3D printers - use for making replacement parts?

[1980s_john]

I was reading an article in Wired regarding 3D printers, which mentioned these are used to make replacement parts. I assume this applies for parts for vintage computers (eg a floppy disk drive latch) as well as anything else.

Anyone done any experiments in this area?

Regards,
John

 

[Ole Juul]

I haven't tried this, nor inspected any parts made with 3D printers, but I am curious about the quality of the material. I too had thought that little parts like drive latches could be made cheaply in a one-of situation. However, the other day somebody posted a link here to a place that does this in numerous materials and after reading the fine print, it became obvious that the material is crap - at least at that place. Probably a great item for covers and things which are basically cosmetic, but a drive latch will take strain. Materials are important. There are good reasons why people use plastics like nylon, or forged (as opposed to cast) metals. I'll look forward to reading what other people know about this stuff.

 

[NeXT]

From what I have seen so far, all the home built units that you find or whose plans are available online have really poor quality results on parts with fine details. They also seem to be more fragile than those made using professional 3D printers.

 

[deathshadow]

I worked for a number of years in plastics manufacturing, and from what I've seen it's mostly a matter of resolution; or more specifically a total lack of it. Most of your sub $10K printers top out at somewhere around 250dpi, but with a minimum dot size of around 0.5mm. The cheap-cheap builds like repwrap are roughly 75dpi lateral and a very poor 50dpi on the z. Variance due to the difficulty in controlling a stream of melted plastic, unpredictability of 'drag' when you turn off the stream, and issues like shrinkage during cooling only further exacerbates the detail level issues.

Meaning for anything truly detailed or precise, it's just the wrong tool for the job. At best you can rough out the general shape, and then hand finish it with files and precision cutting tools, with no two being a perfect match. Really the technology is in it's infancy.

In terms of materials the selection available for 3d printing is not the greatest -- the best of the lot is Acrylonitrile Butadiene Styrene - aka ABS. This stands to reason as a lot of injection molding uses this same stuff. The problem is that ABS is best formed under high pressure (250+PSI) at very high temps (1500F+) and flash cooled. At those temperatures you need to form inside a pressure vessel (like between metal plates held by a 2000+ pound hydraulic press) or there's a massive release of noxious and dangerous fumes. 3d printers operate far, far below those tolerances so the resulting styrene is very very soft and maleable, as opposed to say, the ABS of your typical black drive faceplate.

The other materials go downhill from there, the polyethylenes, polycaprolactone and polypropylenes are also formed at the edge of their temperatures meaning at best you get the consistency of a cheap dollar store dog dish, at worst something more the consistency of hard wax. A far cry from the more durable forms of said chemicals like you'd find in vinyl, polyethylene terephthalate, nylon, etc.

Even polystyrene, the classic easily glued with solvent plastic used in everything from kitchenware to modelmaking, gets widely different results depending on temperature, pressure and cooling... the requirements of operation for a 3d printer and the requirements for best forming the plastics just don't line up; not in the same ballpark? We're talking not even the same country!

Really, given what's involved if I was looking to make classic parts, I'd go out and get a real one, make a casting of it in silicone and then do a resin casting. Temperature cured resins are extremely hard, durable and a great choice for low count manufacturing. Companies like Smooth-on have full on tutorials on their websites on doing this. I like them because they've got stuff that's great for hobbyists, prop makers and engineers alike -- I've used their stuff a couple times to make duplicates of my hand-carved miniatures (excuse the horrifically out of date piss poor website -- I keep meaning to make a new version, never get to it!).

Once you've got your first one made, nothing beats casting for making more of them.

 

[Chuck(G)]

What's true is that 3D printing is a very hot field right now. It seems like every week someone's got a new process.

Here, for example, is 3D printing in stainless steel:

It's not perfect, but it's certainly interesting. However, for onesy-twosy parts, unless they're horrifically complex, it might be better to simply machine the thing out of acrylic.

 

[Ole Juul]

. . . anything truly detailed or precise, it's just the wrong tool for the job. At best you can rough out the general shape, and then hand finish it with files and precision cutting tools, with no two being a perfect match. . . .

. . . at best you get the consistency of a cheap dollar store dog dish, at worst something more the consistency of hard wax. A far cry from the more durable forms of said chemicals like you'd find in vinyl, polyethylene terephthalate, nylon, etc.

. . . the requirements of operation for a 3d printer and the requirements for best forming the plastics just don't line up; not in the same ballpark? We're talking not even the same country!

Great rundown. Thank you so much for this!

This is exactly what I thought. I've seen very durable resin casting, and of course detail can be exquisite with that process. I suspect it is a better way to go than "printing" in most cases.

It seems too, that injection moulding is the way to go for real quality in most cases. I know moulds can be extremely expensive, but make up for it in speed and quality of production. I've always been impressed with the accuracy, strength, and sheer endurance of Lego blocks. If you've ever dug an old one out of the driveway you'll know what I mean. Often not a scratch on then.

 

[Ole Juul]

Here, for example, is 3D printing in stainless steel

Interesting. But why is the type backwards?

How strong is that "stainless steel"? Is it usable?

 

[deathshadow]

What's true is that 3D printing is a very hot field right now. It seems like every week someone's got a new process.

... and the printing in metal is one of the most promising, though you're looking at a quarter million bucks startup cost last I checked.

I've seen wax printers from a company just up the road going for $100K a pop... and wax is a cheap and simple process!

