Cost - 3D printed enclosure vs Project box?

[ChrisLX200]

3D printers are good for making those really awkward shapes that you don't want to spend time/effort making in some other way. They can be a time saver if you're proficient with SketchUp (or similar) and also the design can be quickly modified and re-printed if required. Personally, for something I which intend to keep I prefer to machine metal, but having said that I have an awful lot of those 3D printed enclosures lying around which I couldn't be bothered upgrading to something better. The 3D printer in not my first go-to device to solve a support/enclosure problem, but more of a last resort if I can't think of anything else

Oh, and I should mention Thingiverse - a fantastic resource to trawl through, the 'awkward' enclosure you seek may already have been designed and posted by someone..
 

[brucehoult]

This is about as rough as they get with no effort put into cleaning the bed or levelling it, just draw (15 minutes using Openscad) and print (20 minutes). Control panel for my spot welder.

That's pretty rough. Home (and home made) printers can do far better than that.

Commercial services such as Shapeways produce excellent quality, and can 3D print in metals as well as plastics, and with fewer design restrictions. They do cost quite a bit more than the materials cost though.

I second the use of OpenSCAD. As a programmer -- and for the kinds of things I make --  I find it far more comfortable to use than a click-and-drag GUI design program.

You can find the source code for one of my more complex OpenSCAD designs here https://www.thingiverse.com/thing:135065

[sokoloff]

This is about as rough as they get with no effort put into cleaning the bed or levelling it, just draw (15 minutes using Openscad) and print (20 minutes). Control panel for my spot welder.

Agree with @brucehoult; that's rougher than most of my prints, though perfectly functional.

It looks like you might be over-extruding versus what your slicer is expecting/planning on. Try googling a calibration guide for your particular slicer/printer. You may be able to permanently eliminate some of the surface finish lumpiness with a couple hours of wall clock time (less human time).

[sokoloff]

Took me a little searching to find the guide I used to guide my first calibration.

It's here: https://www.thingiverse.com/thing:52946

[Mjolinor]

That is printed with a 0.5 mm layer height. It was always going to be rough but it is perfectly functional and strong. The point is that it is so easy and does not take huge amounts of time. It has all the supports you need to mount the PCB and tags to screw it to the top of the welder. Such a thing done any other way is a massive cost in £ and in time. Once proven it is just a matter of levelling the bed and setting the height then doing a finished print with a smaller layer height.

[ChrisLX200]

If you print onto a glass bed you at least get one surface that is smooth (like glass ) and that can be useful sometimes if you can orient your print to suit. When using ABS you can also solvent-weld sub-structures together using acetone (or ABS/acetone slurry) which gives a very strong joint. You can use flexible filament to print seals which you can see in this example I designed for an astro all-sky camera: https://www.thingiverse.com/thing:1371445 (it's the white material).

[PlainName]

this example I designed for an astro all-sky camera

That's excellent.

I think trims (the white seal, which you can't normally see but..., and red whatever it is) make a big difference in aesthetics.

[tooki]

Eh, who cares if the  layers are visible, it's funvtional apparatus, not a beauty contest. I print at 0.3mm and i find the finish just fine.  Obviously print orientation needs to be considered, but for my functional prints i don't do cleanup other than removing support. 

For more aesthetic stuff to steal a phrase, filler and paint makes it the printer it ain't.

I care. I'm a perfectionist with a strong eye to the aesthetic, so I absolutely care. If I had the skills (I don't) for my projects to come out so perfect that people asked "how the heck did you make this?!?", I'd be happy! I won't choose aesthetics over functionality, but that doesn't mean I am willing to ignore aesthetics, either. (I will, though, forego some functionality and definitely aesthetics for usability, though the ideal is to not have to make much in the way of compromises with any of them. It just takes more work.)

As for post-printing finishing: of course that's an option, and a good one at that. (Indeed, I don't think many people realize how many of the plastic gadgets we buy are painted.) But that's not what you see a lot. As far as DIY 3D printed enclosures go, we see mostly raggedy things like Mjolinor posted above.

