I'm working on a project that requires programmable motion of a 32" screen within a roughly 4m long by 2m high vertical plane (the dashed rectangle in the attached image). I'm exploring a few concepts but the one shown would utilize a winch attached to a moveable cart on a long horizontal rail. Horizontal positioning would be done via the horizontal track (driven by a belt or pulley), vertical positioning via the winch. moveable screen About a decade ago I built a somewhat similar system using a Parker linear slide ... but the cost of that is way above what the budget allows here. I'm trying to do this for under $5k (excluding the screen and media player). Looking at Open Builds, the available rails are relatively short. I'm wondering if such a 4m long system is something I could achieve? Any other thoughts about the proposed approach (in terms of the winch, sensors, etc.)? Edit: Apologies but I can't get the schematic image to show inline. The link above works though.
for 4m rails have a look at Hilaldam sliding door systems (or similar). for motion, Clearpath servos if you want speed, otherwise high power NEMA23's will do it.
Thanks David, that looks like a nice solution for the slide. Would you suggest "rolling my own" belt drive and position detection system? On a previous project we experimented with a magnetic-encoded strip for the position detection. We didn't have the tolerances though for the magnetic reader and ended up using a wire draw encoder. I guess here we could simply home on a limit switch at one end and gauge position through stepper counts, yes? This assumes though no belt slippage and 4m is quite a distance. I should have mentioned that ideally we're after mm precision since this is about double the size of individual pixels.
homing will give you good repeatabilty, and belts do not slip if implemented properly. look at the OX, the belt wraps around nearly half of the pulley, which is what you need. if you need very high acceleration you will need more wrap and maybe a wider belt. A stock OX with belts will give you much less than 0.1mm per step, and with decent belt tension you will get that quite reliably, so I think you can easily put a pixel where you want it. You could also use rack and pinion in place of the belt. There are some OX builds that use it, just search these forums. Ordinary steppers will probably do the job just fine (through step counts), after all 100's of thousands of commercial routers are doing it every day. Clearpath servos are just steppers with position feedback and some very nice software that figures out the tuning parameters for you to give very reliable and fast operation. Expensive though, so it all depends on just how critical your application is.
