1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

3D printing - thinking of buying a machine...

Discussion in '3D Printing' started by jhiker, Sep 17, 2019.

  1. jhiker

    jhiker Alibre Super User

    Here at work we're thinking of buying a cheap-ish 3D printer to knock up simple prototypes. I'm guessing a 6" cube envelope might be OK but I don't really know much about the technology or how they might integrate with Alibre for ease of use. Materials-wise, ABS or similar I would think and it would have to have good dimensional stability/resolution.
    Any recommendations? Might this one be a decent buy?
    What do I need to consider?

    https://www.amazon.co.uk/ANYCUBIC-P...&pf_rd_p=1790d964-8706-5c28-a213-7c85e9f2cea6
     
  2. jhiker

    jhiker Alibre Super User

  3. DavidJ

    DavidJ Alibre Super User Staff Member

    The photo-curing technology can give some fantastic results. The 'goo' is pretty expensive, and you probably need to have reasonable turnover of parts as I believe the goo has a limited life once opened.

    You might want to talk to a company that sells 3D printers and does bureau printing. Ask John at Mintronics about GoPrint3D (wrong side of UK for you, they are based in Ripon) - they have demo/bureau machines of the major types (should have a metal printer by now too), they also sell and service 3D printers, and have built up a wealth of experience about getting the best results. You can get advice and use the bureau to try before buy, compare technologies, or just to cope with the odd job that your in-house machine isn't suited to.

    Have fun.
     
    Last edited: Sep 18, 2019
    jhiker likes this.
  4. bigseb

    bigseb Alibre Super User

    ABS prints? You want THIS!! The Up300. Definitely one worth shortlisting. And not too expensive for a company.

    index.jpg
     
  5. bigseb

    bigseb Alibre Super User

    If you're looking at the Elegoo Mars then also consider the Anycubic Photon. Read my review.
     
  6. bigseb

    bigseb Alibre Super User

    Here is a recent part I printed on the Photon.

    20190917_205329_resized.jpg

    20190917_212030_resized.jpg
     
  7. neilr

    neilr Senior Member

    bigseb likes this.
  8. Uman

    Uman Senior Member

    The Prusa MK3S printer should also be considered.
    Prusa:
    The magnetic removable flexible build plate with surface textures is a big plus.
    This allows such filament as PETG and its variations; less shrinkage, more stable than ABS.
    You may want to wait until the next batch of new models hits the market in the next several months to get the latest innovations.
    I am hoping for Prusa to increase speed and add high performance hot ends.
    I buy a new Prusa every model year and sell the old one on Ebay....their resell values are very high.
    At $999, it is hard to pass on.
    -Jeff
     
  9. jhiker

    jhiker Alibre Super User

    Seb - is that a vacuum pump? Do you de-gas the resin?
     
  10. bigseb

    bigseb Alibre Super User

    Vacuum pump? No.
     
  11. Toybuilder

    Toybuilder Member

    What kind of parts are you making? And what kind of resolution is important to you?

    FDM (filament) printers are great for overall use if you are typically making parts where features are generally several mm's in size or larger.
    As you get into the ones of mm's in size, it becomes harder to make acceptable parts.
    The cost of filament is just about negligible these days. Print speed for volume is pretty decent. Lots of material choices. When appropriately calibrated, print quality can be superb except for the fact that the extrusions are typically around 0.4 mm wide x 0.2 mm thick (Though 0.1 mm to 0.3 mm thickness are common.) Surfaces not perpendicular to the Z axis will show stair-step patterns.

    SLA/DLP resin printers are great for small parts or parts where sub-mm cosmetic details matter. The material choices are more limited, and they are costly. Also, the material requires more careful handling and requires even more ventilation. I am unfortunately sensitive to the stuff, so I don't own one.
     
    jhiker likes this.
  12. Michael Moore

    Michael Moore Member

    I have not used a printer but I have been researching the MSLA printers this past week, joining user groups (you need to get on the resin groups as well as those for the machines, and you also need to learn to run whatever slicer you decide to use), watching YT videos, etc, so this is my impression of what I've read, take it for what it costs you. Also, it looks like UK/Europe prices tend to be higher than USA prices, and some of the machines/resins may be harder to obtain.

