I run a test routine that moves all axis in many possible directions. As the X-axis transits from left to right (0 to 24") it shimmies. I use a C -Beam with two extra large gantry plates, one on each side with 8 wheels. I still had some flexing at the tooltip so I jury-rigged a third rail (20x20 V-Slot about 3 or so inches above the C-Beam, hoping to stiffen the Z axis. Before I added that change I had the shimmying in X, but I had hoped the stiffening would resolve that problem. It is not the cause of the problem. I can't tell if it has the shimmies on slow moves or positioning but when going 60 IPM or more I can see the shimmy, it almost looks like it is binding and releasing. I use the standard 8mm lead screw for the X and Z-axis. I am including pictures of the portion of the X-axis assembly and hope some suggestions can be made. My Y-axis uses 13 mm lead screws that I had from the teardown of a previous machine made with 1" plywood. I built this one because I could not do inlays and I am concerned that this problem may prevent me from doing so as well.
Easy test: pull the spindle out of the clamp and re-run the program. Does it still do it? Looks like way too few wheels for the weight and cantilever moment that you're putting on that whole carriage assembly. Likely needs a full four-row of doubled up wheels up top, and at least three pairs underneath with eccentrics, which probably need to be a little tighter than normal- at least the middle one, maybe not the outsides- to counteract the big spindle flapping around. You might need to consider the lower rows a wear item. Even if the wheels can take the loading statically, once you start moving there are a lot of dynamic forces at play in multiple directions, and these are still just plastic wheels. You may also need to follow your Y-axis' lead and lose some X travel but spread the forces across more of the beam. The unsupported Z wheels are kinda suspicious too, since they're just shimmed M5 screws. If possible I'd go to a lower profile wheel with an M6 or M8 screw inside. I'd probably also have a high inspection rate on the lead screws and nuts in the gantry- 8mm with a delrin(?) nut moving all that weight is asking a lot. Same on the Y axis wheels; that's a lot of weight that very few wheels are supporting- inspect for flats and grooves regularly. Or just go to linear profile rail, and then your motion components can take 100x the forces you're trying to put on them and what you have to start worrying about is the V-Slot itself flexing- which could well be why the low wheel count here is causing problems; aluminum extrusion looks ultra-rigid, but it can move and twist quite a lot. These machines are designed to be used with cheap trim routers for a reason.
I reviewed your response and went and took a look at the machine. I removed the spindle and as you predicted, the shimmying stopped, at least from what I could see. I also looked at some linear rails. If I: Replaced the wheels on the Y-axis with linear rails, would that take care of the load on that axis? This axis is connected to the bed in 5 places on both sides, which helps it stay straight and not droop. Replaced the wheels on the X-axis with linear rails, but my concern is the stiffness of the axis. Assuming that I want to keep the 2.2 KW spindle as well. Should adding another C-Beam and connecting them together so as to add rigidity to that axis and the addition of a linear rail be enough? I really do not want to reduce the travel of the X-axis if possible. If I did that, do you think the potential twisting of the X-axis would be minimized or eliminated? Would I need a top and bottom linear rail for the X-axis? I did notice that the Z-axis struggled to bring the spindle up, but does not lose steps. On the old machine, it did not struggle as far as I can remember, which had the larger lead screw. I think I should consider replacing both X-axis and Z-axis lead screws with the 13 mm (1/2 In) screws that I have on the Y-axis. Finally, Some of the linear rails came with Ball-Screws, I do not think that they are necessary unless I also used a servo and feedback instead of just step counting. Your thoughts are really appreciated, as this is not my field of expertise, just general knowledge, and a hobbyist. Again, thank you for your input, Regards, Ken Hall
Yes. The weight would be spread out across all the ball bearings, and due to the rigidity of the two carriages, you could bring them closer together and regain some Y travel. Good. That bit's important. I would, but yeah, the stiffness is a problem. I attacked it a little in my (still-desperately-trying-to-get-back-to-it) build thread here: M4: 1510SS Heavy Mk.I Eliminated is never gonna happen. You can see what I ended up with on my X-axis from the later images down the thread- two C-Beams connected by a 3/16 aluminum C-section and backed by 1x3" box steel. It's very rigid- far more so than what you need. It's also very heavy. I [will] have two NEMA 34s driving 16mm ball screws on the Y axis. I think you could get away with a lot less than that for woodwork, but that 2.2kW spindle is not going to be your friend. You'll definitely need two C-Beams. Not sure of the connection, maybe a 1/8" backstrap, and you shouldn't need steel anywhere. Two 15mm rails top and bottom should be fine. Again, you can get away with shorter total carriage lengths due to the inherent rigidity of the square profile rail system. Yes. Tying the two together rigidly with the Z axis plate doubles the amount of work any twisting forces have to do. A larger diameter screw at the same pitch has a shallower slope, and therefore higher leverage. A nice, big, well-greased Z axis leadscrew does a lot of invisible work for a machine. I concur. You might need to make new plates. So don't take the machine apart too quickly! Not particularly relevant. I use ball screws with closed-loop steppers because closed-loop steppers are flat-out better, but I'd still use them with open loop steppers because in general they're better- faster, lower-friction- than lead screws. Not an absolute though, and can cause you issues with backlash compensation and other things. If you have a leadscrew system that works, I wouldn't be in a hurry to change it just yet. Learning from designing and redesigning your own machine is how you gain expertise! It's a sharp learning curve, but I'm generally for it more than using pre-designed kits as-is, because I think that fundamentally understanding the design of the machine is key to knowing how to most effectively use the machine in practice. It depends how willing you are to spend a few months grinding through some esoteric Googling though.
