[mduckett]

Hi, hope there's room for yet another SX2 CNC conversion thread. I bought the LMS Hi Torque 3900 mini mill over a year ago and am very happy with the machine. I always wanted/planned to convert it to CNC, but with no machining skills (at least not metal, lots of woodworking) it took a long time to learn the basics and get enough tooling in place to do anything serious. After reading and learning from the many build threads here, I also need to thank Hoss, DJBird, and many others on this forum--without their excellent work I wouldn't have known where to start. Anyway, this is going to be a sort of "catch up" thread as I am pretty far along on the build. I didn't do very well on documenting while I was doing the build, so I've been doing some catch-up, taking a lot of pictures and will cover what I've done to date, and then to completion of the project. The picture is of my mill, when it was brand new and clean, mounted on it's dedicated bench/flood table.

[mduckett]

The first thing I did after getting the mill (and doing some playing around of course!) was to build a solid bench. Probably jumped the gun but I decided to build a flood table from the start. I almost didn't allow enough room side-to-side to allow for the X-axis stepper, but it looks like it will work out. The finished product will have a angle aluminum uprights and 24" plexiglas sides. The bench is pretty sturdy, with a 1.5" plywood top (two layers of 3/4" plywood) on 2x4 and 4x4 frame. The outside was eventually skinned with 1/2" plywood. The pedestal for the machine is placed offset to the left to allow for the x-stepper motor. Once the pedestal was in place (made of several layers of 3/4" ply), I located and drilled the drain holes and added the stringers and masonite panels to make the sloped surfaces. The top was then covered with overlapping flashing aluminum to provide a sturdy surface. The aluminum was secured with contact cement, and clear silicone calk along the seams. I have yet to run a drop of coolant even though the pump and lines have been run, but hopefully soon! The attached show some of the table construction details and the nearly finished product. I later added the remaining sheet and some aluminum angle around the border to form the bottom for the eventual side walls.

[mduckett]

Ok, with the machine mounted on a suitable work bench, the next thing was the CNC development plan. A primary goal was to do as much of the mill work on this machine for the conversion as possible. Keeping that in mind, here is the rough CNC conversion design plan:

1) Use the proven X,Y ball screw conversion demonstrated by Hoss’s X2 and many others subsequently. Since the saddle needed to be modified to fit the x-axis ball nut, this would require a spare part to be purchased in order to do that machining on this mill.

2) For the Z-Axis, I went back and forth many times trading off the CNC fusion side-mount vs. a more centrally located lift point design for the Z-axis ball screw. I’ve seen a couple of centered designs; one uses a small diameter ACME or ball screw with the nut mounted inside the carriage approximately where the pinion drive shaft normally passes through. The other popular centered design is the StirlingSteele and Hoss rotating driven ball nut that secures the screw to the top of the carriage and uses a combined drive pulley, ball nut and bearing to move the carriage. The centered design finally won out as I really prefer the look and mechanical balance it seems to afford. This isn’t a knock on the cncfusion kits, PLENTY of those out there working fine, just my personal choice. From the two center mount choices, I decided to attempt the embedded ball nut approach, figuring that if I failed I could always fall back to the other proven design. Either of these designs is complicated by the far back positioning of the spindle motor near the column. In order to modify and make parts for the carriage, I would also need a spare carriage part.

3) Identify the basic mechanical components such as ball screws and nuts,
belts, pulleys, motors, and shaft couplers and sources for each.

4) Develop some CAD 3D models to assist in designing the mechanical parts and selecting specifics of the mechanical parts such as bearings, etc. for the X,Y and especially the Z axes. Develop shop prints (well, not really, but good enough for the garage shop) to make the parts.

5) Identify the necessary tooling. After I bought the mill, I had a pretty strict budget for tools so I carefully selected a number of key components to stay within budget but still give good functionality. The essentials were: 3” vise, R8 collet set, end mill set, 115 piece drill bits, boring head and bars, 6” caliper, 1” micrometer, 1” dial indicator, .030” test indicator and holder, and drill chuck as the LMS mill doesn’t come with one. Other tooling needed would be purchased along the way to spread out the costs.

