Atfer Hbot, go to CoreXY.

Hello,

Many time I don't have post there, the last time was to show link, solidworks file, and somes pictures of my own HBot printer.


After 1year with very good result with my printer, some noise appear because the lm8uu don't like moment force. And so, make some groove on my hardened steel rod. But atfer near than 1000hours of printing i think it's "almost" normal. So for me a refurbishment/retrofitting/upkeep of the machine mean change for a better one system

The problem with the HBot, is the rack-force (arc boutement in french), it's not really a big problem. It's only appear, when you want to print parts with some gap (hinge/bearing/...). And the other problem is the dyneema/spectra/fishing line "belt", this solution is cheap and good but I can't get more than 1200mm/S² with my pulley.

So I have draw my own real CoreXY (not hbot) with aluminium part, and Gt2 belt.

Now the most important: This printer can be made with printing part in ABS/PLA, or with "simple" aluminium milling.
I have printed the first prototype in ABS, and after that I milled all the parts in aluminium.

And the teasing:


Tomorrow or for the next week I will post more info, picture, files, etc...

Edit: I have time before going sleep.

My deadly dismounted old Hbot:


All parts have been printed, but have been disassembled after dry mounting and prototype validated for aluminium milling:


Aluminium parts after holes, taps, etc.. Gt2 idler pulley have been made from simple gt2 pulley 8mm bore we can find on the net and 11mm idler pulley have been made with simple aluminium and carbon tubing.
Need many m3x6 Fhc screws, 24x 633ZZ, 8x igus bush (can put lm8uu instead), and some plastic spacer, and printed plastics gt2 belt anchor.


Atfer some screwing, the CoreXY axis is ready and alive:


I know that I need to finish: Dual nozzle mounting or single nozzle mount + Z auto-probe, reorganize the wiring etc... add some bellow on top of the printer.


Sorry for the bad pictures quality, no more battery on my real camera.

Edited 2 time(s). Last edit at 01/20/2014 04:43AM by zelogik.

the issue with the hbot is the forces acting on the horizontal bar are not even so if you dont have a rigid mount the bar will torque and angle it self, however the corexy fixes this by equalizing the forces.

@Aduy: I have seen that the play we can have in bearing with the time is enought to have a less ridig frame.


Now my only concern is how well iglidur bush will last for X/Y axis. Because I have bad review/history with hardened rod vs cheap lm8uu...
And I prefer to change 8pcs of really cheap igludur (100pc at 0.15/pc is cheap for me ) more often and only that, than lm8uu and hardened rod because they have groove/scratch.


So as promise, some best quality pictures:

The complete printer


Corner where the belt crossing


Pulley with gt2 belt path and 11mm flat pulley


You can see the CoreXy path.


I need to make some modification for fixation of nozzle/ Z probe/Air tube etC...


As nobody seem really interested with a "simple" coreXY configuration, I don't spend time for put file on github or anywhere, but if people are interessed, no problem for sharing, pdf/solidworkd file/stl etc... And i think Openscad or inkscape file is really not difficult to make.

You made a very nice printer. I looks rock solid. How is it performing?
I am thinking of building a machine with XY- movements for some time. I was thinking about an ultimaker/tantillus design, but I not not realy like the rotating axis.
The corexy system seems very nice, but I do not have access to a milling machine so the original design is not for me.

It seems like it is possible to build your design with printed parts. Did you actually use the printed parts or was it just for proof of the design.

I would like to have a closer look at the design, so if you are willing to share the design it would be great. Unfortunately I do not have access to solidworks, but if you can export it to a STEP file?

Edited 1 time(s). Last edit at 01/22/2014 02:31AM by jand.

How is performing ... my bowden setup worked pretty well (around 99.9% reliable), my Z system was perfect for my use, and the ramps/rasperry/octoprint/wifi is a perfect setup

And for the news coreXY axis system, it's has only 4-5 hours of dry run without reassembly of the nozzle/Ymin-max contact, air tube etc...
The speed is 350mm/s max, it's the stepper and arduino limit frequency not the frame limit.
Acceleration have been put to 9000mm/s² with 1 stepper motor mounted quickly for test without any problems, so I think I can up the value if needed.

