Buying 3D printer

[metrologist]

I checked there briefly, and another cloud. Do your designs become open source on the cloud? I'll probably try it anyway as I have a CNC mill that should be working.

[Mechatrommer]

3D printed enclosures usually look terrible

I disagree, square flat surfaces 3d printers do quite well.
PS took 10-12 hours to print it. FDM printers are not particularly fast.

that (picture) looks terrible. admit it, i own 3d printer and happily print whatever i desire, and then post processing however i think necessary to my satisfaction, esp weird shape or custom enclosure that is not available in off the shelf market or super expensive to get, but 3d printer only suitable for prototyping or a one-off. if you are thinking selling in mass a 12 hours print per unit, let alone polishing it, you are shooting yourself in the foot. for squared enclosure i still prefer buying off-the shelf, ebay or something and size my project to fit. there is a guy here selling his product in 3d printed "squared" box, lets see how long he will stand by the idea. anyway, ymmv.

[tggzzz]

All talk about the printer, but that is useless without some design to print. How do folks actually create the design, particularly complex designs?

Isn't that most significant aspect of this, and the most difficult hurdle?

That's a good point. The answer depends on what you are trying to create. The tools for creating a head as shoulders bust are different to those for creating a box.

For creating a box, I'd use a declarative computer language where the various dimensions can be specified by parameters p assed to functions. See  OpenSCAD for example.

[Nominal Animal]

For creating a box, I'd use a declarative computer language where the various dimensions can be specified by parameters passed to functions. See  OpenSCAD for example.

Me too. (Although I'd say "OpenSCAD implements a declarative language for defining solid geometries"; it amounts to the same thing.)

I'd really need a small, cheap, quiet, accurate 3D printer I could use for printing small mechanical structures in PLA.

As an example, I've designed a camera dolly for a friend (for GoPros, or cellphone cameras; so palm-sized lightweight devices, not pro gear). The camera stand moves on a linear axis, with the camera holder rotating, both controlled by a small microcontroller reprogrammable via USB.  I'd really want it to have a pan-tilt mount. The forces involved are really small, but I'd like to eliminate any backlash. To do so, I can easily pair cheap geared steppers. However, the stability and usability depends on size and final center of mass, and so practical experimentation is needed. I'm currently limited to wood sticks, Technic Legos, and hobby clays for modeling... When I know the type of the brackets, I can create them in OpenSCAD, obviously, and get them printed even in metal. But for the experimentation, a 3D printer would be nice.

Unfortunately, I can't currently afford a good one (like Prusa i3 even in kit form), and I don't want to risk getting a cheapie that simply cannot produce dimensionally accurate prints. I don't mind spending time to fiddle with the device to calibrate and adjust its performance, though. (I do prefer the new Trinamic drivers, as they significantly reduce the stepper noise.)

Any suggestions to this particular use case?

[brucehoult]

I'd really need a small, cheap, quiet, accurate 3D printer I could use for printing small mechanical structures in PLA.

There's no such thing.

If you can't afford a Prusa then you can't afford a practical and useul 3D printer. Even the Prusa is fiddly. For most people who just want something that works I'd recommend an Ultimaker 2 or 3. They cost much more.

[Nominal Animal]

I'd really need a small, cheap, quiet, accurate 3D printer I could use for printing small mechanical structures in PLA.

There's no such thing.
If you can't afford a Prusa then you can't afford a practical and useul 3D printer. Even the Prusa is fiddly.

Because the Prusa i3 (including MK3) design is GPL'd, there are many derivatives.  I could get a Prusa i3 reprap derivative for under 250€ (which even I could afford), but who knows how good/accurate the hardware parts are?

The Original Prusa i3 MK3 (from Prusa himself) would cost me ~ 800€ shipped. That is over my current spending limit. The reported occasional problems (frame parts needing machining to fix in particular) scare me, because I don't have access to a metal shop to fix any of them, and shipping the parts back and forth (to get them fixed under warranty) would cost quite a bit.  If I had access to even a modest metal shop and dimensionally accurate aluminium extrusions, I'd have built a 3D printer to satisfy my needs ages ago already.

The electronics, even the drivers and steppers I can replace, and I can (re)write the entire software stack if I need to; it's the mechanics I'd have to live with.
I was hoping there would be practical experiences from those using/maintaining multiple printers, especially regarding the hardware side.

