Going full blown CNC

Finally with a makerspace of my own, I can get to one thing that has been on the back burner for a while: Mini-mill CNC conversion.
But of course, I can't just grab a ready to install, set-it-and-forget-it kit. I have to go way over the top...so I did.

I ordered a CNC conversion kit from mbbilici. He is a Turkish guy who makes his kits to order in monthly batches. I ordered on January 1st and got the kit in the second week of march. I chose his design over CNC fusion or heavy metal CNC because of the Z-axis. I like the ball screw not being under the head of the mill where it could gather chips. Ball screws are fairly unforgiving so a few chips develop a knock and knocks add up to scuffs and scuffs turn into binding and binding turns into seizure. Mbbilici's design avoids the exposure by having a driven captured  nut at the top that rotates with a pulley to draw the ball screw up and down. Price cost between the 3 kits is about the same ~$700 no matter which one you go with.

The motors were sourced from China. The ones I am using are 3 phase NEMA 23, 3.5 N-m servo motors, which are just stepper motors with encoders installed and drivers that can accept the feedback automatically. I built my own unregulated power supply because regulated or switching power supplies are known to have issues with triggering safety shutdowns. The drivers can accept up to 50 VDC.

I knew I would be rectifying an AC signal to get the DC voltage I needed for the steppers so I grabbed a 500VA 32VAC toroidal transformer to drop from the 110V outlet to something more manageable.

Based on the rule of thumb that a rectified voltage will be ~1.414*32VAC = 45VDC. Once I ran the 32 VAC through a 50 amp bridge rectifier and into a 22,000 uF capacitor, I got a reading of 48VDC which is even better than predicted. The capacitor was only ~$3 on ebay but it did take a month and a half to get here. The transformer was $55 and I had the rectifier left over.

Since I was already modifying my mill, I upgraded the spindle drive motor as well. I tried to build my own motor controller. I went through a dozen different PWM designs that included power MOSFETS and IGBTs. I cooked them, fried them, locked them, blew fuses, added zenier diodes, RC filters, inductors...everything. I broke down and bought an MC-60 for ~$60 on ebay and it works like a charm. The analog circuit used in the treadmill controllers are really the best way to handle the power these motors need.

I installed the 3 hp treadmill motor by welding a bracket out of 3/8" thick steel plate that I bolted onto the top of the mill. The drive belt is a standard 4A v-groove power belt. On spindle is a steel pulley that I bought on ebay and bored out to fit on the spindle. I tacked it down to the spacer which has a key way groove that meshes into the spindle. I took the flywheel off of the treadmill at first but I noticed that despite tripling the power I was able to bog down 1/2" end mills. It turns out that the treadmill's power rating includes the flywheel weight. I put turned the grooved section down to accept the 4A v-belt. After putting the flywheel back on it now runs like it should. Smooth and strong.

With the upgrade to the power came a change for me to upgrade my spindle bearings to tapered roller bearings. There is a lot of debate out there about which bearings work best and why (angular contact versus deep groove verses tapered...) I choose tapered rollers because they work so well in my lathe. The preload is not as sensitive or critical. Tighten it down and as long as the bearing doesn't over heat, its fine. It doesn't run as high of speeds as other configurations but if I need to run 10k-25k+ rpms, I'll add a separate high speed spindle attachment. I plan on milling aluminum and steel which is more of a 3-5k rpm range though my grease is good up to 8k.

While I had my mill apart already, I figured I would fill the frame with epoxy granite and the head with lead adding about 65 lbs to the mill and significantly more damped. The head along went from about 8 lbs to almost 30 lbs.

Also, in the pursuit of rigidity and minimizing chattering, I welded up a T out of 3/8" x 2" steel plate 2 ft long. That was held parallel to the back of the column and welded down to the 3/4" steel plate that serves as the as for the milling machine. I have two 1/2" bolts that are going to be tapped into the back of the column. I ran a FEA on this configuration and with a 500lb  side load the deflection of the column cuts down from 14 thousandths to 2 thousandths (0.014" to 0.002")