6) Electronics -- Identify motors and controllers, breakout board(s) and a computer to host Mach3. After searching and comparing prices, I narrowed the suppliers for these components down to either Keling or Probotix. Both have good offerings, but in the end I like the bipolar drivers and hybrid motors from Keling a little bit better. I still plan to buy a relay board to control flood coolant pump from Probotix, they have some nice options. I have an older but capable laptop with a docking station with a parallel port so that will be the Mach3 host.

7) Additional Considerations—Most people with the tilting head mills have added some sort of column support to add some rigidity to the system. My plan added a length of 4” steel C-channel to the back of the column to boost its cross section, and also to add some triangular gusset braces between the column and rear base castings. Another area for consideration is the completion of the enclosure, the access openings, placement of the electronics controller, and the wiring scheme, including limit switches.

So that's the plan, much of this has been completed so I'll start posting the design and build details next.

[mduckett]

I decided to tackle the hardest part of the build first, namely the Z-axis. To get started with the effort I purchased the replacement part for the spindle carriage, i.e. the part that has the dovetails and is bolted to the spindle box. Fortunately LMS stocked this part for $65+shipping. Wished it were less, but I couldn’t proceed without it so ordered it and had it in a few days. The first order of business was to clean up the inner casting outlines at the back side opening (dovetail side) so that an insert would have straight edges to fit into. Also worked in from the wider opening (front) side to clean up the shoulder features and basically flatten and straighten up things a bit. Because the casting inner walls aren't parallel to the outside, no attempt was made to do anything with the sides, just the bottom areas where an insert block would be made to fit down against those surfaces. Once the part had at least some flat and parallel areas to work with, careful measurements of the features were made a 3D CAD model was developed. The column was also modeled and by placing the two models together the available spacing and design of the ball nut carrier were much easier to visualize. A 3D model for the ball nut was located on the RBS website, and that allowed orienting and positioning the ball nut for the best fit in the available space. Since there is a 16mm hole in one side for the rack shaft, that was convenient to use as the method to hold the insert. The hole on the opposite side was to be enlarged to match, and a pair of 16mm shoulder bolts would serve as the main fasteners for the ball nut holder. The smaller hole located forward and down would be re-drilled and copied on the other side to provide a second attachment point using some ¼”-20 button head screws.
Using the 3D CAD features, the amount of space available for a ball screw could be accurately estimated, and a model of a shaft opening was developed into the part. The attached rendering gives an idea of what the model provides. I didn't do a complete model, only as much as was needed to determine if it was going to work, and what would need to be done to make it work. Playing with the model, it was clear that while there is almost enough room for a 5/8" ball screw to pass behind the spindle motor, the bearing block to support it would need to be located at least the height of the motor (6 - 8") above the top of the column in order to allow the motor to not interfere with the bearing at the top of its travel. That wasn't an attractive solution, so the unavoidable solution was to move the motor forward. This turned to be simpler than expected, as the modification only required about 1/2" of movement to get the needed clearance.

The attached pictures are of the last model updates including the Z bearing block, which will be covered in detail later. If you look closely at the wireframe you can just see the ball return tube of the ball nut model.

[mduckett]

These pictures show some of the final work being done on the spindle carriage to cut the ball screw pathway through the carriage and the baall nut insert at the same time using a 3/4" ball end mill. The carriage was aligned to the x-axis of the mill using a dial indicator along the dovetail. The aluminum insert/ball nut carrier can be seen fitted into the middle. I'll show more details for the ball nut carrier in the next post.