Eh, who cares if the  layers are visible, it's funvtional apparatus, not a beauty contest.  I print at 0.3mm and i find the finish just fine.  Obviously print orientation needs to be considered, but for my functional prints i don't do cleanup other than removing support. 

For more aesthetic stuff to steal a phrase, filler and paint makes it the printer it ain't.

Sorry, I just like nicely made stuff.

Don't you just hate it when people make it sound as though you have to choose between functionality and quality/aesthetics/fit-and-finish/look-and-feel, as if they were somehow mutually exclusive?

If anything, I'm kinda surprised that engineer-y folks don't have higher expectations on quality, given that engineering is all about making good designs. I would expect engineers to appreciate build quality more than the average joe. 

[Mr. Scram]

Don't you just hate it when people make it sound as though you have to choose between functionality and quality/aesthetics/fit-and-finish/look-and-feel, as if they were somehow mutually exclusive?

If anything, I'm kinda surprised that engineer-y folks don't have higher expectations on quality, given that engineering is all about making good designs. I would expect engineers to appreciate build quality more than the average joe. 

There seems to be an attitude that engineering types don't need nice things. It seems to be partly bravado, and partly the false antithesis you speak of.

Sure, technical people are often more than capable enough to deal with clunky designs, and generally wouldn't sacrifice performance for looks, but why not make something more functional and nice to use by designing and constructing it well?

[tooki]

There seems to be an attitude that engineering types don't need nice things. It seems to be partly bravado, and partly the false antithesis you speak of.

Sure, technical people are often more than capable enough to deal with clunky designs, and generally wouldn't sacrifice performance for looks, but why not make something more functional and nice to use by designing and constructing it well?

Preach!!!

My professional background is in technical writing and usability in the software world, and even there you run into people who think that professional software mustn't be too elegant. It gets dismissed as consumer-y, even if it's just as capable and much easier to use.

Every usability person understands that the usability demands of consumer products are very different from those of professional tools (in that the latter situation can tolerate a steeper learning curve in exchange for greater long-term efficiency), but often times, professional software is just needlessly complicated. It really is an attitude of "It's professional software, we don't need to make it user-friendly." Somewhere on this forum, I wrote a long rant about how maddeningly difficult and error-prone Eagle and KiCad are to learn and use, and of course many people just blamed it on me. But I feel that it's eminently possible to make professional software that is literally an order of magnitude easier to use than the incumbent, because I have literally seen it happen.*


A similar effect happens in scholarly/scientific writing: simple, clear language is actually often frowned upon!!! 

The old HP Journal, by the way, is a great example of clear scientific writing if you ask me. It makes no illusion of pretending to be for the general public, and it delves deeply technical frequently, but the editors did a great job of making it highly readable.




*1. When Apple introduced the original Final Cut Pro, after buying it from Macromedia, it was squarely targeted at pros, and was packed with features. Nonetheless, a n00b video editor could sit down at FCP and learn the basics of editing in 10 minutes. From talking to video pros back then, they said that a novice could learn in a morning or maaaybe full day on FCP what took a week to learn on Avid, the leader of the day.
2. When before-it-turned-evil Adobe introduced InDesign 1.0, it managed to be simultaneously much more powerful and yet far easier to use than Quark XPress, the leading DTP app back then. They really struck a nice balance between learning curve and ongoing efficiency.

[PlainName]

why not make something more functional and nice to use

I don't have any argument with that. My beef was the implication that one couldn't do that with 3D printing. It is possible, and it is just 'finishing' the product as you would via any other manufacturing process.

by designing and constructing it well?

Well, this is the issue, isn't it? There are a couple of simple answers:

1. You (not you personally, of course, but the 'royal' you) can't be arsed. Maybe you've milled and drilled an ally box and just can't be arsed to anodise it. Same thing.

2. You don't have the skillset. I think this is probably the more common reason - you might if you could, but not everyone knows how or has the vision. It is like drawing in that some people can bash off stunning portraits where others can barely manage a smiley emoji. You need to  be somewhat of a creative to imagine how your project box might look nice, and then have the skills to make it like that.

