Hi All, I will start this by stating that I do not know much about CNC routers. I've been watching a bunch of youtube videos and reading threads here and other sites and have decided to build a CNC roughly 1mx1m. Without a lot of planning I dove in and purchased 1m C channel from openbuilds and some linear rail from ebay as well as some 8x11" 1/2" thick aluminum plates an various connectors and screws. These items have come in and I've done a little assembly and mocked up a design and would like for folks who know a lot more about these machines to poke holes in my design so I can improve it. This will be a one-off build for me and I fully expect to tinker with it for quite some time. My goals: 1. 1m x 1m - I'm assuming it will equate to @32"x32" cut area 2. use a standard off the shelf router for a spindle 3. repurpose an old laptop and load linux CNC on it 4. nema 23 motors 5. Use as many off the shelf parts and products as possible 6. Don't customize expensive parts so as I learn more about this area I could re-use as many components as possible if I wanted to take the design in a different direction. here's what I've got so far 1. purchased (3) 1m sections of C channel and have installed linear rail on both sides (I don't want to cut these pieces) 2. (3) 1m pieces of 8020 for cross pieces 3. I figured out the bolt hole pattern for the tops of the MGH15CA bearings and made small piece of 1/4" acrylic 4. I've arranged this in a mock up of the design of the side gantry https://photos.app.goo.gl/6Yg7HeuPXJZRvJoW9 https://photos.app.goo.gl/MkhpP1ubw7sMf39H8 The concept being the C channel becomes a U and the rails facing upwards, a piece of 1/2" thick aluminum will go on top of the bearings and extend outwards @2" to provide a place to attach the anti backlash piece. The ball screws will be installed outwards on the small side of the C channel and the anti-backlash piece will be screwed to the underside of the plate. I realize that I will probably run into some issues with screwing the gantry pieces together and making sure they are 90 degrees as the joint will need some type of reinforcement where I currently have the spray can temporarily holding up the gantry. Going down this path, do you think I could use the nema 23 motor mount plates from this site? Any glaring problems you see? Thoughts on how I could improve it? Thanks in Advance, Matt
Overall looks fine, a pretty good, inexpensive entry point to machine building. Assume that it's going to be a learning process and you're going to damage and scrap parts or decide to go in a direction that makes them obsolete. People (including me) can be reticent about info-dumping because the research process is a critical part of the learning process and can't be short-circuited, so it's usually best to offer ideas that can be pursued without going too overboard. One thing leads to the next which leads to the next, and had you been handed a fish, you'd be missing a good-size chunk of information that would otherwise naturally present itself. Can be any cut size area you like. Wider the carriage footprint, the more stability/rigidity- at the expense of machining volume. CNC and laptops don't usually play nice. Especially when the control is realtime like LinuxCNC. If others have done it, you could try and follow their lead, but I'd strongly recommend a cheap desktop and a Mesa 6i25. It makes the LinuxCNC process MUCH easier. If you don't need the performance level of LinuxCNC but still want a modern control, there are a number of 32-bit grbl platforms to look at. It seems to be a race between grblHAL and Grbl_ESP32, mostly. No, they're designed for very small lead screw and you'd have interference with a large ball nut, assuming 1605 screws. All plates will likely have to be custom. Some plates could be temporary, so the machine can make its own plates later.
You can eliminate the two base mounting plates by rotating your lower c beams 90 degrees facing outward. After doing this, ditch the second rail (per side) and use both blocks on a single rail per side. You can attach the plates directly to the bearing blocks. This will decrease the fabrication work, fine alignment, and material usage. You will gain extra rigidity due to using distanced blocks instead of a single block at the gantry's load pivot point.