    For the small size the Anycubic Photon, Mars Elegoo and Epax X1 seem to be good choices, ranked in that order from low to high. The Photon appears to have a very large user base that does a lot of hacking on them. The Mars seems to be considered a step up in quality and I think it and the Epax area also noticeably faster due to their more powerful LED arrays. The Epax is the most expensive, but it has a lot more sheet metal in the chassis and seems to get the nod as the nicest quality on the design/construction. Other brands that seem popular are Peopoly (they have some SLA machines and now MSLA) and Phrozen. Epax, Peopoly and Phrozen all have midsize MSLA machines and the prices for those are in the US$800-1200 range. The latter two companies have recently released even larger MSLA machines that are running about US$1800, and it looks like Epax has a similar machine in development for release later in the year. Here's some build volume specs from low to high:

    Elegoo Mars 120 x 68 x 155 mm <$300
    Phrozen Shuffle 120 x 68 x 200 mm
    Peopoly’s Moai 130 SLA 130 x 130 x 180
    Form3 SLA 145 x 145 x 185 mm ($3500)
    Phrozen Shuffle XL $1300 190 x 120 x 200 mm
    Photocentric Liquid Crystal HR2 196 x 147 x 250 mm ($2500)
    Epax X10 ($1199) 216 x 135 x 250 mm
    Epax 13.3 inch screen printer -- TBD, likely 2 months after 10.1 inch (shipping now so early Dec maybe)
    Peopoly Phenom (new, $1800 12.5" 4k) 276 x 155 x 400
    Phrozen Transform (new, $1800) 292 x 165 x 400 mm
    Form3L SLA 300 x 335 x 200 mm ($10K)

    It appears even the sub-$300 range machines can deliver quite good prints, but they have smaller build volumes, possibly some corners cut on the materials of the chassis to hit a price point, and they will probably print slower due to lower power UV arrays. There seem to be a lot of those very small printers that are "badge engineered" items, where the manufacturer slaps on a private label and ships them out.

    The FEP (or in the case of the Epax non-FEP) films and the LCDs are consumables. LCDs get rated for 300-500 hours. Of course you can damage one in much less time (usually loose bits of cured resin that get jammed into the bottom of the tank) and some apparently are going strong at 1000+ hours. Prices on the LCDs seem to be $80-140 depending on size and 2K vs 4K and some brands offer them at a lower price and those may well fit in some other brand's printers too. ChiTuBox appears to provide OEM slicer software, firmware and boards for a number of the printers.

    One piece of advice that some people stress (and say that following it means they have had zero problems) is to filter your resin after every print to make sure you don't have one of those stray cured pieces that can damage/puncture the film and/or the LCD. It is extra bother, but that seems to be the voice of experience speaking.

    I'm interested in using one for making foundry patterns, whether masters or actually in the sand (there are some strong/tough resins available -- look at Siraya Blu/Tenacious which may be fine for 1-3, and then there are resins for investment casting "waxes"). It turns out you can mix resins to get a blend of the capabilities, such as 10-20% Tenacious added to Blu to give it a bit of flexibility without degrading the strength. The MSLA resins are often much less than the FormLabs SLA resins, and the latter are pricey enough (along with the machine price) to give me pause about them. $40-90 (with I think a few castable resins at $150)/kg vs $150-300/kg seems pretty significant to me. I've seen photos posted by someone who is making small functional prototype parts pretty much everyday, and on one photo where he was holding a part with the largest dimension of 20mm his fingerprints were much coarser than the layer lines on the part.

    I'm looking at the mid-size at minimum, and might go for one of the new large format models though that means waiting to see what the larger Epax is like when released. The Phrozen Shuffle XL 2019 has been released recently with significant changes in the electronics (no longer using a Raspberry Pi in the machine) so going for the new version may be a good idea. The Transform and Phenom are shipping but it appears those may be later tier Kickstarter order fulfillment machines rather than full retail production parts.

