Aleph Mountain

Loveland 3-D printer company Aleph Objects moves into larger facility

http://www.reporterherald.com/business-top-stories/ci_24388678/loveland-3-d-printer-company-aleph-objects-moves

-Jeff

Thanks for sharing this news, it gives additional info to your Fundable page.

So your printer farm is now exclusively made of TAZes now?

The current one in the old location has 8 TAZ, plus 30 AO-100s (with lots of upgrades such as you see on the forum). We also usually have a few other machines running there, mostly TAZ.

The new fleet will be 144 (maybe 128?) TAZ 2 with Beagle Bone Blacks connected to the RAMBos so it can be remotely administered more efficiently. They will each have a webcam, so they can all be viewed quickly. The room for it is expected to be ready by November 11th.

-Jeff

Wow! I hope you’ll post videos once the setup is working. :slight_smile:

What are you guys running on the Beagle Bone Blacks? I’ve been driving my TAZ with OctoPrint on a Raspberry Pi and it works great (for a single machine), but I’m curious if there is something nice on the BBB as well.

  • Adam

I still want to know if the new office is mountain shaped though!

(also I want to see a picture of your current A0-100 fleet to see if there are any upgrade ideas i missed heh)

Nice view of the mountains from AM http://devel.alephobjects.com/aon/photos/more_exterior/ and http://devel.alephobjects.com/aon/photos/exterior_photos/

Here are some pics of the fleet from June http://devel.lulzbot.com/photos/fleet/

The biggest changes that we’ve made to our AO-10x have been enclosing the whole fleet in acrylic for temperature stability (no more pink sheets!), and upgrading to the stabilized x-carriages on all machines.
https://forum.lulzbot.com/t/ao-10x-x-carriage-mod/273/1

If you haven’t switched yet, I highly recommend it; that change dropped our scrap rate by about 30%!

We’ve considered switching to TAZ style X-ends to clean up the sporadic z-wobble issues that pop up, but we’ve found that recalibrating the x-ends solves the problem just as well with way less down time.

-bam

I’ve already seen a huge improvement by upgrading to the AO-101 x-end-clamps. Could you explain what you mean by recalibrating the x-ends?

I’ve been wondering if the problem doesn’t come from the fact the leadscrews are beefier than the smooth rods (3/8-12 vs. 8mm). I noticed that on the TAZ, the smooth rods are 10mm while the leadscrews are M8.

Maybe that should be split in a different topic?

Thanks for posting the questions, i too would like to learn a response on this.

To recalibrate the X end clamps, basically you start from the lower Z motor mount and work your way up. Are the two motor mounts centered? are the two nema 17 motors positioned equidistant from the side of the frame? Is the threaded rod straight? is the threaded rod coupler centered as best it can be, and at the same height on both sides?Is the bearing rod centered and 90 degrees perpendicular to the frame and equidistant from the bearing rod to the threaded rod and from bearing rod to bearing rod throughout the entire range of travel? Is the Bras ACME nut free to move in the X end clamps properly? Are the X end clamps in too far or out too far on the X axis bearing rods? Is the top gantry mount centered and alligned?

If you line up all of those things and ensure the threaded rod and the couplers are as straight as can be arranged and lined up with the Z axis bearing rod, you can minimize Z banding. In my particular case that was not enough. That was partially due to a malformed idler X end clamp that wasn’t letting the Acme nut slide back and forth properly across the entire range, though even after I cleaned that up it still wasn’t working well enough for me, hence the other projects.

What usually happens with Z banding in the AO-101 style chassis, is that the threaded rod turns with the NEMA 17 motors, forcing the ACME nut upwards, and taking the X end clamps with it. If the rod is out of allignment, the Acme nut is supposed to slide in or out towards or away from the Z axis bearing rod, in theory allowing the path of travel of the X carriage to conform to the bearing rod without interrupting the print quality. I find that with the stock AO-101 design there is a bit of a pendulum efect (on a very small scale) that causes the entire X carriage to wobble back and forth in the Y axis direction as it goes up, especially if the Z nut chamber has any issues. The thinner unsupported 8mm bearing rods will tend to deflect slightly and cause the carriage to conform to the ACME threaded rod instead. THicker bearing rods will alleviate some of that, as will adding a spring disconnect to the threaded rod, which is what the TAZ style ends accomplish.

The other cause of banding in the AO-101 is Hysteresis. Gravity pushes down on the X carriage, the threaded rod pushes up on the X carriage, but there is a tiny bit of play between the Acme Nut and the Threads themselves, and depending on the position of the carriage, nut, etc, you can end up going slightly higher or lower than you intended on one or both threaded rod sides, which can cause layers to be slightly thicker or thinner unevenly. One possible fox to that is to use 2 Acme nuts per side, with a spring between them forcing one nut into contact with the upper side of the thread, and the other on the lower side. Assuming your threaded rod is accurate, that arrangement forces the ACME nuts into allignment and doesn’t allow further variance.

Placing the bearing anchor points further apart, increasing the diameter of the bearing rod, adding disconnect springs and anti hysterasis features all help eliminate that movement.

One additional measure that I haven’t experemented with yet personally, is constraining the upper end of the threaded rod to force it to maintain center on the NEMA motor shaft. It’s a controversial subject, and you can find supporters on either side of the argument. A full scale high end CNC machine will generally have the threaded rod leadscrew supported at either end, but those machines also use very expensive very precise leadscrews that are usually just as accurate and alligned as the bearing rails. In that case the leadscrew is not fighting with the bearings

On the Mendelmax, the rods used on the earliest models were thinner and less accurate than the ACME screws on the AO-100, and tended to have more deflection, so tieing them down top and bottom tended to make banding worse apperently. I plan on trying a supported end at some point, but I haven’t got that far yet.

Awesome, thanks for the information, gives me a few more things to consider.

So with a stock AO-100, do you recommend i go ahead and print the TAZ style X-ends as well as the upgraded (lower) Z-axis motor mounts and forego the X-end clamps 2.0 from the AO-101?

Well, I am somewhat biased since I designed the TAZ style X ends for the AO-10x series. For me, it was a definite improvement over the AO-101 style lower X clamps. That being said, I don’t think I could have printed my AO-TAZ ends without having those lowered X ends in place first since the wobble without them was even worse.

The Taz style ends involve alot of plastic, alot of printing with support and clean up, and heat set inserts. It is definitly a technically challenging project. I think the results are worth it, especially now that there is the Linear rail project you can upgrade to later after that, but I also think you can get very good results, though maybe not quite as accurate through careful allignment and calibration.

I understand, and thanks for your work.