Design questions

Hi !

I'm new to this forum, and also fairly new to the world of 3D printing.

I'm planning to build my own custom 3D printer with my proper ideas and needs. I'm searching a particularly strong and stable design and the best precision possible. The speed doesn't matter and I'm thinking about a build volume of about 30x30x30 cm.

So, I decided to go with the "XZ head Y bed" mechanism (mainly for simplicity) and Openbuilds V-Slot as motion and frame system. However I'm still hesitating about some basic points :

• the drivens (is it the right word ?) : belts or lead screws ? I will certainly use lead screws for the Z axis, for quite obvious reasons (precision and stability). But for the other axis, I was wondering if the precision and the stability provided by the lead screws are worth it, despite the loss of speed and the rise of the final cost.

• the frame design : there are a lot of frame design possibilities with the "XZ head Y bed" mechanism, like the MendelMax 2.0, the Prusa i3, the Lulzbot Taz, the Lautr3k, etc. Which one would be the best (in my case the strongest and the more stable) ?

What do you suggest regarding these questions ?

Thanks in advance !

This page has photos of my own custom built approximately 30x30x30 cm printer, MegaMax.

[drive.google.com]

Print quality is pretty good now (see the photos of the snakebite extruder parts on the above linked page, and my thingiverse items here: http://www.thingiverse.com/search/page:1?q=digital_dentist&sa=), but I can still find flaws in the prints that I attribute to slop in the drive/positioning mechanism. The photos show belt drive in all 3 axes, however, I am currently working on converting the Y axis to screw drive and replacing the round guide rails and bearings with linear guide rails and bearing blocks in both X and Y axes. This is being done in an effort to squeeze the last possible bit of print quality out of the machine. I have no idea if it will actually improve when I am done. I am attacking the Y axis with the screw drive because it has high moving mass and each time the movement starts, stops, and reverses, the belt stretches a little, leaving ripples in the surface of prints near the corners. I am also getting a lot of vibration and noise in that axis that has been resisting all attempts to ameliorate it.

Any design can be made strong, but you have to use the right materials in the right way. I recommend against laser cut plywood. It's hard to make a truly rigid structure and over time it will warp and delaminate. If you want rigidity and stability over time, use steel or aluminum bolted/welded to more of the same. Minimize the use of wood and plastic parts in the structure of the machine. Plan on building the machine into a box to control temperature- it will prevent warping and delamination of ABS prints. That means design the machine with the box always in mind and put the electronics outside the box. If you can, put motors outside the box, too. I added a box as an afterthought and have not yet moved the electronics out.

One thing I can recommend it that you use one motor and a belt to drive as many screws as you will use in the Z axis. Using separate motors is cheaper, but you will curse every time one of the motors gets out of sync with the others, tilting the X axis boom or the print bed. It is also easier to make adjustments on the machine when you can raise/lower the X axis or bed by simply pulling on the belt, knowing it won't cause any alignment problems later.

Quote
the_digital_dentist
motors gets out of sync

What do you attribute two motors getting out of sync too?

Tks.

@the_digital_dentist : Thank's for your answer !

Quote
the_digital_dentist
Any design can be made strong, but you have to use the right materials in the right way.

So, you think that a kind of Lulzbot Taz design can be made as strong as a Lautr3k design, using the appropriate materials ? I'm asking that because, for me, the Lautr3k design seems a bit stronger, with its triangle structure.

Otherwise, any other opinions ?

I have been using a two motor Z axis for a long time, without any sync problem.
Both motors attached to the same motor driver.
If the Z-Axis gantry gets wonky -- a couple of turns of the leadscrew on one side,
with motor power OFF aligns the gantry with the bed.
I would think leadscrews locked together with belts are much harder to adjust.
Pulling on a belt --- I don't think there are designed to do that.

