Hello all, First post here. Some background (will try to keep this short). Building an OpenBuilds frame based DIY machine. Its mix of C-beam, off the shelf hardware, a few custom aluminum plates, and HGR20/15 linear rails/slides. I have been running an MPCNC from V1 for 4 years. Loved it, but wanted to upgrade on stiffness. The build has been progressing great so far, super happy with it. I love that there is no plastic or wood to be seen, and feels like a tank compared to what I was coming from. However, I've run into a brick wall with the long (Y) axis. I was originally planning on running 15mm wide GT belts, but the more a researched the more I realized that may not be the best option since I want to cut aluminum/brass on occasion. The Y is 1500mm long C-beam, 1500mm slides, and will give me a 48" travel. I've looked at rack and pinion but the cost and backlash scared me away. Budget is always a concern. I feel like my options are: 1) 15mm belt and deal with it. <$100 solution, but performance issues. 2) 1500mm ballscrews. Don't have a great place to put them, though im sure I can figure it out. moderate cost $200, but I would be pretty much stuck with 16mm diameter due to space constraints and availability, so I'm afraid they will whip. 3) Try to come up with a DIY double belt-on-belt solution. I have the tools and stuff to get it done, but will take time/experimentation/and is a relative unknown in terms of performance. Here is a quick look at the build so far to give a general idea of what I am working with. I was originally going to tuck the belts under the Y carriages on either side, but there wont be enough room under there for screws. Also, I want to enclose this machine to keep dust in, and it isnt much narrower than the table so I want to try and avoid hanging things like motors off the side. I know I have backed myself into a corner here. Not enough planning, too much building. And the "build as you go" incomplete CAD model: Thanks!
This is the best I can come up with for ball screw placement. It doesnt increase overall width, allows the steppers to stay stationary, and is easy to implement. I would probably have to add some dust shielding mounted to the inside of the C-beams (thinking thin plastic or aluminum "L" channel going up and over.
I have spun 1500mm 16mm ball screws at 1000rpm and they were ok, but that is approaching the critical speed: that is, the total screw length was 1500mm so between the fixed and simple bearing was more like 1430mm. The worst-case is when the carriage is all the way at one end (i.e. homed) but in practice the ball screw nut is also about 40mm long, and does sort of act as a support point too, reducing the length that can whip. Although 1605 ballscrews are common, if you can find a 1610 in that length, it can make a huge difference as a 1605 at a max of 1000rpm is only 5m a minute for fast travels.
Thanks for the reply! As a matter of fact, I was going to use 1610s. I have a 1605 on the short axis already and am kicking myself for it. Maybe one day I will swap it out. For now I use a gear ratio on short run of closed gt belt to get the rotation speed up. Hopefully some more opinions come in, but the more I think about it the screws look like the best option.
1605's are great for the Z T8-8 acme thread lead screws are also a lower-cost option of course and can probably be made to fit in your design within your c-beam channel (especially if you turn your C channel around). T8-8 with one end fixed and the other in a floating bearing at 1500mm should do about 3m/min while still being under the critical speed where whipping is likely, and with tensioning, may get up to about 5m/min safely. A lot really hangs on how fast you want to push your rapid moves really.
Just a quick update on this. The 1610 x 1500mm long ballscrews arrived. I just installed them and made one of the plates to connect the nut to the gantry. I used my cordless drill to turn the screw, and even at the highest speed (1750rpm), whipping was pretty minimal, with a little bit when the gantry was at the extremes of the table. I've done the math and this is probably about 2-3x faster than I will spin the screws in practice, so I think it will be ok. On to spindles!
Yeah, for sure. Though I have narrowed it down to either 1.5 or 2.2 kw 220v air cooled unit. I have the power available and really want the versatility of a spindle vs a router....I'm just not sure about the weight. My machine feels solid so far, but I just don't know if it is solid enough to hang 13lbs on the z axis.