[ChrisLX200]

To me it's a matter of convenience, and whilst I prefer machined parts I'm more than willing to use a 3D printed part if it solves a tricky problem. Take the stepper motor bracket (below) which attaches a motor to a large focuser. To reproduce that using machined parts would have been quite a chore - and that's if I had managed to get the dimensions right first time (which I didn't quite). Using SketchUp I could calculate radii and intersection points then print the thing out, then make corrections if it wasn't quite right. The control box (green) was modded to include a step-down voltage converter (yellow version) simply by adding an extra compartment. The finish is 'acceptable' without further work - at least, I didn't feel inclined to reproduce the parts in metal.

[innkeeper]

what about diy box vs 3d printed

[CatalinaWOW]

If the objective is a mirror smooth finish, 3D printing is at a disadvantage.  Many other technologies come with a smooth finish for free, or nearly free.  While you can post process or use other techniques to improve the smoothness of the finish another approach is possible, and far easier with 3D than with other technologies.

Give the finish a deliberate texture.  Appropriate choices of texture can mask the roughness from the 3D process and enhance the beauty and functionality of the box.  Things like better grip, non slip stackability with minimal volume penalty, and so on. 

Our preferences in boxes and our definitions of quality have been shaped by the fabrication technologies we use.  They are not immutable.  Think of the hammered metal paint finishes that were popular toward the middle of the last century (partly because they hid manufacturing defects).

[Mr. Scram]

If the objective is a mirror smooth finish, 3D printing is at a disadvantage.  Many other technologies come with a smooth finish for free, or nearly free.  While you can post process or use other techniques to improve the smoothness of the finish another approach is possible, and far easier with 3D than with other technologies.

Give the finish a deliberate texture.  Appropriate choices of texture can mask the roughness from the 3D process and enhance the beauty and functionality of the box.  Things like better grip, non slip stackability with minimal volume penalty, and so on. 

Our preferences in boxes and our definitions of quality have been shaped by the fabrication technologies we use.  They are not immutable.  Think of the hammered metal paint finishes that were popular toward the middle of the last century (partly because they hid manufacturing defects).

You have to work with the qualities and shortcomings of the material, indeed, rather than ignoring or even denying them.

[ag123]

hi all,
I'm bouncing this old thread to ask a different question and to get a feel of the state of the art. It has been quite a number of years since this thread started.
Today 3d printing is a rather mature tech, many people probably have a 3d printer compared to back then.
However, my biggest peeve about 3d printing is that it can take a long time (hours) to print even rather modest enclosures.

Does anyone here routinely literally 3d print your own project boxes?

take for an example, this instructable
https://www.instructables.com/3D-Print-a-Custom-Project-Box-Without-CAD/

the author literally runs a service to do just that, generates the STL model for you to make a project box (free)
https://lightningboxes.com/
and if you don't have a printer, you can purchase it from them.
https://lightningboxes.com/product/basic-box/

you can literally download the generated STL file from the same link.
first the specs, the default selections generates a 100 mm (length) x 75 mm (width) x 50 mm  (height) box
https://lightningboxes.com/product/basic-box/

The thing is the basic model as given there, I run it through pruslicer
and for my very 'average' 'low cost' 3d printer (a Creality Ender 3 variant), it gives an estimate of 6 hours 18 minutes to print that.
This is a pretty long time considering that it isn't a 'big' box after all. I tried my own designs and the times run into roughly the same zone, 5-6 hours and more easily on average.
What could be worse is if you consider failure rates, e.g. with the 'difficult to print' ABS, that needs a heated bed and is prone to warping, 'complicated' 'large' models are particularly prone to warping.
There could be other failures, you could imagine worse situations if failure occurs.

Hence, I'd like to hear from those who literally 3d print your own boxes.

The thing about 3d printing boxes of course is, if there are a lot of custom openings and particulars, all that (windows and customizations) can go into the model. That is a big win vs conventional boxes. But the lengthy print time is still inevitable for the common FDM printers.