Repeatability will be tested at the end of the week, when the head setup will be finish (nozzle mount/wiring fixation/etc..), and the belt tensionning system will be validated.

FYI I debated a long time between ultimaker/tantillus and hbot or coreXY for my new one.
The problem with ultimaker/tantillus is the rotating and linear movement on the same axis ... I don't like it even is its seem to work well.
Hbot problem is the rack-force, you need to have a really rock solid frame, axis, and with the time bearing take more play and need to change more often bearing/rod...

CoreXY, 2 belts shorten than with Hbot for the same size area. So theoretically less stretching/damping etc...
The 'head' is took with 2+2 points and no more with 1+1 point like Hbot (pull+push)
You can compensate with the perpendicularity of the frame with belt tensionning.


I have tested the printed part running only 2 hours on dry test, look good, but I think you need to make some easy modification for stiffen the assembly (m3 tapping in abs is not so good, need insert or longer bolting system).

For the file I will post, PDF file for milling now, stl for printer, and solidworks file and Step file will be posted this evening at home

The speed and acceleration numbers are impressive. I like it. I know it is still a work in progress, so it might change a bit.
Does the speed/accel include a stepper motor on the carriage?

The more I read about corexy the more convinced I am that my next printer has to be a corexy system.
Thanks for the PDF files. These and the pictures give already a nice overview of the printer.

Keep up the good work and please keep us updated on your progress.

Yes speed and accel have been tested with a stepper motor on the carriage, (last photo on the first post). it's was a simple nema17 40mm longer.

Normally is all go right, I will put the printer ready to print for the Week-end.

For me, the more I use and see my new printer, the more I like it
The only downside is that it's not so cheap with 10 gt2 idler pulley + 2 gt2 motor pulley when I compare with my old Hbot.

This evening I will post step file + solidworks file + stl file If is found all of them ...

This meens I can get reasonable speed even when the extruder is mounted on then carriage.
Looking forward to view the step file. Take the time you need to get the files in a publishable way

I will publish it on github near my old Hbot printer file, I think it's a good place for that.

Yes I hesitate now that I have seen the really good performance, to mount a bowden extruder like on the current setup, or mounting a direct drive extruder on the carriage...

I found your github, so that's fine by me. I noticed you made your Z-stage in inventor. I have access to inventor so I will be able to have a look at that too.

So I have updated all the files on github:
[github.com]

This printer work really well, rigid frame, fast, perfect for 3D printer, and I think it's no bad for light milling.

For the moment I have just printed 2-3 stuff, because the nozzle support is in progress and not really good....

Beautiful machine!

Been playing with a coreXY design for a little while now. Your results are great. Any chance you could post a video of the motion at higher speeds/printing?

---

- akhlut

Just remember - Iterate, Iterate, Iterate!

[myhomelessmind.blogspot.com]

Thanks,

I will try to make a movie, but only when I will be the winner against Marlin or Repetier!


Marlin don't like X and Y max endstop with Hbelt system, I have hardware so i want the feature

And Repetier don't like min feedrate,and my poor "not finished" extruder jams when it's extrude at less than 30mm/s ... So i will play with EXT0_MAX_START_FEEDRATE ...

a special thanks to Repetier for his help.

to Do list:
- Going to put some Dibond plate for make a beautiful enclosure for the printer and big spool holder
- "Betterize" nozzle mount and cooling system.
- improve cable management with cable chain.
- make a bellow for keep heat.
- make a Z-probe system.

Just out of curiosity, is your printer capable of having 4 hot-ends via bowden style extruder along with the z-probe for auto-leveling? Also, what is your printers build volume? Would it be possible to scale your design to have a build volume of 24"x24"x24"? Does your printer have a heated bed? How quiet do you find the dry-tech bearings? Are the Igus bushings better than LM8UU bushings? Any particular reason why you decided to go with GT2 belt/pulley system instead of cable-drive system? Last but not least, roughly how much did your printer cost you in parts/materials?