Again much information that I will certainly consider. I have a lot of designing to do before I do anything. I got the original lead screws from McMaster years ago. I will see if they can look up my order so I can get a lead screw for both the X and Z-axis. I already had to make plates for both sides of the Y-axis and I have the bearings for the ends a; already. I think the entire gantry will have to be redesigned to accommodate the changes that will stiffen and improve the machine. Again thank you for your thoughts. BTW I went to M4: 1510SS Heavy Mk.I and I realized that I had been there perhaps 6 months ago. A ton of information. If I used two C-Beams for the X-axis, would it make sense to have them tig-welded to form one unit, or would attaching them to a steel plate be better? Just an FYI, I started building this contraption in early April 2022, I did it because I could not get a good inlay with my original machine and I figured that this machine would do that, IT Does Not, but it will, I can be sure of that because I will not quit till it is functional.
Rob, Your machine is massive and looks so very solid. It is obvious that not only do you do great work, but you are knowledgable and dedicated to getting it right. I did not realize that all that was back in 2013. I am not a machinist, I used to repair Kerney and Trecker horizontal mills and Warner and Swasey lathes, both companies CNC brands, and unfortunately for America no longer around. I of course watched a lot of machining happening and was always fascinated by the process but didn't learn much about the actual process of feeds and speeds and material machining. I broke 8 1/8 in single flute bits just trying to cut out the column supports that hold the X-axis, I finally cut them with my table saw and a carbide blade. I think the bits galled up and could no longer cut, thus broke. If I remember correctly, I think my depth of cut was about 1/16 in and the feed rate was probably 25-50 IPM, spindle about 12000, best bet from my memory. I think all of those numbers were wrong. This was for 1/4 in aluminum. Just an FYI concerning my lack of machining. From what I have read, I should keep my depth of cut shallower, feed rate much higher, and spindle speed more like 18000 rpm. Getting back to the machine, so many thoughts and questions, not enough knowledge. I am thinking of putting rails on all axis' and getting rid of all the wheels. Never thought they were a good idea, but thought Open Builds knew better than me. Parallel rails for the Z-axis, I think the rails should be mounted back of the Z-axis, and also the 13 mm lead screw, which I have and will change to C-Beam as well. The X-axis will also have a 13mm lead screw, that I will get from McMaster Carr along with a matching nut. I will add an additional C-Beam for X as well, not sure if back to back or up and down, think up and down should reduce the twisting of X when cutting. Although I think there is a very small amount of lateral stress due to cutting, I think the weight of the spindle is a bigger problem. I think the rails for the X-axis should be placed one on top and one near the bottom on the front. If I go up and down, I think the two C-Beams should be joined. perhaps facing in opposite directions with the lead screw in the lower C-Beam. I think that will also reduce the potential twisting of the X-axis. Y-axis will also receive two rails on each side with one on the top and one on the inside surface at the bottom of the C-Beam. I have a dozen 30mm V-Grove wheels from my old machine, which match the V-slots in the C-beam and V-Slot extrusions I have. They have a 9.5 inner bearing opening and a 3/8 machine screw fits them perfectly. If I were to use them, would not the V-slots wear rather quickly? I still like the rails, but trying to see if the V-Grove bearing had any future in this. Columns are another thing to consider, I am not sure if there will be side-to-side movement of the X-axis. Looking at your build, your Y-axis is massive and should not allow for any X side-to-side movement. Do the end plates help stiffen the Y? Lastly putting it all together. I will be moving one side out to the full 1000mm thus the X-axis will be extended about 30-70mm. Although I realize that this will add some risk of additional twisting, I do think that my changes will make that a minimum. When I get into the final design and assembly I will post pictures of my efforts, yours are so helpful, even though I am not building such a strong machine. Again, thank you for your assistance.
