Resin 3D Printers for the Masses

[Builder 2010]

I didn't search the forum to see if this topic has been addressed. If it has, please forgive me. If not, please take a look since what I'm describing (to me) is really mind-blowing.

Up until about 2 years ago, high resolution resin 3D printers were about $3,000 USD and up…way up. Along comes a new approach using a high resolution LCD panel illuminated by 405nm UV LEDs. Sitting atop of the LCD is a resin reservoir with a transparent teflon film bottom that is in direct contact with the LCD. The only moving part is the Z-axis which which rises one layer at a time as the part is being formed out of the liquid. I bought an Elegoo Mars LCD Msking printer from Amazon for $359.00 a month ago. The UV light cures a photopolymer in about 8 seconds per layer. The object to be printed is sliced like a salami into specific layers. The resin printer is capable of resolving layers as fine as 10 microns. That's 0.0007". The average filament (string) printers are capable of 100 microns and is too coarse for model scales that we typically work in. This machine is so amazing that it transfers the challenge of modeling from "How can I build it?" to "Can I draw it?" While it can theoretically print anything within its size constraints—in this case 2" X 4" X 6" high—there are physics to consider. The print is created upside down. The first layers to form stick to the aluminum platen connected to the Z-axis lead screw. The platen touches down at position zero and then moves up one layer thickness. The LCD, which is effectively black and clear, shows the image of the first layer (like a cat scan), the LEDs light and the resin trapped under the platen cures to a solid. The platen moves up about 5 mm to break the suction with the teflon film and then repositions one layer higher and the process repeats, only this time the LCD displays the image for layer 2. At 40 microns which I'm using, most objects are between 800 and 1,800 layers. The time to print is not based on complexity since the entire LCD is exposed at one time. Instead, it is based on three factors, the height (= number of layers), the layer thickness, and the exposure time. The default is on minute for each of the base layers (about 10) and then 8 seconds per layer.

I found out about this through one of my build threads, and after researching, bought one. My son said, "Jeez Dad, for $350 how can you go wrong?" The resin is not cheap, but the Elegoo resin is one of the least expensive at $30 per liter.

The hardest thing to master is understanding how to design for this kind of machine. Adhesion to the platen is the critical factor. If the model being built separates partially or completely from the platen, the model ceases to build. If the part can't be lifted from teflon film, then a new layer can't be formed under it. It just sits as a blob stuck to the teflon. The slicer software that came with the printer can build a labyrinth of supports that hold all parts of the model to the platen. What I had to learn was how to fine tune this support network. You can customize the support structure at a very granular level, but there are three easy choices: light, medium and heavy. I was using the default light setting, but realized after some failures that medium and even heavy worked better. The more mass the object has dictates using a heavier support structure.

i'm building a massive 1;48 engine house with an attached machine shop. I wanted to populate it with some O'scale machine tools, but none were commercially available. Then along came the printer. I downloaded a 3D image from SketchUp's 3D Warehouse and after some fiddling to make it print capable, printed out a very respectable engine lathe. It's about 2" long.

It was my second attempt, but was very encouraging. I also want to build a 1:48 model of an ElectroMotive Division 567 Diesel locomotive prime mover. I was thinking for years how to scratch-build one of these and wasn't making any progress, and then along comes the printer. After a couple of tries with different print configurations, I produced a pretty successful start.

Here's the real thing.

And here's where I am so far. I'm having to draw this entire thing in SketchUp first since no one has done a SketchUp model that I can pirate. This is how far I've gotten. The detail is astounding. The cam shaft is actually sitting proud of the head deck as is the injector throttle shaft. Strangely, some of the details didn't print that were facing in one direction. I'll have to do some research on that. I'm in the process of drawing that complicated maze at the front end of the engine. EMD diesels had four exhaust valves. Under the two side rocker arms was a bridge rocker that pushed two valves on each side of the injector, which was operated by that middle rocker arm. There are actually valve springs under those, but they're hard to see.

