanics - very basic explained. Do you have any question, don't hesitate to ask. I'm more than willing to share my experience and of course also looking for feedback - e.g. how to do things different. It's always a better way... This is where I ended up - so far..

Back

Controller Unit & stepper

The controller is the easiest part of the CNC project if you don't decide to build it your self - which is absolutely possible. A lot of drawings/schematics is available out there, but with a price of approx. $100 for a driver-kit you can't say no!!! Check HobbyCNC and pick HobbyCNC Pro driver board kit with 4 axis.

Next would be to find a box, connectors, cable and a power supply. This is how I did it. I had an "old" 5-1/4 " disk cabinet laying around. Keeping the fan and 220 input, power switch & fuse. Removed the rest of the crap and installed the HobbyCNC board and a simple homemade power supply based on components from Elfa

Like this:

 

The final unit:

There are three connectors and cables
- one to each stepper (6 wires in each) and
one parallel port connected to the controlling PC.

 

 

There's a lot of software tuning/configuration I only would like mention:

• Steps / rotations / mm etc.
• Feed rates / accelerations / path / inside / outside / climb cutting /......
• Spindle speed / end mils / engraving tools ...

Use the cnczone and links to the software vendors. Read forums and ask - even me if you believe I could help out!!

Back

Testing, tuning & testing .....

After firmly squaring up the machine (very important) and checking wiring and power-/controller-unit, I was ready for the first test runs. In stead of the router-unit I installed a ordinary pen and could in that way do some mistakes without harming anything. The first "drawings" came out of the unit - like straight lines and circles. These kind of figures shows very clearly if there are misalignment in the construction - like not squared and if you have backlash in X & Y axis. I had to do some more adjustments, but finally things was just fine and I could advance to next step with more complicated figures....

 

 

 

 

 

 

 

After several runs and even moretuning of both software and hardware (mechanics) the real test was just around the corner....The Korry - 389 switch was my first real test - check TechTips.

Back

First upgrade

I didn't say I was finished and absolutely satisfied with the unit.

The spindle motor ( Christmas gift) is too weak for the upcoming work so a new motor was ordered from CNCplus Germany.

Kress FME 800

And what a difference. From 125W to 800W is like from 2CV to Ferrari. MDF & plastics act like butter. I have to say some new and may be better endmills were ordered at the same time.

 

But before I retired the old unit, one last job had to be done; brackets for the new unit.
The old ones helping the youth as usual.
It's not an elegant design, but strong. Made in MDF.

 

 

 

 

Second upgrade

Jan. 2009.

My Z-axis have always been the weakest point so finally I got time to do something with it. Two 32cm linear bearing slides come one day from Ebay. $44 for the set - not bad. And what a difference. The whole construction became suddenly very rigid. Now I could also increase the vertical speed from 140 to 400 steps and the milling head seems to stay rock solid. (May be it's time to try milling alu).

 

 

 

 

 

 

 

 

Still using mostly 25mm MDF but strengthen with 90 degrees brackets in all corners.

Back

Index wheel / mini-lathe

Finally got my fourth axis - an index wheel. With this "tool" I can work on the peripheral edge of a tube-like object - e.g. a knob which can actually have any contour. The main mission is to be able to rotate the object 360 - either in small steps or eg 45, 90, 180 degrees for each operation.

I will show more pictures when some samples are ready.

 

 

 

 

 

The design is based on a mini-lathe chuck "found" at Ebay - $26. I had some old 3A steppers (1,8 degrees/step) from a Printronix printer on the shelf. One of these is connected to the last driver circuit on the HobbyCNC PCB and off we go. Made 2 gears on the CNC-machine with a gearing of 20:72 (teethes) - 1 full step is 0,5 degrees on the lathe and that should be more than enough. Do I need better precision, I just add another set of gears. Besides I'm running with 1/8 steps. I have also tested the speed and with a slow start the stepper managed to give me about 700RPM - that should be approx. 200RPM at the lathe. And still, If I need more, I just swap the gears and I will se something about 2000RPM, get a sound like a Boeing 767!!!! and things starting to get dangerous. The chuck is firmly attached to a 40mm MDF block with two ball bearings - one at each end and bolted to the table, so it seems very solid but you never knows......

The only thing now missing to have a "real" mini-lathe is a tailstock to firmly secure the working piece and some tool holding mechanism.

Here is also a short video of initial testing.

If someone need drawings or any other info - just ask!

Other tools