[mduckett]

Early on, prior to buying the mill, I bought an 18” length of ball screw made by Rockford Ball Screw (RBS) from Automation4Less.com so that I could assess whether I could learn how to machine the ends of the ball screws, and if my little 7x10 HF mini-lathe would be up to the task. I’ve seen mixed opinions on this, but decided to give it a try. After having a pretty hard time, and consulting my machinist friend several times, I finally was able to get a decent 10mm journal and even cut some metric threads (10 x 1.0 mm). I had to resort to grinding off the hard threads, which was a nasty, dirty job as I don't have a dust trap system on my bench grinder. Still, the RBS screws seemed to be a satisfactory choice and it was very easy to order them. Or at least it used to be. For whatever reason Automation4Less dropped the RBS screws by the time I bought the mill, so I had to decide whether to stay with those or change to Roton or some other brand. To be truthful, the machining for the one screw I had already done was so unpleasant that I wanted to minimize the number of additional pieces I needed to turn, so I contacted RBS directly and they were very easy to work with to get two additional 2’ lengths of 5/8” ball screw and the 3 square ball nuts I needed. I also requested that two of the nuts be pre-loaded with oversize balls for reduced backlash (they can’t do the oversizing without the screw, so I had to order one standard nut for the piece I already had). Note that these ball nuts appear to be nearly identical to those sold by Roton, Thomson, Nook, and McMaster-Carr. The RBS screws have the black oxide coating however. The comment I will make on the RBS screw quality is generally good, however one of the 2’ lengths was not as straight as I would have expected. I discovered this after I had already cut a piece for the Y-axis. Since the Y-axis is a short length of screw, that helped to reduced the error significantly, but it still required some straightening. Had I realized it before cutting I would have exchanged it for another piece. Given the experience I’ve had with RBS, I believe that they would have exchanged it without issue. I’ll post the end machining details in the next post.

[mduckett]

With the X,Y bearing block and mounts finalized, the ball screw journaling details were established, so I got ready for the torture of turning 2 more ball screws. In the initial attempt, the tools were mostly brazed carbide tools of the 3/8” shank BR variety. Several of these were broken as the tool dug in and made a loud bang—not pleasant. They also needed frequent re-sharpening. A triangular 3/8” shank insert tool with C2 carbide inserts quickly broke or lost their edges too. My ignorance of proper cutting angles and amateur grinding skills were likely part of the problem, however my experiences seemed to be worse than others’ I’ve seen. I looked for other inserts for the 3/8” tools, and bought some TiN (gold color) inserts, and also some TiAlN inserts (shiny black color, EM YBG202) from Shars. I followed Hoss’ example and made a split collar to center and protect the ball screws, although the 3-jaw 3” chuck is too small to pass it into the spindle so it is much shorter. However, it still worked well. Additionally, a slip fit Delrin plug was turned for the opposite end of the spindle to keep the screw from whipping. Started turning the next ball screw with some new, sharp brazed BR tools and immediately broke one. Then tried the TiN inserts and found they would hardly cut at all. About this time I found what was causing most of the problems I was having with the cutters digging in and breaking. Not adjusted properly, the carriage had a small amount of vertical play that was allowing the cutting tool to move very slightly when a heavy load was applied. The movement caused the tool to rotate deeper into the cut, causing it to bind up, chatter, and either break or stall the machine. The carriage was tightened, and the BR carbide tools began to make consistent cuts. The TiAlN inserts were tried next, and man, what a difference. These inserts are very sharp and tough and they really cut well. I couldn’t believe how easily these inserts were cutting that hard stuff. Now it only took a few minutes to cut through the case hardened layer, and these inserts leave a nice finish too (look at the bushing support journal of the x-axis --this is the raw machined finish!). The cutters do wear out fairly quickly though; the swarf is very abrasive and hard to keep clear from the cut so it tends to wear down the cutting edge. I tried to keep it brushed off, but it tends to stick in the cutting oil. A flood coolant setup to keep the swarf flowing might help the inserts to last a lot longer. I think it took 1 insert (3 points) to get through the threaded area, and one more point on a second insert to complete the roughing. At about $4 per insert, this is maybe a little pricey, but well worth it. The ball screws were roughed down to within .015” or so of the final size, then switched to a HSS tool fashioned like that recommended at 5bears.com for the finish cuts. I stopped at about .001” to .002” over size and carefully sanded them in for a tight slip fit in the bearings. On the first ball screw I used an E-style triangular insert cutter to cut threads with satisfactory results. For the last two screws I bought a thread cutting tool that uses a thread cutter insert to cut the threads. This produces better threads and cuts very cleanly.