This afternoon I cut one of my chunks of aluminum in 1/2 to make base panels https://photos.app.goo.gl/Uf77o5ACCqEwoS2D7 then I tried making a piece of mdf with the bolt pattern for 2 bearings per rail. I printed off the bottom 1/2 of this drawing and taped it on a piece of 1/4" mdf and used a center punch to mark the centers and then drilled out the holes. I put the bearings on the rail and then screwed them down, I quickly realized that my drawing, marking and drilling was not accurate enough and then went back and drilled out the holes one size larger. The larger screw hole worked and all the holes lined up https://photos.app.goo.gl/HReToVa6AvbvHfAG7 Once on the rails slid ok, I could push it and it would go 7-8" and stop, It required a small amount of force to start it moving and moved fairly easily but it is not as smooth as others that I've seen videos of. To try to make it better, I loosened the bolts holding down the rails and put the gantry right near a set of bolts and tightened them down moving the gantry along the way. This made the travel go easier but still not as good, I'm guessing I'll be learning a lot about this process It did seem to slide better when I put weight on the bearings. As for the comments about making the gantry as wide as possible, I have 8" wide aluminum stock for the gantry and was placing the bearings at the outside edges. When I ordered the linear bearings, I got the kits with 2 bearings for each rail so I have enough rails and bearings to do 4 per gantry surface. I understand the point about putting the bearings directly on the side of the gantry to cut out the # of pieces and K.I.S.S. but I would like to understand how to get a larger ball screw inside of the C area. Most of the ball screw kits I've seen have components which don't fit inside the C. BallScrew SFU1605 Antibacklash 250-1550mm+BK/BF12+DSG16H+BF 6.35*10mm Coupler | eBay I did my design idea to get around that limitation of the space of the C channel, I think I am going to keep that T shaped gantry to give plenty of space for ball screws outside of the rails but I am open to the idea of not using 2 linear bearing rails per side.
It sounds like you're designing around an impossibility as a 16mm ballscrew will never fit inside the rails. 12mm fits, but you don't want to use 12mm due to whipping concerns. If you are willing to custom make a lead nut, you should be able to fit up to a 3/4" (19.05mm) or a 16mm lead screw. The resistance you're feeling has a large amount to do with the fact that you are pushing on four rails instead of two. You have the obvious extra resistance due to more blocks, but also extra resistance due to micro level inconsistencies in the rail construction and alignment. I don't want to assume, but considering that authentic HiWin rails cost an arm and a leg, I'm guessing you went with the China special (like the rest of us). Due to different manufacturing tolerances, the rep rails are inherently going to have a higher resistance. Two other things I can point out would be cleanliness. Having the channels cupped upward would lead to dust build up. This would hurt any lead screw that you can fit. You can also fabricate all of your parts in MDF to make sure everything fits and operates as expected before finalizing in aluminum.
After the conversation about the track not being that smooth and the question of how many bearings I went back and looked at the MDF piece I put together and determined that the hole patterns were not perfectly aligned and tried another time to do a print out and tape it to the part and transfer the center of the holes with a center punch. In the end it was not an accurate method so I watched a video of an old-timer machinist step through a few ways to go about getting an accurate and consistent hole and over lunch today followed the guy's basic technique and measured out everything and scribed scratch marks into the aluminum then used a center punch to mark the center followed by hand drilling a tiny drill bit in the center to better define the center punch hole followed by drilling out the holes on my drill press. I don't have the appropriate bolts to hold it down right now and have posted a question about bolt types in another thread. I am happy by the way it turned out (other than the one errant scribe line when my machinist square shifted) and all the bolt holes line up and the group slides nicely with very little effort. I did have to adjust the rails because on their own they want to not sit 60mm center on center. https://photos.app.goo.gl/G9Nepn4GdPHRSTnN6 as I will need to do this precision layout a few more times, I will be shopping for some more precise tools, I have an old school machinist square and a collection of micrometers passed on from my Dad but will need a few more tools to do this work properly. Its funny, I grew up with my old man being a machinist, he made taps and dies for a living and over the dinner table I heard non-stop of pitches, thread and what it took to get tolerances down to 1/1000th of an inch and the one thing he really didn't teach me is machinist skills. I knew his work was tough but am now just getting the slightest glimpse of what he was talking about.
I received in the 16mm ballscrews that I ordered and have started to draw out the components in sketchup and make a design for the side plate. I started by drawing the bearing block for the ball screw and then trying to arrange it to match up with the C channel and linear bearings. I setup the ball screw on a countertop and it seems that the nut housing support is 5mm higher up than the bottom of the bearings. i drew in a line 5mm from the top of the bearing housing and tried to line up that with where the bearings will sit and for the life of me I can't get sketchup to let the lines match. It seems that sketchup is snapping to a 1mm grid I am only designing in 2d to figure out where drill holes are required and also drew out the nema 23 motor mount. In the drawing I temporarily overlayed the centers of the motor and the center of the ball screw shaft to make sure nothing would interact with the thought process. the nema 23 holes would be tapped and drilled from outside, and the bearing mount screws would be tapped and drilled from the inside of the plate. The whole plate will sit on top of a bearing cart plate and ride on the rails and will have screws coming up from the bottom of the bearing cart plate. I'm no pro at sketchup like a lot of folks here so please excuse my inexperience in the drawings. I am primarily concerned with getting distances to make a grid to center drill holes. Any suggestions?