Working with Fusion 360 is not a strong suit for me, but this gives a general idea of what I am thinking about the X-axis. C-Beam on top of each other, top reversed from the bottom, 8 mm aluminum plate the entire length of the X-axis using t-nuts every 20 mm, top slot, bottom slot and the two middle slots, I think this should stiffen X up pretty good. End caps are also of 8 mm aluminum as is the motor mount.
Nice! I tried to document as much as possible to show both the machine and my thought process behind it. Not much point in just showing a replicable machine build if nobody knows *why* I made the decisions I did (at least to me). Nah. Let the anodization do its job. TIG aluminum moves even more than steel, and I'm a little leery of non-heat-treated steel weldments for CNC purposes. Bolt together everything you can. Also makes things more adjustable and configurable in the end. Welding is much more permanent. That's my join date. M4 construction occurred between spring 2020 and autumn 2021. I'm determined to finish it, because I have plans for it, but life has gotten in the way more than anticipated, and I'm currently fighting the mill over some oddities in the calibration. Even if I have to brute-force it by buying another manual bench mill, I'll be re-starting it in the next few months. Nice! I learned machining at the university of YouTube. There's nothing you can't learn from Tom Lipton, Adam Booth, and occasional others for specifics like Joe Pieczynski and John Saunders, you just have to be willing to sit and watch hundreds of hours of video- which I did, before I even bought my first machine tool. Once you can predict what they're going to say and why, you're ready to do it for yourself (without hurting yourself or the machine!) A single tooth at 12,000rpm running 50IPM is over 0.004" per tooth chipload. That's a decent load for a bench mill, but it's hugely overkill for a flimsy hobby router. Needs to be about a quarter of that. Double the speed and halve the feed to 24,000rpm at 25IPM gets you to 0.001" per tooth, which is appropriate to the size of tool and a router should be able to deal with the cutting forces. In aluminum 24,000rpm in a 1/8" tool is only around 800SFM, which isn't a problem with carbide tooling. Cutting dry or with air/mist shouldn't, in theory, get chip welding unless you have insufficient chip evacuation. Definitely spend more time researching feeds and speeds- look up calculators, tables, as much as you can. Understand the relationships there. Plenty of this trips up even experienced CNC machinists (which I wouldn't call myself any time soon), so don't expect to get up to speed overnight. Also, why 1/8" tooling? For inside corners? Or to minimise material removal? Don't try to preserve material too much, that was a fatal early error of mine. OpenBuilds knows better than most brand new hobbyists who may not even yet know what they want to machine, which is a whole lot of customer base out there. Wheels can also work fine for printers, lasers, and weird "clean" gadgets that university research departments build like auto-testers and dosing units, which modular motion components are great for. Moving rail vs moving carriage. You can deep dive that topic, there's a lot of conversation out there and it depends on what you're trying to do. I went fixed rail because I wanted the highest rigidity with the spindle at its lowest position. Probably not, depending on the weight applied. Could be good for a moving cover/door/enclosure/whatever. The steel end plates (though as I said in the thread, they probably don't need to be steel) definitely help stiffen the Y, thanks to all those screws plus the inner rectangular boss that fits precisely inside the C-Beam cavity. The lower C-Beam is rigidly tied to the table, and then the dual rail also largely prevents them from moving in relation to each other. So there's nothing to flex unless the lower C-Beam is literally folded... Somehow. Columns are important though. I liked my discovery of the idea of keeping the gantry beams long so they could also sit on top of the column plates. More area to rigidly tie things together is a good thing. More gussets, more better. The smaller you can make the frame, the better. You just have to compensate for whatever adjustments you make. Always pay it forward. Someone else might be looking at your build in future. Sounds promising, but I don't think the pics posted right.
