1

Topic: HBOT printer (or YAHP?)

Hey everyone. When I was waiting for my SD3 to ship, I was busy purchasing all the fun stuff to modify it with. GT2 belts and pulleys, DRV8825 drivers and motherboard/LCD, LM8UU bearings and hardened shafts, you know.

So what does one do when they get their SD3 and the biggest fluke happens? (This fluke being that the unit actually prints without banding, perfect circles, etc). I have all these parts that will maybe marginally improve my print experience. So instead of ripping a good printer apart, I build a new one.

http://imageshack.com/a/img703/5070/1zj9.jpg

So I purchased $55 in extruded 10 series and went to town.

http://imageshack.com/a/img826/3982/k1yr.jpg

I have used HBOTs at work for pick-and-place projects, and I like them. The idea that your motor is not attached to the moving parts is so invaluable. I also did some things that most people try to fix on their Solidoodles. For instance, removing weight and backlash from the X and Y axis for more precise movement and cooler stepper motors. I even switched to a Bowden setup. One of the main design goals is to get a really fast machine put together. For instance, here is the first test print at 180mm (3 times faster than stock SD3 speeds):

Here is the untouched output:

http://imageshack.com/a/img706/1765/v823.jpg

I have been finding that Bowden setups take alot more attention to get right. The Octi I printed at 60mm/s I had tuned the retraction to get rid of hairs and perimeter blobs, but any changes (especially speed) to other settings needs a re-tune of retraction. I am sure I will be able to get this dialed in better as time goes by. Of course, it means even more slicer profiles (uggh).

So far, here are the pros and cons of this printer:

PROS
-Perfect circles, there is a single belt tension and it is not that picky about how tight to get it.
-Stepper mounts are 'oval'ed, making belt tensioning really easy.
-Bowden setup has steppers running at 0.8 Amps and corners look decent since there is not alot of inertia to overcome for fast turns.
-The total X carriage weight including hotend is less than 40 grams. Its less than the E3D without fan.
-MK2A heatbed and 40W cartridge heater - very fast heating times.
-GT2 belts and 6MM Z axis screw: Steps in firmware are super easy. Everything falls on whole numbers.
-I got a 0.35mm nozzle hotend, and due to that I adjusted standard extrusion width down to 0.37mm and 0.25mm layer height. Its nice to see 4 layers per mm.
-LM8UU and LM12UU bearings are nice.
-I modified a Wades geared extruder assembly for 1.75mm filament and a NEMA23 stepper motor that has alot of low voltage torque. Its a heavier motor, but since it is a Bowden setup I dont care. This gives 39:11 gearing on the extruder so I can keep 1/16 stepping and I dont have the slightest moire effects.
-X axis has LM8UU on one rod, and uses 623ZZ bearings on the other rod as followers (wish SD did that instead of delrin blocks).

CONS
-Wades extruder uses the infamous hobbed bolt. Its temperamental. Combined with a cheap (less than $30) hotend, it leads to increased torque which leads to occasional stripping. I would like to either design a dual toothed extruder or something.
-Bowden setup has its issues getting just the right retraction. You can move the filament upto 2mm before it begins building pressure, or I should say that you need to retract 2mm to relieve the pressure. This number is not static. Changing speed and sometimes the type of print job you have means changing the retraction value. And it has to be fast (40-60mm/s).
-While the HBOT is super cool, there is a problem with the X gantry remaining square to the Y gantry. Even with the linear bearings, you can 'twist' the gantry. So far, I have not seen a problem as the single tension in the single belt holds it in place. Really what this type of printer design needs is linear blocks for the Y axis (super expensive) to hold it true. Or, a COREXY where you have 2 belts, but I dont want to go back to a 2 belt system.
-I tried a 4-screw Z axis system where a closed belt turned 4 screws to move the bed platform. While this was great when it came to leveling the bed, I found out that this created a Z-wobble on steroids. I printed out a cylinder and it looked like a screw. I trashed the whole thing and installed your basic Z axis (with all of its problems). I still need to work on this.
-Extruded T-slot does not play nice with printed plastic corners. No matter which direction I put the layers, any stress will begin tearing the layers apart. The entire machine actually shifts (as in, looking form the top you can push on the corners and bend it into a diamond shape). I think I will be getting some metal corners to hold the frame together better.