Sorry for all my questions. I really appreciate you taking the time to reply. Very interested in learning more from you and seeing your printer in action. 350mm/s sounds like a super fast print speed.

How so much questions:

So first:

I'm planning to drawing/making an easy and fast quick change tool head with:
- 1bowden and nozzle + Z auto probe + 2 fan duct (cool nozzle and print for small part)
- 2bowden and nozzle + 2 fan duct (same)
- 1 small cnc cutting motor for pcb milling + ( vacuum tube)
- 1 direct drive motor mount with nozzle. + 2 fan duct (same)

With some good engeneering and small nozzle (like mine that need forced cooling), mounting 4 nozzle is possible, and with good placement 1 Zprobe more, don't seem impossible to mount too.

This printer volume is 230mm x 220mm x 215mm, easily scalable in each axis.

You only need to make a bigger hole (10mm or 12mm) to put a bigger rod if you want extrem longer printer, and for longer distance, just add lenght to your rod and frame
Rod lenght = 20mm aluminium profile + corner mount lenght - 30mm, so I you have a 350mm aluminium profile, you need to have a 320mm rod
And moving distance = rod lenght - 80mm. So with a rod lenght of 320mm you can have 240mm of total travel maximum for each axis (less X/Y min and max switch and some gap)

I like easy math and simple calculation formula ;-)

Yes of course this printer have heated bed, and I just putting some panel for make a total enclosed printer.

For the moment the dry-tech bearing is really interesting, the most annoying noise come from motor and stepper driver frequency, I really need enclosure and motor cork mount.

I have take gt2 pulley to have more accel, with my old system and cable-drive system, the maximum accel was 1200mm/s² before loss grip/step. Now with this system with 9000mm/s² is amazing ;-) and for small cnc milling, have some torque is better.

Depend If you have acces to milling machine or not, It's the same price as i3, and all reprap machine. Now if you don't have milling machine, I thing it's cost about more or less 400€ only for aluminium milling parts.

350mm/s is the limit of the arduino with marlin, I have seen that repetier firmware can be more fast

Regards,

Quote
zelogik
As nobody seem really interested with a "simple" coreXY configuration, I don't spend time for put file on github or anywhere, but if people are interessed, no problem for sharing, pdf/solidworkd file/stl etc... And i think Openscad or inkscape file is really not difficult to make.

Me! me! me!
I am interested in your design... I am looking a HBOT/coreXY design that seem robust, and your is the only one...
I tried to download from Github but the assembly is not complete... Can you share the SW files? or at least update the github with the a finished design?

Thanks in advance!

regarsd

Some news. I have finished to found a good solution for mounting my "old" nozzle. And the first print was an army of t-nut holder.

Cura estimated time: 2h15min, and pronterface response :
"Print ended at: 19:28:57 and took 1:50:51"

Youpi! 20% time less, so printing time took less time than Cura estimation, It's another big victory for me :-D

@akhlut: I don't forget you, I hope to have some time for make and upload a good video showing travel speed/print speed.

@fleming, SW files are ALL on github for X and Y system(corexy system), only the SW assembly file don't work. now I you want, you can make a updated version of SW assembly file, for update the github tree :-D, but for the moment I prefer "finish" my printer, so much work to be done...(post-it for me: Cable fixation, Dual nozzle or Zprobe mount, enclosure, bellow, switch support, Led mount, add harting plug for the power, clean my desk....).

Normally with github file you can make the "structural" part, but I don't have SW file for: pulley assembly and idler, and belt tensioner.

My army of t-nut holder: (from thingiverse, sorry don't remember the name)


Howto make a good tooth belt idler: 10mm spacer, 2* 3x8x4 bearing, cheap 8mm bore and 16tooth(if i remember....) from robotdigg with some filing for make a small pulley, M3 Extra large washer (14mm large), and 2x0.2mm thin spacer printed.


The new bowden nozzle mount for PLA, ABS, Nylon....



Last photo (because I'm proud of it), of my home-made and jams free nozzle

How do you create a jam-free nozzle???

