Re: Tek Knob Project - scan and make

Kuba Ober

21 dec. 2017 kl. 00:24 skrev stefan_trethan <stefan_trethan@...>:

Thinking about it the smooth surface after vibratory grinding might be
just the ticket for knobs?
Or maybe that solvent polishing that they do on some materials.
Loss of detail would not be critical I think.
I’d have to try that. The grinding medium would need to be small enough to deal with the grooves/knurls in the knob.

Alternatively, for visual consistency’s sake, a knob design “manual” more amenable to printing could be made and then an instrument would be completely retrofitted. The thing to understand is that 3D printed knobs are essentially free if you have a 3D printer. Once a process is set, getting more of them amounts to a few clicks.

Solvent polishing is a simple process that works wonders in ABS and PS. And that’s that, since IIRC no other plastics can be smoothed with solvents you can buy in a hardware store. Or am I mistaken?

Another thing worth considering for the material is plain old PLA. After annealing it becomes a very good performer in terms of softening temperature - better than ABS! And it’s ubiquitous and every printer will print it!

The rotary mechanism would not be too difficult to make, but I would
surely pull out all my hair over the software, no thank you.
I will work on my printer over the holidays to get it set up again.
Normal Marlin firmware will deal with the gcode of course; it doesn’t care much as long as you’re outputting gcode in coordinates of the printer’s axes. The issue is getting the non-Cartesian gcode. Since the knobs have simple shapes, a straightforward parametric gcode generator script can do the whole thing for output in cylindrical coordinates. I like vagrearg’s gcmc for that.

Another simple idea is to generate the 3D model unrolled onto the hub, i.e. using pseudo-cylindrical coordinates directly: X->axial position, Y->circumferential position, Z->radius. Then the gcode output from a slicer can control such a printer directly with no modifications!

If one were to generate the 3D model in true cylindrical coordinates (Y->angle), the slicer output will have to have the flow factor set at the beginning of each layer to a value proportional to the layer number, and the speed set to a value that’s an inverse of the layer number. That’s to accommodate the growing angle->circumferential position scale the radius increases.


On Thu, Dec 21, 2017 at 4:04 AM, Kuba Ober <kuba@...> wrote:

19 dec. 2017 kl. 23:52 skrev Dave Seiter <d.seiter@...>:

That would be an interesting project. I have a Type W plugin that has a split white DC balance knob. IIRC, this is a common problem. I believe there is a similar knob used elsewhere that has the same problem. They should be simple objects to print, but the set screws are really small and I don't have a hex key that will fit.
(It's smaller than .05", which is usually the smallest you see on Tek gear). I was also wondering if it would be possible to print the 7K series backplane connector covers.
I haven't tried printing with ABS yet, although my printer is capable of it.
I’m printing all the way up to high temp materials that print in the 350-400C range. A slow polycarbonate print with thin layers (0.05mm) will look like if it was rough turned :)

But for knobs I might bite the bullet and set up a lathe-like 2DOF mount that could be used to get the nozzle to iron the surface to a presentable smooth texture. It’s not a very complicated implement and save for the stepper, it can be printed! I have the controller board with extra motion channels left precisely for such uses. The DOFs would be rotation along the axis of the knob and tilt so that most any convex surface with off-radial normals can be ironed smooth.

With such a mount it’s also possible to print directly on aluminum inserts!


Join to automatically receive all group messages.