I tried printing parts with them sitting horizontally in the slicing software. This works for parts with small surface areas, but as the surface area in the horizontal direction increases, the suction on the teflon is also increased and they start breaking away from the supports. It's best to situate the model at 45 degrees. This minimizes the horizontal contact area and produces a better print. It also increases printing time… a lot. The machine is very quiet and just goes about its business. You can start something at night and it's done the next morning. Unlike the string machines, it doesn't have a nozzle that can clog with molten plastic, or have a filament feeder that can jam. Everything here is optical and electronic except for the Z-axis. The orange acrylic cover keeps UV light out of the printer to keep the reservoir from curing by ambient light.

Here's what a part looks like coming out of the reservoir. The engine block was almost perfect. I wasn't careful enough cutting away the forrest of supports and damaged the thin bottom flange.
You'll notice the angular interconnecting supports that lock the entire network together. It's really tough! I used the heavy support setting and nothing broke loose. That whole network is completely and automatically created by the slicing software. It's really easy.

The resin is somewhat toxic and you need to use nitrile gloves when handling it. You wash it in two baths of alcohol, either ethanol (190 proof grain alcohol) or 95% isopropyl. You wash it in bath one and a final wash in bath two. The part out of the machine is not fully cured and is still a little rubbery. That's when you cut off the supports. I then purchased an inexpensive LED 405nm UV curing light and assembled a foil-lined cardboard box to contain the light and evenly cure the parts. It takes just a few minutes to finish the curing and then part is very hard, almost ceramic-like. It's similar to the resin used by dentists with photo curing dental resins.

All in all, it has completely changed how I'm approaching my model making. It's not whether I can build it. It's can I draw it. They say any technology sufficiently advanced appears as magic. To watch a fully formed 3 dimensional, real object rising out of a bath of liquid not even an inch deep, appears miraculous. You don't see how it happens since the LEDs are hidden by the platen.

Just for fun I prepared a list of all the technologies in my lifetime that had to be there to have this machine appear on my workbench. I came up with 13, but there are more. I didn't consider things like AC generators for electricity, but I did include transistors, micro-processors and micro-computers, solid state memory, high accuracy stepper motors and their controls, photo-curing polymers, LEDs, LCD panels and control circuits, Teflon, 3D software, Slicing software, machine code, and the lists goes on.

 

[Airframes]

Impressive - too complex for my ageing brain, but impressive !

 

[vikingBerserker]

I so want one of those!

 

[Builder 2010]

I'm going to be 74 in two weeks. If my aging brain can take it, so can yours. My philosophy is to keep pushing the envelop. I learn new software as often as I can. While the 3D printer was a bit daunting, I've been able to logically analyze what's going on and work to improve it. I wake each morning building stuff in my head. This morning I reconstructed the process by which the printer is producing the engine block and discovered why the lower edges of the access hatches weren't there. I drew them as they are in scale, with the lid overlapping the base and overhanging a bit. The printer, because of how the part was aligned, had the LCD expose the little bits of overhanging material before it exposed the solid part of the model that it would be connected to. So that bit just formed in free space. Resin printers can't form anything in free space, so the little bit solidifies and then floats away. It isn't until the exposures got to the part where the rest of the engine block was being formed too that the overhanging flange took shape. So I learned that you can't have parts where unsupported overhangs get exposed before the body of the part does. They will just drift away one very thin layer after another. It's logical and can be learned.

 

[Crimea_River]

The last pic reminded me of the new resin Beaufighter cockpit parts seen here: Model Monkey 1/32 Beaufighter Mk.VI & TF.Mk.X Cockpit by Brett Green

 

[Airframes]

Oh, I'm sure that I'd figure it out if I had one. It's always much easier to understand things when the object in question is in front of one - although, of course, there are exceptions, such as trying to figure out how a woman's brain works !!!!

 

[mikewint]

how a woman's brain works !!!!

Never be too disturbed if you don't understand what a woman is thinking. They don't do it often.

 

[Gnomey]

Cool!

 

[Donivanp]

I am on the "so want one list." I've been drooling over one for the last three or four years.

