Gathering information for DIY Large Printer build - any help much appreciated!

Printable View

09-22-2019, 03:31 PM
MystiqueEngineer

2 Attachment(s)

Gathering information for DIY Large Printer build - any help much appreciated!

Hello everyone!

I am going to build a large format 3D printer, and I have some questions that maybe some of you have the answers to. So any help is much appreciated!

The size of the print-area will be 1600x600x800mm. I have built a printer before out of a parts kit from china. It was fairly large, with a print-area of 310x310x420mm. Needless to say the Chinese seller didn't speak much English, and their "technician" wasn't much more knowledgeable than I was after a week with Google - so I ended up using a new Marlin software, and editing it until it functioned with my printer.

My printer ended up functioning quite well. But being a open frame design, I wasn't able to use any of the ABS I bought with it. (I wasn't read up properly on the needs of ABS.)

Since then I have been employed for a while as a CNC-machinist, and the need for a printer for prototyping disappeared. However, now it would be very nice to have a 3D printer again at home for easy prototyping of designs. If my build succeeds, and it can give pleasing results, then my employer wants me to build one or two at work for the same use.

I'm going to use 80x80mm T-Slot extruded aluminium for the frame, as this is easy for me to cut and assemble accurately. The table will be mounted into the frame. The table will be screwed down to a welded and end-machined frame, with PEEK spacers in slots underneath the table. This is to allow the heated 15mm grinded aluminium table to move the small distances it will do due to heat distortion. The plan is to utilize a pre-grinded aluminium plate, which is treated from the manufacturer in such way that gives minimal tension in the material, and also as little heat distortion as possible.

The printer will have 4 extruders, mounted on two custom machined water-cooled blocks in aluminium or copper. The printer will get a fully heated enclosure, so water cooling is necessary. All movement will be with ballscrews, and linear rails like SBR-20 or HGR-20. I plan on keeping the X, Y and Z -axis assembled in the back of the machine, so that I can keep all high precision movement parts cooled down, without affecting the 3D printing in the enclosure. The open space between the two will be separated by bellows.

All movement will be controlled by servos, and not steppers. I know this is a controversial thought for a 3D printer, even though it has been done already. I've done some research on the subject, and I know that steppers are good enough in most scenarios. However, I want to make sure the printer is not able to skip steps, and most importantly; I need it to be quiet. The prototype for this machine will be placed right next to me by my computer. And I want to make sure that I can remove as much noise as possible. I figured the best way to start would be to replace the steppers with servos. I am not really worried about the cost of servos over steppers. I am more than willing to spend 1000 to 1500 USD on servos with encoders if that is what it takes. But when it comes to the process of making a 3D printer controller normally using steppers, to utilize servos instead, I'm not sure where to start.

Here are some questions I am hoping we could discuss:

1. What would be the best controller for me to use for a build like this?

2. How would I go about using servos with encoders to control movement from a controller normally using steppers? Would the best way be to make a 'translator' with Arduino?

Thank you for reading! Right now I'm doing design and research, but I promise to post progress when the building is commencing!

PS: Sorry for the very, very simple models I have attached pictures of. I've only been designing it after work on a couple of days to start somewhere. I need to have a plan for servo-drives etc before I continue.
09-25-2019, 05:11 AM
Trakyan

The best controller depends on your needs. If a simple atmega based board fits your needs, just use that. No sense in using a sledge to swat a fly. As for servos, there are some servo which can be driven with a step/dir signal just like stepper motor drivers, so your control board can treat them the same. Making your own arduino controller to control a servo using step/dir isn't worth it unless you're very familiar with control loops and tuning them.
Also, servos aren't necessarily quieter, I'm not sure what's convinced you of that. The noise people typically associate with steppers is more an artefact of the stepper drivers commonly used in 3D printers. If you want servos the cheapest option I can think of is a mechaduino, which is a NEMA17 fitted with an encoder and arduino to provide closed loop feedback, this can be run with a step/dir signal from the control board like a regular stepper.
The O-Drive control board is another option, it lets you use a regular RC brushless motor, mount an encoder to it and drive it like a servo. This is a bit more pricey but the motors can be a lot more powerful and a lot faster for the same form factor, but a 3D printer won't really be going fast enough to take advantage of that. On the flip side brushless RC motors have pretty weak holding torque, not that it really matters for a 3D printer.
Final suggestion; don't make it that big. Even industrial machines don't really go that big. Even with a heated chamber you'll get distortion and warping. Not to mention with a build chamber that size you're looking at print times upwards of a week or even more, the chances of nothing going wrong during such a long print are small. What are you even going to print that large? Plastic parts, especially 3D prints, at that size are probably too weak to be useful.
sorry for all the weird line breaks, my post keeps getting formatted to remove new lines/paragraph breaks for some reason.
09-26-2019, 03:52 PM
MystiqueEngineer

Had to edit this post 4 times due to the problem you were describing. All paragraphs etc gets mashed into one big cluster of words. Really annoying.
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Quote:

Originally Posted by Trakyan

The best controller depends on your needs. If a simple atmega based board fits your needs, just use that. No sense in using a sledge to swat a fly. As for servos, there are some servo which can be driven with a step/dir signal just like stepper motor drivers, so your control board can treat them the same. Making your own arduino controller to control a servo using step/dir isn't worth it unless you're very familiar with control loops and tuning them. Also, servos aren't necessarily quieter, I'm not sure what's convinced you of that. The noise people typically associate with steppers is more an artefact of the stepper drivers commonly used in 3D printers. If you want servos the cheapest option I can think of is a mechaduino, which is a NEMA17 fitted with an encoder and arduino to provide closed loop feedback, this can be run with a step/dir signal from the control board like a regular stepper.