[mduckett]

These pictures show the completed ball nut carrier insert and the ball nut fitted into the carrier. To make that carrier was a lot of work. The first part involved fitting a chunk of aluminum into the carrier housing. I began by marking out the areas and depths that would need to be removed to clear the shoulders in the bottom of the carriage. These were just approximated and then many, many test fits and removing of material until it fit down into the opening and flush with the back surface. Once that was complete, the shoulder bolt holes were drilled and tapped to 10mm x 1.0mm. The insert was installed into the carriage with the shoulder bolts (giving an appropriate Frankenstein look), to get ready for the ball end mill cuts to make the ball screw shaft as shown in the previous post. When this was completed, the cut out for the ball nut was made with the same setup. Here a 1” 5-flute end mill was used to make a 1” channel centered on the ball screw channel, and 1” deep, with length equal to the length of the ball nut. Then the insert was removed and rotated upright so that the 1” cutter could cut down from the top side into the channel to square up the end of the channel. When completed, this left the inner corners of the ball nut channel with fillets in the bottom corners that needed to be removed. I used a smaller end mill to remove more of the material, and in the end just scraped and chiseled out the remaining bits of material. A sharp 5/16” HSS lathe tool bit shaped roughly like a chisel was used to do some of the work. Finally, since the ball return tube faces downward, an end mill slightly wider than the ball tube was used to make a channel for that feature, and also made some plunges to accommodate the screws and the metal clamp. It just took a lot of patience but finally got it done. By this time I was wishing for a CNC machine to make this part! If I did it again, I would simply drill 1/8” holes centered in the corners for the ball nut to eliminate the need to make sharp corners.

[BAMCNC.COM]

Great build! Great Pictures and I love the bench!

[mduckett]

Great build! Great Pictures and I love the bench!

Thanks! Lots more to come.

[mduckett]

By this time I wanted to get started on the X and Y axis ball screw details, but I couldn’t start working on those until the X,Y bearing details were figured out. I spent a lot of time trying to decide on the axis bearings, and had even bought some of the thrust type bearings thinking at one point that I had it settled. Then I saw where a lot of setups were using angular contact pairs. In the end, I got realistic, stopped over thinking it, and settled on using double row angular contact bearings. These have the additional advantage of being sealed units too. I bought a bearing from vxb.com to measure for use in the CAD design. The unit is a 5200-2RS (10mm x 30mm x 14.3mm) double row angular contact bearing. Although not rated as high for load carrying capacity as a pair of ACs, these seemed to be adequate in my estimation for this size machine and the forces it will generate on these axes. With that finally decided, the designs for the bearings were pretty easy. The Y bearing block assembly is basically a modified Hoss design with a spacer and the bearing block. The X bearing block assembly consists of a new table end piece with the bearing block attached. The bearing blocks consist of two pieces: a bearing holder and a lid. Both the holder and lid are designed with recesses for shaft oil seals, although since the bearings are sealed these will probably not be used. This is a typical design that you will see described at 5bears.com and other places. This design calls for the bearing recess to be a couple thousands shy of the true bearing width to allow the lid to apply some load to hold it in place. Shims can always be used if the machining doesn’t accomplish this so no worries there. The motor mounts were inspired by djbird’s design that uses two flat plates for the motor mount. This gives a nice appearance of a one piece design and also provides some motor heat sink as djbird mentions. The attached are some of the CAD drawings that show the general design and a rendering of the Y axis. I couldn’t find the X axis rendering but you get the idea. In the next section I’ll cover how these were made.

[TroyO]

Nice work! I'll be reading along..... superb so far!

[mduckett]

Nice work! I'll be reading along..... superb so far!

Thanks TroyO!

[luv2ride]

Nice write up and great looking parts. Keep up the quality of work and documentation.

[mduckett]

Nice write up and great looking parts. Keep up the quality of work and documentation.

Thanks!