I noticed that you mentioned on your "to-do list" that you are looking to "improve cable management with cable chain", does this mean replacing the current GT2 belt/pulley system with a metal-chain/pulley system?

Also, does anyone here have any experience with the Kraken 4-nozzle hot-end from E3D?

On a complete side note, Happy Chinese New Years to everyone in the RepRap community! Wishing everyone new and innovative ideas to help strengthen the growth of the reprap program, strong personal health, abundance of wealth, and many happy moments in the new year of the Horse!

For a Jams-free full metal nozzle, polish the inner tube, polishing a little more the inside of the tube and polish again, and cooling the barrel with "fresh" air with a 9mm OD, 7mm ID silicone tube.

Yes cooling the "inox tube" with hot air from the heated bed is not a really good idea for me like everybody does.... I prefer take outside air outside the printer with a small DC blower fan and silicone tubing...

No GT2 pulley are really perfect, cable chain management is for the wiring of electrical/air/plastic:
http://commons.wikimedia.org/wiki/File:Cable_drag_chain_sliding-guide_channel.png

DC Fan blower:


You can see a small gap between PTFE sheet and inox tube, just for let some air cooling a little more the small section of the inox pipe.


PushFit connector, Silicone Tube and Events:

This is exactly what I'm working on - pulling fresh air from outside my heated chamber to cool the nozzles. Glad to see the squirrel cage blower works well - my first attempt with PC fans failed miserably.

And I didn't get why you would contain the thermal break inside the X-carriage. But now that I see you're cooling it it makes total sense.

What do you think about scaling the machine? Beefing up to 10mm rods will add significant weight to the moving mass. What steppers and drivers are you using?

---

- akhlut

Just remember - Iterate, Iterate, Iterate!

[myhomelessmind.blogspot.com]

Repetier and Marlin have both a very good feature, "when nozzle temperature is more than 50°C, enable fan" :-D
I have tested a big one 7530 blower fan for put 2 silicone tube... to much pressure, to much noise ... This little $1.54 5015S blower fan from china are amazing :-D


Add 10mm rods will maybe add some weight, but I don't have high torque nema 17, and that printer work really well (no one step lose with 9000mm/s² acceleration).
I think the best think to do (If I want to improve a little more this printer), is to put 10mm rods for Y axis (the axis where X axis move), and lets 8mm or maybe 6mm rod for X axis, where there is only a very light carriage. Anybody for test ? :-)
I have draw this design exactly for that, possibility to change easily rod diameter or build size without affecting the belt path.


So the sorry-bad-quality video...

It's just a travel & print speed test. 4 cube on each corner at 0.1mm layer height, 350mm/s travel, 80mm/s perimeter, 150mm/s infill.
I have stopped the print because of the bad adhesion problem. But I'm impressed by the layer alignement (better than a print I have seen from a 2years old and 25000€ commercial FFF machine from Str****sys)...
It's just for saw speed.

Video & Photo when I have stopped the print

I am looking forward to see if somebody will make my design printer :-)

Hi,

I downloaded the step files and made an assembly. A few parts seem to be missing in the github files. The part for connecting the Y-bars to the frame and the part for connecting the X-bars to the XY-ends. I decided to make all bars 300 mm The Alu extrusions are made to fit. For now it looks like this:



Unfortunately I do not have access to a milling machine so I decided to make the parts more printable.
I started with the XY-block. In the assembly I replaced the left XY-part with my variant.



So it looks like at least one person is going to build your machine.

Github files have been updated and added

It's "guide X X4" and "guide Y X4".

Nice printable XY-block, I like the way you make the part symetrical, just be sure you have 24mm at least between m3 hole for bush/lm8uu holder.

If you need any information, don't hesitate.

Hi,
Stil missing the iddler plate, rood and bearing.
Can you also add the Gt2 pulley?
Thanks in advance.