[mduckett]

After finishing the journals, the screws were cut to length with an air driven cut off wheel. The screw was chucked in the lathe and the cut-off tool zip-tied the cross slide. Set the lathe to a slow turn rate, and feed the cutting wheel into the screw with the cross feed until it parts. I put a wet cloth beneath the cut to protect the lathe ways from the swarf. (When turning I also used those heavy blue paper shop towels secured with some small round magnets to catch the swarf during machining.) The final operation was to adopt an idea mentioned on this website (Will's Mini-mill), but mine is a removable fitting to bolt onto the end of all of the ballscrews to assist in loading and unloading the ball nuts onto their cardboard tubes. This is made from 5/8” CRS, but aluminum should work ok too. The retaining screw is a ¼”-20 button head, and the non-bearing end of each ball screw is tapped ¼”-20 about 3/4” deep. The smaller diameter section of the fitting makes a slip-fit onto the ball nut cardboard tube, making it easy to slip the ball nut on/off the screw without worrying about holding the tube against the end—one slip and the bb’s go scattering. It works great, but make sure to tighten the screw before removing a ball nut so that it doesn’t unthread while removing the ball nut—the voice of experience.

Commentary: If I did this over I would change to 12.0mm bearing journals and 12.0mm x 1.25mm threads as this would require less material removal, and nearly the same size bearings are available with 12mm ID as for the 10mm versions (e.g. 5201-2RS). I wasn’t sure when designing the bearing blocks whether a 12mm diameter would be below the case hardening layer which would complicate threading, but it seems that it is, so threading wouldn’t be a problem.

[mduckett]

Some pictures of the ball screw end fitting to help install/remove the ball nuts.

[mduckett]

As planned, I bought a replacement saddle from LMS for $65+shipping. They do a good job keeping parts in stock. Although I can’t completely justify it, at this point I also purchased another base (and another $65 spent) so that I could completely fit up the Y-axis, and by removing the two screws from the X axis bracket and sliding the table off of the mill, Icould mock up the entire X,Y assembly independently of the mill without much effort. It also allowed me to get the new saddle and base working smoothly together prior to the final machine teardown and CNC buildup. I went ahead and also bought a new set of x,y, and z gibs from LMS too. After making some measurements and determining the height of the ball screw, the usual material was milled out of the saddle to allow the ball nut and the ball screw to pass through. The measurements showed that only a small amount of material would need to be removed to allow the ball nut to be in position vertically, maybe .020”. The cast iron mills easily and so this didn’t take long.
The X and Y axis ball nut holders are now very commonly seen parts in the many X2/SX2 builds, but for completeness I’ll post my copies. These are pretty much to Hoss’ prints, slightly wider and a little beefier to fit the SX2 slots. These are made from 7075 aluminum, only because I had some scraps available. The scraps were round stock, so it took a little more machining (the radius was built-in for the top of the X-axis holder) but I needed the practice anyway. End mills were used to plunge out the holes to ¾”, then ran a Silver&Deming 7/8” drill bit, then tapped with the 15/16”-16 tap (bought this at wttool.com). I didn’t have a large enough tap wrench so I used two end wrenches, and also the spring tapping center held in a collet. All steps were done with the same setup to keep alignment, but this operation was almost beyond the vertical reach for the mill, and I had to remove the stop block that keeps from running the mill off the rack to get the extra movement needed. When installing the nuts I shimmed between the nut and the holder to get the nut to be tight in the orientation I needed, and then installed nylon tipped set screws per the drawings.
In doing the X axis ball nut fit-ups, it became apparent that the wide, shallow machined slot on the underside of the table wasn’t quite wide enough to clear the ball nut and would bind up in different spots. Since I had no way to mill it, I used a dremel with a grinding wheel to knock off the corners of the machined slot, basically putting a chamfer on at 45 deg. This worked, but the clearances were still tighter than I liked, so later on, when the mill was dis-assembled I took the table to my friend and had the slot milled a little wider and a little deeper so I could stop worrying about it.