    At this moment I'm tending towards the Phrozen, Peopoly and Epax. Epax seems to have a "heavy duty build" rep and they have a USA partner that can do sales/support instead of having to deal with Taiwan/Hong Kong/China direct.

    It would be nice if the screens, which are basically phone/tablet stuff, were more squarish instead of the long rectangles. I'm not sure that I'd need more than the 200-250mm X travel (400mm in the new big machines) but they all seem a bit scant in the Y direction. Maybe the problem is that the parts I've been looking at tend to be more square than long rectangles, so on some of the mid-size machines I'm wondering if they might end up needing to be done as two parts and then glued (which you can do with the resin, just cure it with some UV) together.

    Use of high levels of anti-aliasing can swell file sizes a lot, and not all models benefit from it. Too much may also degrade fine details, it is something that needs to be experimented with on your particular parts.

    One thing to keep in mind is throughput. A small machine may need to spend more time on each layer for curing, slowing down the height/hour speed. But if you need to print more stuff than will fit on the small machine then you get to double or triple the print times for the multiple jobs, where if you have a big machine you may be able to do them all in one print. I've seen photos of these printers being used in dental labs and they've got the entire build volume packed full of false teeth prints, maximizing the number of prints for the print run. I guess time is money.

    YMMV, always wear a helmet, and verify things you read from random people on the Internet. But I hope this at least gives you some more information than you had before.

    cheers,
    Michael
     
  13. bigseb

    bigseb Alibre Super User

    Good post. Some points to that:
    • The Anycubic Photon is all sheet metal and built like a tank. The Anycubic Photon S is all plastic. Two very different machines with different firmware. Be aware of this if considering purchasing. The Photon now also has new firmware that allows anti-aliasing although I haven't done the upgrade as the prints were looking amazing anyway.
    • FEPs are more consumable than the screens. I'm on my 3rd one already in about 9 months. They cost about 10 quid each and are easy to change. Screen is still going strong. A new one is about 60 quid.
    • I can count on one hand the number of times I have filtered the resin. And old resin thats left over from the last print I simply leave in the vat, sometimes for weeks. When I print again I top up the vat and stir the resin with a (gloved) finger.
    • Join the groups, absolutely, but bear in mind a lot of people that buys these machines are ordinary people (idiots maybe?) without a shred of mechanical understanding. They make it a lot more difficult than it needs to be, don't understand the physics behind the printers and generally make it a much bigger chore than it needs to be. My point is it helps to sift what you read through the engineer's filter.
     
  14. dmckee101

    dmckee101 Alibre Super User

    We've been using an Ultimaker 2 for a few years now. While I'm satisfied with the performance for product development, we are looking at manufacturing specialty parts. We've examined our designs using three different vendors, Ultimaker S5 (with soluble supports), Formlab ('white', 'rigid', & 'tough' materials), and Markforged Onyx.

    Ultimaker 2:
    Don't assume that the printing process is absolutely correct. My prints have a bit of under-size variability, typically <1%. If you deal with tight tolerance that is something to be aware of.
    Another point about the UM2 is the finished results. Support sides are not as smooth as the finished side.
    Lesson learned: position the part on the stage to minimize the supports and help hide the layer lines. Here are examples of the difference orientation makes:
    Ultimaker S5:
    We've purchased samples of our designs printed on the newer Ultimaker S5, which has a soluble support material. The result was that both sides, (finish & support) had a good finish. Although the finish quality of the S5 printed part was good the typical layer lines were evident and a smooth finish would require post-processing and/or orienting the part on the stage to reduce the number of supports and layer lines.
    Clean up after build is just letting the supports dissolve. These pictures show what is left after support material is dissolved. We bought this from a print-hub and the operator really was terrible, note the oval holes and layering due print orientation:
    Formlab 2:
    We also had sample parts made on a Formulab 2 using different materials. The parts are near perfect and would be ideal for prototyping and proving concepts. We could never settle on the material with the right properties. The materials we tried were too brittle or were too compliant. A redesign may have been helped but that is not our goal. Pictures of perfect parts:
    Markforged Onyx One:
    This is a fused filament design using a proprietary material, Onyx, a mixture of nylon and carbon fiber. Our sample design we had printed came out very nice. The operator was experienced enough to orient the parts for minimal supports. Clean-up after is simple clipping of the supports. Dimensionally very stable.
    Things that gave me pause with this were: the proprietary material, the slicer (program that takes the STL and outputs GCODE for printer) is a cloud-based product.
     