Here's some photos to try to express how I will setup the ball screw. The center span is 1000mm and I ordered 1050mm length ball screws. The non-motor side bearing will be bolted to the outside the side gantry plate and a hole will be drilled to allow the ball screw shaft to go through. https://photos.app.goo.gl/ywa9Qxa9FK54S2j56 The motor side will be bolted on the inside of the gantry plate and a hole will be drilled to provide me access to the set screw on the coupler https://photos.app.goo.gl/MrqqSYLsnh4V3WzK6 The linear bearing rails and the C channel from openbuilds are both supposed to be 1000mm but both don't match. In this case it is to my advantage as it will be easier for me to cut a few millimeters off the C channel extrusion and it will match up nice with the bearing holder and should make a nice neat installation. https://photos.app.goo.gl/WxM53hLM6xojxBF56 This same technique of using the ball screws and mounting the bearing carriers will be used on the 2 side rails.
To be honest, I'm not picking up on what you are asking, but I did make a file with parts that may be of use to you. I don't trust most parts in the warehouse or they have too much irrelevant information, so I just remodel everything I use based on OEM specs. There are NEMA 23 (86mm), all M5 low profile screws, nuts, washers, spacers, shims, 16mm ballscrews and accessories, and HGH15 rails and blocks. As for working in SKP itself, try this. You can set it to whatever you feel is comfortable. In the top bar: Window > Model Info > Units
a couple of prizes showed up this morning ( dial height guage and digital caliper ) and after work I put them to use making my z axis. Having proper tools makes getting precise and parallel lines a lot easier. I didn't trust myself with the dial height guage (sae) and converting to mm so I doublechecked each height with the digital caliper. I drilled out all the holes for the bearings with one drill bit which was a little oversize for a M4 bolt to give me a little leeway and did the holes for the linear rail with the drill bit which was spec'd out for the m4 tap. I didn't have high expectations when my first test hole in a piece of scrap I snapped the tap but I took my time for the 2nd test hole and was able to complete out the balance of the tapped holes. All holes were kissed off with a 1/2" drill bit to get rid of any burs. I left the z axis plates long and will cut them to size later on and did a temporarily assembled the setup to see how it did. I didn't put all the bearings on but I did test them for all the bolt patterns and they all fit fine. Z axis all the way down, gantry does not have side plates and is just resting on the rails https://photos.app.goo.gl/Z79VnbHG4kjyfKRd9 z axis up https://photos.app.goo.gl/HQ6vS2X7K59oEzd96 a somewhat side view https://photos.app.goo.gl/rC12rRHMQqVx766A8 here's the fun seat of the pants engineering It seems that the gantry is going to be very front heavy and that front weight is going torque the side plates and I am questioning the ability of the 6 bolt holes in the side of the c channel to resist all that weight and the torque from the router. I am contemplating how I could counteract some of that torque and the thought I have is to use a setup like the reduction stand off plate set sold in the part store and hang the motor off the back side of the gantry. I would need to fabricate a piece of aluminum to bolt that part to Reduction / Stand Off Plate Set The other thought is to find some type of 90 degree plate which could attach to the bottom of the C channel and pick up some of the load. I also don't have a lot of space between the plates it measured out at 27.95mm and my head is spinning trying to figure out how to get around the lack of height. I don't want to put strips of aluminum under the rails due to effort to do so and adding more weight. I am going to have to think about this some more. I'm pleased with my progress so far but it seems like the more I get done the more problems I generate to solve. I'm an IT guy and it is nice to work with my hands and learn new techniques and how to work with metal. To others - buy someone's kit if you want to be making dust more quickly
I ended up going from a "sandwiched" gantry to a "pillar bridge" gantry with a longer gantry resting on the columns for this exact reason. A little trickier to attach, but not too much more, and far more space for joining plates and gussets and angle brackets to really cinch that connection down.