For most work, a 1.5kW would be plenty, especially when using small diameter bits for doing either fine work, or to minimise material wastage. How fast you want to move the spindle around is also a really important consideration, with heavier spindles needing more motor torque to accelerate them quickly, and the associated changes that need to go along with bigger motors. Where the larger units are useful is if you want to have a high top speed (e.g. 24k rpm) for the small diameter bits, but also need lower RPM and still have some power available for cutting with much larger diameter tools. Much of my work uses 1/8" and 6mm bits, but I also have 50mm and 63mm diameter face cutters with inserts that I use for milling off slabs of wood for coffee tables. The large milling cutters have a 12mm shaft, so having a spindle with an ER20 collet is important for me. I used to use a 2.2kW spindle that was total overkill for the small bits, but struggled at 10krpm with the 63mm cutter. I have gone to a 3kW spindle now but only so I can mill more easily at 10krpm with the big cutter. Milling aluminium with an 1/8" bit now hardly gets the spindle off cold and it can go for hours, but the spindle weights 10kg even before filling with water! Having to move the extra weight around quickly has not only pushed me to Nema34's, but I am going to have to rebuild the bench the machine sits on as fast acceleration of the 10kg spindle (now attached to 35kg of X gantry) moves the bench in the opposite direction.... My ideal would be something like a much smaller 800W water cooled spindle for up to 1/4" bits, with then an option to switch to a low-speed power head and separate motor for the larger milling tasks, but I have not worked out how to fit that in yet I have an attachment that allows me to mount my palm router horizontally so I can mill the end of long bits of wood; apart from the crazy amount of noise when compared to the spindle, the 'about 700W' router works fine for all but the face-milling tasks.
Thanks for the Info! I don't see myself doing much facing, though I do have a 1" diameter bit on a 1/4" shank that I used to plane the spoilboard of my last machine. However, I do want to cut aluminum and brass, so I am trying to build with that end in mind. I am using NEMA 23s and wouldn't like to go any bigger if it can be helped. So if the 1.5kW would be sufficient for aluminum it might be better to save the weight? I didn't even think to weigh my gantry before installing it. I could probably estimate based on the length of C-beam and the two X-rails....
My go-to for cutting aluminium and brass is a 1/8" single flute, ideally with a DLC coating, but uncoated can work too. I normally use the 1/8" cutter as although my machine can easily handle much larger ones, the material wastage goes up too rapidly for my likeing. For wide pockets, I run at 24krpm and 1500mm/min (250mm/min plunge), with a 1mm depth of cut. For deep slotting I run at 12krpm and 600mm/min (120mm/min plunge) again with a 1mm depth of cut. I cut dry, but with a hefty air blast to clear the chips out. The reason for running slower with slotting is just so I can get the swarf blown out in time for the next pass, especially for really deep slots (i.e. 15mm+, if I am cutting deeper than 20mm, I make the slot wider and use trochoidal milling so there is more space to get the swarf out). Running at 1.5mm per cut is very possible, but I find the tool life reduces quite quickly. I have tried running 2-flute cutters and although they can go at twice the feed rate, I cannot get the chips out fast enough. It was a similar story with coolant in that wd40 or a mist spray made it harder to clear the swarf than with just a strong airblast. Running at the 24krpm has more risk of uncleared chips welding to the cutter, but it makes most use of the spindle available power. With a gantry that is rigid enough to hold a spindle, cutting soft metals at a useful pace should not be a problem.
Thanks once again for the reply. Between this and some other reading I am doing, I am leaning back toward the 1.5kW. Although I really think my gantry is strong enough for the 2.2 (i also found the C-beam deflection calculator that backs up my thinking), the conventional wisdom seems to be that I wont ever really reach its full capability anyway in a machine like this. The weight savings then will only help avoid missed steps and other issues with the motion controls. On my previous machine (V1engineering MPCNC) I did some aluminum and brass successfully, with a far less powerful spindle (~550-600W) on a far less rigid machine. FWIW, I had the same experience you did with a 2-flute 1/8" cutter. the flutes are just too small and the feeds were too slow to not get jammed up.
If its one our machines, those with a single X-axis rail (LEAD1010 etc) we'd not go over 1.5kw. 0.8kw is usually fine for most. With the double X-beams (LEAD1515 and LEAD1010 High-Z you can go with the 0.8-1.5kw air cooled 65mm body spindles. The 80mm body water cooled ones tend to be a bit heavy.
