LulzBot TK0

Have you got any example prints at 250mm/s 0.1mm layers? What acceleration do you use? The speed I can believe, but I'd like to see how minimal a resonance pattern you are getting!

My printer will happily move that fast, but to keep prints nice at even say 100mm/s I have to have quite low acceleration, and I'm trying to figure out where your machine gains it's performance advantage.

how much faster is the TK0 compared to something like a Stratasys Mojo? or even a Stratasys Dimension?

sls2go Wrote:
-------------------------------------------------------
> how much faster is the TK0 compared to something
> like a Stratasys Mojo? or even a Stratasys
> Dimension?

I believe all the commercial printers are pretty slow compared to what many people get out of their repraps.

I'm curious why you use such oversized threaded and smooth rod for this design. wouldn't 1/4" ( or 7mm) be better suited to such a small footprint design? And wouldn't the $$ savings be better spent on a precision rod rather than general purpose ACME to reduce backlash?

xiando Wrote:
-------------------------------------------------------
> I'm curious why you use such oversized threaded
> and smooth rod for this design. wouldn't 1/4" ( or
> 7mm) be better suited to such a small footprint
> design? And wouldn't the $$ savings be better
> spent on a precision rod rather than general
> purpose ACME to reduce backlash?

Over 500mm the deflection in standard reprap 8mm smooth rods is quite significant just due to static loads (mass of bed and extruder) , the deflection is proportional to the rod length squared so you are significantly worse off than with 300mm spans. With my printer, the 500x8mm smooth rods on my bed are noticeably not stiff enough when I press down on the bed, although I seem to get away with it for printing, at some point I will probably move to something thicker. I moved from 8 to 12mm on my z-axis and the results are much better.

Ah...thank you for the explanation. I'm a bit surprised, but maybe I shouldn't be, and that would be compounded by dynamic loading if the print speed is high.

I presume the same can be said for the threaded rod as well? you've presumably experienced some sort of impulse driven flexure due to discontinuous accelerations (directional reversals)?

Are you still using the ~40-ish oz-in motors or have you had to bump up to the heavier duty steppers (like are usually recommended for the extruder) to accommodate the increased mass and inertia of the larger threaded rod?

I just use M6 threaded rod with 40kg/cm steppers. I don't see why the lulzbot would need to use anything beefier. I think it has proper leadscrews for precision.

Oh, ok. I wouldn't think so either, but I thought that I'd read somewhere that the spec for this design called out 10mm diameter, general purpose trapezoidal screw, hence the question... thanks for the clarification.

Lots of questions piled up, I'll try to get to them all:

@kakaroto: Ya, dual extruder would be rad. It won't have it when released, but perhaps down the road. The AO-100/101 had a nice dual system done by a user (sjkelly), so maybe the same thing will happen here. For the TK-0 it should be even easier.

> at least the stl for printing the new head adpater

I'm not quite sure what you mean there. All the STLs should be up. If something is missing, let me know.

Tracy is COO of Vergent. We hired them to do research on improving our printer from the factory's perspective (design for manufacture) and they came back with the TK-0. But the TK-0 isn't wholly new, as it obviously borrows lots from the past (mendelmax, wade's extruder, the list of names could go on & on). It was also a very back & forth process: they'd show us something and we'd say "foo" and/or go build it, we designed some of the parts that they then used, etc. etc. In sum, we had experience running machines and they had experience building things in a factory and the TK-0 is the result of that. We both own rights to the files, which we long ago placed under the GPL. We are at present working with three other contract manufacturers doing various aspects of the machine and the next generation. I'm not sure what exactly they are doing, but I know they used it internally quite a bit as well (e.g. building prototypes for their other customers, printing parts to fix their other fancy expensive machines, etc). I don't think they will be doing a separate version though, except maybe in house.

@ambiguousphoton: We will be selling them fully assembled before we do kits. We may offer some of the harder to source parts first (e.g. extrusion cut/tapped to certain length, igus parts, etc).

@willingc: When it is ready? That said, factories are working on it right now. But there is still a long ways between that and having something in a box in a warehouse ready to ship. We don't sell things we don't have available in the warehouse. We only take money for things we actually have.

I think we calculated making 32 TK-0 sets/week on 40 printers printing 24/7. That should be a conservative estimate.

Get an AO-101 now, print a TK-0, and give the AO-101 to your local HS.

@roygpa: we've switched to igus bushings now. They are way quieter than the metal LM10UUs. We switched because they are quieter. They are also similar to what we had in the AO-100/101, so we know they work well.

[www.igus.com]

@roygpa: awesome job, btw.

Edited 1 time(s). Last edit at 03/04/2013 10:09PM by jebba.