Hi Rob, Here's where my inexperience comes in to play, my design criteria is to cut wood, acrylic, starboard, and some aluminum ( I heard you on using the machine to improve the machine). It seems to me that you are designing and building to a much higher standard than what I need so I would like to learn from your experience and dial back the capabilities. In that thought, I see that you've used 2 C channels and are separating them by a distance, I am assuming that any time you add distance between the 2 c channels you are increasing strength and you are also using that space to run your bearings in the middle. The whole concept being an I beam, the greater the height of the center portion the stronger the I beam. Also the increased distance acts as a moment arm to resist torque (its been 20 years since I took physics in college - please excuse me if I confuse terms) I am thinking about how I could improve my initial design with the least cost and least amount of rework so the idea I have now is to sandwich 2 C channels together thus the distance on center between the bearings should be the same but I know in practice I may have to deal with a not precise fit. I was going to have my bearing exposed on top of the gantry and if I kept designing that way I would only add @$40 for an additional C channel extrusion and loose 40mm of gantry height. 12 bolts on either side have to be stronger than 6 and I could also add some thin aluminum to join the 2 c channels to make a more consistent piece across. I think in essence it is what you are doing but a dialed back, less robust solution. Not adding distance between the rails won't add additional resistance to torque but the load per C channel has to be 1/2 of what it would be by using just 1 c channel.
here's a quick and dirty drawing of my concept - make a square out of the 2 Cs and put the linear rails on the strong side. Theoretically the center of the rails would be 60mm in either direction so my Z axis plate could stay the same. I didn't realize it when first thinking about it but I loose no gantry height either with this concept. essentially only adds 6 bolt holes to each gantry. Another side benefit should be the additional flat space on top for the wires to rest on. Thin alum plates could be added to join the 2 pieces and make it stronger with the only consideration that the plates on the front side do not interfere with the travel of the bearings, I'd have to plan out how to make this happen. https://photos.app.goo.gl/jax93W1arQD6kdad9 thoughts?
Essentially, yeah. But there are multiple degrees of freedom to think about here. Not just vertical bending in Z, but also bending in Y as well as torque- both positive and negative- about both X and Y. So I separated them to increase rigidity in Z and about X, but kept the orientation flat so as to maximize inherent rigidity in Y and about X. The biggest weakness is twist about X due to my large moment arm (ie. Z axis height/travel). The filler 2060 beam and tie plate at the back means the entire thing has to twist as one unit rather than being able to simply pull one beam forward and push one backward- which is extremely easy with 1500mm beams, C-Beam is quite flexible. You can see that in the simulation posts in my build thread. It ended up being necessary, to reduce the twist to an acceptable level, to add the heavy steel tube reinforcement top and bottom- per Fusion's simulation, anyway, we'll see how it does in the real world later. For most people, lighter tube or adding another 2060 to the front to close the extrusion up and having the screw outside the beam, between the rails at the front, would be a better option. Dual C-Beams but separated is basically the LEAD high-Z mod, which seems to work well for a lot of people. Sandwiching them together could be fine, as long as they're tied together well along their length. If they're only fastened at the ends, you end up with a flexure mechanism and the center can move a lot. The idea in your image should work ok, I would think.
40MM square steel tube fits snuggly inside the C beam.. and would help with rigidity. I saw a video (cannot find it now of course), but that guy was also using rack and pinion, so no lead screw and mounting block to worry about running up and down the center of the c channel.
How critical is the spacing between the ball screw coupler and the back of the z axis? is .5mm over 1 meter good enough or does it have to be tighter tolerances? In this pic, I aligned the ball screw bearing holders across the back of the gantry to determine how much space is offset. It looks like it is 10mm offset https://photos.app.goo.gl/akq2Uf8aqmZb6wH88 I've got all the lines scribed out on the gantry pieces for the bolt holes for the bearing holders and the center line of the bearing, I just need to figure out the offset because the bearings need to be further forward to account for the difference in the coupler. https://photos.app.goo.gl/4ALVakxi3MKizNqN6
I f'd up bad today. let me count the ways.... 1. When I did the sketchup design the day before I had the bearing plates going vertical in orientation and that gave plenty of space for the motor to mount on the outside. Not thinking clearly I changed the orientation and did not account to leave enough space above the bearing to mount the motor. Essentially I would be able to bolt it on the bottom 2 bolts. 2. I layed out the lines in the previous post to have the ball screw holder be tight up against the back side of the z axis. I'll admit I have a tough time with left hand / right hand parts and making pieces backwards. Well that brain fart hit me and instead of moving the bearing carriers further forward to account for the offset, I moved them backwards the 10mm to create a 20mm offset. 3. When drilling out the holes for bolting down the bearing carriers I didn't think and used the wrong drill bit which is too large for me to tap. So how to recover? here's what I am thinking 1. 2 pieces of 3/8" aluminum plate to make up the difference between the z axis and ball screw carrier https://photos.app.goo.gl/Q7Ytwcm9nqb2WNXDA 2. through bolt the bearing carriers and recess the heads - temp screws in this pic https://photos.app.goo.gl/r9th29nf7SFir61F7 3. I am thinking of turning the motor on a diamond pattern so as I could mount 3 of the 4 bolts Successes - the layout of the ballscrew between the plates is perfect, I did cut ~1/4" off the C Channel as I was describing in an earlier post to make the rails match the C Channel and the setup fits perfectly. The outside bearing fits nicely on the side https://photos.app.goo.gl/hY7dJ3EEnNqdfQ7r5 and the motor side bearing sits nicely inside the gantry https://photos.app.goo.gl/ZZrTPVD9x5J9Mygy6 I'm not ready to junk the side panels yet as I am sure I will mess something else up but I think I'll be able to use what I've made even though it is not perfect, it will be good enough for right now. I did put a cordless drill on the ballscrew shaft and ran the z axis back and forth with the drill set at the weakest clutch setting and it didn't have any issues moving the z axis and didn't engage the clutch.