    Last edited: Oct 8, 2019
  15. MikeHenry

    MikeHenry Alibre Super User

    If you've looked at the Prusa SL1, what are your thoughts on that? It appears that there aren't enough of them out in the field for objective reviews. I was initially set on the Form3, but the user reports on their forum are not very encouraging. I was also had the Uniz Slash on my short list, but users on their forum are downright angry, so its off that list now.

    I was tending toward the larger format resin printers, but it occurs to me that most of my parts may be sized more toward the medium or smaller range and I'm wondering if there is a practical downside to having to load far more resin than the part requires.

    One of the other criteria I'm thinking about is the "round" dot size of true SLA (laser) vs "square" dots on the MLSA printers. The former apparently have a much finer resolution along the X and Y axes than do the latter, which are limited to increments of the dot size. Hope that makes sense. For those little statues, it probably makes no real difference, but one of my uses would be for microfluidics where resolution could be more important.

    Once one starts to delve into the subject for serious use, the issues become a bit complicated.
     
  16. Michael Moore

    Michael Moore Member

    Hi Mike,

    I'm looking at the printers for aluminum foundry patterns for various motorcycle projects and expect that whatever I buy will be purchased without having tried one out or even seeing one in person. That can be more of a crapshoot, but I don't see it as a deal breaker.

    As an example, one thing I'm considering right now for a pattern is a liquid-cooled two-stroke single motorcycle cylinder and the build volume on that is roughly 130 x 130 x 175mm. So I'm largely excluding anything from consideration with a smaller build volume, and that includes the Prusa and things like the Photon, Epax X1, Elegoo Mars etc. I saw someone post a photo of a model they did that had a 400mm build height (Transform or Phenom printer), and they said it took 30 hours so that's about 1.3cm/hour print speed. I can deal with a printer running for a long day (10 hours or so for the cylinder) but if it has to be pieced together from multiple print runs it sounds like it could be multiple long days of checking on the printer periodically and I'd like to avoid that. I don't want to have something running at night and risk running out of resin or some similar problem. With a big print volume you run the risk of a big failure but getting something done and in my hands more quickly appeals to me. And having to do multiple small long prints gives more time for Mr. Murphy to visit and adds wear to the LCD and tank film so there's a trade-off there too.

    I'm not sure about Prusa's tilt tank. There are much larger build-volume MSLA printers that seems to get by fine without that feature and to me the tilt seems like something else to go wrong. I presume there's an extra stepper motor, driver, a hinge, some sort of lever/cam to move the tank and maybe some other widgets added to the Prusa.

    If you look at https://formlabs.com/blog/3d-printing-technology-comparison-sla-dlp/ the Form3 appears to not run a continuous line with the laser but instead does more of a raster sweep so it isn't a smooth path, but the spot size seems smaller than the Form2 so maybe it all works out. But the Form is just too expensive for me as are their proprietary resins. I think they disable some of the features in "open" resin mode so using less expensive resins may negate convenience things you've paid for. Anti-aliasing on the MSLA printers seems to be one of those "it might help your print, it might not too" deals and depend on the printer, firmware and software being used.

    I've attached the photo I mentioned above with a the small mechanical check-part done on an Epax X1 at .05mm layer height, and that surface looks pretty nifty to me for being smooth enough to pull from a sand mold, but it may not be good enough for everyone's projects. I'd think for critical micro-sized projects you'd want to get a test part printed on the different printers first. With the number of these printers that seem to be around I'd think it would be pretty easy to find someone willing to print a small part in for a modest fee to cover their time/resin if it is an hour or two print job. Many people are happy with their printers and probably enjoy recruiting someone to get one too. Provide a clean .stl for them to slice and see what you get.

