After reading hundreds of threads on CNCZone I figure that it is about time to post about my machine build. This will be my third machine build and hopefully not the last. I have plans to convert a Craftsman wood lathe to full CNC (It has nearly 200 lbs of cast iron in it's base), and to build a CNC mill completly from scratch. But first, I need to stiffen up and improve the accuracy of my CNC router so I can use it for all of the other projects I have in mind.
My workshop is in the corner of the garage. Or perhaps I should say, if I kept it clean and organized, I could get a car in there, but for now, it is taking up one of the bays in the garage, and sometimes we are lucky to have a pathway to get to the back door.
I got interested in CNC about 7 years ago when teaching telescope making classes. One of our students was the lead engineer at a local medical manufacturing plant. The plant was being shut down and he was being sent out of country to oversee the transfer of the manufacturing to the "Out Sourced" manufacturing. To make a long story short, before he moved, he said I could come over to his place and pick up all I wanted or he would just have to throw away all of his "Junk". In that junk was the start of a desktop CNC machine, and a bunch of components. I came away with the desktop components, some linear rails and other misc. parts that seemed too good to go to the land fill.
My first build was the desktop CNC machine. It had a cutting area around 15x15x6" or so and had ball screws on the x and y. I made the Z axis out of plywood, aluminum, and 3/4" round rails. To make the parts for the Z axis, I used a couple of shelf brackets, plywood and a router lift designed for a normal router table as a temporary setup. I rigged up a stepper motor to the lift screw and used that to mill out both plywood and aluminum parts for the Z axis. The y axis used 3/4" unsupported rails while the X axis was fully supported. There was a bit of flex in the unsupported rails, but I was able to cut aluminum pretty well. At one point I lost steps in the Z axis and cut 1/2" deep, 1/2" wide at about 20 ipm. I was surprised by the PC 7518 router. It bogged down a bit, but the speed control brought the speed right back up again. It plowed about an inch through the aluminum before I managed to hit the e-stop button. For controllers, I bought a linistepper kit and put together a 24v power supply. I was on a strict budget and had to use what I had on hand. But I learned a lot from that first project.
My second build is the router table shown in the first picture. I had a bunch of 1.5" x 2.5" x 0.093 wall rectangular steel tubing left over from another project laying around, so I welded up a steel table. The vertical uprights were cut from a 13" x 1.5" "C" channel. I cut it into triangles and welded the two halves back together to make a "T" shape that was about 11" deep at the bottom and 2" deep at the top of the uprights. I welded this to some 3/8" steel plate and used a second 3/8" steel plate on the other side of the table frame to bolt down the uprights. This allowed a little bit of left / right and front / back adjustment to square up the "X" axis.
The "X" axis was built with 3/8" plate as connectors to the uprights and rectangular tubing to support the profile rail. The "Y" axis uses some supported 1.5" round rails. I mounted the bearings at the center of the table. When tightened up, I could stand on the edge of the table with minimal deflection. The rails are mounted to 3/8" steel plate welded to the table frame. I bought some 3/4" x 5 TPI Acme screws for the X & Y drive screws and used some angle iron bolted to the frame to hold the bearing supports I got with the screws. The motors were mounted with carriage bolts as stand offs.
The "Z" axis is probably overbuilt. I found some 9/16" steel plate at a surplus store and proceeded to go through about a dozen sawzall blades trying to cut it to size. I ended up having to use a cutoff wheel on an angle grinder to cut through this stuff it was so hard. I have had it for a few years now and not a sign of rust or corrosion, so I think that it must have been some form of magnetic stainless steel. The rails on the Z axis are 1" round rail with 4 SPB 16 pillow block bearings. I bought a 1/2" x 10 TPI precision acme screw for the drive, and used smaller bits of the 9/16" steel to mount the drive screw bearings to. The aluminum router mount was taken from build #1. Overall, I think that my "Z" axis assembly is close to 100 lbs with near half of that being lifted by the drive screw. Even with the unsupported rails, it is very stiff, but I can move the axis up / down just by twisting the motor connector with my fingers.
My controller box is an old server case. Since I kept frying out the linistepper controllers, I mounted some drivers I got from Kelling and put together a hodge podge of power supplies, including a Linear, PC, and Laptop power supplies. (Talk about using what you've got). This gave me 5v, 12v, 38v, and 62v for everything. Then the breakout board I got had all kinds of problems so I ended up cutting the end off of a printer cable, identified and marked all of the wires and direct connected wires to the controllers as a "Temporary" fix that ended up lasting for a couple of years.