[CJay]

Flash forge finder (We have like 10 of them)
-High quality prints, is easy to use. You’re locked to its slicer, has a small build plate and not heated bed

I'm not convinced you are locked to their slicer, there's at least one commercial offering that supports their machines and it seems there are a few open source/free packages that work too (I've not tried them yet, Flashprint works for me so far but I did look when I got the machine)

Unless you know different?

[exe]

I could get a Prusa i3 reprap derivative for under 250€ (which even I could afford)

It will be nowhere as good as Prusa without putting a lot of efforts, imho. Just quality electronics, direct extruder and a few other crucial parts will cost 250 alone.

I saw people having good printing results using junk parts, but I wasn't able to achieve the same with my cr10 (I spent 3 or 4 months tuning it). Now my printer costs twice the original price, yet doesn't print well because controller board is not good. So, if I want to remove printing defects I need new brains (like duetwifi, but it costs 160euro), or I can spend another few weeks tuning firmware, playing with drivers' trimpots, etc trying to repair the original board. The funny thing is, after half a year of owning the 3d printer I printed just a few useful parts. The rest was calibration cubes, etc.

So, my advice don't underestimate cost of your time. Now I'd pay three times more than what I paid originally for the printer just to get a decent device I could trust.

[Eka]

All talk about the printer, but that is useless without some design to print. How do folks actually create the design, particularly complex designs?

Isn't that most significant aspect of this, and the most difficult hurdle?

I used OpenSCAD. You program your design rather than use a GUI to form it.

[Sudo_apt-get_install_yum]

Flash forge finder (We have like 10 of them)
-High quality prints, is easy to use. You’re locked to its slicer, has a small build plate and not heated bed

I'm not convinced you are locked to their slicer, there's at least one commercial offering that supports their machines and it seems there are a few open source/free packages that work too (I've not tried them yet, Flashprint works for me so far but I did look when I got the machine)

Unless you know different?

It might be so, I think Simplify3D has support but I’m not sure. At least with the older flash forge you were locked but it’s probably changed by now.
The slicer is good and works just fine, i just miss "advanced" features.

I really like these printers, they print 24/7 and have been doing so for a year or two and are holding up just fine. Only thing we have had to replace is the nozzle but after several thousands of hours of printing it’s kind of expected from a brass nozzle.

[Mechatrommer]

I could get a Prusa i3 reprap derivative for under 250€ (which even I could afford)

It will be nowhere as good as Prusa without putting a lot of efforts, imho. Just quality electronics, direct extruder and a few other crucial parts will cost 250 alone.

I saw people having good printing results using junk parts, but I wasn't able to achieve the same with my cr10 (I spent 3 or 4 months tuning it). Now my printer costs twice the original price, yet doesn't print well because controller board is not good. So, if I want to remove printing defects I need new brains (like duetwifi, but it costs 160euro), or I can spend another few weeks tuning firmware, playing with drivers' trimpots, etc trying to repair the original board. The funny thing is, after half a year of owning the 3d printer I printed just a few useful parts. The rest was calibration cubes, etc.

So, my advice don't underestimate cost of your time. Now I'd pay three times more than what I paid originally for the printer just to get a decent device I could trust.

my earlier comment on your printout crappiness was nothing to do with the lacking you mentioned above, you certainly dont need new printer brain unless it is a china mcu blob on a pcb, i suspect your printer brain will be some atmel avr cpu variant which i think adequate (prove me wrong). my printout is just as crappy as you are, same avr brain and self modified firmware, imho its just the nature of layer deposition mechanism that make its crappy (visible layer by layer on the finished product) and the dynamics of the motors (acceleration and decceleration commands) in the firmware will provide plenty of processing power for other stuffs and if you still think its lacking, try to print even slower, and i suspect the dynamics is coded as interrupt based mechanism inside the fw, so excuses such as lack processing power is moot as we should see lagging in GUI if that is the problem.

please do show me the best OSHW/reprap printout using some GHz raspberry pi brain i suspect its will be just a little bit better if not as the same crapiness. there are more technological involvement (price) if we want a better print, heat chamber, sensories etc. thats what make $$$ unit apart from cheap OSHW unit. accept what it cant do and enjoy what it can. i agree dont waste your time too much on it, if you really concern, there are few used stratasys unit sold on ebay for fraction of the original price, get those, but then, it will not be guaranteed CNC moulding grade product output. yes i've seen the printout result of stratasys, they are still "layered visible based" product, just a little bit shiny due to better fusion inside the heat chamber.

