CNC milling machine assembly

Hello,

with all the quarantine stuff going on right now I was not able to mill the rest of the parts on my university's milling machine. Thankfully my friend is a cnc operator, milling mostly in plastics and plywood. I asked him to mill a few plywood plates for me so I can continue the assembly. This way I could learn operating the machine on plywood parts and then replace them with aluminium sheets.

I like to show a lot of photos so here you go:

Plywood parts 

Assembly of the machine under a watchful eye of the cat-supervisor

assembled frame without Z axis

For not the milling machine is located on my desk

Currently the machine is partially assembled and is able to mill. I'm still not sure about the housing and protection for linear rails and ball screws. During the build I used the dial indicator with a resolution of 0.01mm and I was able to set the rails paralel to each other with a +-0.02mm error. I find it a quite resonable error considering that the rails were mounted on regular aluminium profiles. All the measurements were made relative to these big 90x45mm aluminum profiles side planes, so this could introduce some error too. I had to put a a couple of paper sheets under one side of Y axis linear rail, as it was about 0.1 mm lower than the other rail. For now it seems to work, but I will inspect the heights after some time as the paper can be compressed to some extent.When I checked the table height the indicator showed about 0.05mm difference in X axis and about 0.03 in Y axis. That's quite acceptable, especially when the table is made of plywood.

The biggest issue is the Z axis which has about +/- 0.15 mm of play in Y direction. However I kind of saw that coming and ordered additional two 300mm long 20mm wide linear rails. These are going to be mounted on a new plywood plates and only when it works I will mill the final parts from aluminium sheet.

Milling the first piece of laminate was very satisfying. After taking it off the table my calipers showed that the hole is perfectly 10mm, while the contour was about 0.07 off. For a first run it was fine, but when I mount new Z axis I expect to achieve better accuracy :). Milling in aluminum wasn't very demanding, but the dimensions were off by 0.15mm. It turned out that the machine was settling down and all the screw connections got a bit loose. After screwing them once again the dimensions are off only by 0.05mm in Y axis (due to Z axis play) and about 0.02 - 0.01mm in X axis. Hopefully I will get down to 0.02 in both axes when I mount new Z rails.

So far I've tried milling laminate, PA9 aluminum and a soft unknown aluminum. I'm really happy with the PCBs. Compared to the traditional thermo or photo methods this one is the fastest and most predictable. Even two sided PCBs are easy :). After getting the machine to work I milled a LED lamp for the spindle:

Milled traces and outline

Assembled lamp

Working setup (actually it could be mounted on the spindle itself, because when the spindle is too low it causes a shadow to appear just in the place where the milling bit is mounted) 

assembled double sided PCB for USB<->CAN converter

 I think that's all I wanted to show for now. Next I'm going to try milling some plastics and try a bit more aggressive aluminum milling.

CNC milling machine project

Hi, 

Recently I got interested in diy cnc mills. We have a mill at my university, but usually I have to wait a few days for the operator to become available, as he has a lot of work to do. So basically this slows everything down. As I design only small parts for the brushless controllers I decded to make a small "desktop" cnc mill. The only assumptions I made at the beginning were: low overall cost and ability to mill in aluminium. 

With that being said I started to search for different designs of cnc mills in order to pick one as a reference. I stumbled upon many 3d printed mills (which I automatically rejected due to the limited stiffnes of the parts and overall durability) and very professional cnc mills that cost a fortune. I wanted something  in between. That's why I contacted my friend from electrical engineering departament, who had already built one in the past. His design was rather simple but at the same time it ws very stiff. It was a combination of aluminium 40x40 profiles and mostly 15mm thick aluminum plates. I decided to make a silmilar machine, with a bit larger working area. 

While planning I came across an auction where I could buy a used 2.2 kw hy02d223 VFD for like 50$. I contacted the guy and eventually I bought a spindle mount, VFD and supported shafts for the z axis (is that the proffesional term?) with bearings for about 80$. That was the starting point. 

Fig. 1 Z-axis, 80mm spindle mount, and the VFD

Since then I made a few attempts to prepare a 3d model of the mill. After many different approaches and ideas I eventually decided to make a structure similar to my friend's mill. It consists of a lower frame (y axis) made of aluminium plates and profiles between them. The upper part is mounted on two 15mm thick pieces of aluminium plates. Two profiles perpendicular to the plate make the x axis. X and Y axis are equipped with linear rails (Chinese hywin), and the Z axis moves on supported shafts which I bought. 

Fig. 2 3d model of the milling machine

X and Y axis are driven with 1605 ball screws, whereas the Z axis is equipped with 1204 ball screw. I plan to place a 65N furniture gas spring (the one mounted in the furniture fronts, to make it easy to lift), so the screw will not be loaded a much. 

I started the construction from buying the 2.2kW water-colled spindle, dm542 motor drives,  nema 23 3Nm motors, motor power supply, ball screws with bearing blocks and linear rails, a PC and LPT port control PCB. Aluminium profiles and plates come from a company that disassembles machines and other pieces of equipment and sells the parts. They were much cheaper than new profiles and aluminum sheets, so I decided to purchase them. There was a few additional things as cables, screws, induction sensors, lightning etc. Right now the overall cost is 4000PLN (currently about 1000$). I do not think it will change much, as I have the majority of the parts needed. 

The assembly started from milling the 15mm aluminium plates and cutting the profiles to 670mm. After cutting the profiles with a saw, they were milled in order to have the exact same length and perpendicular surface. 

Fig. 3 Lower frame after milling

Having done some online research I found that people use to fill the profiles with epoxy and sand, concrete, or some different filling. It is supposed to make it heavier and damp the vibrations to some extent. The epoxy option was quite expensive so I decided to fill the profiles with special concrete for leveling the floors. The producer claims it has low linear contraction and easily fils the empty spaces. With a little help from my university club's colleges we managed to fill the frame. We put some water in the holes and slowly poured the concrete while shaking the whole frame. If was quite a mess, but eventually the frame was filled. I struggled with air bubbles pushing the concrete from the holes for some time, but breaking them with a piece of wire seemed to help. 

Fig.4 Concrete filled frame

The only thing that I realized after filling the profiles with fluid concrete is that I could fill them with loose concrete and tamp it inside so that it stays solid. Maybe this way I could manage to put more concrete in the same volume, and the damping properties might be enhanced. I will try it with the  profiles for the x axis.

That is all for now, 

see you ;)