 

[Red 3]

Very interesting, exactly what do you mean by "drawing"? I have read about scanning things for other 3D printers, is it something similar?
Thanks.

 

[Builder 2010]

I've given up figuring out women's minds. I've been married to one for 51 years and still haven't gotten it right.

Drawing… A 3D printer is driven by a 3D drawing file. This file can come from three sources. You can obtain STL files from others (STL= Standard Translation Language; a file type for conveying a 3D image) or you can draw your own in a 3D drawing program. Some are free, like Blender, others are free for some versions like SketchUp Make or for fee like SketchUp Pro. SketchUp is very elegant and is easy to learn to draw simple things, but does take quite a bit of learning to master more sophisticated things like contoured surfaces… kind of like learning to play guitar. Play a few chords = One week. Play like Jimmy Page = a lifetime.

The third way is scanning parts with a 3D laser scanner which will generate an STL file, but the good ones are very expensive. I you wanted to replicate an existing piece from a model, for example, scanning would be the way to go. My route, so far, is drawing them on SketchUp. SketchUp has a vast library of 3D drawings done by thousands of users all over the world. I'm finding that that some of these are not fit for 3D printing. They look good but they're not "solids" as defined by 3D software. To print an object is must be a solid. It must have all of its faces and edges connected to one another with no extraneous lines or shapes. If it is just a wall, but unclosed, the printer can't understand it. The wall in SketchUp has no depth. It can't be made since it couldn't exist in physical space. When you produce the STL file, it displays on the screen as a 3D object that you can spin around a look at all sides. If it has unprintable areas you will see them as missing things on the STL display.

SketchUp doesn't natively produce STL files, but there is a free extension that converts the SketchUp (SKP) file to STL.

The next step is slicing. 3D printing works like a cat scan in reverse. A cat scan is an Xray that slices up a physical thing into a series of thin cross-sections images so they can be viewed. The 3D printer takes these thin cross-sections and piles them, one on top of another, to create a physical object. Therefore, the STL file needs to be sliced into these individual printable layers. Depending on the kind of printer and its resolution, the slices are made to specific thicknesses. Filament Additive printers generally have resolutions (layer thicknesses) of 100 microns or larger). Resin printers, like the one I now have, generally work at layer resolutions of 50 microns are less. I'm printing at 40 microns (about 0015"). It can go down to 10 microns which is ridiculously small. The Z-axis step motor is capable of moving the lead screw by those tiny distances and doing it repeatably.

Here is the STL file of the entire Gantry Hoist assembly that I'm building for my large engine house project. I originally had just the motors and gear boxes printed by Shapeways, a for-profit 3D printing house, for $40. Now, with my own machine, I'm going to attempt to print the entire hoist in one go. I didn't originally draw this. The entire gantry crane was downloaded from the SketchUp warehouse, but it needed extensive re-drawing to make all those interesting shapes into printable solids. It took all of my SketchUp skills to pull it off. The resin printer will resolve all those fins! The facets on the curved surfaces are the direct result of having SketchUp producing circles with 24 segments. You can dial up any amount you want, but the higher the count the more CPU time the software needs and the slower it works. Since this is going to be inside the building and won't be visible unless the roof is removed, those segments don't matter. If I was building something that would be viewed more closely I would increase the segment count.

Here's the engine house all this in going on. It's really big.

I'm 3D printing four air conditioning units for the roof. I'm doing them in four parts. It was amazing (to me) that the printer produced the door handles projecting out from the surface. You can't do this with injection molding without very expensive slide molding systems.

Right now my challenges are learning how to effectively design for 3D printing. There are stresses created in the resin when it transitions from liquid to solid and you need to design the part to reflect these. This means that some parts, even though you can draw them, may not turn out as you expect.

 

[Red 3]

I've given up figuring out women's minds. I've been married to one for 51 years and still haven't gotten it right.