[mduckett]

The Z axis bearing block evolved into a combined bearing block and motor mount as a result of playing with the 3D CAD models. I wanted to stay consistent with the X, Y bearings and use the 30mm OD dual row angular contact bearings for the Z axis also, so that helped define the size of the front portion. I also wanted a clean look, which meant that the block would need to match the column width and extend at least to cover the column stiffener. Once extended, it was easy to see the natural place for the Z stepper motor was to fit down inside the C-channel. With a little more modeling the distances between the motor shaft and the ball screw were known, so timing belt pulley and belts could be figured out. The first place to look at was Econobelt, but they don’t carry the GT-series profiles that are specifically designed to reduce backlash, so the pulleys and belts were purchased from B&B Manufacturing (bbman.com). Their online selection tool allowed specifying the pulleys and shaft center distances, and the tool gives the correct belt length, which was later verified perfectly. They also had 3D models for the pulleys, so that helped to improve the model. I also found a model for the bearing, so more fidelity was added. The final design is probably overkill, but I wanted to have two bearings to ensure stiffness in the belt drive, especially with the relatively small 10mm journal size. A modular approach was also desired to eliminate the need to machine a large blind pocket for the bearings. Therefore, the final design consists of a top, middle, and bottom layer. The middle is the bearing holder and basically holds everything together, and is two bearings widths thick. The bearing holder is a through-hole, so at least it is straightforward to make. The top piece retains the bearing outer race and at the back has the bolt pattern slots and motor shaft cutout for the Z axis stepper motor. The motor is mounted to the underside of the top layer and hangs down through a cut-out made in the middle and bottom layers. The bottom layer also retains the bearing outer race and forms the base that interfaces with the column top insert. Three @ 3/8” bolts are used to secure the assembly to the column top insert.

[mduckett]

The construction of this part is unremarkable except that it consists of large hunks of 4” wide aluminum that needed to be faced. I ended up buying a 4” vise, which looks ridiculous on a mini-mill and requires a mounting plate to lift it up to clear the rear column bracket, but it works great. I used 1 ¼” 6061 aluminum for the middle, ½” for the top, and 5/8” for the bottom. All were faced with a 1 ½” R8 carbide insert end mill, which passes for a face mill on this machine . The 30mm OD hole was plunged in the middle layer piece with end mills to 1”, and then I got my first experience with the boring bar. Except for needing to sharpen the boring bar, this went well by taking light cuts and checking often. I got really lucky though, as my cheap boring head has a lot of slop in the adjusting screw and near the end instead of taking off .002” as intended I got about .008” and somehow the hole ended up about .0015 oversize—perfect fit! Pure luck but I’ll take it. The rest of the parts had their features milled and the unit was assembled. A .020" shim was placed between the two bearings to relieve inner races, and another .002” shim was needed to make the bearings tight against the top/bottom pieces. The shims were bought from McMaster Carr. With the top layer removed and the middle/bottom firmly bolted together, the motor cutout was made. The sides of the cutout were band sawed and the end was then cut free using a ½” end mill. It took a while to get through almost 2” thick. Once separated, then the band sawed edges were cleaned up. With the functional machining complete, the front corners were band sawed with a radius and then smoothed on the belt sander.

Edit: I try not to revise history, and that means recording the screw ups that happen no matter how dumb. On the top side of the block (fourth picture) that recessed cutout is not supposed to be there--I got carried away and marked it out on the top side and then machined it out. It adds a little clearance for the pulley, but it's not supposed to be there--easy mistake to be on guard for--practice for the bottom side .

[BAMCNC.COM]

Great documentation of this!

[mduckett]

Great documentation of this!

Thanks for reading. I know some of the posts are painfully long, but hopefully they are useful at some level. Thanks again!