[mduckett]

I wanted to beef up the column, as much for appearance as function so I settled on some 4” x .187 steel channel. I bought a 2’ length at a local machine shop, but since I couldn’t machine this length, I took it to my friend to face mill the sides and bottom. He did a great job, but this material was far from flat and was not easy to fixture and keep from chattering. Lunch was definitely on me for that one. The column outline was traced onto the channel and cut to match, but I needed to use a bigger machine to cut the large pivot hole. The stiffener was clamped to the column and a pattern of 1/4"-20 holes were drilled and the column tapped, and the holes in the channel opened up for clearance.
The insert in the top of the column is a chunk of aluminum cut to fit down into the column by approximate 1.5” to provide a solid mount for the Z-axis bearing block. This took a lot of effort to make, and is probably not the best way to go, however since I started it I just kept going till finished. This would definitely be re-designed if done over. The insert was clamped down tight and the holes drilled with the ¼” tap drill, then separated for the tapping. The channel and column holes were then enlarged to clearance size as necessary. This approach ensures great alignment and fit of the screws. Assembly uses ¼”-20 button head socket screws. I had ¾” length screws on hand so that’s what I used throughout.
The pictures show the insert finished while the mill was together, and then later attached to the column as part of the assembly with the stiffener. I also had the insert milled perfectly flush with the column at this point since it was made a little oversized to allow for finishing.

Edit: Note that the column insert is primarily there so that the spindle belt/pulley cover can rise above the height of the column and allow the spindle to rise to completely to the top of the column before meeting the Z bearing block. You can see that in the second picture.

[mduckett]

As mentioned in the design plans posting, I decided to go with Keling for motors and electronics. Overall their price was a little better than Probotix, but it was close. Here’s what I bought for a little under $400:

1 x KL23H286-20-8B 381 oz-in stepper dual shaft
2 x KL23H276-30-8B 282 oz-in stepper motor, dual shaft
3 x KL-4030 Stepper Driver
1 x KL-350-36 36V/9.7A power supply
KL-DB25 Breakout Board

The motors are NEMA 23 frame and I picked the dual shaft versions for two reasons: 1) it would allow manual use if desired by attaching a handle (I know, probably never happen, but us newbies need to have the option like a security blanket) 2) In the future I would like to look at adding encoders to the motors to support closed loop (servo) control. I recently was reading about the DynoMotion KFLOP controller, and it has the capability to close a servo loop around steppers that have an encoder installed.
I also started looking for motor shaft couplings and it seems that the overall consensus is to use the Oldham style. I bought one unit from McMaster-Carr and it is very nice but a little pricey ($28) so I gave Amazon a try. I found what I thought were Oldham half couplings for about $5 each so I ordered 2 pieces for each side (5/16 for ball screw side, ¼” for motor side) and some of the spacers. These are the clamp type instead of set screw. When the order came there were four complete couplings in the bag, 2 @ 5/16” on both sides and 2 @ ¼” on each side and including the spacer. The bags have part numbers from smallparts.com and apparently they were dumping these (Huco mfr.) couplers so I got some extras out of the deal, just needed to swap parts to make the coupling I needed, and all for $26 with free shipping.