I finally found a Fusion 360 design of the extrusion of the C-beam. I am looking for some very inexpensive sources of Fusion functional components to work with in helping me do some drawings, which will also help me improve this machine design. I hope you don't mind a bulleted list, helps me keep my thoughts and separate sections of them. Up until now, it has been, I'll put this together and see if it works. So far I have assembled and redesigned the Y-axis columns two or so, times and still, I am not totally happy with it, wheels will be gone as part of the solution, and perhaps doubling the C-Beam. For what I plan on using the machine for, I do not want to overbuild it either. The rest of the machine, especially the X-axis, I am really not happy with it either. I am contemplating adding a second C-beam on each side of the Y-axis which will raise it another 80 mm from the bed, which can be compensated by shortening the column. I want to keep the working travel of Z to about 3-4 inches with as close to 3 as possible. Although if everything is stiff, that is less of a concern. The most concerning problem is everything related to the X and Z-axis. Adding rails to Y will stiffen both sides and as you said earlier allow me to decrease the width of the support column giving me more available Y travel. I may lose this extra in the design of the X-to-Z transition, which will be deeper with the addition of the 13mm lead screw and C-beam, which is ok with me. I will double the height of the X-axis thickness and add a plate to the rear portion of it to further stiffen and connect the two C-beams together, making them as close to one piece as I can. Additionally, I will install rails on top and bottom of that assembly, which seems it would further stiffen and improve accuracy as well. I need to design a good transition for the X-to-Z-axis carriage. Assuming that there is no discernible twist of the X-axis, which, previously, I had tried to compensate for with my current flimsy design, a wider and more robust Z-axis design is important. Both X and Z will have the 13 mm lead screws. As I only plan on generally cutting hard-woods and most of what I have been doing generally involves depths of 1/8" to 1/4" there is not too much stress, but repeatability and accuracy are important because I want to do some inlay work, so far, I have not been able to do so. Errors of .005 were not unusual and the gaps in the inlays were pretty big. Whole reason that I built this machine and it is currently no better. (Try to explain that to a wife, you mean you spent $xxx and it is no better?) Your thoughts and suggestions opened my mind up to other solutions, such as the rails, which up until now were not in my mind. I have learned a lot from the experience and now your input has been super in recognizing what needs to be addressed. I will do better at making some drawings that represent the direction I expect to go. I used to have Autocad, but my license expired and it was too expensive for the little I used it. 2-D drawings were much closer to what my brain could envision. And who knows, I may get better with the 3-D stuff. I hope that if anyone follows this trip, they will see an evolution from a simple, not-so-worthy design to a much-improved process and end product and might take some of the efforts that you have provided and I have used to help them in their own efforts. I hate to use up your valuable time, but you are so helpful and provide insight, not just answers, which is rare when asking for input. On one programming forum, I got half the answers to my questions, I think the person was trying to make me research the question and I was criticized for what I wanted to accomplish. Eventually, I did resolve my problem and his input was a little helpful, but the solution was not even near his point, actually, it was a whole slew of problems I had to resolve. But I degrees, just trying to tell you how helpful you are not only to me but to many that come across your work. I know how that happens, life can be so unpredictable, best to your efforts, and again, Thanks, Ken
This link might help with the openbuilds components - step format though. Thers is a library of Openbuilds parts in Fusion format somewhere here - I'll post it when I find it. STEP Parts Library Alex.
I found the Fusion one but it only had a few items in it, I think I tried using the Step items but somehow it did not work, that was many months ago. I will look into that again. none of these areas do I have any sort of expertise, but I keep learning. Perhaps I misunderstood GladCad, and that they wanted a fee, which unfortunately at this point is way out of my ability. If there is a hobbyist availability for their stuff, I just plain missed it. I don't show the motor or motor standoffs in these drawings. I also do not show the coupling or the two bearings for the lead screw. Now to start thinking about how I want to make the columns and mounting them to the Y and X-axis. I also need to place the linear rails into the drawings. I am Still not sure if I should have a top rail and a side rail near the bottom or just a top and bottom rail on those surfaces. I see advantages and disadvantages to both thoughts of how to do that. I think I saw that there are 16mm, 20mm, and 25mm sizes I could use. This machine will probably not do any heavy lifting, so I think the middle ground should be adequate 20mm. Currently, my gantry wheels are approximately 22mm apart to prevent the X and Z from moving multiple Millimeters. Still not good, but at least the machine is usable and I will be able to use it to make some parts. Both Y and X will be1000mm, Z is currently 250, but with the redesign, it may be different. Stiffness and precision is my goal, and I will not give that up to get another 25-50 mm out of Z. I added a rail above my C-beam for X to stiffen Z and it worked to a degree, not great but better, it is like a plumber doing carpentry on a NASA space probe with a shoemaker providing the materials. I have some additional work to research, including the STEP image compatibility and GladCad. It has been almost a day's worth of these images for the X-axis. BTW, I don't know why I keep seeing double images when I only upload one each. Thanks, guys, appreciate the insight, suggestions, and help, Ken After much thought, this is changing , see below comments
Additional information on STEP files, A common question our community asks us is: can you import STEP files into Fusion 360? The answer is, yes you can! See https://www.autodesk.com/products/f...question our community,answer is, yes you can!