 
 
I was thinking about getting the Duet Wifi or a Smoothieboard. I guess they both would do the trick - so I'll just see what I end up with.
 
 
When it comes to the step/dir compatible servos, could you give me a link to an example? I've tried to find some, but without luck.
 
When it comes to the sound I want to get away from; I'll admit my prejudice is based on other 3D printers (My China kit in particular). But even the Stratasys we have at work (apparently a very expensive printer according to my superior) is noisy as hell. It's a pretty noticeable motor sound from it - a sound I have never heard from servos. Therefore my presumption that servos in most cases should be quieter. But of course, AC/DC or Brushed vs. brushless is also a factor. I was thinking going for a AC brushless servo. Also, the videos I've seen of guys converting to servo-drive on their 3D printers seem to gain quite amazingly low noise levels.
 
I was thinking of going for something like this; https://www.ebay.com/itm/JMC-100W-0-...gAAOSwAbpdZNwc
 
I'm starting to think making my own Arduino controller is my only shot to make it work with something like that though. I came across something called ServoStrap here the other day, and it seems to be exactly what I need. I just need to play around with it for a bit, and see what I can make of it. I'll have to build a test-bench of similar controls first though.
 
ServoStrap: https://github.com/danithebest91/ServoStrap
 
It recently came to my attention that we have quite knowledgeable contacts on the subject through my employer. A collage of mine will introduce me when possible - so maybe I can get some help from that end on how to attack this.
 

Quote:

Originally Posted by Trakyan

The O-Drive control board is another option, it lets you use a regular RC brushless motor, mount an encoder to it and drive it like a servo. This is a bit more pricey but the motors can be a lot more powerful and a lot faster for the same form factor, but a 3D printer won't really be going fast enough to take advantage of that. On the flip side brushless RC motors have pretty weak holding torque, not that it really matters for a 3D printer.

 
I definitely see the potential in this for a lot of hobbyists, but I would rather put the extra money into a propper servo with encoder. Thanks a lot tho, I am sure I can put that to good use in other projects.
 

Quote:

Originally Posted by Trakyan

Final suggestion; don't make it that big. Even industrial machines don't really go that big. Even with a heated chamber you'll get distortion and warping. Not to mention with a build chamber that size you're looking at print times upwards of a week or even more, the chances of nothing going wrong during such a long print are small. What are you even going to print that large? Plastic parts, especially 3D prints, at that size are probably too weak to be useful.

 
I don't have any specific parts that needs to be that big. However, I am planning to have quick exchangeable extruders, so I can switch materials, and nozzle sizes faster. This includes a Volcano setup with 1mm or 1.2mm nozzles. I've never printed with these before, but I presume they would make it easier to build large things with less warping. For example a chassis or larger constructions. For the most part regarding the size, I want a big build area for printing large series of different parts at once. Print times doesn't really worry me, but I agree on the increased chances of something going wrong. I am willing to try nevertheless. ;)Thanks a lot for your input!
09-26-2019, 04:29 PM
Trakyan

Quote:

Originally Posted by MystiqueEngineer

When it comes to the step/dir compatible servos, could you give me a link to an example? I've tried to find some, but without luck

The O-Drive and Mechaduino (type either of those into google) I mentioned in my previous post are the open source/hobbyist level ones I'm familiar with. Some of the name brands usually seen in CNC routers and higher budget projects like clearpath have products that can do this as well. These days I'm not convinced spending money on a clearpath or similar servo is worth it, projects like mechaduino and O-Drive tend to be more feature rich and customizable as a result of being open source.

Quote:

Originally Posted by MystiqueEngineer

When it comes to the sound I want to get away from; I'll admit my prejudice is based on other 3D printers (My China kit in particular). But even the Stratasys we have at work (apparently a very expensive printer according to my superior) is noisy as hell. It's a pretty noticeable motor sound from it - a sound I have never heard from servos. Therefore my presumption that servos in most cases should be quieter. But of course, AC/DC or Brushed vs. brushless is also a factor. I was thinking going for a AC brushless servo. Also, the videos I've seen of guys converting to servo-drive on their 3D printers seem to gain quite amazingly low noise levels.

Again, this has to do with the drivers commonly used for steppers. I believe it's something to do with the frequency or chopping of the waveform, it's closer to the audible spectrum which gives that typical high pitched stepper whine people hear. Have a look at a printer using some TMC drivers and you'll notice the lack of motor noise. The prusa i3 mk3 is a common example if you want to watch some youtube videos on it, people tend to say the fan is the loudest part.

The main reason to switch to servos I think is speed, reliability (especially at speed) and maybe precision. I don't think most of these are relevant other than reliability. The extrusion speed will limit your printing far sooner than the stepper's max speed. And I think the bottleneck for accurate parts is also the fact we're squeezing molten plastic out of a 0.4mm or bigger bore. Servos will probably let you use higher accelerations though, which would help avoid blobbing with sharp corners and intricate details. On super detailed parts with lots of direction changes it might make the print noticeably faster.

As for type of servo, I'd ignore DC, they are noisy both regarding EMF (which can mess with endstops and other electronics) and audibly which is the opposite of what you're going for. AC and brushless DC are the same thing as far as I'm aware, at least as far as the motor is concerned.

I don't recommend making your own servo with an arduino. It is not easy an will require a good knowledge of control loops, and even then it will take a while. There's a reason so many printers use stepper and that open source servos have only really arrived in the last few years. Go with one of the existing projects, it'll be cheaper and if you want to improve or modify it from there, you can because they are open source.
https://odriverobotics.com/
https://tropical-labs.com/mechaduino/