Conclusions: End goal is to get 0.1mm layer prints at 300+mm/s speeds. Retraction is constantly a problem. I know I need a quality hotend, but I dont want  the weight of the E3D. I already have plans of making my own hotend, but not right now.

I am also planning a modification to the LISA SIMPSON delta robot. I may work on this more than tweeking the HBOT over the next month.

Opinions?

Chuck Bittner is a quadriplegic gamer who is petitioning the major console developers to include internal button remapping in all console games. You can help.
Sign Chuck Bittners petition

2 (edited by adrian 2014-01-11 01:15:52)

Re: HBOT printer (or YAHP?)

nice job mate.

Some random suggestions:
You can ditch the wades concept and direct-drive the filament.. checkout the AirTripper v3 Extruder  - uses the same setup as the direct-drive on the original SD.. You just need to make sure you have a torquey enoguh motor.. a 1.7amp job or something...

Or you could use a BullDog XL and that would give you the 5.2:1 gear reduction (and thus torque improvement, critical on a bowden!) and still avoid the whole wades extruder... it also would allow you to run 3mm filament if you desired (as it has the required torque for 3mm in bowden.. thats a lot more plastic to push than 1.75mm wink ).

You could also fix the entire torsion issue by using some v-slot as the Y-rails... using some wide makerslide you would end up with solid y-ends that will not move if the makerslide is attached firmly... the delrin v-slot wheels would have to explode/deform for it to shift from perpendicular to the y-axis...

For your end pieces - pay attention to print orientation. If you rotate the piece it will be possible to print with the filament lines going perpendicular to the t-slot (thats how I print all the accoutrements that go onto the makerslid t-slots on my Ords). If you leave it orientated such that the filament 'threads' are parrallel to the t-slots, then yeah, it will fail along the thread lengths.

And I would also dismiss any issues with the E3D.. stock hotend is 21g's, a J-head is 17g's and a E3D is only ~40g's.
The 20g's will be the least of your concern in terms of weight... many other things will come into play loooong before the extra 20g's on the hotend causes any actual problems... When people talk about shedding weight from the hotend, they mean the 300g's of the stepper, not the 20g's of a hotend smile

Anyway - just some thoughts....

3 (edited by Hazer 2014-01-14 04:20:51)

Re: HBOT printer (or YAHP?)

adrian wrote:

nice job mate.

Some random suggestions:
You can ditch the wades concept and direct-drive the filament.. checkout the AirTripper v3 Extruder  - uses the same setup as the direct-drive on the original SD.. You just need to make sure you have a torquey enoguh motor.. a 1.7amp job or something...

Or you could use a BullDog XL and that would give you the 5.2:1 gear reduction (and thus torque improvement, critical on a bowden!) and still avoid the whole wades extruder... it also would allow you to run 3mm filament if you desired (as it has the required torque for 3mm in bowden.. thats a lot more plastic to push than 1.75mm wink ).

Thanks for the tips, that was what I was looking for. I am OK on the Wades, my extrusion is actually not bad, but I do know that a hobbed bolt has a bad tendency to strip quickly when there is a problem. I have a direct drive wheel (actually 2) that I would like to switch to. I have been looking at a printed planetary gear setup that I think I will do. This gives me the same gearing as the bulldog, but I can mount a NEAM23 on one end and have printed NEAM17 holes on the output end, thus making it easy to print out any generic direct extruder print to mount to and use the better toothed filament drive. As for the motor, I am using a NEMA23 that has more torque than most NEAM17 motors but runs at a nice cool 0.7 Amps. Its only drawback is its weight and size, which means nothing on a Bowden setup. I did have to make my own models though since all the others were for NEMA17 specs.


adrian wrote:

For your end pieces - pay attention to print orientation. If you rotate the piece it will be possible to print with the filament lines going perpendicular to the t-slot (thats how I print all the accoutrements that go onto the makerslid t-slots on my Ords). If you leave it orientated such that the filament 'threads' are parrallel to the t-slots, then yeah, it will fail along the thread lengths.