Drawing… A 3D printer is driven by a 3D drawing file. This file can come from three sources. You can obtain STL files from others (STL= Standard Translation Language; a file type for conveying a 3D image) or you can draw your own in a 3D drawing program. Some are free, like Blender, others are free for some versions like SketchUp Make or for fee like SketchUp Pro. SketchUp is very elegant and is easy to learn to draw simple things, but does take quite a bit of learning to master more sophisticated things like contoured surfaces… kind of like learning to play guitar. Play a few chords = One week. Play like Jimmy Page = a lifetime.

The third way is scanning parts with a 3D laser scanner which will generate an STL file, but the good ones are very expensive. I you wanted to replicate an existing piece from a model, for example, scanning would be the way to go. My route, so far, is drawing them on SketchUp. SketchUp has a vast library of 3D drawings done by thousands of users all over the world. I'm finding that that some of these are not fit for 3D printing. They look good but they're not "solids" as defined by 3D software. To print an object is must be a solid. It must have all of its faces and edges connected to one another with no extraneous lines or shapes. If it is just a wall, but unclosed, the printer can't understand it. The wall in SketchUp has no depth. It can't be made since it couldn't exist in physical space. When you produce the STL file, it displays on the screen as a 3D object that you can spin around a look at all sides. If it has unprintable areas you will see them as missing things on the STL display.

SketchUp doesn't natively produce STL files, but there is a free extension that converts the SketchUp (SKP) file to STL.

The next step is slicing. 3D printing works like a cat scan in reverse. A cat scan is an Xray that slices up a physical thing into a series of thin cross-sections images so they can be viewed. The 3D printer takes these thin cross-sections and piles them, one on top of another, to create a physical object. Therefore, the STL file needs to be sliced into these individual printable layers. Depending on the kind of printer and its resolution, the slices are made to specific thicknesses. Filament Additive printers generally have resolutions (layer thicknesses) of 100 microns or larger). Resin printers, like the one I now have, generally work at layer resolutions of 50 microns are less. I'm printing at 40 microns (about 0015"). It can go down to 10 microns which is ridiculously small. The Z-axis step motor is capable of moving the lead screw by those tiny distances and doing it repeatably.

Here is the STL file of the entire Gantry Hoist assembly that I'm building for my large engine house project. I originally had just the motors and gear boxes printed by Shapeways, a for-profit 3D printing house, for $40. Now, with my own machine, I'm going to attempt to print the entire hoist in one go. I didn't originally draw this. The entire gantry crane was downloaded from the SketchUp warehouse, but it needed extensive re-drawing to make all those interesting shapes into printable solids. It took all of my SketchUp skills to pull it off. The resin printer will resolve all those fins! The facets on the curved surfaces are the direct result of having SketchUp producing circles with 24 segments. You can dial up any amount you want, but the higher the count the more CPU time the software needs and the slower it works. Since this is going to be inside the building and won't be visible unless the roof is removed, those segments don't matter. If I was building something that would be viewed more closely I would increase the segment count.

View attachment 545240

Here's the engine house all this in going on. It's really big.

View attachment 545241

I'm 3D printing four air conditioning units for the roof. I'm doing them in four parts. It was amazing (to me) that the printer produced the door handles projecting out from the surface. You can't do this with injection molding without very expensive slide molding systems.

View attachment 545242

Right now my challenges are learning how to effectively design for 3D printing. There are stresses created in the resin when it transitions from liquid to solid and you need to design the part to reflect these. This means that some parts, even though you can draw them, may not turn out as you expect.

Many thanks for the info, it's very intriguing. Somewhere along the line I just might have to try this.

 

[herman1rg]

Where I work we have a Formlabs resin printer, I think I shall do some experimenting (especially as it's now summer holidays for our college).....

 

[Builder 2010]

I solved the warping problem with the large HVAC box. I built in cross bracing (overkill), but it produced a dead straight output. It's very similar for designing parts for metal casting. You can see no inward bends on the access doors.

I made the bracing too heavy, but it worked. All that garbage inside is the remains of a forrest of support rods that the software installed. The box is very strong. You could stand on it.

Any dings or divots can easily be filled with Bondic, UV curing filler. It's the same chemistry as the resin in the model and produces invisible fixes. It's a steep learning curve, but I'm figuring it out.

