My own H-bot with Fishing line concept.

Hello,

I have design and draw my own H-Bot 3d printer.
The overall dimension is 390x340x340mm for a build size of 245x200x200mm.

So for some characteristics:

-Bosch-rexroth 20x20 aluminium profile for the frame.
-Fishing line breaded Dyneema, with cheap Vgroove bearing bought on Ebay.
-And Two toothed pullet from an Qu-bd extruder. (they don't grip on ABS/PLA filament but are very good for fishing line)
-Some LM8uu, and 8mm rod.
-The "ACME" F8 screw for Z was from Zentoolworks.com.
-Homememade Dual bowden extruder, with homemade mk7 clone nozzle and for the ABS feeder it's from: "http://www.thingiverse.com/thing:34757" with many modifications for add bushing/bearing/spring tensioner, with and Arcol 5mm mini hyena.

All plastics parts have been draw be me on Solidworks for the X/Y h-bot design, and on Inventor for the Z, all "specials" parts was draw with Freecad.

A big Thanks for crispy1 for the help to found the firmware bug in marlin with a H-bot design. And to MakeDave for the Very Compact MK7 upgrade feeder. And to my 'old' 3d printer I have made on a zentoolworks frame.


The video link:
H-bot Fishing line design

If you want more details/file/pictures I will be happy to send to you/put there :-D

ZeLogik.

Edit: On the video it's my second print, the first one was the same Yoda bust but not scaled to 250%, with Green ABS filament. (no counting calibration object printing)

Edited 2 time(s). Last edit at 04/06/2013 02:58PM by zelogik.

This is really really awesome.

I bet that think can really fly, HBot with a bowden extruder having very little moving mass.

Are you worried about the pulleys wearing the UHMWPE braid? (I suppose UHMW is super wear resistant...)
Would be interested in releasing the files on thingiverse?
What belt-path did you use? CoreXY? Or convention H-Bot?

Again, congratulations - this is an awesome printer, and you have taken on many battles such as dual-extrusion, and H-Bot all in one go, for your first machine!

Thanks you very much :-)

The thinks I like the most is the ratio overall size and total build size.



For the moment the dyneema breaded fishing line are OK. But there is only 30hours of printing + 40hours of dry movement.

For releasing all files no problems at all, I just need time for learn git process, I don't like the way thingiverse goes...
Or another options for releasing some solidworks, inventor, stl...

for the moment the Z is my main issue, to much noise when moving.... Constrainted...

It looks very nice.

Github is really easy to use. I'm sure many here (myself included) would like to see the files.

Between an H-bot and a BerryBot I'm going to be quite busy this year... But first I need to finish my small CNC table.

Too many projects.

---

- akhlut

Just remember - Iterate, Iterate, Iterate!

[myhomelessmind.blogspot.com]

OK I will try to look for github the next week.

And its an hbot, not a core xy, I need to found the advantage of corexy instead of simple h-bot.

My first Tantillus still has its original 65lb spectra cables and has been running for a year and a month. It has worn out a LM8uu but not the cable.

---

FFF Settings Calculator	Gcode post processors	Geometric Object Deposition Tool Blog
Tantillus.org	Mini Printable Lathe	How NOT to install a Pololu driver

I have only 50lb, error when I have ordered.. So i need to change it maybe.

@sublime: Yes thank you (for idea and tantillus) or maybe not( for the time I have passed since 3months)... Your tantillus is incredible, just the build area is too small for me....

well done. nice big yoda printing there.

zelogik Wrote:
-------------------------------------------------------
> And its an hbot, not a core xy, I need to found
> the advantage of corexy instead of simple h-bot.

I think the point of the corexy crossover setup is to counteract racking forces.

Awesome work. Could you post some pics of finished prints?

Nice machine, looks excellent! Any idea how much tension you have on the dyneema line?

@billyzelsnack: yes i have understand that, but I not sure the gain you can maybe add with the "racking force" is better when you see the complexity/misalignement when you cross the belt/line...

@maddox: Too much for the first test, after a printing of 20hours long for the Big Yoda, the dyneema line have break. The dyneema (clone), was to small, (50lb/0.36mm), and there was a misalignement of the geared pulley.


So image of the first print without to much calibration of the extruder, speed fill 100mm/s, speed perimeter between 40 and 60mm/s and 0.12 or 0.15 layer (don't remember)





The Big yoda with the failure near 60%... at 150mm/s infill and 80mm/s perimeter at 0.18mm layer.:



So i need to change:
- To a more strong dyneema, I waiting for a 100lb/0.5mm real dyneema but i think 50lb with less tightening is enought finally... (i have really tensionning too much...)
- And the geared pulley is maybe to aggressive from qu-bd, I waiting for cheap mk7 aluminium pulley from reprapworld.


