Hi all, original SM printer owner here, just used it yesterday to make some nice little parts. I quite love the thing. Near-zero complaints.
So in real life I’m a laser physicist and optomechanical engineer at a small laser/lidar company in Colorado. I’m thinking about adapting an F350 machine, to act as a sort of ‘x/y gantry’ at the heart of a custom optical component alignment setup; i.e., we would remove the 3D printer head and replace it with one or more very custom modules that would have degrees-of-freedom appropriate for holding (vacuum chucks probably, ‘tweezers’, etc.) and manipulating numerous very small optics in the system that are eventually UV- or other adhesive-set into the assembly permanently. My question is whether anyone at SnapMaker, or otherwise, have any specifications and/or data on the Snapmaker x/y/z stages’ spatial resolution and accuracy? I figure they’re surely ‘quite darn good’ or I wouldn’t routinely get such nice prints from my machine, but hopefully someone might have actual numbers I could ponder? Thanks–C. Hale, Lafayette CO
If you need precision you’re going to likely fight against some of the design issues of the carriage internals. In my opinion you should be looking at a machine built around linear rails which will achieve far higher precision.
Those numbers posted by Snapmaker are not wrong, but they are optimistic. Many people have issues achieving that. My machine operates at that, and I’ve only had to rebuild 2 linear modules, replace the bed and carriage, and enable backlash compensation.
Thanks, Brent113. I hear you on the caveats you note, thanks. The thing is so cheap, though, it truly might be worth a try. I expect that the ‘holding/manipulating modules’ that we’ll need to spend serious design labor/time developing, would be largely transferable to a more sophisticated ‘x/y gantry’ if need be (and/or I could make sure they are!) The dimensional tolerances on placing all these small optics (mirrors, beam splitters, lenses, waveplates, and so on) will be ‘pretty high’, but not ‘crazy high’, I’d say. If the stages can come pretty close to such behavior as listed in that January forum post, I do believe we’d probably be happy. It’s an interesting thing to consider for this job. --Charley
As long as your expectations have been managed, then you’ll likely be satisfied.
If you don’t mind humoring me, what will control the custom toolhead? Controlling automatically via gcode or USB commands is not impossible, but the few people who have attempted have not gotten very far.
You will need to keep a toolhead attached to the machine or it won’t boot, so the 3DP or some other toolhead will need to be set aside, but still connected to the controller.
Totally good question, and you’re getting me to think about this a little more : ) So, this is not a production-tooling type situation, quite the opposite in fact, a true one-off thing (it’s a prototype instrument being developed on a NASA SBIR effort, if you’re familiar with those). There’s way more emphasis, at this early development stage at least, on extreme compactness and alignment ruggedness in the face of operating temperature variation and vibration and shock, than there is ‘how easy it is to build’. That said, I think our custom heads would probably be ‘man-in-the-loop’, manually-actuated stages and actuators. Thorlabs’s OEM piezo motors, with basically ‘joysticks’ for controlling them, come to mind. Cheap but multiple video cameras/optics to look really closely at the given optic/assembly being placed and manipulated, will be key I think. To some degree I’m basically envisioning a sort of ‘poor mans hexapod’ or Stewart platform sort of functionality, but cheaper (within the limited budget of the SBIR program) and more geared to a one-off, not-production type situation.
Yea, that all makes sense, and I think it would work.
That being said, I have worked with other Marlin, GRBL, and Repetier based machines and there’s something to be said about the simplicity. Everything is driven by a single controller. If you need to add or modify functionality there’s only 1 place to look.
If you ultimately deciding you need to extend functionality of this system, you will need to a) abandon the snapmaker philosophy of distributed microcontrollers speaking via CAN and add direct functionality to the controller with a new custom cable (the connectors are not off the shelf available) or b) use the base module firmware from here and build an all new set of hardware to drive it.
A similar project on a RAMPS board would be just grab one more wire, throw it in a screw down terminal, and write some code.
Brent, do you know of accuracy and precision values for the SM stages out there anywhere? I might simply jig my existing 1st-gen SM machine up with a tenth-reading dial indicator I’ve got, and probably glean some useful real-world values that way (and on ~ 3 yr old, considerably-used stages, too). Again, I think the ‘basic/coarse x/y gantry’ functionality doesn’t necessarily have to be spectacularly good; the small, very short-range linear and rotational stages that we would glom onto the unit would be the ones that would have pretty darn good resolution and accuracy. The SM x/y stages would mainly just get you fundamentally to the right place for placing the next optic, is the nominal idea…
Sorry, can you clarify this? By ‘stages’ are you referring to the linear modules? And for SM is that the original that you have?
I guess if you’re asking for the same data list I shared earlier, but for the original instead of the 2.0, I haven’t seen one. I would imaging the numbers would be similar as the underlying hardware is similar in theory.
Sorry, yes, by ‘stages’ I mean the linear modules that move the SM in x/y/z. They do seem pretty impressive to me, so far, at least in terms of printing impressively decent parts. (I’ve yet to delve into the laser head or the CNC spindle head. I’m knee-deep on a Tormach 770M machine at present, and the other SM modules fell by the wayside at least for now) Yes, my SM machine is their original, small guy. If the values you directed me to earlier are for the latest/greatest linear modules, that’s good, as I’d buy a new F350 machine for this for sure. I’ll keep my little guy at home where it’s safe : )