Profile cover photo
Profile photo
Preston Bannister
133 followers
133 followers
About
Posts

This is going to come a bit out of left-field...

To be clear, I am a software guy, speculating about appropriate hardware, so adjust your expectations accordingly. :)

The core problem of a 3D printer is four dimensions of motion (X, Y, X and extrusion). Even a cheap 8-bit controller CPU can twiddle four stepper-drivers, mostly fast enough for present-generation 3D printers.

On the upside there are outfits like Trinamic who come up with up with stepper-drivers that offer more smarts, when controlling a single stepper. Though most folk most of the time are not using most of the features.

What bothers me, as an algorithms guy, is that in model a 3D printer is a singular, simple state machine. In practice we have a slow CPU controlling each stepper individually via each distinct stepper driver. That means skew, in the motion of steppers. Maybe small enough we do not care. Maybe.

In my day job, recently joined an outfit that builds advanced experimental radars for the military. They use a couple of large/fast FPGAs to do very fast signal processing. As my present exercise, re-writing the Linux device driver for controlling the FPGAs. (The box has a bunch of high-end Xeons for more complex processing.)

So ... learning about FPGAs. :)

Wondering ...
Is the usual partitioning of the control logic for 3D printers ... wrong?

Older generation FPGAs are pretty cheap. An old/cheap FPGA looks more than adequate for controlling motion in a 3D printer (or other like robots...).

Need to bump up the small voltages (~1-2 volts) to the relative massive voltages required to drive stepper motors (~`24 volts and an Amp or so). Seems the requires simple "driver" chips. (For jargon, called up my father, whose work spanned from the first integrated circuits, up to current cell phone chips, ten years back.)

Maybe the optimal partitioning is a simple FPGA (for exact and integrated motion) hooked to a bank of simple "drivers" (for needed power)?

Maybe "smart" stepper driver chips are ... silly?

(Watching videos about FPGAs. Seems they freak out hardware folk. As an algorithms guy, least bothered by the part that most bothers hardware folk.)

An FPGA is programmable hardware. Cool. :)

Give me a board with a Xilinx Spartan-6 and adjustable drivers sufficient to power stepper motors, and I suspect we have an optimal motion controller.

(Or maybe something simpler/cheaper.)

Are there hardware folk in this group, who can comment?

Has use of Chromebooks in secure environments been sorted?

Started a job recently at an outfit that makes experimental radars for the military. That means our development systems are air-gapped, and entirely separate from those systems (Windows desktops) connected to the Internet.

As a bit of a security geek, I would rather have Google's team managing security. They have massive in-house expertise.

I know that Google and Amazon went through substantial exercise to ensure secure cloud services to government agencies. But did this go far enough to offer pre-packaged services to defense contractors?

Is anyone else having trouble getting orders from +MakerGeeks? Been wanting this US maker of filament to succeed, but near to losing hope.

I see my August Geek Box was charged on the 17th. As the July box has not yet arrived, this is not good. Also two other orders are not yet fulfilled, and no response from two emails to their support address.

Missing in order of billing date:
July 12 - missing 4 from 6 spools of Crystal PLA ordered.
July 14 - missing 1 spool of HTPLA Raptor ordered.
July 17 - missing the July Geek Box (2 spools?).
August 17 - missing the August Geek Box (though not really expected, as yet).

Might be time to go to my credit card company, and dispute charges. Not wanting to do this. :(

Post has attachment
Experiments with annealing PLA. Not a new subject, but going at this a bit differently. Started on accident - left a print in the trunk of my car, parked in full sun in Summer, in southern California, and - of course - it was after a bit warped.

Baked another of the same print in a 180F (convection) oven, to sad effect. (The orange rocket fighter-like things.)

To go at this more methodically, popped a delicate single/thin walled print into the oven for an hour at 140F. No apparent change. Another hour at 150, no change. Hour at 160 ... and visible change (more opaque?), but no obvious change in shape. Hour each at 170, 180, 190, 200, 210, 220, 230, 240, 250, 260F with no apparent change.

At 270F, the first sign of failure - a small shiny spot near the nose, that grew slightly a 280 and 290F.

Starting to look like we can make temperature-resistant more delicate prints. Would want to repeat this exercise, to take less time.

Pictures are most recent, first.
PhotoPhotoPhotoPhotoPhoto
8/7/18
5 Photos - View album

Again, about extruders, which are more properly called "feeders"...

This started as I had fabricated a filament-out sensor, but felt this should be an integrated part of the filament feeder. Also I am starting to suspect the Bowden feeder on my printer with a 0.8mm nozzle on a V6 hotend might just be running out of torque.

(And I have not yet added the "Volcano" kit.)

So I need a better feeder, with an integrated filament-out sensor.

If you are buying stuff off the shelf, buy a Bondwell "extruder" and just deal with the not-integrated filament sensor.

My interest is in something better.

Digging back, I found two interesting examples.