@konwiddak: no samples at the moment, other than random shots from the TK-0 dev page. Seth recently turned down jerk quite a bit I think, and that really helps with the resonance.

@sls2go: no idea. I don't really follow the proprietary printers nor have I ever used one.

@xiando: we are using M6 threaded rod not ACME. We are using the same NEMA 17s we've been using for a couple years now or so, they are great. I dare not switch manufacturers even. We are using nice Misumi 10mm smooth rods, but they are spendy.

-Jeff

jebba Wrote: @xiando: we are using M6 threaded rod not ACME. > -Jeff


DOH! That's what I get for starting my build before the final BOM is out.

Anybody need a couple pieces of 3/8 inch by 12 acme threaded rod?

Should be an easy change over.


Roy

Hey now ... don't blame it on me... I read a similar BOM, which is why I was asking about the use of under-performing standard-grade, oversized 10mm diameter trapezoidal threaded rod (the value I saw in the lulzbot BOM I viewed) .

BTW, just for the record, many distributors and manufacturers still refer to metric trapezoidal as "metric ACME" , (I suppose that's since ACME was invented first and metric followed afterwards) which is why I originally referred to it (admittedly erroneously) as metric ACME.

@jebba: Thanks for the info. Between your info and Roy's on printing time for a TK-0 kit, I'm working with the assumption that 2 weeks would be a reasonable estimate to print the parts on the AO-101. Struggling a little with sourcing and building a less expensive open source reprap versus supporting your company's strong commitment to open source sw/hw. I think that your company's commitment to open and "transparent" design positions you well for the technical engineering education market and enlightened engineering design teams since you are not locking one into a proprietary product and an expensive road to obsolescence. I could see these markets riding with you up the learning curve of rapid innovation since there is incentive for collaboration and info sharing. Send me a PM if you have any thoughts of how to drop my start up cost with the AO-101. Thanks

@roygpa: I laughed out loud at your comment about a final BOM and threaded rod availability. Pretty cool to be building it now

Just so nobody thinks I am a total freeloader using the Luzlbot designs... I have purchased 3 of their Budaschnozzles over the course of the past year.

I like 'em.

Roy

Roy,

I think you are awesome Thanks for sharing your knowledge about Lulzbot, their designs, and their products.

It's folks like you that take the time to answer others and share your positive experiences that make open source what it is. I doubt I would even be thinking about teaching these high school students how to build a 3D printer without the valuable info in this community. I can teach them the EE/software stuff and following design/build docs, but their willingness to learn from the community and give back is the greater lesson that I am hoping they will learn.

Thanks for your answers so far. I am pretty certain there are some Lulzbot purchases in my near future

Cheers!
Carol

@jebba: Thanks for all the interesting info and backstory
What I asked about having the STL, I was referring to an STL for hooking two extruders to it. As far as I know, the only printing head STL available only allows one extruder to be mounted on it. A dual head mount STL would be cool so we could just buy the extruder, print the new head, and mount it there.
Of course, this would only be possible once dual extruder support is working and tested, and I meant this as a "until you get a 'dual extrusion kit' on sale on your store".
So don't worry, nothing is missing at the moment, and there's nothing for you to do until Tk-0 is released and you start (or not) working ona dual extrusion head.
Thanks again.

FYI, the first draft of a dual head extruder mount for a TK-0 was printed today (or maybe over the weekend). It doesn't quite work yet, but it is getting closer. There's really not that much to it. It us going front/back instead of side-to-side as the TK-0 has more space on Y to put one.

This looks like some really great work--I can't wait to see the final design!

I especially liked the presentation file and the stated design goals. I have been pondering some designs of my own with some similar goals. Specifically, switchable/modular head design (extruder/laser/mill/etc.), swappable build platform (since conveyor belt designs seem inherently flawed, IMO), and possibly dual extrusion. The goal of a small home factory is something I find quite exciting!

One major point that differs is this: I'd like to avoid moving the build platform back and forth. Perpendicular to gravity only (Z-axis) movement is a point where I might relent, but ideally, I'd like to keep the build platform stationary. My reasoning is this: I've currently got a Printrbot Plus with Super-Z (possibly a dumb upgrade purchase, I freely admit) and it seems to me that the bed flying around is the cause of jerking printed parts loose quite often (hair spray helps a LOT). I've been intrigued by the Rostock Delta design for this very reason (stationary build platform), but I feel that design suffers due to height issues and is probably not going to do well with swappable heads or multiple extruders (but it does look like a fantastic single extruder design).

The counterpoint to this goal however, is I'd also like to keep the amount of moving mass (and thus any problems related to inertia) to a minimum. Still, it seems like if it is designed properly, at least the moving mass will be known, whereas moving the bed around leads to the moving mass being a variable function of the part being printed.