Wax printing has some interesting advantages -- low melt temp,fast cure time -- fast enough that you can reduce shrinkage issues, and really high (300dpi+) resolution -- but best of all, it's wax, meaning you can just get two or three base renders done, and then use them to make hard casts for metalwork like bronze or pewter... where you cast your mold in clay or sand, then 'melt out the wax' before dribbling in the metal.

The gear on that page you linked to is proof enough of why it's still a very rough process -- the striations inside the gear are completely unacceptable in an operating environment; you might as well use sand as lubricant for all the mechanical soundness of such a design -- which is why as you said, it's probably better to just machine an original. Read up on their process though -- it's loose steel lattice they still have to hand clean before they give it a bronze dip, so structurally it's more of a loose braze. Given the low accuracy (look at the inset fonts to see just how bad we're talking) and bead size, again it's probably alright for prototyping, but not something I'd even consider using as a production master or as a working part... They even pretty much say it on the site:

The minimum recommended wall thickness for Stainless steel is 3.0mm and minimum detail size is 1.0mm

-- ouch. 1mm detail accuracy is pretty much useless for anything 'functional'; hand done machining from the 18th century was more accurate.... but it is an emerging technology; right now what we're seeing is, well... reprap is the technological equivalent of the Altair 8800 and some of the other ones are closer to the Elf... The big industrial printers being the equivalent of a PDP-11. You have to look at where it's at in that light.

 

[Ole Juul]

-- ouch. 1mm detail accuracy is pretty much useless for anything 'functional'; hand done machining from the 18th century was more accurate....

Indeed. People were making trombone (sackbut) slides in the 15th century. Most musical instruments require working to 1/1000 inch or less which skilled people can do by hand using various tricks.

 

[Tor]

3D printing clearly isn't there yet.. but CNC has come a long way. I hadn't really thought it could do this though:
http://www.youtube.com/watch?v=reBPyJeyPac

-Tor

 

[tipc]

I've heard of this. W/regards to hobby stuff, more often then not the process is used to make a prototype (I think). Anything made of or cast in (most) plastics can be turned into say brass (or any other metal) w/comparative ease using lost wax or investment type casting. I'm going to guess Chuck's example was not made directly w/3d printing (though I could be wrong on that). You make a prototype out of plastic (or wax, hence the name) encapsulate in a plaster-*like* substance (sometimes referred to as _investment_), put that in an oven, cavity facing down, allow the cavity to be evacuated due to the plastic/wax melting, then invert the mold, and eventually the cavity will be *entirely* evacuated, the residues of plastic/wax being vaporized by ~1200 degree farenheit temperature. While the mold is still kind of hot, pour molten whatever into that. Toss it in a container of cool water, and the original mold will shatter, revealing the cast metal part. Hard machinable waxes are ideal for carving/shaping w/hand tools. Bring that master to a caster and have it turned into something far more durable, from which you can make a hard rubber mold for hundreds of repros.
I hear precious little about this stuff, but I was thinking you can make a metal mold this way, the reverse of making the object. I dont' want to go into laborious details, but you should be able to figure it out.
Get acquainted w/Lindsay publications. They have books on all sorts of crazy stuff, even how to make your own cheap hand operated injection molder.

 

[tipc]

also w/regard to making small parts, don't rule out the trusty old file, saw, drill, etc. You can make separate assemblies, and join them w/adhesives. It helps to have some experience w/stuff like this, but the only way to get there is to be taught it, or sit down and think about it (sometimes for a long time). I'm 45, and recently I got to wondering how a rubik's cube operated. Wasn't that hard. W/i 10 minutes I figured it out. I could show the drawings I did, but they would mean little to anyone but me (at this point they might not even mean that much to me).
You could jury rig a small milling machine, for wood and plastics, even a bit of metal (aluminum, yellow brass) w/a router or a drill press. You could conceivably make a small die (metal mold) for making repros in plastic. A bit tedious, but look at the other stuff people are doing - making cd-roms and modern hard drives work w/Commie 64s and whatnot.
You can do anything you want. Divide and conquer. Break a task into simple constituents. You might actually be surprised how relatively simple the solution is.

 

[tipc]

3D printing clearly isn't there yet.. but CNC has come a long way. I hadn't really thought it could do this though:
http://www.youtube.com/watch?v=reBPyJeyPac

-Tor

There is some very sophisticated stuff coming out of Germany. The machine in the video is very sophisticated (look at the clothesline of toolholders!), but nothing it's doing there is all that demanding. That's a large object milling out of aluminum, for the most part. You can make dies out of aluminum, drifting back on topic, and expect a few thousand faithful repros. But it's going to get tricky when you're making a die for say a Z-scale model locomotive or freight car. To date I don't know how much of that is done w/CNC. Mainly by hand I would suspect.
I was told by an engineer that your eye can see irregularities down to say .0005 inches (1/2 of 1/thousandth of an inch, or 5 10-thousandths, a sheet of paper is about 2.5 thousands thick). Generally that's the minimum repeatability (locatability, my word) you would expect from a cnc mill. Part of the superiority of CNC equipment is the use of ball screws instead of threaded rod and nuts. I don't know if floppies use ball screws or not (I should know shouldn't I!). It seems like some do, given the spherical grooves in the feed screws I've seen occasionally. The same principle applies, that is the ability to _locate_ with a stepper motor. Imagine a floppy disk or hard disk didn't have to spin to be read or written, and you had not only the head being driven backwards and forwards by a stepper motor, but side to side also, to locate any point on the platter using rectangular coordinates. That's how cnc works. Don't know offhand what the resolution is of a floppy mechanism. I tried doing the math once. It seemed pretty freaking precise. Maybe I need to do it again. That's how cnc works.
There's also water jet...