And there are various limits as well. One rather 'common', but I'd guess overlooked issues is, FDM 3D printing is a layered technology, with one layer stacking on top of another.
This means large vertical windows are *overhangs* and they are a problem to print. It would either need supports or that the whole model need to be redesigned to eliminate supports etc.
Possibly a multi-part print which needs to be glued together etc.

as to 'quality', 3d printing today is simply deemed *different*, some people literally likes those 3d printed 'layered' look, it is more of a preference / (personal prejudice).

[xrunner]

I've 3D printed the entire box for some of my projects. But another idea I use is to utilize an aluminum shell and just print the inside support, front, and back.

I'm showing a project I did last year which utilizes an Arduino platform to decode NMEA data from my GPSDO and display a custom time display, # of Sats, HDOP, and Alt. I used an aluminum project box and custom designed the inside support, front, and back. Turned out really well and saves a lot of time (and looks pretty professional).

[PlainName]

Neat 

[tooki]

Funny that this thread got resurrected recently. Nearly 5 years ago, I wrote:

There seems to be an attitude that engineering types don't need nice things. It seems to be partly bravado, and partly the false antithesis you speak of.

Sure, technical people are often more than capable enough to deal with clunky designs, and generally wouldn't sacrifice performance for looks, but why not make something more functional and nice to use by designing and constructing it well?

Preach!!!

My professional background is in technical writing and usability in the software world, and even there you run into people who think that professional software mustn't be too elegant. It gets dismissed as consumer-y, even if it's just as capable and much easier to use.

Every usability person understands that the usability demands of consumer products are very different from those of professional tools (in that the latter situation can tolerate a steeper learning curve in exchange for greater long-term efficiency), but often times, professional software is just needlessly complicated. It really is an attitude of "It's professional software, we don't need to make it user-friendly." Somewhere on this forum, I wrote a long rant about how maddeningly difficult and error-prone Eagle and KiCad are to learn and use, and of course many people just blamed it on me. But I feel that it's eminently possible to make professional software that is literally an order of magnitude easier to use than the incumbent, because I have literally seen it happen.*
…

Well, over the past 2 years, I’ve become quite proficient with Altium, and I think it falls into the category of “complex because it’s professional software, but still fairly easy to learn”. They clearly put a lot of effort into making it easy to use. (Still lots of room for improvement, but it’s not terrible.) I haven’t revisited Eagle since it’s not free for hobbyists anymore, KiCad still makes me homicidal, and what I’ve heard about other EDA programs (Cadence, Mentor, Eagle, Orcad, Pulsonix, etc) from fellow students and on here make it clear they’re much worse than Altium from a usability standpoint. So I feel vindicated that my gut feelings about Eagle and KiCad’s interfaces were right: they ARE more difficult than they need to be. Altium proves it.

[thm_w]

hi all,
I'm bouncing this old thread to ask a different question and to get a feel of the state of the art. It has been quite a number of years since this thread started.
Today 3d printing is a rather mature tech, many people probably have a 3d printer compared to back then.
However, my biggest peeve about 3d printing is that it can take a long time (hours) to print even rather modest enclosures.

Yes, in Cura on my settings (0.3mm layer, 0.7mm line, ~60mm/s) thats about 4.5hr print. But, its also quite a large box ~7x10x5cm.
I don't really care about time to print. Make sure your printer is fire safe, and let it run overnight if you need to.

What could be worse is if you consider failure rates, e.g. with the 'difficult to print' ABS, that needs a heated bed and is prone to warping, 'complicated' 'large' models are particularly prone to warping.
There could be other failures, you could imagine worse situations if failure occurs.

Use PLA or PETG, in place of ABS.
Use a heated bed, $200 printers come with heated beds now.

And there are various limits as well. One rather 'common', but I'd guess overlooked issues is, FDM 3D printing is a layered technology, with one layer stacking on top of another.
This means large vertical windows are *overhangs* and they are a problem to print. It would either need supports or that the whole model need to be redesigned to eliminate supports etc.
Possibly a multi-part print which needs to be glued together etc.

OK, nothing wrong with using supports in a print where necessary.

Better off designing your own box instead of using this generator, if you are interested in that.

[sleemanj]

and for my very 'average' 'low cost' 3d printer (a Creality Ender 3 variant), it gives an estimate of 6 hours 18 minutes to print that.

