I'm by no means an expert, but I'm currently trying to design a CO2 laser machine for engraving an item that is 6 metres long. The actual length that I will be engraving along the Y axis is approx 4.5 metres. It's a fairly short travel required in the X axis, this is not a problem. I'm just designing/drawing the thing using CAD software, trying to source components at a reasonable price. The electronics will be a colleagues department. The short X axis of the machine will be running at a high speed and will be belt operated, moving the laser head. The X axis assembly is pretty light weight. Made out of aluminium extrusion (Rexroth), incorporating the laser tube. The X axis assembly sits on a 6 metre long aluminium extrusion frame on SBR20 linear guide rails for the Y axis, which is going to be moving very slowly. My problem is what to use to move the X axis assembly along approx 5 metre travel in the Y. I have a feeling that this is too long for a belt drive, stepping motor setup. I've been looking at rack and pinion drive systems, but I'm concerned about backlash. This has lead me to look at the Nexen RPS roller pinion system, but I'm concerned about the cost, and wish to keep this down. There's not a great amount of load required, but I require accuracy. The machine will be etching a high value item, and it's permanent, so the last thing wanted is any issues caused by slop in the drive system on the Y axis. I've also been looking at regular rack using an anti-backlash spur gear. Sorry if this sounds a bit vague, I'm just basically looking for a linear motion system, for a fairly long travel length, not carrying a lot of load, travelling at a very slow speed, but it has to be accurate. Cost is a factor, but it has to be accurate, if it costs for that, it will have to. Thanks for any help.
We got very excited about roller pinions on this forum 5 years ago, when I knew very little about machine design: ...But you're looking at probably somewhere in the $1500 range for one axis, if the prices have held up. When accuracy matters over speed, positional feedback is usually the way to go over trying to over-engineer a perfect drive system with super low marginal positional error and backlash. Something like the Renishaw Tonic system with a stainless steel tape encoder that you can just stick to your machine somewhere. Or this: High Accuracy Magnetic Linear Tape - magnetic encoding isn't quite as precise and robust as optical, but it is usually much cheaper and at 0.02mm accuracy I'm imagining that's plenty good enough for a laser dot. Reader head's less than $200, tape looks to be in the $400 range. Not super cheap, but allows you to use, say, a $50 drive system and have more peace of mind to boot. You electronic system could then go with "move until I see this point" instead of "move to this point and hopefully that's where you actually are". There's no substitute for closed-loop when looking for accuracy.
Thanks for taking the time to reply Rob, it's much appreciated, it also makes sense. I'm kind of new to all this, when you say a £50 drive system, I take it your referring to belt drive? It's just with that the belt in would be in excess of 10 metres long, unless we use a static belt, which I have seen on some laser cutting machines, although information is pretty sparse when I search for these. I just haven't a clue if a belt as long as this would be feasible, but I'm sure that there are people on here that no a lot more than I do..
You could use a static belt- either looped or as a rack, but with very low feed speeds and relatively low forces you could use a leadscrew (if you can find one that long), a cheap plastic rack and pinion, even a friction drive- just a rubber wheel on the motor shaft bearing on the extrusion with spring tension or something. Maybe a pair of wheels pinching the SBR shaft- think tennis ball machine- or something. Literally any method of pulling the gantry along is fine (assuming you've taken care of racking and other off-axis rigidity problems) when you have a DRO scale mounted to the machine. It would be excruciatingly slow but you could even manually index it for each X pass if something went horribly wrong, because your controller would be telling you where the machine physically is. Handy backup for a can't-fail expensive project. So you COULD go with a fifty quid drive, but that might not be the way to go in reality. Just don't discount "absurd" options when you have zero cutting force and closed-loop positional feedback. Static chain might be the best way to go, all things considered. Advantages of belt- including relatively low price- but less stretchy.
I wonder what the 'static belt' is. Is that when two belts are used, one glued/fixed to the rail, acting as a rack, and the other meshing into it?
According to Google that's the "Bell-Everman belt drive". I'm mildly dubious, but who knows. I'd call a "static belt" any version of a drive system where the belt is fixed in place at each end and doesn't directly get driven by the motor. Meshing, non-meshing, or otherwise. Even a weird and unnecessary contraption like say something where there's two belts and one winds up as the other one unwinds on a motor spool, the belt itself is still static- any given point on the belt, while not being directly interacted with by the motor, corresponds to a fixed position on the machine frame.