And I would also dismiss any issues with the E3D.. stock hotend is 21g's, a J-head is 17g's and a E3D is only ~40g's.
The 20g's will be the least of your concern in terms of weight... many other things will come into play loooong before the extra 20g's on the hotend causes any actual problems... When people talk about shedding weight from the hotend, they mean the 300g's of the stepper, not the 20g's of a hotend smile

Anyway - just some thoughts....

Thanks for that. I should have been on the ball better since I have already redesigned a few of the pieces for exactly that reason. Here is the new and functional corner brackets:

http://imageshack.com/a/img836/8636/67l2.jpg

And for comparison, here is a couple of prints: The one on the left is a magnet clip I printed on my SD3 with all of its glorious moire. The middle is the new bracket printed at 60mm/s on the HBOT, and the one on the right is the bracket printed at 180mm/s on the HBOT. The only bad thing about removing the moire is seeing the Z banding. Threadless ballscrew is probably going to be another modification at the top of the list.

http://imageshack.com/a/img802/8897/yt0h.jpg

Oh and thanks about the E3D. I will most likely get one soon.

Chuck Bittner is a quadriplegic gamer who is petitioning the major console developers to include internal button remapping in all console games. You can help.
Sign Chuck Bittners petition

4

Re: HBOT printer (or YAHP?)

Beautiful work Hazer...
I'd like to hear more about what went wrong with running the Z with four separate screws.  I've been working on a design that incorporated that same idea.  Although I want to do it for different reasons, you have me rethinking my thought process.

Printit Industries Model 8.10 fully enclosed CoreXY, Chamber heat
3-SD3's & a Workbench all fully enclosed, RH-Slic3r Win7pro, E3D V6, Volcano & Cyclops Hot End
SSR/500W AC Heated Glass Bed, Linear bearings on SS rods. Direct Drive Y-axis, BulldogXL
Thanks to all for your contributions

5

Re: HBOT printer (or YAHP?)

It creates a perfect Z wobble. My single walled cylinder test print looks more like an ACME screw.

PROS:
-easy bed leveling
-virtually no banding since the 4 threaded rods would normalize each other. Every Z movement is averaged between rods.
-no need to level the bed even if there was a printhead crash.

CONS:
-tensioning of the 4 pulleys to eliminate backlash also 'bowed' the threaded rods, even when using bearings to hold them straight
-Unless you press-fit the pulleys to the threaded rods (this includes threading them which I did) any slight offset from the machining of the hole, slight angle change of tapping the hole, or pushing the rod out-of-center when tightening your set screws will create Z wobble that translates to all 4 rods.

The only way to eliminate the Z wobble with this design IMO is to use threaded rod thick enough to not flex under great stress, and machine the pulleys on professional equipment (which if I had available I would not be printing parts in the first place).

The simple 2 shaft axis that Solidoodle uses is easy and reliable. Although I did spread apart the distance on my machine to stop the sideways movement. Currently my bed has the ability to tilt downwards without twisting sideways. It can handle a print collision just as well as my SD3, but instead of the print shifting sideways on itself, the bed bounces back into position like a diving board. This is due to using alot of thin cross sections to give the Bed base support without printing a solid object.

Chuck Bittner is a quadriplegic gamer who is petitioning the major console developers to include internal button remapping in all console games. You can help.
Sign Chuck Bittners petition

6

Re: HBOT printer (or YAHP?)

Thanks for the detailed reply.  It sounds like I have some re-designing to do.

Printit Industries Model 8.10 fully enclosed CoreXY, Chamber heat
3-SD3's & a Workbench all fully enclosed, RH-Slic3r Win7pro, E3D V6, Volcano & Cyclops Hot End
SSR/500W AC Heated Glass Bed, Linear bearings on SS rods. Direct Drive Y-axis, BulldogXL
Thanks to all for your contributions

7

Re: HBOT printer (or YAHP?)

Update: tweaked a few more settings. I also decided to play. The idea with this is to get speed. I wanted to see how 0.1mm prints would look ....... at 180mm/s print speed.