Anyway, now that I've gotten that out of the way, my question is this: have you considered using a bowden tube extruder design? If so, what are your thoughts on it?

(I've read hysteresis is a concern with bowden tube extruder designs, but I wonder if it might be possible to compensate for this in the slicing software, e.g. Slic3r, as this seems to be a way to reduce the moving mass.)

-James (Leorex/SirGeekALot)

---

1st working printer: Printrbot Plus v1 (built from kit, not pre-assembled) with Super-Z upgrade.
2nd printer currently under construction: Printrbot Plus v2 frame, with RAMPS electronics and TrinityLabs.com Magma hot end.

@jebba:
Great, I can't wait to see the final results.
It would be great if you could upload a video showing various heads, like Tracy had done. But it seems he disabled his youtube account and all videos are gone.
This also means that your link for the tutorial on PEM nuts here [devel.lulzbot.com] is now invalid. So I suggest someone from lulzbot reuploads the tutorial and updates the link.

@leorex: We've built Bowden extruders before but we never liked them since we could never perfect retraction. I'm kind of surprised there are so many people using Bowdens, actually. I do notice that the Bowden machines frequently use vases for sample prints.

@kakaroto: Ya, the text on that page is really showing its age. I'll go through and clean it up. Just last night I did the first print on a TK-0 dual extruder. I didn't really like the setup, but it is a start.

I built a clone of the TK0 to suit an MDF frame instead of the aluminium extrusions.

It uses NEMA17 motors, 8mm stainless rods, 8mm threaded rod for Z and some 8mm aluminium rod for the Y axis "stands".

I'm pretty happy with the prints, although I was more worried about dimensional accuracy than appearance in tuning the settings.

The parts are in OpenSCAD here:

[github.com]

- some of the parts need some minor fixes
- there are no X or Y carriage parts yet (I just used some craftwood pieces with LM8UU holders)
- no endstop mounts
- there are some templates for drilling holes which I will upload later

If anyone is interested in building one I can do some rough instructions and a BOM, maybe STLs.

darkpaw Wrote:
-------------------------------------------------------

>
> If anyone is interested in building one I can do
> some rough instructions and a BOM, maybe STLs.


here

looks very nice your mixed design.
bom & stl. would be great.

@darkpaw: Nice! Thanks for uploading the files.

No worries

I'll have to think of a better name for it now I suppose.

I will do a wiki page and upload STLs etc over the long weekend.

Some general build instructions would be great.

I am thinking about trying to put one of these together. It will be my first printer so I am trying to read up as much as possible in hopes of being able to tackle everything by the time the parts are ordered.

Is there any interest in a rough solidworks model/assembly from anyone else? I am not very familiar with the scad models that seem to be the preferred method in the reprap community so I am thinking about trying to pull something together in Solidworks. Hopefully if I can assemble it in Solidworks I will have a general idea of where everything goes when I do it in real life.

There is a FreeCAD file that shows it all assembled:

[devel.lulzbot.com]

How is the heated bed going to be insulated from Y drive?
What is the design for bed leveling?

X-Y plane connection to Z plane seems wimpy ( like Ordbot)
How to make / keep Z axis perpendicular to X-Y?
Just may be better with large square base. (portable model + Fixed model

There seems to be NO way to adjust belt tension and keep it tensioned properly?
How are the belts attached to slider -- how are they kept tight?

Can I get in on Beta build?

Tnx
Cozmicray

The BOM calls for 10x500mm Misumi smooth rods, which are hardened stainless steel. Misumi's site prices that part number (SSFHR10-500, IIRC) at ~$35.50, but your BOM refers to them as ~$15-18 if I remember correctly. Am I looking at the wrong part number, or did you guys get a massive volume discount?

BTW, currently printing the parts for the TK-0, I gotta say I'm pretty stoked

Thanks!

bradym Wrote:
-------------------------------------------------------
> The BOM calls for 10x500mm Misumi smooth rods,
> which are hardened stainless steel. Misumi's site
> prices that part number (SSFHR10-500, IIRC) at
> ~$35.50, but your BOM refers to them as ~$15-18 if
> I remember correctly. Am I looking at the wrong
> part number, or did you guys get a massive volume
> discount?
>
> BTW, currently printing the parts for the TK-0, I
> gotta say I'm pretty stoked
>
> Thanks!


In another thread there was someone building a TK0 and asking about the rods and lead screws. Link

I am planning to use a PN: 6112K47 from McMaster for my build. They are hardened ground steal (not stainless) 10mm X 500 for about $15 instead of $35.

When will a kit be available? Beta builds? cost?
Is there a Beta build forum?
Really like all the data out there. Much better than TrinityLabs A1.