Rob, After much thought, I am going to use 4 C-Beams for the X-axis, 2 x 2 configuration. I think this will stiffen X in both the Y and Z directions. Hopefully, it will significantly reduce any twist in the X-axis due to the weight of the spindle motor which is almost 12 pounds. I have ordered linear rails for both the X and Z-axis, which should arrive end of September. I will use two blocks for each rail as well. I thought about only using one block per rail in the Z-axis, but decided that two would help stabilize that axis better. For the Y-axis, I have decided to use the V-Groove bearings I already have. Dimensions of them are OD is 30 mm and ID is 9.525 mm and width of 11mm. Although rails may allow me to reduce the spread of the column, the 210 mm separation of the V-grove bearings will be sufficient. It is similar to what I have now with the Open Builds wheels. These, however, being steel will be stiffer and stronger with a decent surface contacting area with the V-sots. That's where this is now, I am working on the design of the Y-axis and column. Ken
Went there, and found I could download the files, at no charge and was extremely helpful, thank you, Peter.
Have you looked at other methods of stiffening the gantry? This sounds heavy; I'd run the calculations for the Y-axis motors and see what kind of accelerations you'll get out of the setup you have. They'll also wear the aluminum extrusion- there's a reason the standard wheels are plastic. What kind of timeframe you're talking I'm not sure- a member who was building back in the day with steel wheels might have an idea- but with your gantry weight, making the actual frame of your machine into a wear item might warrant reconsideration.
You certainly know how to take the wind out of the sails, but that was a concern of mine as well, just looked like a cheaper way to go and I already had the bearings.. I just was not sure how much wear that might be. As far as your previous concerns about the Y-axis and if the motors could handle the extra weight, I am not sure how to run the calculations, but my previous machine was made out of 1" plywood the X-axis was 8" x 10" x 26", the Z-axis was about 15"x6", and held the 11# spindle motor. The columns were about 1.5 sq ft as well. It was much more weight than the proposed 4 C-Beam and I was lucky to get 100-150ipm. The weight with respect to the Steppers is my least concern, the wear you mentioned, is my greatest worry, as in the old machine the aluminum angle did not do well with the V-groove bearing riding on them. I do not need excessive wear to be my enemy. I may be able to push that up some more. My motors are Hi-Torque steppers I remember the 1.8 degrees and 3 amp. Partnumber on the stepper is 24H-290-35-4B 3Nm. I found specs for the 4A but it had 8 wires instead of the 4 I have. I never attempted to push the speed up of any axis. Didn't even think I could. When I read that some people were at 3-400ipm or more, I just figured they had some commercial unit. I just checked Mach 3 will let me adjust ipm and accel to anything I want, the motors are the limiting factor. With this current (Open Builds) configuration, I am seeing about 175ipm now and the extra weight of four C-Beams vs two is a little less than 4 Kg. So are you in favor of the 2 C-Beams held together along their length with the aluminum strip or the 4 C-beams? The two will be easier for me to work with. I think the main concern will be the twisting of the X-axis. The whole reason I tore the other machine down was 1. Wear and always tighten the opposing V-groove, which was riding on the aluminum angle, bearings to stiffen the machine. 2. Could not do an inlay without gaps of up to .013 along the edges. 3. I want to do some 3-D work which requires more than the old machine could possibly do. If I seem impatient, I am, at 75, I don't have that much time to do all the things I want to with this stuff and I have already been messing with this for almost 6 months. I still have a lot of work to do, and I want to pretty much finalize the design ideas so I can get on with trying to make it happen. Your insight and help have been tremendous and I do appreciate your efforts. Ken
This is for anyone. The machine's purpose is to machine wood and rarely some light metals. I am trying to use linear rails on my Y-axis and on all axis to improve the precision of the plastic wheels and make the machine. I have a 2.2 KW spindle that weighs over 5 Kg. I currently have a C-Beam with a 13mm acme lead screw in the channel, it is proud of the channel about 25mm to accommodate the lead screw nut bracket. The axis is 1000mm. I am also trying to keep my expenses down as adding the linear rails will add about $100. the linear rails I have chosen are from Ali-Express and are 28mm high, 34mm wide, and 39mm long. I wanted to mount the rails on the open side of the C-Beam above and below the lead screw, away from the spoil board side, there is enough clearance, but I would have to install the screws for the rails in the channel of the C-Beam. I am hoping to keep the rail configuration verticle so as to not need something to extend from my column support which is on the outside of the C-Beams. If I move the rails on the top of the C-Beam, I have the same problem with the screws, as they would be in the channel as well. I think I have included all the pertinent information. On my X-axis I am using two C-Beams, in a vertical configuration. I am tieing the upper and lower beams with a steel 3mm thick the length of the axis of the joint tieing the C-Beams together. I am hoping that it will stiffen the X and Z axis minimizing tool walk while cutting. Any thoughts would be greatly appreciated. I was in the middle of designing the support column and realized I had no idea which was a good way to do this. I included the X-axis as information, it is the Y-axis that is my concern.