 

[Ralph Haus]

Builder - Have you by 1 year on age and 4 on marriage (still haven't figured out the female mind!) and I'm too still learning (both engineering advances and women!).

Anyway, I'm still doing engineering design using SolidWorks, which will export STL files, and a device like this would come in handy both work wise and for home. I'm fairly sure that the 'safety' powers at my location would nix having this on site, BUT, if I can provide them with a toxic free parts which are created off site, they may spring for one? I will need to research the resin for structural characteristics (use as prototype parts) but in general how 'robust' would you consider a part to be?

I am impressed with the layers not being visible. Years ago (30 odd) I was involved in getting an early design like this into the development lab; the 'coarseness' of the part kept it from being a useful tool. The image I have attached is of a bottle which has a common top and bottom. Fun to design. It will hold liquid. Just hard to empty! It illustrates the issue with the coarseness of the layers.

I see ELEGOO on Amazon but only the Neptune (filament). Where did you buy yours?

 

[herman1rg]

Ah the Klein Bottle always fun

 

[N4521U]

Will it also print from a 3D PDF file?????

 

[N4521U]

We in Oz do live on an island.... doesn't come Close to your 3 hundred something deal.....
$482 AUD
ELEGOO Mars UV Photocuring LCD 3D Printer with 3.5'' Smart Touch Color Screen Off-line Print 4.72(L) x 2.68(W) x 6.1(H) Printing Size Black Version | Buy Products Online with Ubuy Australia in Affordable Prices. B07K2ZHMRF
$1200 AUD
LAST 2 LEFTBRAND NEWElegoo Mars SLA Printer Black Resin UK Photocuring | Buy Products Online with Ubuy Australia in Affordable Prices. 323874180475

 

[Builder 2010]

The printer won't take a PDF, but you could import that into something else and then work upward. The STL file is an accurate description of the surfaces of an object and how they relate to one another. A PDF is none of those things. You could import the PDF into a vector drawing program. From that you could export three views into a 3D drawing program; SketchUp for example. In SU you would redraw the object in full 3D by using the 2D views as guides. It's a long convoluted process.

BTW: in the two months since I got the Machine I've done some seriously amazing things. Here's are some examples: This ia 1:48 model of an Electromotive Division 567 diesel locomotive engine. It is a reject print as noted by the misshapen rear exhaust muffler. But you can see the valve gear with the injectors which is exceptional. The model is comprised of 7 pieces. I've got another partially done that fixes all the errors in this one. No model of this engine in this scale exists or has ever existed. Now it does.

All the parts on the roof of my scratch-built engine house are 3D printed.

Inside the engine house is going to be a machine shop. Some of the machines I was able to download from SketchUp's 3D Warehouse and were highly detailed, but needed extensive editing to make them acceptable to 3D printing. Others have been designed by me. The first is a large locomotive wheel lathe that is a freelance design. I couldn't find enough images on a complete machine so I made my own. The model consists of bed, left headstock assembly, right assembly, left and right tool carriages and a cross-shaft with gears under the floor. The faceplate cleats are also separate parts.

This is a radial drill press. Again, there was none available on SketchUp so I found a prototype and drew my own from this one. I only had one dimension on which to base it… a 17" column. From that I was able to draw it.

This lathe was a SketchUp model. I erred on the editing leaving the chip pan scale thickness. It printed as thin as tissue paper. I had to reinforce it with some styrene. If a free standing part is smaller than one inch in 1:1 world, it will print too fine to stand up in the 1:48 world. You have to go in and selectively beef up parts to make them sustainable in resin in scale. I'm learning every day how to manage it. The stock in the chuck is an actual piece of drill rod. The Machine printed the part with a hollow spindle. The repaired chip pan actually came out looking a bit damaged like a real machine would be if you rammed it with a fork lit truck.

Needless to say, 3D printing capabilities have dramatically change how I approach model making.

 

[N4521U]

No pictures?
Or are they part of another alter relm???????