Possible NSFW:

http://www.youtube.com/watch?v=D0ki5LIvK70

This model has been shrunk to 50% original size. Slicer settings are 180mm/s on everything. 4 perimeters with 0.34mm extrusion width (its a 0.3 nozzle). 0.1mm layer height with 10% infill every second layer. I brought the cooling section down to 5 seconds. You can see the first layer without infill getting slowed down. 580 layers to finish this, and it took 1:15. In the video, you can see my camera was set for low quality capture. Also, the stills at the end of the video got pixelated from the youtube conversion. The mp4 I made was much better quality.

Anyway, I printed the model on my SD3 for comparison at regular 60mm/s speed. In the following pictures, SD3 is on the left, HBOT on the right:

http://imageshack.com/a/img194/6456/9q4w.jpg

Second photo is attached. It has more of the front view. There are some slight artifacts on the HBOT one. A problem I see here is that the extruder fan shroud is way too big, and doesnt reach the print often enough for something this small. Overhangs begin curling and at these speeds, causes a blob to get pushed around. I will redesign and print out a new shroud.

As for Z artifacts, it matches the quality of the SD3. I do plan to upgrade to threadless ballscrew soon.

EDIT: I just finished another print at 300mm/s with 2 second minimum layer time. While the print time was cut in half, I did not see a noticeable speed difference than what is seen in the video. I think something else is holding back the speed. I do not know if the curved shapes in this print slow thing down alot. I have seen videos of other printers set at these speed and they seem to move alot faster than what I see here. I guess I will do a larger structural print soon to see if the speed looks faster with straight lines. Oh, and the print looked bad in three places (overhangs). I really need a better fan shroud.

Post's attachments

20140123_211421.jpg
20140123_211421.jpg 1.12 mb, file has never been downloaded. 

You don't have the permssions to download the attachments of this post.
Chuck Bittner is a quadriplegic gamer who is petitioning the major console developers to include internal button remapping in all console games. You can help.
Sign Chuck Bittners petition

8

Re: HBOT printer (or YAHP?)

hello would you be willing to share your printer files with me?

9

Re: HBOT printer (or YAHP?)

How much do you have into building something similar? I want to try for a 12x12x12 build area if possible but give me the cost of your unit.

10

Re: HBOT printer (or YAHP?)

Stevos758 wrote:

How much do you have into building something similar? I want to try for a 12x12x12 build area if possible but give me the cost of your unit.

My first iteration was a 20x20x14. I built the X-Y of it and quickly realized the problem with the X assembly shifting out of square on the Y rods. You can easily build one that big, but your going to have to use either true linear square beams, or what adrian suggested by using v-slot carriages. With a 8x8 build area, torsion is not an issue even at 300mm/s speeds. But I have yet to tweek good quality at those speeds. I am having issues with curling on small objects and I need better fan ducting. I also have RAMPS 1.4 with an MK2A bed, and going any bigger than 8x8 will require a unique solution. My suggestion at that point would be using metal brackets for joingin the frame, MIC6 for a bed plate, and 120V band heaters bolted directly to the MIC6 and an SSR to drive it.

Anyway here is the current sketchup file. I will post a zip of stls later.

http://www.mediafire.com/download/87bs9 … HBOT2B.skp

Chuck Bittner is a quadriplegic gamer who is petitioning the major console developers to include internal button remapping in all console games. You can help.
Sign Chuck Bittners petition

11

Re: HBOT printer (or YAHP?)

Hazer wrote:
Stevos758 wrote:

How much do you have into building something similar? I want to try for a 12x12x12 build area if possible but give me the cost of your unit.

My first iteration was a 20x20x14. I built the X-Y of it and quickly realized the problem with the X assembly shifting out of square on the Y rods. You can easily build one that big, but your going to have to use either true linear square beams, or what adrian suggested by using v-slot carriages. With a 8x8 build area, torsion is not an issue even at 300mm/s speeds. But I have yet to tweek good quality at those speeds. I am having issues with curling on small objects and I need better fan ducting. I also have RAMPS 1.4 with an MK2A bed, and going any bigger than 8x8 will require a unique solution. My suggestion at that point would be using metal brackets for joingin the frame, MIC6 for a bed plate, and 120V band heaters bolted directly to the MIC6 and an SSR to drive it.

Anyway here is the current sketchup file. I will post a zip of stls later.

http://www.mediafire.com/download/87bs9 … HBOT2B.skp

I know this is a dated post but is there any update on this?
thanks