Different styles of printers
Posted by pineapple
|
Re: Different styles of printers March 02, 2015 11:53AM |
It isn't, a printer's architecture is different as long as the math to control it is different. 
Collective intelligence emerges when a group of people work together effectively. Prusa i3 Folger (A lot of the parts are wrong, boring !)
Collective intelligence emerges when a group of people work together effectively. Prusa i3 Folger (A lot of the parts are wrong, boring !)
|
Re: Different styles of printers March 02, 2015 01:44PM |
Quote
Zavashier
It isn't, a printer's architecture is different as long as the math to control it is different.
Fair enough!
Quote
dc42
Don't worry about the math, that's taken care of by the firmware. The delta has some more advantages over Cartesian printers:
* More regular construction. You build 3 identical towers and carriage assembles, instead of building 3 different axes. Get one of them right and you get them all right.
* No issues with bed levelling, because the bed is fixed. Once you have done the delta calibration (which takes less than an hour even if you don't have a Z probe and a good auto calibration facility in the firmware), the machine is stable. It's much more difficult to keep the bed of a Cartesian printer level and stable because it moves.
* Smaller footprint for the same build area, because of the non moving bed. The delta needs more height, but that isn't usually an issue.
I have a Cartesian printer (Ormerod) and a delta (Mini Kossel, about to be enlarged). Based on my experience with them, I recommend the delta for its easier build, smaller footprint and faster print speed for the same print quality. I have no experience with CoreXY or Scara.
If you choose a delta then I strongly recommend you choose a 32-bit controller board for it. The 8-bit boards have barely enough power to do the delta calculations in spite of the approximations that firmware for 8-bity boards uses. Both my printers use 32-bit Duet electronics.
That makes a lot of sense. I noticed the symmetry in the Delta style printers but I didn't link it to easier manufacturing.
So then two follow up questions (if you don't mind). Why would someone choose a Cartesian printer over a delta? Is there a reason polar isn't as popular as Cartesian? Polar and Cartesian would appear to be almost the same complexity to me, but then again I'm just a noob here.
Edited 1 time(s). Last edit at 03/02/2015 02:50PM by pineapple.
|
Re: Different styles of printers March 02, 2015 04:35PM |
Well that's my humble opinion (and facts) about deltas :
1. This can be an extremely fast machine, initialy used for pick and place applications. The FDM technology limits the max speed anyways. When you build a delta 3D printer, you use only 25% of the speed that architecture can provide.
2. On a delta, the footprint looks a bit smaller. Honnestly, it's only smaller compared to mobile bed printers. Scara or CoreXY can have a comparable footprint. Throught, the delta needs height. The overall volume is bigger on a delta.
3. Some says the delta print volume is a cylinder while carthesian is a parallepiped. It's not exactly a cylindre* but who cares ? Except if you're printing vases or lamps once at a time only, it will be easyer to manage multiple parts on a carthesian build area.
4. Unlike carthesian printers, on a delta the resolution is not constant on the build plate. You've got less resolution on the centre, and more at the edges.*
*see the picture above.
5. The build is more simple on a delta, but the arms must be cutted accurately, and the build ask for a better precision.
6. Except if you build a well documented kit, the calibration of the machine can be a nightmare ! Especialy when you scalled up or down your delta.
7. The delta needs a bowden, so you probably won't be able to print exotic filaments, like filaflex.
8. The dela's head has a limited space, and that head is limited by weight too. So you probably won't be able to add a lot of features to your printer, like multiple extruders, a laser head, fans...
The delta have some advantages too, like a supported and fixed bed.
Edited 1 time(s). Last edit at 03/02/2015 04:36PM by Zavashier.
Collective intelligence emerges when a group of people work together effectively. Prusa i3 Folger (A lot of the parts are wrong, boring !)
1. This can be an extremely fast machine, initialy used for pick and place applications. The FDM technology limits the max speed anyways. When you build a delta 3D printer, you use only 25% of the speed that architecture can provide.
2. On a delta, the footprint looks a bit smaller. Honnestly, it's only smaller compared to mobile bed printers. Scara or CoreXY can have a comparable footprint. Throught, the delta needs height. The overall volume is bigger on a delta.
3. Some says the delta print volume is a cylinder while carthesian is a parallepiped. It's not exactly a cylindre* but who cares ? Except if you're printing vases or lamps once at a time only, it will be easyer to manage multiple parts on a carthesian build area.
4. Unlike carthesian printers, on a delta the resolution is not constant on the build plate. You've got less resolution on the centre, and more at the edges.*
*see the picture above.
5. The build is more simple on a delta, but the arms must be cutted accurately, and the build ask for a better precision.
6. Except if you build a well documented kit, the calibration of the machine can be a nightmare ! Especialy when you scalled up or down your delta.
7. The delta needs a bowden, so you probably won't be able to print exotic filaments, like filaflex.
8. The dela's head has a limited space, and that head is limited by weight too. So you probably won't be able to add a lot of features to your printer, like multiple extruders, a laser head, fans...
The delta have some advantages too, like a supported and fixed bed.
Edited 1 time(s). Last edit at 03/02/2015 04:36PM by Zavashier.