It is not a LEAD machine, but it is a C-beam based custom design. The CAD up in the first post is missing alot of parts, but shows the general idea. Although the X-rail C-beams are oriented in the less stiff vertical orientation as opposed to a LEAD machine, the steel linear rails will help stiffen the beams. Also, there are 4 short, vertical uprights tying the top and bottom beam of my design together which should also cut deflection. There are flat plates at acting as caps to the vertical C-beams and securing to the top X-beam, as well as 8 off-the-shelf flat corner plates tying each x-beam to upright connection to prevent racking. (I should just take an updated photo. once I got far enough in my CAD design to feel confident I just started building and didnt finish the CAD) The flexibility and affordability of the C-beam and extrusions from OpenBuilds helped me work through a design block I had on improving my old machine (or in this case, starting over). However, I wanted to utilize HGR linear blocks/slides along with ball screws on at least X & Y. I will probably use 8mm lead screw for Z still, especially if I end up with the lighter spindle.
In light of Peter's comment, and looking back at your first photo, having the double X-beams does help a lot. One feature in your first photo that is potentially very important for when you add the spindle is the two vertical braces between the beams. They will help to distribute the load, and also, help reduce torsional twist. Short sections of C-beam added between the two rails are fantastic for increasing the overall rigidity if you can find a way to add them in (i.e. adding another one may be worth a try too). When cutting aluminium in particular, if you have deflection due to torsion, then when cutting a pocket that runs to the edge of the material, the bit may 'ride up' a little on the level of the pocket. When the bit goes off the pocket edge and into free-space, the weight of the spindle may cause the tip of the bit to droop relative to the edge of the pocket. When the bit re-engages with the metal, it is now hanging potentially quite a bit lower than the level of the pocket, and can try to take a hefty bite for the first cut back into the pocket..... I broke a few bits on my old machine that had a torsional sag this way. Edit: just seen your post that came in while I was typing!
So here's where I'm at, for reference. Tonight I made the end ball nut connection plate and finished mounting the bearing end blocks fully. Next up is the x-plate, and steppers come tomorrow.
looking good! Your Z looks like it will have a useful size travel to it. Are you making the Z actuator as a classical 'fixed rail, moving block' or more rigid (but more awkward and heavier) 'fixed block, moving rail' design? I notice in the one picture too that the wire on your steppers is not twisted yet: are you planning to change the connectors over and twist the wires to reduce the EMI, or are you going to sleeve them/ wrap them to get the wires as physically close as possible to reduce the current loops radiating? In my first design, I had the 'loose' wires like in the picture too, and had an EMI nightmare with false triggering of limit switches happening really often. Just twisting the wires together to minimise the current loop area slashed the problems dramatically. In the end I went the route of trimming all my wires out of the motor to be much shorter, then twisted the A-/A+ and B-/B+ into pairs, then twisted both 'twists' together into one 'rope' of wires before putting a connector on. I then have sheilded cable running back to the controller.
I've allowed for a minimum of 4" of useful travel in Z. Some of the space seen will of course be consumed by the spoil board, which will sit on some leftover melamine, which will sit on the 2020 cross beams. The Z will be a traditional fixed rail, moving block. The plan is plate on the X blocks, then mounting some extrusion on the plate, then the Z rails to the extrusion, then mounting the tool either directly to the Z blocks, or with another plate. I am considering maybe skipping the extrusion layer in that stack-up, as it will save me almost 1" of projection, but I thought the extrusions would give me mounting flexibility for things like endstops, dust shoes, Z-stepper, etc. Whether or not that extra 20mm makes much difference in terms of moments exerted on the gantry, I don't know. Admittedly, I didn't approach this project super analytically (and I'm an engineer by trade, for shame!). When it comes to passion projects like this I always start out super thoughtful then get too excited and jump in with a "that looks about right" attitude. I suppose if it doesn't work I can just re-configure it down the road. I also thought about fixed blocks moving rails back in the design phase, but it was going to make the attachments overly complicated. I've never had issues with EMI on my MPCNC, and none of the wires in that setup were twisted or otherwise shielded. Of course, that machine didn't use endstops. This one will. I can give them a twist once it comes to installation time.