Rob, you said earlier "I'm not sure- a member who was building back in the day with steel wheels might have an idea-" WireMonkey and M90 Ranger had some excellent ideas and I will look into their suggestions. Perhaps the two rails separated enough for the lead screw in between would be a good solution with a steel backing plate behind the two beams to tie them together and stiffen them. The end of September is when the rails come in for the X-axis, so I have some time to think this through. I usually run wood in popular but have cut oak as well. My cutting is generally with .1 deep, 30ipm, with a 1/8" 2 flute end mill between .03" and .05". I have found if I go much faster than that the bit breaks. I have thought of making a clearance cut first, and that would probably make the cutting faster, but have never tried that process. Other times I may be using a 90 deg V-bit. I plan to rework all my files and will test a clearance cut idea if I ever get my machine up and running. I expected 4-6 weeks to make the transition but now it is 6 months, so I am anxious to move forward. With that in mind, I want to do something, but not at the expense of the precision I want so I can do 3-D projects and inlays. Haven't tried 3-D yet but inlays failed miserably, even with this machine, so it needs to perform the way I want it to. Up and down is not as big a concern as the material is soft, but the bit walking, especially when plunging is especially important, of course, that could be the up and down motion with a not-so-stiff X-axis. With the previous machine, I do not think had a problem with stiffness as it had two 8" by 1" plywood for the entire length of X, but the Z was probably the weakest point, as it relied on the connection to X by the V-Groove bearings. all motion was on V-bearings on Aluminum, which wore excessively. I continually adjusted the movable bearings, but the wear was uneven because a majority of my items were the same size and I placed them in the same location every time, so when I went to a different size project, the depth of cut was different in the less used areas and both X and Y were impacted. My finances are a little tight, so I need to take inexpensive corrective action with what I have. I think the linear rails will be an improvement and I am thinking that placing a 1/8" steel runner underneath the rails will further stiffen the X-axis, as well as the steel in the back of the X-axis. I think that could bond the one under the rails to the C-Beam as the screws for the rails should go through the steel into the C-beam. I think that the Z-axis is short enough that it won't need the additional steel, it is the transition from X to Z that will be my concern and I haven't even thought about the design of that yet. nor have I finished the columns as I haven't decided on the layout of Y and X. Amazing, what I thought would be a quick move from one machine to another design, is nothing but.
That's what I would do. An 1/8" steel plate or 1/4" aluminum plate. Just get ready to drill a lot of holes! Roughing-finishing is a much faster and more efficient process, it's just tricky on hobby routers because they can't have tool length tables, so you're probing TLO more than would be ideal. A reliable, automated process for that would be a good plan. Always the way with non-commercial machines, unfortunately. I can't move forward on M4 until I get the mill up to speed again- it's capable of sub-thou machining, once properly calibrated, but it's fighting me. If you're doing moving rails, you'll need to screw the carriages to a plate and then screw that plate to the X axis carriage, unless you come up with a way to be able to mount the carriage plate to the X carriages with the Z carriages already installed. Easiest columns would be 1/2" aluminum plate, but that might not be the cheapest way.
I found something I could use if I wanted to switch over to Gerble and an Arduino, it has the ability to add a TLO, but I am not proceeding in that mode. I checked out some prices of 1/2 plates and they are above my pay grade at this time. Rob, again for your thoughts. I am not going to have moving rails, trucks will be riding on the permanently attached rails. I am expecting that the rails attached to all of the axes will provide some strength and stiffness, and with the steel backing as well as 2 high C-Beams that even with the heavy spindle, twisting will be at a minimum. Additionally, I and changing to the 13 mm lead screw for both X and Z. Y already has them. As I think about this, with the weight of the spindle motor pulling down on the X-axis, twisting it some, especially in the previous configuration of one C-Beam, as it entered the material, the weight of the spindle would be compensated for by the support of the item being machined, or pushed up some, causing the tooltip to move from an angular penetration to a more perpendicular position. The hardness of the material would determine the change in charecteristics of the penetration and also how long a tool may be viable. Another spindle I own is 500 watts a sure cry for the 2.2kw I now have. I am researching a smaller lighter spindle, but it has to run on the VFD I have which is a 220VAC one. Now, Z-axis is more critical in my mind, sure X could sag or twist, but I do not think that will be as significant to the tooltip walking as it penetrates the material and alter its center as it changes direction. I actually watched this occur and is the reason I am going through all this effort to try and make sure each axis is stiff enough to do the jobs that I normally do and I hope it may be better than I now expect. The jobs I have done did not require precision nor repeatability, cutting lettering into popular all looks the same compared to doing inlays, which I have been unable to do. I have the X-axis designed except for some screw and bolt placements and I need to start on the transition piece that will attach the trucks to the yet-to-be-designed Z-axis. What should take me, in my mind, about 4 hours takes 12-16 in CAD, haven't used CAD for 12 or more years and have to learn a new program as well. so that is slow going but gives me a chance to see what I am missing in the design. So that is the status for now.