Quote

the_digital_dentist
One thing I can recommend it that you use one motor and a belt to drive as many screws as you will use in the Z axis. Using separate motors is cheaper, but you will curse every time one of the motors gets out of sync with the others, tilting the X axis boom or the print bed. It is also easier to make adjustments on the machine when you can raise/lower the X axis or bed by simply pulling on the belt, knowing it won't cause any alignment problems later

Look at using CoreXY system

[www.corexy.com]

My experience with z-axis motors getting out of sync comes from the makerspace where we have different people with different levels of understanding using the machines. I've seen people manually twist one of the Z screws on the TAZ, thinking it would lift the X axis boom. Wiring problems can also cause the motors to get out of sync. Having kids around the printer often causes things to get out of alignment, too. From using my own printer a lot, I know that being able to grab the belt and move the X axis up without screwing up its alignment, which can be done with power on or off, is extremely convenient compared to powering up the printer, selecting the z axis movement through the menus on the LCD controller, and then moving the axis. It typically costs a little more to buy 3 pulleys and a closed loop belt than it does to just add a second motor, but for me the time saved in literally never having to realign the X-axis was worth 10x the cost of the pulleys and belt. Initial alignment is simple- you loosen the lock screw on one of the screw's pulleys, align the X axis boom, then lock the pulley down.

With two motors you're making an assumption that the steps the motors make are always exactly equal. I would trust the native steps the motors make to be pretty close, but with microstepping the motors won't necessarily both stop at exactly the same angle. The errors result in slight wobbling of the X axis boom (or the print bed) as it moves up or down. Is it enough to cause visible errors in the print surface? I have seen a lot of prints from a lot of printers that use two motors and a lot of them seem to have print quality problems in the Z axis. I don't know if it's because the motors aren't matched, the screws aren't matched, or other mechanical problems, but I'm always in favor of trying to eliminate sources of error and unreliability. When in doubt, fix it before it becomes a problem.

The lautr3k design is good for rigidity (and looks similar to the Mendel Max), but I would get rid of the plastic parts in the frame structure. All you need to bolster and attach the frame pieces to each other at an angle is a couple small aluminum plates with a few imprecisely drilled holes in them. You can fit button head cap screws or carriage bolts into the slots in the frame pieces. The frame parts that meet at 90 degree angles can just be bolted to each other. The center holes in the extrusions are always sized for a standard tap so they can be threaded to fit common bolt sizes. My machine uses 1515 series 8020 and the center hole is sized to fit a 5/16-18 tap. I use button head cap screws to attach the pieces- the screw and washer fit into the slot of one piece and screws into the end of the other piece. You have to drill a hole through the extrusion to access the cap screw head with a hex wrench, but the hole doesn't have to be precise. The slots in the extrusions are likewise designed to accept standard carriage bolts that will slide along the slot but won't rotate. For those you just use a wrench on the matching nut to tighten the pieces you are attaching.

I'm a bit of a fanatic when it comes to the mechanical stuff. The reality is that a lot of printers produce good quality prints without the expense, bulk, and weight of using 8020 frame pieces. We are talking about positioning liquid plastic that comes out of a nozzle and gets laid on previously printed layers, and shrinks as it cools. How precise can it be? I don't know. Right now I'm doing a lot of mechanical upgrades to my printer to try to squeeze the absolute maximum positioning precision out of the machine. I will be pleasantly surprised if I see an improvement in print quality when I'm done. It's entirely possible that the screw/ball nut will have some backlash that will give me worse results than I get now. The fallback plan in that case is to go back to belt drive while keeping the linear guides and bearing blocks.

All that said, I think the corexy or whatever you call the Makerbot type system has a lot of merit, especially with a large print envelope. In the RepRap style machine like mine you're moving a massive print bed (and print) back and forth and that causes some print quality problems that I am going to extreme lengths to rectify. The Corexy system is more complicated mechanically but has a lower moving mass that is better suited to belt drive (and a big print bed). The corexy machine also makes it pretty easy to move the electronics and motors out of the thermal enclosure. I think if I were going to do it all over again I would go with corexy. I also think that I would use three screws driven by a single motor and belt to lower the print bed.