The table was made by welding some of the same rectangular tubing then bolting a 1" thick slab of HDPE to it. To get a flat surface (since neither the Welded steel table or the HDPE were flat) I used the router to surface the entire table top and mortise in some grooves where I screwed in some "T" slot guides I got from WoodCrafters to use as hold downs. Most of the time, i used a MDF sacrificial surface and clamped my work to that.
This machine worked fairly well, but had two major flaws. The table surface was constantly changing shape. I do not know if it was the HDPE surface or the welded steel frame underneath it, but I could end up with 1/8" or more difference in "Z" depth from one part of the table surface to another from month to month. The second flaw was in the welded steel table base. When the machine was in motion, the legs would flex and spring resulting in cuts being wiggly instead of straight. The end result was that I hardly ever used the machine, I was just simply frustrated with it and unhappy with the results.
The turning point was when I found a 3' x 4' optical breadboard on eBay. It was a local seller so i could pick it up instead of paying for freight. It has bolt holes on a 25mm grid over the entire surface and according to Newport Optical is flat to within .001" across the entire surface and can support a lot of weight without flexing. I think it was spec'd to flex less than .001" with 200 lbs in the middle of the table when supported out near the corners or something like that. In addition, and optical breadboard is designed to dampen vibrations between the top surface and the bottom. However, when I went to pick it up, I found out that it weighs over 260 lbs. O.K. that is good and bad. Like cast iron, weight helps to minimize vibrations in the machine, but my table base is springing around with my current table that probably weighs less than 40 lbs. It would be all over the place with a 260 lb table moving around.
So, last December, I tore apart the machine and started rebuilding it. I read through much of the epoxy granite threads and decided to stiffen up the table base by filling up the steel tubes. Since I was just filling the tubes, the EG was not going to be the only structural component, I use fiberglass resin as a binder instead of the more expensive epoxy. This worked fairly well until...
First lesson learned. If you are going to weld on a steel tube, do not fill it with concrete first. Weld it first as the heat from the welding boils the polymer concrete. I melted a hole in the rectangular tubing at one point and ended up with a blow torch blasting out of the hole where the super heated fiberglass resin out gassed and blew through the hole. I am just glad it did not explode.
But welding up all of the tubes first makes most of the tubes closed so I cannot get concrete inside any longer. So I thought about that for a bit and decided to just go ahead and weld up a bunch of triangular supports to stiffen up the frame as much as I could. I sort of designed with a cut off saw and welding rods. I would put in some angular supports, then step back and look at the frame for any weak points or points that could flex or had become weak points due to other supports I had welded in. By the time I was done it looks like I have gone totally overboard with angular supports and what not. My wife jokingly says that this is my "Leg Lamp" (from the movie "A Christmas Story" 1983) while I think it is looking like a work of art. I have finished up all of the welding, which takes me a while since I just bought a cheep harbor freight stick welder and taught myself to weld by reading a book, watching you tube instructional videos and just going out into the garage and doing it. My welds look horible. Thanks goodness for angle grinders so I can grind off some of my first, second, and third attempts before I get something that has a chance of holding the two parts together. Happily by the end of this project, I am finally able to produce some welds that could be passed off as "Presentable".
The next step will be to fill in the entire top of the table frame with concrete. While at the local Home Depot, I met the "Rapid Set" distributor stocking shelves. We had a long talk about the product and what I was going to use it for. He said that the Rapid Set concrete uses a different chemical process than portland cement in that the water is actually consumed by the setting process reducing or eliminating any water from seeping out of the concrete over the following days, week, months or even years. This means that the steel will not have such an issue with rusting from the inside out if I were to use normal concrete. The other thing I really likes is that this stuff sets up in 1 hour and is rated at over 3000 psi in one hour and over 6000 psi in 28 hours set time. They have a second product that reaches 12000 psi in 28 hours, but I went with the less expensive stuff since I am filling steel framing and am using the concrete for vibration dampening and as a solid mass to help counteract the movement of the 260 lb moving table this frame will be supporting.
Well, it is now 1am and I need to get some sleep before going to work in the morning. I will try to get some pictures of the frame and post them tomorrow.
This rebuild will probably take me a couple of months to complete along with the other chores and projects around the house, but I plan on documenting everything along the way. Especially since I just received in the mail a 55" THK C7 25mm ball screw with 4 start 25mm lead and a 43" long Kuroda C5 precision ground 15mm x 15mm lead ball screw with a 5 um accuracy along the entire screw length and 5 um backlash. I have designed a fairly simple rotating nut design I will be building with just a drill press, cross slide vice and home made jigs. I am very excited about this since those Acme drive screws I was previously using had about .015" backlash, but even worse had unpredictable travel along the entire length of the screw. I.E. .2" per rotation at one point, but .21"+ at another and .19"- at a third with anything in between along the length of the screw.
Later - John Z