[mduckett]

After completing the Z bearing block and the column insert and stiffener work, the next thing to do was to drill and tap the three 3/8”-16 holes into the column insert to hold the Z bearing block in position. I had the locations of the holes from the CAD models, but I wasn’t comfortable just making the holes without doing a fit check, so the carriage assembly including ball nut and ball screw were slid on to the column and the gib adjusted. The ball screw was inserted into the Z bearing block and the 10mm jam nuts tightened against the bearing. The Z bearing block was then pushed down against the column top and clamped snugly with a long pipe clamp that reached to the bottom of the column. (The column was resting on it's back with the bearing block hanging off the edge of the table). The bearing block was snug but still able to be moved with some effort. With the right side of the bearing block aligned with the side of the column, the ball screw was dead straight down the channel in the column (this verified the CAD model and the location of the bearing center w.r.t. the ball nut center and axis for the side-to-side location). Next the bearing was moved up and down to adjust the ball screw parallel to the front of the column. With the carriage at the top of the travel I put a 1-2-3 block next to the ball screw and made a reference mark at the contact point, then checked the screw at the bottom end and adjusted the bearing position until it was the same in both places. The clamp was snugged and the ball screw could be turned by hand to move the carriage. It seemed to move smoothly, but it was taking too long so I chucked the end into my cordless drill and slowly ran the carriage to the bottom and then back to the top. The drill torque seemed pretty consistent so I called it good and used a transfer punch to mark the hole locations in the column insert, which was then removed, drilled and tapped for the three 3/8”-16 holes. With the insert re-installed in the column, the Z bearing block was then bolted on, but this is where I found that the bolts were too long by about 3/8”. I had tapped the holes to the limit of the tap, but they weren’t deep enough. So order new bolts, or cut these down? It took five minutes to cut the bolts down and then the Z bearing block was firmly attached to the column. One tip here—I used the minimum tap drill size because I wanted these to be very tight threads, and they are, but with aluminum there is the possibility of galling or seizing, so I used a little anti-size grease on these bolts. It’s nasty stuff, so use a small amount, but at least the bolts won’t get stuck. The attached show the Z bearing block and column with the belt drive and motor attached also for a fit check. Later, I had the left side of the bearing block machined to the exact width of the column as it was about .060" over size.
Edit: When I finally got around to trying the spindle with the motor attached on the column, there was interference between the Z bearing block and the motor. This wasn't a problem to correct as there was plenty of margin at the front of the bearing block. The CAD showed only about 1/8", but I made it about 5/16", no need to make it too close. That's why the front profile of the block is flatter than shown in the CAD drawings.

[mduckett]

With the work major work completed on the painted parts of the mill, I had to work in some time to re-paint the parts. The replacement spindle carriage was in that wonderful HF red and the new base came in with the cream color instead of the LMS blue. I liked the light/dark theme that LMS puts on their machines so I went to look for some similar colors at Lowes. They really didn’t have anything that close, so it was either mix and spray, or just buy spray off the shelf. I didn’t feel like a painting adventure so I picked out a cream color and a metal flake blue in Rustoleum. I have no idea if this paint will stay on over time, but it can’t be that inferior to the stock paint. The parts that needed paint were the spindle carriage and spindle box, base, round column support, non-crank table end, and the spindle belt/pulley cover.
All painted areas got lightly sanded, then everything got cleaned with Simple Green and rinsed well, and dried with paper towels. When dry the machined surfaces and any place that paint didn’t belong were masked off with blue painter’s tape. The tapped holes got rolled paper towel screwed in and sliced off flush with a razor blade. Some 2” wide blue tape really helped out with the larger areas on the base, and the rest used ¾” tape and some masking paper. I didn’t paint the top of the spindle box because it was very rough and really need some major work. It’s hidden anyway…. I used gray primer on the carriage because I was painting cream over the HF red, but elsewhere I just painted over the existing paint. Everything got two coats, and I let it cure for several days as it was pretty cold and my garage is only warm when I’m heating and I wanted it to be hard before touching it.
The new steel motor plate also needed some treatment but I didn’t want to paint it. The original came with what looks like black oxide coating. I remembered that I had bought some gun blue for a project so I tried it. I didn’t get anything that looks blue, but the resulting grayish tone actually looks ok and you only see the edges anyway. The attached are the results. I already scuffed the spindle front a little but oh well, it looks a lot better than before, especially getting rid of that HF red. At some point I might need to paint the electronics box, we’ll see how bad the mismatch is. May need to do something with the column stiffener too, but I’m done painting for now.