J-Max
http://www.j-max.fr/?p=632
https://www.youtube.com/watch?v=LU_pFnTfH9M
https://www.thingiverse.com/thing:1579394

B2B Extruder
https://www.thingiverse.com/thing:2289203
https://www.youtube.com/watch?v=JjQ7FrQe8s4

In both cases, a belted (geared) feeder, that could potentially generate a lot of torque with a smaller (and lighter) motor.

Wondering if I can design something more compact, using the existing examples as inspiration...

Post has attachment
Was not going to do this exercise, but saw yet another repetition of "plastic bearings do not work" (and one too many).

Our desktop FDM printers bear light loads and move fairly slowly, compared to the more demanding world of machine tools. We do need precision, but are otherwise undemanding of bearing design.

The (otherwise interesting) video at the point in question:
https://youtu.be/7x-3ruD9Kl0?t=4m6s

At this point, was more(!) than a bit annoyed, as the design offered is not suited for making with a 3D FDM printer. We should use designs suited to the materials and method of manufacture.

Posted an example Y motion bearing to Thingiverse, with link to the design in OnShape. The example is not meant to fit any specific printer.
https://www.thingiverse.com/thing:3026175

I am not inventing anything, here. These are old ideas from examples I saw very long ago (though probably not made of plastic). Too far back to recall the source.

Most 3D FDM printers are simply not accurate enough to fabricate bearings that call for close tolerance. But you can achieve close tolerance in the end with careful design.

There are two design tricks here.

First, there are three lines of contact along the rod. Three points is sufficient to force exact placement of the rod within the bearing. As the two upper points are offset, the weight of the build plate is also forcing exact placement.

Second, the slot/plug in the bottom allows fabrication with less precision. When assembled, a hand clamp (or any substantial weight) forces the plug tightly against the rod. The bolt locks the plug in place, but does not press against the rod. This is how we achieve very tight tolerances, when our method of fabrication is less precise.

Note the design is for a 10mm rod. The test rod (savaged from an old printer) is actually 9.5mm. The bearing works fine as it is tolerant of small difference in dimension.

I soaked these in a light spray oil (something generic from the local hardware store, that also claims to contain Teflon and Silicone). The slightly porous nature of FDM prints is of advantage, in this case.

Tried both longer and shorter variants. The longer had more static friction, and I do not see any advantage, so uploaded only the shorter. There is no detectable "play" at all in this bearing, and it slides with little effort.
PhotoPhotoPhoto
7/29/18
3 Photos - View album

About extruders...

Specifically the part that grips filament and forces it into the hotend. Off the shelf, the Bondtech extruder (should it be called a feeder?) seems to be pretty much the best thing out there. The E3D folk have their own equivalent (I have a Titan Aero), though with spotty history.

The flood of cheap-Chinese (mk8?) extruder/feeders are ... not bad, with some failings.

Adding a filament-out sensor to my (cheap-Chinese) printers, and realized I could not just feed filament through the funky extruder / feeder without a lot of manual fiddling.

So what is better?

I know there were more designs, before the massive Chinese wave.

If I want a fully enclosed filament path, and feeding without fiddling, are there better designs?

Post has attachment
An attempt at a filament sensor...

First, clearly I do not know how to spec a limit switch. Got limit switches off Amazon by searching on "limit switch" and buying a couple likely looking popular choices (not having other criteria).

The below is meant to be a minimalist filament-out sensor. The end goal is to detect if filament is physically present, not visual appearance. So a mechanical switch makes more sense than optical, to my mind.

Teflon tube protects moisture-sensitive filaments. Spring in switch presses roller against filament. Hole in side allows visual inspection, and serves as drill guide for putting a notch into the tube for the roller. (Exposure to air only at that notch.) Twist fitting to clamp tube. Adjustment screw to set spacing of switch.

Seemed workable, in theory.

Problem is, this switch needs about 2mm of extra travel to open once closed, and close once open. This extra travel makes sense for many uses, but not so much with 1.75mm filament.

Need to find an appropriate limit switch...

... Just turned around to find a print head waving around in the air, as filament ran out while typing the above. :/
Photo

Post has attachment
Diagnosing a suddenly troublesome hotend. This is a basic mk10(?) hotend, with a PTFE tube far into the heated area.

Had a long print fail about half way through, with plastic simply not extruding. Cleared the plastic out of the hotend without trouble. (But felt a bit of initial resistance pushing out the melted plastic with a metal tool?) Manually pushed plastic through the hotend w/o trouble (a few tens of mm). Started the print. Some plastic was extruded, then stopped.

Repeated the above, twice. Suspect the bit PTFE nearest the melt zone is ... sticky? Marked and measured the filament pulled from the hotend, as a check.

Waiting for the hotend to cool, so I replace that bit of PTFE tube.

Make sense?

Photo
Photo
7/16/18
2 Photos - View album

Post has attachment
Some past and present examples. Bit limited at present with a CoreXY printer 300mm^3 print volume (or a bit more, sometimes). Building a printer with 400mm^3 volume, and meant to be fast. Past that I am of the opinion that slicing up models and using multiple printers is probably more effective.

PhotoPhotoPhotoPhoto
7/12/18
4 Photos - View album
Wait while more posts are being loaded