- The Z plate(heated bed) have to much gap in x y, so I need to modify the system spring/bolt style...
- Maybe change almost completely Z axis, 8mm rod is to small... to for 12mm rod
- Need to add more strong spring. (the Big yoda is heavy ...)

- Put an enclosure for keeping heat ...
- Change the kind of cooling on the nozzle (when the enclosure will be there), I have made an water cooling system made with copper/"bronze" block, I have a very cheap water pump ... but don't have so much time... Or put a external fan with hose ...
-continu to clean all differents files I have when I have done the R&D process for put to GitHub .... (Approx 1kg of part for the R&D process.)
- Negotiating with my near "perfect" wife, because now than the machine is almost done and working she just a little less patient, and it's most difficult for me to found good argument :-(

ive got a water cooled hot end from ebay its just been great, coming from the qu-bd it was a major upgrade.

So as promise...

Github folder,

[github.com]

-So there are some solidworks files (2012) for XY axe with stl, and for some accessoires.
-Inventor 2012 files for "Z axe" and stl.


And yes git rocks on Linux ;-)

hey so ive got a new idea for an hbot configuration. i find it very hard to get belts that are over 4 meters long, so i have come up with my own version of the hbot that uses two continueous belts and one regular belt, instead of one huge belt. heres a sort of cad model of it, tell me what you think.

That's an interesting idea, but you should look at the core X/Y mechanics it ends up with two open belts of half the length give or take.

---

___________________________________________________________________________

My blog [3dprinterhell.blogspot.com]

yeah my problem with the core xy is it uses so much belt, and extra pulleys too, for which there is no place for on my printer.

@aduy: Yes the concept is interesting but I think you need to be carefull with gap and backlash. (more pulley ...)


ie: I have changed the dyneema with real dyneema 100lb and I made a dry run since more than 50hours without any problems for the movement in X/Y axis, I thinks the problem was too much tension and misalignement of pulley the last time..

Any pics of the "Two toothed pulley from an Qu-bd extruder" that you're using?

ZeLogik,
Where can I find the bugfix for the Marlin H-bot firmware that crispy1 found?
Thanks

[github.com]

This version working well, but i don't think there are all news features from the last marlin

I was looking in the firmware last night, and corexy is part of the latest build of marlin, just search for corexy in the configuration.h file, then comment out the one line.

Yes i have try marlin two week ago, but there is a strange bug in x/y when you try to print non symetric parts.

And somebody know the max frequency pulse/step atrduino with pololu can reach?
I can't find anywhere good answers.
Because i think the "bloquage" at 200mm/s i have can be that.

10.5mm pulley, 1/16 ustep, 1.8deg motor, if my math are correct, i'm near to 20khz pulse for 200mm/s.... I maybe need to change to 1/8 ustep but i will have only 0.02mm precision.....

yeah 20KHz is close to the limit.
If you use Repetier, by default he does two steps/interrupt with an option for 4. but with 10mm pulleys you'll be limited for top speeds without going down to 8uSteps or even 4.
The bug I had with CoreXY in Marlin was it didn't correctly enable both steppers for single axis moves. Just doing a tiny jog in each direction was enough to work around it.

---

___________________________________________________________________________

My blog [3dprinterhell.blogspot.com]

Quote

The bug I had with CoreXY in Marlin was it didn't correctly enable both steppers for single axis moves. Just doing a tiny jog in each direction was enough to work around it.

There was a worse bug that caused layers with high numbers of segments to be slightly offset in one direction or another. This would lead to "tilted" parts and drove me crazy looking for the source of slip that didn't exist.

@crispy: Have you tried to push your patch for this worse bug to the main repository of Marlin?

Quote

@crispy: Have you tried to push your patch for this worse bug to the main repository of Marlin?

I didn't write the patch, I just raised the original issue. I don't think it's been pushed to mainline Marlin yet. If you search through the marlin issues you can probably find the issues and the changes associated with them.

Today I have made some change, DRV8825 from Reprapdiscount, and 1/8 microstep, so the theorical precision is 0.02mm (not so bad for my use), and the maximum speed before the firmware/electronic bottleneck is 500mm/s :-D

And I have changed the arcol v2 drive gear for a mk7 clone from ebay for the filament, because I have somes problems with the arcol ...

I think 500mm/s is not so bad with fishing line :-D. it's more 512mm/s but I prefer to round the value.

With what I have seen today if I remove some precision for go to 0.04 mm with 1/4µstep, I think the machine can go more fast.

does it go that much faster with lower micro stepping? i always thought it smoothed everything out and made it faster, but i guess at high speeds thats not the case.