[mduckett]

I didn’t want to spend much on an enclosure to house the power supply, breakout board, and stepper drivers. That’s pretty hard to do, but I found a barebones PC case at geeks.com for $13 with $12 shipping—sold! In hindsight, I should have just gone to the local landfill where they have a dropoff for electronics, and there are always tons of junk computers in there (I’ve put two in there over the years) and it would have been better than what I bought and free. The case arrived and I think the cardboard box was stronger than the case. It is really flimsy, but I had to add some structure for mounting the various hardware anyway so it worked out ok, and at least it looks good. I saw a guy on YouTube (CNC G0704 Part 2 - Assembling the Control Box - YouTube) doing almost the exact same thing I was planning so I copied his idea of attaching Plexiglas to the case bottom as a mounting surface, and I also installed an inside layer for the whole back panel. I had also purchased the same 4-pin XLR connectors shown in the video for the x,y,z drive and limit switch connections. The connectors came from UGRA CNC, but later found they can be had for about $1.50 each or less on ebay straight from China with free shipping (search for "aviation connector" on ebay). For wiring, I had access to a lot of scrap CAT5 bury-ready cable and it looked good, with black, tough cover, but alas it turned out to have solid wires. I made up some cables with it before that fact sunk in, but after reading some other posts on cabling I finally realized that with flexing/vibration solid wire is very prone to breaking over time. Also, the ampacity of the 24 AWG wires, even paired up as I was planning, is very nominal at the upper end of my setup (3 Amps/phase). So after a lot of searching I found some 4 conductor 18 AWG stranded wire with a tough black EPDM sheath for $0.43/ft at wesbell.com and I ordered 40 feet. The jacket is called SJOOW, which means oil, solvent and water resistant, and its extra flexible—seems perfect for the CNC environment. This is not shielded cable, so we’ll see if that becomes a problem. The connectors strain relief clamp will be tight, but should be ok. If too tight I will cut it back, heat shrink the clamp area, and then heat shrink the cable and back part of the connector with large tube to seal and stiffen it up.
The build up in the box is pretty straight forward, just install the Plexiglas bottom and back pieces, then layout the locations for the components. I drilled most holes in place, marking with a Sharpie. While the Plexiglas is convenient, it can be trouble to drill, especially larger holes. I used a 5/8” Forstner bit to drill the connector holes. Use some oil and compressed air to cool it (or just blow on it like I did) and clean the bit between cuts as the plastic will melt onto it. I used a hole saw for the large hole for the power strip, make sure to drill half way from both sides for a clean cut. The breakout board slot through the Plexiglas in the back was a real pain—I drilled holes in the ends and used a coping saw. If you have a scroll saw this would be a simple cut. I also had a large round hole too big for a hole saw to cut out for the fan, also done with the coping saw. Making straight cuts for length in Plexiglas is easy, just score and support along a sharp table edge with a piece of wood or a straight edge and snap off smartly. You can also cut it with a saw. The pictures are the box with the CAT5 wiring, soon to be changed out as I got the new cable a few days ago. The power supply is mounted on two painted wood rails in the bottom using some aluminum angle and the wood is screwed down from outside into the bottom with black 1” drywall screws. Went to Best Buy and there was an open box LED case fan for $3, so why not?, it fit perfectly and jazzes it up a little. To accommodate the the various items that need AC power, I decided to install a power strip onto the top-side panel and place it so it can be operated by reaching a finger in through the same hole that its cord comes out. This way I only need one AC hookup to the outside. I built a small 5 Volt regulator to supply isolated 5V power to the CNC breakout and drivers and plan to have it share power with the fan's 12 volt wall wart, which is also plugged into the power strip.