Flash forge finder (We have like 10 of them)
-High quality prints, is easy to use. You’re locked to its slicer, has a small build plate and not heated bed

I'm not convinced you are locked to their slicer, there's at least one commercial offering that supports their machines and it seems there are a few open source/free packages that work too (I've not tried them yet, Flashprint works for me so far but I did look when I got the machine)

Unless you know different?

It might be so, I think Simplify3D has support but I’m not sure. At least with the older flash forge you were locked but it’s probably changed by now.
The slicer is good and works just fine, i just miss "advanced" features.

I really like these printers, they print 24/7 and have been doing so for a year or two and are holding up just fine. Only thing we have had to replace is the nozzle but after several thousands of hours of printing it’s kind of expected from a brass nozzle.

if the printer can accept standard gcode command/file, i believe any slicer producing descent gcode file can be used. i use KISSlicer that came with Repetier Host.

[metrologist]

This a very typical dead end experience. How to convert stl to gcode.

stupid article stops half way, just gets the stl design into fusion and says good luck with the rest.

https://support.bantamtools.com/hc/en-us/articles/115001671634-Converting-STL-Files-to-G-code

the best masheen in the world can't do anything if the idea can't get into just the right kind of machine language.

So slicer is analogous to cam? always the wall.

[exe]

you certainly dont need new printer brain unless it is a china mcu blob on a pcb, i suspect your printer brain will be some atmel avr cpu variant which i think adequate

Sorry, but saying "better brains" I didn't mention lack of computational power. I meant the whole board. I did some research about seemingly random defects on my prints, I found other people have them as well. What else I found is one guy reported that replacing original board with duetwifi solved it. Rumors say the problem is in cheap drivers. Duetwifi has TMC2660 which are presumably much quiter and don't suffer from this.

May there is not enough current to steppers, I dunno. I adjusted current pots so the printer does not produce horrible loud noise when driving steppers, while yet able to move them (looks like I have resonance problem as well). That's why I just want a controller board with good drivers -- to forget about noise and missing steps.

PS I'm not really proficient in 3d printing, so whatever I write here can be wrong.

[CJay]

I could get a Prusa i3 reprap derivative for under 250€ (which even I could afford)

It will be nowhere as good as Prusa without putting a lot of efforts, imho. Just quality electronics, direct extruder and a few other crucial parts will cost 250 alone.

I saw people having good printing results using junk parts, but I wasn't able to achieve the same with my cr10 (I spent 3 or 4 months tuning it). Now my printer costs twice the original price, yet doesn't print well because controller board is not good. So, if I want to remove printing defects I need new brains (like duetwifi, but it costs 160euro), or I can spend another few weeks tuning firmware, playing with drivers' trimpots, etc trying to repair the original board. The funny thing is, after half a year of owning the 3d printer I printed just a few useful parts. The rest was calibration cubes, etc.

So, my advice don't underestimate cost of your time. Now I'd pay three times more than what I paid originally for the printer just to get a decent device I could trust.

my earlier comment on your printout crappiness was nothing to do with the lacking you mentioned above, you certainly dont need new printer brain unless it is a china mcu blob on a pcb, i suspect your printer brain will be some atmel avr cpu variant which i think adequate (prove me wrong). my printout is just as crappy as you are, same avr brain and self modified firmware, imho its just the nature of layer deposition mechanism that make its crappy (visible layer by layer on the finished product) and the dynamics of the motors (acceleration and decceleration commands) in the firmware will provide plenty of processing power for other stuffs and if you still think its lacking, try to print even slower, and i suspect the dynamics is coded as interrupt based mechanism inside the fw, so excuses such as lack processing power is moot as we should see lagging in GUI if that is the problem.

please do show me the best OSHW/reprap printout using some GHz raspberry pi brain i suspect its will be just a little bit better if not as the same crapiness. there are more technological involvement (price) if we want a better print, heat chamber, sensories etc. thats what make $$$ unit apart from cheap OSHW unit. accept what it cant do and enjoy what it can. i agree dont waste your time too much on it, if you really concern, there are few used stratasys unit sold on ebay for fraction of the original price, get those, but then, it will not be guaranteed CNC moulding grade product output. yes i've seen the printout result of stratasys, they are still "layered visible based" product, just a little bit shiny due to better fusion inside the heat chamber.