Regards

@Fleming: The idler plate show on this assembly was a prototype, I have decided to change completely for make a more easy and robust idler system:

Bearings side:


Motor side: (remember for me: Need to down a little bit the motor pulley)


For each idler you need: Plastic or Metal 8.4mm spacer, Simple U aluminium profile 11.5mm, 2x 3x8x4 bearings, FHC m3 screws, bolt, and use the same bolt than the motor mount, m3x40


I don't have draw rod and bearing/gt2 pulley, as all was supply material.
Maybe I will make a more complete material list with all screws/nuts/washer/spacer/pulley ;-)

Need:
-qt:28, 3x8x4 bearing (thanks to cheap china bearing :-D).
-qt: 8, 20teeths gt2 pulley and 8mm bore from robodigg (and no 16teeth as say before), you need to dissassemble it and file it to have 8mm height pulley.
-qt: 4, 8mm rod 320mm lenght for 220/230mm travel distance.
-qt: 4, 11mm od, 8mm id, tube (ie: I have used a 8mm id and 10mm od aluminium tube inside a 11mm od 10mm id carbon tube.) with 8mm height
-qt: 2 16/20 what you want teeth gt2 pulley with 5mm bore for motor.

Edited 1 time(s). Last edit at 02/02/2014 04:40AM by zelogik.

If I am building a similar design machine but want a print area of 24"x24"x24", should I be using 20x40 v-slot aluminum extrusions instead of 20x20 for better stability of the entire frame? Also, what size smooth rods would you recommend for such a big print envelope? Is it better to go smaller and lower the overall moving mass, or would it be smarter to upgrade to something a little thicker so to ensure no slop in the smooth rod? I am thinking of using a Kraken 4-nozzle hot-end that has water-cooling, would I still need to integrate "cool-air" cooling to the hot-end nozzles to ensure no jamming?

Thank you for taking the time to reply to my questions.

@NewKidOnTheBlock: 24", or 600mm is a big printer, mine printer is more designated for build printer with a size between 100mm and 300mm on all axis.

You don't have the same problematic when you design a big printer or a small/medium one.
Nema17 will be to light, because you need 12mm or even 16mm rod diameter.
This printer can take 6 to 12mm rod without any problems and igus bush (only need maximum 15od for the igus bush), just need to change 1 hole size on 2 parts(easy).

I hope the kraken extruder only need water cooling for working well, so don't need cool-air.

If you have chance, try to scale x2 all the parts, I may work with 4040 vslot profile.

Rod will be 16mm, belt will be 12mm width, and all parts can be scaled without problems.
Only for the motor, nema17 have 31mm between screws fixation, nema23 have 47mm and nema 34 have 69.6mm, so you need to change the motor fixation hole.

I hope for you that it's not your first engineering project for build a 3D printer.

Quote
zelogik
@NewKidOnTheBlock: I hope for you that it's not your first engineering project for build a 3D printer.

This is actually my first build project... Hence why I want to get as much information before I purchase any parts.

Would you suggest NEMA23's for all motors if my machine is going to be so big (print volume of 24"x24"x24")?

Would 12mm smooth rod be strong enough to support a 4-nozzle metal-extruder-hotend like the Kraken or is 16mm a must? Is there a 14mm option?

Would 20x20 extrusions build a strong enough frame as to eliminate vibrational errors while printing? Would 40x40 work better towards eliminating vibrational errors while printing? Is there a 30x30 profile available? Where would you suggest a good supplier would be for V-slot aluminum extrusions and for the 12mm belt/pulley system? Is V-slot better than T-slot? Do they even make 12mm GT2 belts? Also, because of the CoreXY design, would the area where the belts cross touch or cause any problems because the belts are so thick?

Thank you for taking the time to help me in this journey.

Edited 1 time(s). Last edit at 02/03/2014 05:20PM by NewKidOnTheBlock.

Just finished watching your machine in action and it looks super fast... Is it just my eyes or does the frame look like it's shaking a little when at higher/faster frequency?

@Newkidontheblock: So have so much question, that only you can response with your experience and calculation etc...
The more easy to you is to draw your project/frame on inventor/solidworks/cao program and simulate deformations, moments, constraints.

For the shaking, it's because I have put too soft vibration dampener for the foot, and it's the complete "tower" that shake.