    Phrozen has just announced the Sonic which seems to be the same size as the standard Shuffle but with a monochrome LCD panel they say has a lifetime of 2000+ hours and that may allow shorter layer cure times for quicker prints. I haven't seen them announce a price or a lot of details though and you can probably put off a decision forever waiting for the next greatest thing to come out "real soon now".

    I think as long as the resin is not being exposed to UV when you add it to the tank filling up the tank isn't an issue, and you'd do the print with the lid/door closed so there's no worries about external UV there. The issue would be that if your tank springs a leak, you've got more resin to flood the printer.

    But since the small and medium printers look pretty similar in a lot of features I think that there's a good chance that if you are doing small parts you might be able to mount a smaller machine's smaller tank and build plate to the bigger one. It seems like you just have to watch that the build plate is centered in the tank and the -Z home position works out the same as the standard tank. You might need to mask around the edges of the small tank with the larger screen, but that sounds like a job for cardboard and scissors. But that might be an answer in search of a question.

    So get on a few FB user groups or the mfgrs forums and see if someone will give you a hand with the print. The guy who printed the small part in the photo is on the Epax FB group and since it sounds like his parts are more like what you'd be doing than some gamer who does figurines he may be a possible way to get a small test print done on an Epax machine that is already printing dimensionally accurate pieces. Maybe Seb could do a print for you on his Photon, though since I think you and he are separated by an ocean you might want to find someone closer to home unless he can evaluate it for you. The printer user group folks seem about as helpful to each other as the Alibre forum people are, someone should be willing to lend a hand.

    I added another image I saw on the Epax group that looks like a nice surface finish. I think it is some sort of vaping accessory.

    cheers,
    Michael
     

    Attached Files:

  17. Michael Moore

    Michael Moore Member

    Mike, your micro-fluidic parts might benefit from a smaller machine. A 2K screen is common, but the larger screen on the larger machines gives a larger voxel for that resolution. Some of the smaller machines offer a 4K screen version/option, and I think I saw that results in a 30% or so smaller voxel. e.g. the Phrozen Shuffle XL is 75 micron XY, the smaller Shuffle is 47 micron and the Shuffle 4K is 31 micron. But they'll relieve you of another $400 or so for the 4K screen.

    https://ultimate3dprintingstore.com/ appears to do a lot of Phrozen support/parts for the USA, as does https://www.intservo.com/ for Epax. Maybe they could do a test print for you.

    cheers,
    Michael
     
  18. Michael Moore

    Michael Moore Member

    I asked the person with the standard Epax X1 for a bit more info on his experience with the small mechanical parts he prints:

    Q: Did you have to do much fiddling with machine settings or scaling in your part models to get them to reliably print with dimensional accuracy? Will the same model print to the same dimensions anywhere in the build space or did you have to map dimension shifts and then adjust the model depending on where it is being printed? Also, what kind of tolerance are you holding on the models, +/- .1mm, .01mm or ? Do you see a change in the tolerance depending on the relative sizes of models? thanks for the info

    A: No fiddling with scaling - just eliminating resins that printed with bloom or shrinkage (I don't see the point in Siraya Sculpt at all). Same model prints the same all over the bed. For little parts as pictured, I often print 3 or 4 all at the same time. Tolerance is closer to +/- 0.01mm on the hole. can't say that I see a tolerance shift based on the size of the model with the mix I am using.
     
  19. MikeHenry

    MikeHenry Alibre Super User

    Thanks for all the feedback Michael - that's going to help quite a bit as I winnow down the short list. It's encouraging to hear that parts generally print accurately as there doesn't seem to be much consensus there from what I've read so far. Objects about the size of that 1st picture are one thing I'll probably focus on at first. They are small parts for a 3:1 scaled working model of a Curta calculator and are too small to print reliably with an FDM printer and more fiddly than I'd like to machine.