Flash forge finder (We have like 10 of them)
-High quality prints, is easy to use. You’re locked to its slicer, has a small build plate and not heated bed

I'm not convinced you are locked to their slicer, there's at least one commercial offering that supports their machines and it seems there are a few open source/free packages that work too (I've not tried them yet, Flashprint works for me so far but I did look when I got the machine)

Unless you know different?

It might be so, I think Simplify3D has support but I’m not sure. At least with the older flash forge you were locked but it’s probably changed by now.
The slicer is good and works just fine, i just miss "advanced" features.

I really like these printers, they print 24/7 and have been doing so for a year or two and are holding up just fine. Only thing we have had to replace is the nozzle but after several thousands of hours of printing it’s kind of expected from a brass nozzle.

if the printer can accept standard gcode command/file, i believe any slicer producing descent gcode file can be used. i use KISSlicer that came with Repetier Host.

It might be so, I think Simplify3D has support but I’m not sure. At least with the older flash forge you were locked but it’s probably changed by now.
The slicer is good and works just fine, i just miss "advanced" features.

I really like these printers, they print 24/7 and have been doing so for a year or two and are holding up just fine. Only thing we have had to replace is the nozzle but after several thousands of hours of printing it’s kind of expected from a brass nozzle.

I'm going to have to check when I get home but there seems to be an option to display advanced print options in Flashprint. I'm happy with the software and I'm not seeing any reason to change it but of course if there is better out there...

I bought mine faulty from eBay and when it arrived the status display claimed it'd only got 7 hours printing on it so I'm really happy they last for thousands of hours.

if the printer can accept standard gcode command/file, i believe any slicer producing descent gcode file can be used. i use KISSlicer that came with Repetier Host.

I think the files the FLashprint utility uploads are .stl but I'm really not certain, I'll have to check later.

[Mechatrommer]

This a very typical dead end experience. How to convert stl to gcode.
stupid article stops half way, just gets the stl design into fusion and says good luck with the rest.
https://support.bantamtools.com/hc/en-us/articles/115001671634-Converting-STL-Files-to-G-code

  the link you provided is relevant only if you want to modify the model, you really need an answer that is relevant to your question... "How to convert stl to gcode" is a simple step as... https://www.repetier.com/ perhaps with many other OSS softwares.

[Monkeh]

Unfortunately, I can't currently afford a good one (like Prusa i3 even in kit form), and I don't want to risk getting a cheapie that simply cannot produce dimensionally accurate prints. I don't mind spending time to fiddle with the device to calibrate and adjust its performance, though. (I do prefer the new Trinamic drivers, as they significantly reduce the stepper noise.)

Any suggestions to this particular use case?

Please ignore the naysayers who can take a perfectly functional machine and fail to make it work.

It's quite easy to start off with a basic machine like an Ender 3 and get good results - dimensional accuracy is basically a non-issue (always has been). You can certainly improve on a machine like that, especially noise-wise, but they do print just fine.

[Mechatrommer]

if the printer can accept standard gcode command/file, i believe any slicer producing descent gcode file can be used. i use KISSlicer that came with Repetier Host.

I think the files the FLashprint utility uploads are .stl but I'm really not certain, I'll have to check later.

i think it would be impractical to upload raw/native 3d format (stl) to printer and ask the printer to do realtime slicing to produce gcode that a machine mechanics can understand. its like sending jpg format to photo printer and ask the printer to decode and translate it into printhead's bitstream codes. i have a suspicion the spawn of cheap to not so cheap 3d printers (other than full engineering product like stratasys) throughout the globe including prusa inc was originating from OSS firmware such as marlin firmware et al (developed back then based on earlier cnc machine motion planner i forgot the name is by some good people you can search the history or learn from the firmware source code) and from some RAMP board hardware, and later each brand name close the door for their custom tuned FW and circuit. we can tell by the familiar LCD2004 GUInterface. the original fw is reading gcode for printing. if you have the chance to save the output file to sd card or sniff your printer usb comm, it should looks something like this.