The 4" travel sounds good. On my old machine, I could just scrape about 3" of actual Z travel, and found it to be ok for actual milling, but I was forever adjusting the spindle up and down in the clamp and an extra inch of travel would have made a world of difference. I have some long-shank indexed cutters for roughing and had one job where I had to raise the spindle in the clamp to give me enough clearance over the work piece just to get the cutter out to change to a smaller bit! Over 4" of travel, the difference between fixed block or fixed rail is probably not going to be noticable, and I agree, working out a simple way of doing fixed-block is a challenge for another day! As you say, the nice thing about building your own CNC is that you can make the parts to try out a different idea Yes the EMI issues I had were with the end-stop cables. Although I had opto-isolators to remove ground loops, I did not have any filtering on the signals initially, and the combination of careful wire layout/screening and then some filtering on the switch lines solved all the problems. I have had two different types of VFD for driving spindles now too. One was surprisingly clean, but the other one is rough from an EMI stand point and I ran a sheilded cable from the VFD to the spindle otherwise the workshop radio was just static (although roughing out makes a racket, when doing finishing cuts on 3D work, the steppers make the most noise overall and I can still hear the radio).
Made my x plate today. I got the opportunity to make it on a CNC mill, so I put two sets of holes in it. One for using 20x100 extrusion to provide mounting for the z stepper (not pictured), the z slides, and future accessories. However I also added holes to mount the slides directly on the x plate. This would save 20mm of projection of the spindle from the gantry. Is it worth it? Or should I soldier on as pictured?
Nice. Mounting directly and saving the 20mm will keep the forces that are acting to twist the gantry as low as possible, but it is a trade-off really for how easy it is then to make the rest of the Z and use the machine (especially when it comes to mounting other 'stuff'). How significant 20mm is also depends on the diameter of the spindle and how far the spindle mount pushes it away from the Z assembly, and then also on the depth to the centre of the X axis that the twist occurs around. I went through many iterations of the same question when I decided to upgrade my machine as I knew I wanted a much heavier spindle. For example, on my old machine, the spindle was 80mm in diameter, and the back of the spindle clamp was 15mm, so no matter how thin I made the Z, the centre of the spindle would always be at least 55mm from the face of the gantry. In the end my Z took another 72mm, so the centre of my spindle was 127mm from the face of the X gantry. The X gantry was 80mm deep, so I had another 40mm to the centre of the X gantry acting as a turning moment. That meant that I always had to live with the 40mm+55mm=95mm as a minimum, even if I could have made the Z paper-thin, but practically had a turning moment arm of 40+72+55=167mm. If I could have removed 20mm of my 72mm of Z actuator, then going from a moment arm of 167mm to 147mm would have reduced the twisting forces which would have been nice, but only by about 14%. On my new machine, the spindle diameter is a larger 100mm, the clamp is again 15mm but that all sits only 15mm from the face of the X gantry (which is 80mm deep again). The turning moment I have now is 100/2+15+15+40=120mm, and adding an extra 20mm would still only be about a 17% increase in the turning moment. When I went from 167mm down to 120mm I saved nearly 40%, so it was a significant change for me; the bigger spindle would actually have sat at 177mm, which would have been over 47% moment arm reduction. A lot depends on the depth of the spindle and mount, and how far the Z is from the centre of the X gantry as they are factors that are often not controllable.
I'm trying the other configuration now. It's not just ease of mounting, it's also order of operations. To make this work, I need to drill a bunch of clearance holes in the extrusion to access the screws that hold the plate to the slides/ballnut. With direct mounting, I was going to mount some 2020 on the plate edges outside of the rails. Turns out I did the math wrong and didn't space things out right. I either have to drill some MORE holes in the plate, or remake it..... As for the spindle, I'm thinking to go with a 65mm unit.
Ah,yes, I find 'hidden' screws are the ones that even with thread lock on them have a habit of working loose too! I have a box of shame under the bench that I call 'offcuts' but are really the plates I messed up
Hey Evan, here's what I came up with. I ended up saving the thickness of the extrusion and mounted the z rails direct to the plate. Had to move them inboard about 10mm, so I could mount some 2020 on the plate edges. While I have to pre-place some of the carriage screws, I can tighten them in place. It makes it hard to mount cause I have turn each screw only a little at a time, but it works. I could probably mill a half circle into the 2020 so I can pass the entire screw. Will do that next time it's apart. Then I mounted a crossbar of 2040 above the plate to give me a place to mount the z motor. This should work for now.