Grbl is using a TLO regardless of whether you set a number, just like it's using a WCO regardless of whether you probe a work zero. If you're using multiple tools on one job, effective use of a TLO is critical. This is a much easier build and has slightly more consistent rigidity throughout its travel range- depending on your Z rail support and travel length- at the expense of a little bit of reach in some circumstances. They will to some extent, but it's not recommended that you include them in your calculations. The biggest improvement comes from the rigidity of the carriages; four carriages connected to a small plate riding on two linear rails is an extremely rigid system- to the tune of kN. I guarantee you can push on the bottom of that tooltip in the +Y direction with your Z axis near the bottom of its travel and move it somewhere between 5 thou and a millimetre with relatively little effort, due to the twist in the gantry. It's not the Z travel that you need to worry about, assuming your gantry carriage and Z axis system isn't too cantilevered out from the X rails, it's offsets in Y. It's less critical in wood, but you won't get accurate inlays as the cutting forces pivot around the tip during pocketing. If you can hold the gantry rigid and vertical, Z basically looks after itself as long as there's no lash in the screw and your Z motor is strong enough. X is always the weak point in a gantry system, but it produces variances in Y, not X itself. The further apart vertically you can put your X rails, the more rigid it'll be since the less leverage the tool will have against the imaginary "non-pivot" rail (obviously in reality they both twist/bend in some way). Basically, every significant failure in "Z" is actually a failure in the gantry itself if you're using rails- there's no plastic Z wheels to flex, after all.
The X and Y-axis will have 25mm rails with 2 carriages per rail. The Z-axis will have 20mm rails with two carriages on each rail as well, of course, as each carriage is almost 60 mm long they will be reasonably close to each other on each rail. The Z axis in total will be 300mm or less. I don't see that as a problem. I must have been brain dead when I wrote the following paragraph which is in italics. I am running into a physical constraint in accommodating the lead screw and the lead nuts connecting to the moving gantry, as well as the Z-axis. I was hoping to not have to add a spacer to the C-Beam to move the rails further away and provide space for the nut and holder. It's actually the width of the trucks that provide the interference because I am placing them on the sides instead of above and below. I am not yet locked into either but thought the side would be easier for me to fabricate last night I realized there was an interference problem. Completely Wrong no interference at all the nut moves with the column, duh! An image of the intermediate design is attached. You can tell that I am not an engineer or draftsman. The drawing does not show the plate on the back of the X-axis or some other details which I know but are not critical to my drawing for my use, not production. Hell, it took me a day and a half to just come up with this, which included, going out to the shop to make measurements, moving things around, and redrawing it as well. I expect the gantry will be pretty rigid as I am giving up some Y travel to spread the carriages out. I also moved the X-axis toward the back of the gantry to centralize the impact of the weight of the motor, so it will be Over the front carriage rather than overhanging in front of the carriage. Unfortunately, even so, adding some twist, that with the current situation I will not be able to absolve, just limit it some. If I keep X rails on the sides, which is not a problem as X is twice as high as Y and there should not be any interference, I will have about 150mm spread between the rails, about twice as much as on Y. I don't think it will matter if I place the rails on the top and bottom of the X-axis or the front side. I think I already am stiffening the X axis by doubling its height and adding the plate to the back side of it to tie the two C-Beams together. Furthermore, I have read that a thin coat of epoxy significantly tightens the fabrication as well.