Code: [Select]

G92 E0
G1 Z0.200 F7200.000
G1 X35.655 Y-7.957 F7200.000
G1 E2.00000 F1800.000
G1 X35.728 Y-7.502 E2.01450 F2880.000
G1 X35.961 Y-7.426 E2.02222
G1 X36.929 Y-6.711 E2.06009
G1 X37.044 Y-6.673 E2.06392
G1 X45.189 Y-6.673 E2.32028
those are gcode that a printer/cnc machine can understand. stl is just a bunch of vertices location of the model all over the place in 3d space. we'll need quite powerfull processor and memories to translate it to layer by layer bottom up polygons/motions in gcodes.

[exe]

stl is just a bunch of vertices location of the model all over the place in 3d space. we'll need quite powerfull processor and memories to translate it to layer by layer bottom up polygons/motions in gcodes.

I think slicing on printer is a bit impractical. I change slicing settings depending on filament, nozzle diameter, temperature, complexity of model, etc. So, the printer needs to be aware of that somehow (or user needs to specify the profile).

PS octoprint can do automatic slicing and stl upload.

[Mechatrommer]

btw. fwiw i built my own GCode Editor (which i received few decent donations from good people out there so i can buy stuffs cheaper in ebay ) to edit vertices or positional commands in gcode graphically, because my 3d modelling SW is aging, it cant do  modern NURBS (i think FreeCAD or Blender will do a better job on this), so i have to do modification in gcode. remember gcode is supposed to be an endpoint machine code, like CAM file we are not suppose to edit in there. not as in native 3d model data (like stl etc) that we are suppose to edit our model in 3D SW such as Fusion 360, SolidWork, Inventors etc or even OSS free FreeCAD or Blender. i think i'm the only soul (or few other? i dont know last time i checked its nowhere to be found, or not free) who insisted on the workflow (edit in gcode) and some other people have found use to it.

[CJay]

those are gcode that a printer/cnc machine can understand. stl is just a bunch of vertices location of the model all over the place in 3d space. we'll need quite powerfull processor and memories to translate it to layer by layer bottom up polygons/motions in gcodes.

Yeah, you're correct, I've just remoted into my machine at home and had a look at the backup of the internal SD card from the finder, it's full of .g and .gx files

[Cherenkov11]

I do not have a lot of experience on 3D printing, but in a video of Marco Reps I met the Cetus3D and it looks pretty solid and reliable, so it could be an option.

[metrologist]

btw. fwiw i built my own GCode Editor (which i received few decent donations from good people out there so i can buy stuffs cheaper in ebay ) to edit vertices or positional commands in gcode graphically, because my 3d modelling SW is aging, it cant do  modern NURBS (i think FreeCAD or Blender will do a better job on this), so i have to do modification in gcode. remember gcode is supposed to be an endpoint machine code, like CAM file we are not suppose to edit in there. not as in native 3d model data (like stl etc) that we are suppose to edit our model in 3D SW such as Fusion 360, SolidWork, Inventors etc or even OSS free FreeCAD or Blender. i think i'm the only soul (or few other? i dont know last time i checked its nowhere to be found, or not free) who insisted on the workflow (edit in gcode) and some other people have found use to it.

OK, I will look at your software. I write G-code manually for my machine and have used Excel to help by using formulas to expand point coordinates and ultimately toolpaths. Or I have used scorchworks to create paths and modify the code for 3D milling.

[sairfan1]

anybody else used Cetus3D printer?

[beanflying]

Depends what you are trying to a achieve. If you're selling to anyone but hobbyists with low expectations, a 3D print will just look like a joke. There's a huge range of off-the-shelf cases, and even the cheapest will have a better looking surface finish than a 3D print.
Obviously you need to start by selecting the enclosure and designing around it, rather than looking for a box to fit an existing PCB etc.

Clearly you just don't get it so perhaps the joke is on you and you miss the point? Ease of Rapid prototyping and proof of concept with custom parts and or enclosures prior to looking at volume production it is one of the better options available.

A tacky cheap enclosure can also look like a Hobby Project from the 70's and equally unprofessional IMO!

All talk about the printer, but that is useless without some design to print. How do folks actually create the design, particularly complex designs?

Isn't that most significant aspect of this, and the most difficult hurdle?

Fusion 360 once you get into it is Free and powerful you just need to get over the 'Cloud and Autodesk' mental barrier.

Enjoying the results and the ability to get from Pencil Sketch to Design to product quickly with a practical result for one offs and proof of concept.

Currently tossing around a Prusa MK3 kits or some other sort of direct extruder printer for a second printer for flexibles and higher temperatures than my CR-10S offers.