    I'm starting to wonder if it may be better to go with a cheaper and smaller printer for the time being as the market seems to be changing quite a bit right now and more capable or feature rich printers seem to be on the horizon, though I suppose that is always the case with this sort of technology. One reason is the larger voxel size on the larger format printers as you observed. With time, the voxel size may come down.

    One possible issue with the Form3's smaller laser beam size is that it might take longer to print objects. At least that's the conjecture from someone on Form3's forum and seems to be born out by the longer print times some have observed with the Form 3 vs the Form 2 on the same object.

    I will probably drop from consideration any company that only offers FB support - that medium seems difficult to search and usually full of fluff and short on technical info that is easy to find. I've started bookmarking manufacturers support forums and gradually going through them.

    Another aspect I want to dig into is the cost for replacement supplies, native slicer support for 3rd party resins, and the maturity of the slicer provided or recommended by the manufacturer. I'd like a slice that is pretty much plug and go, at least initially. Automatic support construction with manual override would be very nice as well as drainage hole creation. I'd rather not work through 2 or 3 apps to do all that.
     
    Eddy So likes this.
  20. Michael Moore

    Michael Moore Member

    Sure, all it takes for the voxel size to come down for a given panel size is going from 2K to 4K to 8K ad infinitum. If you start with the small machine for small parts then even the 2K panel makes a pretty small spot. Not spending on a printer bigger than your needs seems a good plan and that can make going for a higher quality printer/higher resolution screen more affordable.

    On the Form I would expect with a smaller laser point and the same resin they have to spend more time traversing to cover a given area, and the resin needs X amount of energy to cure so maybe the dwell time could be offset by a higher energy laser so they could speed up the feed, but maybe they haven't done that. To me the price of the Form printers needs a commercial justification/payback. $300 for a small printer is easier to classify as "fun money". In my reading I was getting an impression that some of FormLab's customers feel that they've "sold out" from a customer-based start-up and are going for a corporate mindset where maximizing the money/growing bigger is coming at the expense of some customer satisfaction. Of course, a commodity-class machine may not return enough profit to pay for much support, but if the installed base of machines is large enough and it works well enough then the users may take care of their own problems with community support. FB can get clunky, but unfortunately that seems to be the way a lot of things are going and there seems to be some allowance made for "this is a repetitious question you've seen a bunch of times, but I can't get FB search to give me an answer in the group" issues. Some groups do have file sections, and if you are already on FB then it is easy to join a group and ask questions and download what they have.

    Some spares are on Amazon, and at leaset Epax and Phrozen have affiliated large dealers that do support/spares here in the USA in addition to the factory support. I think when you look at the bottom line these MSLA machines are not very complicated and some share a lot of components, it isn't like you are dealing with a 3-5 axis mill with an ATC and one or more rotating axes/spindles.

    It appears that a lot of slicers do have a list of common aftermarket resins built in, but be prepared to adjust those default settings for your machine and parts and also the environment/resin temperatures during the print. The resin manufacturers often have tech sheets with base settings to get someone started without a lot of failures. You can start looking at some of those at https://www.siraya.tech/

    There are some YT tutorials on slicers and supports and some of those are pretty uncomplimentary about the value of a lot of the automatic support generation. It sounds like you want to plan on maybe using it only to populate a bunch of supports, and then you adjust them to better locations. Inadequate supports appears to be a big cause of failures, but once you get your eye trained for the "why didn't they put a support in this obvious place?" issues you'll be able to sort that out. More smaller supports (that can be adjusted to be even smaller where they contact the part) may pay off vs a few large ones. So even the free OEM slicers can work fine with some assistance, I think I'd be more concerned with having a slicer that generates good code with no dropped/duplicated layers. Supports and hollowing and drain holes can be done in CAD before dealing with the slicer, and that might (or might not) give a better printing part.

    One of my engineer friends has an FDM printer and he seems to have an eye for identifying handy small shop and motorcycle bits to make with it, and it is obviously a valuable addition to his infrastructure. It would be nice to see more of those types of things in the groups/YT vs yet another small figurine.

    cheers,
    Michael
     

Share This Page