I'd aim for around 0.1mm at worst. A thou or two, ideally. 0.5mm is gonna start flexing/jamming things as you go, especially near the mounting points.
Welcome to the club. I have my X axis stepper mounted with only three bolts with no problems. I did this to have access to the coupling without having to disassemble anything. That link seems to be dead.
It's not consider a mess up when you are the builder, and you catch the flaw, and fix it yourself. Well that what I tell my self.
Since I am taking quite a bit of time to figure out the mechanicals, I'd like to take it easy on figuring out the electronics and would like to purchase them sooner than later and start learning how to make them work. Is the black box a simple way to go? also if I am interpreting right, I wouldn't need separate stepper motor drivers if I went that path
just thought of another idea - was the difference in bearing rail length and c channel length intentional? to deal with the differences in how the metals expand and contract due to temp? so my question is if I made the rails fit tight and it was 65 degrees at the time, am I going to have issues putting in my basement over the winter where it might get down mid 40s?
My stuff is in my garage.. which gets down the the 40s sometimes, and 100+ in summer.. and have not had any issues with thermal expansion/contraction. As for the Blackbox, it is an all in one solution, just be sure to pair your steppers appropriately, the Blackbox has a max peak current that it can support, so if you get monster nema23 or nema34 steppers, you may end up under driving them which can damage them. Pick your electronics stuff carefully.
Can't really damage a motor by running on lower amps. The opposite can burn them out though. People tend to forget the steppers' Max current rating is not a goal, but a limit (;
My project has taken a slightly different path, a couple weeks back I was browsing my local craigslist and came across an ad for a 6040 in my town for short $. I went and purchased it with the thoughts of designing and making the parts for my CNC on the 6040 and when I get my CNC up and running I would sell the 6040. The previous owner did the same thing, he built a better CNC with the machine. I've spent some time getting a PC built up with Linux CNC and interfacing with the 6040 via the parallel port. I was able to get it up and running and it seems to work fairly well. I've done some carvings of my kids names in wood as well as converted a few images to basic carvings. I am reasonably happy with the setup. I've also gone down the path to purchase components to build a control box for my CNC with the thought of doing the same wiring setup as the 6040, get it all to work, then transfer it to my cnc when it is finally ready. I've purchased a 7i96 mesa card, stepper motor drivers and a pair of 7.3amp 48v Power supplies. Back to my original project, One of the questions I have is the design of the screw holes and using fusion 360, it seems no matter what tool I choose I cant get a tool path for a m4 sized hole. How do folks go about doing this? Is this something you build a recess for the head of the bolt and that tool path is fine, then use a centerpunch with the right diameter to mark the center of the hole and drill that out the correct diameter?
So that's awesome. Congrats. I have read many people in other forums that have used the 6040 cnc and got them work reasonable well with a longish run time. As to your question about Fusion 360 and M4 holes I use the Bore function and a 3.175mm or1/8th inch end mill. It does a spiraling dig in to the part.
Thank you for the info, I'm guessing that I would run one program to do all the Bore functions then bolt the panel down to the machine base, through the bolt holes then come back with a larger endmill to cut out the panel As you can see from the photo the previous owner made his gantry plates on the machine. BTW the guy had an amazing machine shop in his basement and his new CNC was sweet. https://photos.app.goo.gl/aFyv1ByK8W3XSADR9