The more I thought about the columns and the spindle weight. I think that keeping the carriages spaced as they are, but getting rid of the angle to the top on the right and making that perpendicular to the Y-axis will allow me to center or be close to the center of the carriage spacing with the spindle. This will allow me to move the X-axis C-Beams further back. I am "guessing" that this will improve the stiffness of the X-axis in the Y direction by moving the center of gravity within the envelope of the carriages and spindle motor, just guessing. I am sure it can't hurt, but how much it helps, I don't know. As I have yet to design the X/Z system, I am not sure at this time how much the spindle will actually be from the X rails. I think the Z plate that the spindle will be attached to will be about 40mm from the face of the X C-Beams if I place the rails for X on the face of the C-Beams, otherwise, they will be about 20mm or so., right now the spindle center is about 60mm from the Z plate and about 105mm from the X C-Beam. But that is with the Open Builds modules, which will be improved on with the rails. Just some thoughts. Other thoughts, as I go through this are the following. If I decide to place a rail on the top or bottom, or both on the C-Beams I will have to add a mechanism of sorts, that is stiff and adjustable for the carriages to be mounted at 90 degrees to the column or X-axis plate. 1/2" by 12" by 1.5" aluminum is about $65/ft and I would need 3 feet just for Y. I could use a 90-degree angle, which is much cheaper, but I would then opt for steel as it is just stronger for the size, but heavier. 3/16" wall thickness 1-1/2 by 1-1/2 x 6' is only $22 I am thinking 3/15 wall thickness over about a foot of spread is plenty stiff enough for any of the axes, perhaps even overkill. Ken
Sounds like you're generally thinking along the right lines, except: I hope you got them on an amazing deal, because it's gonna be hell on your motors trying to accelerate all that mass back and forth on a little engraving job. I'd have gone with 20mm on Y, 15mm on X and Z. On a wood cutting system, you could probably technically get away with 12mm rails all the way around, though I wouldn't personally. You may have to make your motors bigger simply to help inertial-match the system you've designed. And I hope they come straight, because pushing them around as you tighten the screws isn't gonna be fun, and they may even force the C-Beam to conform rather than the other way around.
My error. The X-axis is 15mm, Y (I still have to order), which currently I am thinking 15mm as well, and Z will have 12mm. As best I can tell with the first set (X)I received, they are true and straight. Sorry for my error earlier. My motors are Hi-Torque motors Nema 23. Considering that the prior machine was all 1" baltic birch and actually had bigger dimensions, than the current design. I am sure my motors will be able to handle the loads. I just finished one of my jobs with the "plastic" wheels and had increased my acceleration profile by about 150% with seemingly no degradation in performance. I expect the carriages should reduce drag, but I may be wrong. The gantry now is probably 25% of the previous weight and moving that previous weight seemed ok. As I think of it, that kind of mass may have been interfering with repeatability for things like inlays, even though my acceleration profile was much lower, as well as the known misshapen aluminum rails they rode on. I checked the rails on my cast iron, and ground table saw surface, and without actually measuring them, I saw no light on the bottom or sides, I am assuming that they are pretty straight and should not cause a problem, I hope! I watched a YouTube of an update of a machine from a [spam] to a [spam] machine which added rails and other items as an update. [spam link removed] I was impressed with their columns, as well as the X/Y interfaces. They looked like 1/2" aluminum and were all predrilled.. I am inquiring if those items may be available for sale as they would help me in not having to make something like them myself. I also would like to purchase some other items such as their end caps and motor mounts. We will see. Again, thanks for your input. Ken
Gotcha. Much more reasonable. They'll look undersize but they should be fine even for the 2.2kW spindle. They're not- you'll likely need a dial indicator to install them if you want them to stay within +/-0.001" whilst being tightened down. Long-body motors tend to be slow motors; you get torque at the expense of higher inductance. Might still be ok though They should- unlike wheels which have to "pinch" and slightly deform to work, rails have little to no stiction and move very freely- just make sure you strip the shipping goop off of them and lubricate properly before putting into service. That's certainly a good start!
I have a dial indicator, .001 increments, so that is a plus. Yes, that could be a problem, I have never run at high speeds because most of my bits were .0472-.0512 in diameter 1.2-1.3mm to get into the smaller areas of the lettering. I could have used a hog bit of 125" but I actually did not think the time and effort for tool changes were worth it. I now know that it was but just didn't bother to rework the V-Carve. Most speeds were 17000 at 15 ipm. I probably could cut my machining time in half. Many of the bits I used were for PCBs and designed for plunging, not side stresses, so faster feeds broke them. As I have more plans, if I ever get this functioning, I will be doing a lot of different types of cuts. I have watched a few youtube presentations on cleaning the rails and the carriages and re-lubing the carriages and saw how much gunk was in them to protect them in shipping. Now I know what steps and materials to ensure they are as good as they should be. My old machine was massive compared to this one, even with the heavy spindle, so I probably no longer need the torque that I used to. But I am not going to invest in motors at this point. The last time I messed with the profile for the motors I was able to increase it significantly, like I said, the mass being moved is probably 25% of what it once was.
On hold right now, will order Y-axis rails end of October and expect them in 3-4 weeks after that. In the meantime, I am going to try and machine the uprights out of aluminum. Probably next week or so.