Edited 1 time(s). Last edit at 11/28/2014 01:26AM by the_digital_dentist.

One more thing... the main argument against a large print envelope RepRap type machine is the mass of the printbed that has to be slammed back and forth. The mass can be reduced significantly and the printer simplified by printing on room temperature PIR foam instead of a heated metal or glass bed. I have done it, it works, but you have to give up the super shiny, smooth bottom surface of the part and print on a raft.

PIR (polyisocyanurate) foam is used for building insulation. You can buy 4'x 8' x 1" thick sheets of it in the US at home improvement stores for about $15. It can be cut to the needed size using a razor knife. One piece of foam can be printed on over and over, so the cost is negligible. A single 4' x 8' sheet would probably last a year or two, depending on how much printing you do.

If you print on foam- something Stratasys has been doing for years- you don't need a mechanical system to level the print bed- the foam can just sit on the undercarriage that rides on the bearings/rails. You just bury the tip of the extruder nozzle about 1mm into the foam when printing the first layer (use the z offset option in slic3r). You don't need a bed heater or the power supply for it. You don't need any wires connecting to the moving bed.

Do the parts stick to the foam? See these videos for the answer: [vimeo.com] , [vimeo.com]

If you're going to print large parts with ABS, you still need a thermal cabinet to prevent delamination and warping. You will still have to provide some sort of heater for that cabinet.

If you bury the nozzle tip wouldn't that either rip up the foam or clog the nozzle? Or both?

---

greghoge.com

HUGE 3D PRINTER PARTS SALE!!!

The motors are driven in parallel from the same driver on RAMPS,
exact same steps/voltage going out to both motors

The teeth on belts have NO slip or slop?

The nuts (Backlash - slop) on the Z-Axis lead screw cause errors in polymer layers.

If you want super precision --- need linear encoder and feedback loop to feedback position within 0.000001mm???
and optics to measure extruded polymer size and speed

Perhaps a $500,000 machine

Quote
Digital_dentist
With two motors you're making an assumption that the steps the motors make are always exactly equal. I would trust the native steps the motors make to be pretty close, but with microstepping the motors won't necessarily both stop at exactly the same angle. The errors result in slight wobbling of the X axis boom (or the print bed) as it moves up or down. Is it enough to cause visible errors in the print surface? I have seen a lot of prints from a lot of printers that use two motors and a lot of them seem to have print quality problems in the Z axis. I don't know if it's because the motors aren't matched, the screws aren't matched, or other mechanical problems, but I'm always in favor of trying to eliminate sources of error and unreliability. When in doubt, fix it before it becomes a problem.

Mendalmax has less plastic parts?

You will get 6 different recommendations from six different people on (the strongest and the more stable) all of them with little experience.

If you don't bury the nozzle too deeply the effect is to push the foam down without tearing it up. In the first video you can see an area in the upper left corner of the foam where I buried the nozzle too deeply and it tore the foam up a bit. The foam is soft enough that it gets out of the way when the extruder is pushing plastic against it. I did not have any problems with clogged nozzles- nothing goes back up into the nozzle. You can see that I printed the skull on the same piece of foam, and also see that when I finally managed to break the skull loose it took a little chunk of the foam out, too. Stratasys machines that print on foam have a rounded profile at the nozzle tip to ensure that it compresses the foam instead of tearing it up. That would be a good idea for printing on PIR, too, but not really necessary.

Edited 1 time(s). Last edit at 11/30/2014 01:22AM by the_digital_dentist.

Gotcha. I think I'll pick some up once I get my large scale printer up and running. Should help tremendously with keeping the bed weight down.

---

greghoge.com

HUGE 3D PRINTER PARTS SALE!!!