[tahustvedt]

Here's my first scetch for a CNC-router constructed from MDF. I haven't bought any components yet but am watching Ebay for useable parts. I will have to change the drawing to suit whichever components I get.

I planned to use ball screws for the X and Y and a simple threaded rod for the Z axis where precision isn't that important (yet), but then I got to think about belts. I didn't find much info on belts when I searched the forum. Is belt drive a viable solution?

How are ball screws supported when they are direct-driven by stepper motors? I just can't seem to find out by searching. Are they supported on one end by two bearings and the "floating" end connected to the stepper motor, or can the stepper motor support the screw on its own and not cause backlash?

I will use it mostly for 1-2 mm sheet fibre glass and up to 6 mm plywood. Will a Dremel be able to cut that?


Thanks!

[rustamd]

the sketch looks pretty cool. The best way i found on how to find ideas for my build is go browse project log forum. I know it can take few weeks to look through those threads, but then its possible to see what has worked/not for different people. Most of the thread have very many pictures which makes it even easier to get ideas from. Good luck

[Guldberg]

MDF and ballscrews imo doesnt go hand in hand. A chain isnt stronger than its weakest link, so personally i would think you would be better off buying something else for the difference between an acme/trapezoid screw and ballscrews, a good router i.e. The same thing goes for the slider bearings, i would have a look at all the DIY concepts in here, instead of buying expensive stuff. Of cause you can reuse all the items on your next and improved router (which will come:-D ), but for a first learning go cheap (well, it isnt really cheap no matter what you do). I would suggest you started with a well proven design with plans and everything, a solsylva i.e. You learn a LOT building one of these, which you can use designing your next machine.

Just my 2 cents:-)

[tahustvedt]

Ok. I will keep reading the forum. I can't believe I haven't joined here before as I've wanted a CNC router for years.

Is it optimistic to hope for better than 0.1 mm (0.004") accuracy with Acme screws and MDF construction? I will probably feel the need to upgrade the design later like you said.

Making it myself is half the fun to me. I have access to lots of waste sheet aluminium cut-off at work so I can make a MDF/aluminium sandwich construction in some places which should be possible to make very stiff and light if I make lightening holes inthe MCF inside the sandwich. I had some 1.5 mm fibre glass parts cut out by someone with a CNC router in USA and they were nice but had 0.2 mm inaccuracies so I would like to achieve better than that if possible.

[tahustvedt]

I added alu/MDF/alu sandwich construction and some drive components. I think I should be able to lighten the MDF between the aluminium skins a lot and still end up with a super-stiff structure.

Since ballscrews are a little overkill I'm going with Acme rods instead. There are dual bearings and nuts for preloading on the end of each acme rod.

[tahustvedt]

I did lots of small changes. I made the sideplates straight because I don't need to center the spindle over the rail trucks with good rails like these I think. I'll turn the acme rods down to 6 mm on one end and use thrust bearings. I think the gantry will be super-rigid.

[Madclicker]

I think it is optimistic to hope for .004" accuracy with an MDF machine using low end hardware. I just measured a dado i cut that was programmed to be .50". It was .488". .012 is great for what I make. Also gotta split that in half because the dado was cut in 2 passes...left then right, so .006 accuracy on the right and the same on the left.??? Anyway, when I built this i thought I would be happy with .050" accuracy. I was wrong, but have tweaked my machine to perform well for my needs.

[Fer1526]

I want to build a similar machine like this, but I´m searching for cheaper an eficient parts, wath do you thinck aboth the cable system used in this machine??
http://cnc25.free.fr/planscnc25.htm
(look the palns under "Machine de Guillaume COQUERY (Version à entrainements cables)")

[tahustvedt]

Here's a simple drawing employing the table-integrated machine idea. I think I'll go this route because it's a lot less work to build.