Collective intelligence emerges when a group of people work together effectively. Prusa i3 Folger (A lot of the parts are wrong, boring !)
|
Re: Different styles of printers March 03, 2015 03:55PM |
Well I'm not a noodle expert, what about a review ?Quote
jaguarking11
Quote
Zavashier
7. The delta needs a bowden, so you probably won't be able to print exotic filaments, like filaflex.
8. The dela's head has a limited space, and that head is limited by weight too. So you probably won't be able to add a lot of features to your printer, like multiple extruders, a laser head, fans...
The delta have some advantages too, like a supported and fixed bed.
I Disagree on the multiple extruders sentiment. You can design a delta with multiple extruders and it will function properly. Each extruder will have to be bowden fed.
I also disagree on the flexible filament sentiment. That is more of a function of the extruder than anything else. A proper extruder would not have issues pushing flexible filament down a bowden tube, unless the filament is sticky by nature. I am fairly sure that with a proper extruder you can push a wet noodle through a bowden tube.
Some people says you can't extrude with enough precision flexible filaments on a bowden extruder, because of the compression applied to the filament. I must confess that seems logical to me too. That's why I did not tryed flexible filaments on my bowden setting yet. I also don't need flexible prints very often. Honnestly, I just give my spool some trys, then it's on the shelve since 4 months now.In absolute... you're right for the multiple extruder. BUT, read well, I said "So you probably won't be able to". I meant, it's not impossible, it just won't be easy, you'll probably need to start over a reworked design, and arrange maths/configuration which can be tricky on a delta. The huge majority of deltas kits and/or BOM/plans are engineered for a single extruder plate, with limited space to install multiple hotends side by side. Even the Chimaera takes a piece of space. My opinion can sounds distinct to your ears. At last, appreciate that on other machine types, it's much more simple.
Edited 1 time(s). Last edit at 03/03/2015 04:02PM by Zavashier.
Collective intelligence emerges when a group of people work together effectively. Prusa i3 Folger (A lot of the parts are wrong, boring !)
|
Re: Different styles of printers March 04, 2015 11:12AM |
The problem with rotational axes are very simple.
In a linear system, mechanical errors are directly reflected in the position of the extruder. If the belt is out by 0.1mm, the extruder will also be out by 0.1mm.
In a rotational system, any mechanical errors are multiplied by the length of the arm. A tiny error at the motor shaft becomes a major problem 6 or 12 inches from center. I have attached a crude drawing to illustrate this problem.
Note that amplification of error goes two ways: any torque applied to the end of the lever gets multiplied by its length. With a linear system, a 200g extruder is just a 200g extruder. On a rotary system that 200g is suddenly on the end of a 12" lever and might be applying ten times that force to the bearing.
So you have two major problems - you need more accuracy and more rigidity. Both requirements scale multiplicatively with build area.
This already sounds pretty bad, but to fully understand the implications you have to realize that rotational accuracy is harder to achieve than linear accuracy. The only viable option for a hobbyist is timing belts, but you run into rigidity/stretch problems due to the huge torques involved and it is difficult to get sufficient mechanical reduction in one stage. Most people end up with a servobelt clone.
This problem isn't new and has plauged the CNC and telescope mount guys for a very long time.
For an example of "budget" rotary stage construction you can look at the pocketnc. Note the dual stepper motors and massive 200+ tooth pulley (not an off-the-shelf part). If you dig through the rest of their posts you can also spot the massive 60mm tapered roller bearings. Their current design may still be insufficient for milling. They started with a single motor and smaller bearings and have progressively added to the design with each revision.
Ironically, 3d printing might actually be the best way to achieve accurate rotary positioning (at least with low loads) since you can easily print geometries that are otherwise impractical. A differential planetary gearbox using spring-loaded tapered planets could provide zero backlash reduction, and with 6-8 planets any geometrical errors in the print will be averaged out.
In a linear system, mechanical errors are directly reflected in the position of the extruder. If the belt is out by 0.1mm, the extruder will also be out by 0.1mm.
In a rotational system, any mechanical errors are multiplied by the length of the arm. A tiny error at the motor shaft becomes a major problem 6 or 12 inches from center. I have attached a crude drawing to illustrate this problem.
Note that amplification of error goes two ways: any torque applied to the end of the lever gets multiplied by its length. With a linear system, a 200g extruder is just a 200g extruder. On a rotary system that 200g is suddenly on the end of a 12" lever and might be applying ten times that force to the bearing.
So you have two major problems - you need more accuracy and more rigidity. Both requirements scale multiplicatively with build area.
This already sounds pretty bad, but to fully understand the implications you have to realize that rotational accuracy is harder to achieve than linear accuracy. The only viable option for a hobbyist is timing belts, but you run into rigidity/stretch problems due to the huge torques involved and it is difficult to get sufficient mechanical reduction in one stage. Most people end up with a servobelt clone.
This problem isn't new and has plauged the CNC and telescope mount guys for a very long time.
For an example of "budget" rotary stage construction you can look at the pocketnc. Note the dual stepper motors and massive 200+ tooth pulley (not an off-the-shelf part). If you dig through the rest of their posts you can also spot the massive 60mm tapered roller bearings. Their current design may still be insufficient for milling. They started with a single motor and smaller bearings and have progressively added to the design with each revision.
Ironically, 3d printing might actually be the best way to achieve accurate rotary positioning (at least with low loads) since you can easily print geometries that are otherwise impractical. A differential planetary gearbox using spring-loaded tapered planets could provide zero backlash reduction, and with 6-8 planets any geometrical errors in the print will be averaged out.
Sorry, only registered users may post in this forum.