On my ancient printer with my current settings it would take 2h 42minutes according to Prusa Slicer, so probably about 3h in my experience on my printer.  And that's with ironing.

The difference is I use a 1mm nozzle and default to 0.65mm layer height.

That box also has pretty thick walls, if I was designing it for my printer I'd make the walls a single perimeter probably, and that would be a lot quicker to print since an entire 17m is spend on gapfill alone due to the wall dimension, not to mention the extra perimeter.

Now certainly at 0.65mm layers the layer lines are very evident but personally, I like the aesthetic and the speed gain is worth it.

[CatalinaWOW]

I design and print my own boxes.  They usually are designed to snap together and have appropriate apertures for connections, switches, displays and the like, and sometimes I design in incised or raised labels.  Cost is hard to evaluate.  They use a dollar or so of filament and a few cents electricity.  And take several hours to design, usually involving a couple of iterations to get them "perfect".  At standard engineering labor rates that makes them incredibly expensive.  But the time is usually less than I would have spent purchasing and modding a standard box.  Long print times don't matter to me.  I don't;supervise the oven, coffeemaker, clothes dryer, dish washer or furnace and see no real reason my printer is more hazardous than those devices.

The real magic is that when I need a second copy of the project the enclosure takes only a couple of minutes of my time.  I now rarely do anything other than print project boxes.

[ag123]

Thanks all !

I think xrunner's hybrid approach is a nice touch.
https://www.eevblog.com/forum/chat/cost-3d-printed-enclosure-vs-project-box/msg4184563/#msg4184563

It looked nice and 3d printing fills in for things that are flat with openings etc. in particular the panels part. And as well as the supports.

I'd also attempt a print for a full box to see how things goes, the dimensions are similar to that posted earlier, but that I've included all the windows and openings (e.g. drill holes) that i needed for my box.
long print times is inevitable with this approach, but that the openings are built into the design.
long print times and *large parts* significantly increase the risks of failed prints, hence these are concerns. This is in addition to the irritation of long print times.

There is one thing i might explore is to place more ventilation openings in my design as holes are literally 'free' with 3d printing, as compared to machining drilling etc.
It seemed with 3d printing, the more 'holes' there are in your application / box, the greater is the benefit, e.g. small support structures etc.

[mawyatt]

I think xrunner's hybrid approach is a nice touch.
https://www.eevblog.com/forum/chat/cost-3d-printed-enclosure-vs-project-box/msg4184563/#msg4184563

Agree, that's the same approach we took with the LCR Bias Adapter. We cut the Al extrusion in half (~65mm) and 3D printed the end plates with text. The supplied Al end plates were drilled for the 4 BNC connectors and mounted under the 3D printed plates. These Al plates aren't necessary but provide a higher degree of shielding for the adapter. The PCB slides on the extrusion guides and held in place by the BNC connectors.

Overall this approach provides a rigid mechanical and electrical environment for sensitive electronics

https://www.eevblog.com/forum/projects/bias-network-for-lcr-meter/msg4170607/#msg4170607

Best,

[sokoloff]

hi all,
I'm bouncing this old thread to ask a different question and to get a feel of the state of the art. It has been quite a number of years since this thread started.
Today 3d printing is a rather mature tech, many people probably have a 3d printer compared to back then.
However, my biggest peeve about 3d printing is that it can take a long time (hours) to print even rather modest enclosures.

I don't really care about time to print. Make sure your printer is fire safe, and let it run overnight if you need to.

Same here. I regularly print project boxes, RPi enclosures, RPi DIN rail mounts, and a custom box for a low-volume tool I sell.

I've almost never worried about the print time, because my involvement in the process is around 5 minutes total and the robot (3D printer) can take as long as it needs; I can check on it from my phone if I'm really worried about when it will be done (which is almost never). Once the first layer goes down correctly (on one of my printers), I can walk away. On two other printers, they are dialed in and reliable to the point that I just take a look at the webcam pointing at them [to make sure the bed is clear], and send a print job to them without even walking downstairs. 5 or 10 minutes later, I'll look at the webcam again just to make sure everything started correctly, but it almost always does.