[Fer1526]

Yes but look at the plans not for building, the french mahine uses a steel cable intead a timing belt and a small piece of trheaded rod instead a pulley, wraping the cable inside of the tread! similar to this:
http://www.cnczone.com/forums/showthread.php?t=38727
, do you think that this method could work for a cheap router!

P.S.: Excuse my very bad english!!

[tahustvedt]

It seems to work well for him. I'm not experienced enough to say how well it would work compared to screws. It's an interesting idea.

[Buzz9075]

IMO #7 the extension of your Z axis is to long and deflection will increase as this length is increased. IMO you should try and keep cutter as close to the parts that hold it to the Z axis, the more you move away the more you magnify any deflections.

[tahustvedt]

Ok, thanks. I see what you mean.

[Alex2005]

My First post.

I'm looking forward to sharing idea with fellow members of the forum.

If something is slightly in accurate is there any compensation built into the major control softwares available that may help to reduce this problem when in practice.

[tahustvedt]

Here the spindle axis is just 7 cm from the Y-axis rails. I can't get it much closer than that without putting the Z-stepper offset to one side next to the wide bearing rail, I'll try that later.

I ordered a stepper motor drive and steppers from www.bergan.se today (Xylotex).

MDF is proving to be difficult to find so I may have to find an alternative. I know plywood can warp, but this will be a box construction so I don't think it will warp if I used plywood for the construction, especially since I will use aluminium skin on some parts. It's becoming more and more an aluminium machine, mostly because I can't find MDF anywhere.

I think I will try to route the wires inside channels in the frame so that they run down to the underside of the table inside the gantry. It'll be very neat and the cable can hang in a natural U under the table near the wall to allow the gantry to move back and forth.

[Alex2005]

I had a quote from Motion control products in the UK for I think 400+ oz-inch NEMA 23 bipolar for £17.00 + Tax each. Of course for that price they are Chinese.

Has anyone experience of the quality of Chinese Stepper motors?

[tahustvedt]

Even more aluminium. :rainfro:

I was lucky enough to find a few pieces of useable extruded aluminium 100x50 U-profiles with 8 mm wall thickness. I'll use that as the basis for the gantry torsion box and the Z-axis mount. I have also tried to figure out ways to stiffen the gantry sideways should that become necessary. I though about building a triangular box structure on the outside of the sideplate and use a third rail on the top surface of the table, but since I might build a housing the the machine to reduce noise I may take a different route where I mount a rail higher up to support the dise of the gantry. Only if it's necessary though.

I'll probably end up getting a Kress 800 W spindle first. It seems to be a good machine and not too expensive.

[tahustvedt]

I have just about everything I need now. I have to make some more parts but that's not a big problem. All the linear rails I have are in perfect working order. The set I bought for the Z is actually new, the other four rails are used. The ball screw for the X has a little backlash so I'm going to try re-balling it. I have ordered another ball screw if re-balling it doesn't work as well. The Y has a C1 ball screw in great condition and the Z wil either get a Universal Thread lead screw or a 12 mm ball screw that I have.

I tested the controller and steppers with Mach 3 and it all seemed to work fine. It was fun to see things move. I'm putting together a PC with Win XP to use on the machine and will use a USB memory stick to transfer the code from my CAD/CAM workstation.

Today I spent a few hours in the workshop and I'm about halfway done with the aluminium parts for the gantry. I finally found a source for MDF which will be used as the filler in the sandwich composite construction and I will buy that some time this week.

It should start taking shape later this week.

[tahustvedt]

Here's the partially assembled Y- and Z-axis. It was a lot of work drilling and tapping all those holes. I don't have all the bolts and screws I need so it's just mocked together with a few. It fits well and slides nicely though.

[tahustvedt]

Drilled and tapped more holes and prepared the side frames for the MDF filler. If the weather is noce tomorrow I'll go buy the MDF and finish the side plate sandwiches. I screwed on one sheet for the side plates on the gantry to get a feeling for the size.