Tuesday, September 3, 2013

How My Form 1(s) Stopped Forming, & Formlabs’ Awesome Support

20130830_StoppedForming-Mess

So you may have noticed my posts have kinda trickled off a bit. I sincerely apologize, it wasn’t intentional. In the end weeks of July, my printer started producing some very hideous prints, such as the one above. The bad prints were consistent; nothing was printing nicely at all, and I hadn’t seen prints fail so badly up until that point. During an open support conversation with Formlabs, the printer gave up completely, refusing to stay powered on without tripping the power supply’s internal fuse. Something was wrong, terribly wrong...

Let it be known that I have very bad luck when it comes to technology. It doesn’t matter if it’s cheap or expensive, new or used, if a problem is going to happen, it usually does. Perhaps it goes without saying that the Form 1 isn’t immune to my curse.

The first problem I had was apparently with the power supply. I pressed the power button, and the Formlabs logo showed up on the screen, then went blank. My computer immediately followed this event with the “USB device removed” sound effect. Pressing the power button again had no effect. Neither did holding the button down. The printer simply would not turn on, no matter what I tried. DynamicCoyote, it would seem, was dead.

I emailed Formlabs support about the situation, and I was very worried. This was my first problem with the printer, and I didn’t know if I’d have to ship it back, let alone out of my own pocket. I received a reply shortly after saying that there’s a safety in the power adapter, and that sometimes it trips. It could be reset by leaving it unplugged for about 30 minutes. Sure enough, 30 minutes later my printer was back from the dead, forming along as though nothing had happened.

The next problem came in the form of a nerve-wrackingly loud buzz. Up until this point, the printer sounded like this:

The odd new sound coming from the printer was this:

Yeah... not good. Another support ticket to Formlabs helped put my mind at ease, if only by a little, when they replied that I shouldn’t worry about it for now if it’s not affecting my prints (which it wasn’t). Still, the awful, loud sound was not sitting well with me, especially since I tend to run the printer at night while I’m trying to sleep. On the bright side, at least it was still working.

I continued to print things like parts for some prospective Form 1 buyers and some testing on individual parts from the Form 1 Stress Test, and things seemed to be going somewhat well. Prints weren’t perfect like they were in Formlabs’ photos, which bugged me, but I kept pressing onward in the hopes that some updated software would fix things.

20130830_StoppedForming-GunAndBarrel

One of the things I built to print was a part I had planned to make since before I ordered the printer. It was a replacement Nerf gun barrel for a modified pistol. The original part (the orange cylinder) has a small hole in the barrel that most modders just tape up. With a 3D printer, I could just make a new barrel without the hole. I was going to recast the part in some sort of clear plastic, but since the resin was clear, recasting was no longer necessary for the prototype.

I had printed two or three of the barrels before it became very apparent that something was wrong: the parts were coming out worse. They were coming out with a slightly lumpy surface, and there was more and more surface separation. The inner surfaces of the barrels were consistently coming out of the printer with “flakes” of resin falling off them. Since the parts were still somewhat decent, I kept printing. That’s when the parts below started to come off the printer.

20130830_StoppedForming-FailedBarrels1

20130830_StoppedForming-FailedBarrels2

20130830_StoppedForming-FailedBarrels3

20130830_StoppedForming-FailedBarrels4

20130830_StoppedForming-FailedBarrels5

As you can see, nothing was coming off the printer without a failure. I tried different positions, different layer heights, different angles... they all failed. I decided to shoot Formlabs another support ticket.

After a friendly phone call from Formlabs’ Jory, it was decided that I should try “printing” their laser spot test to see what the laser spot looks like. He sent me the file, and I opened it up in PreForm. It was pretty much just a small rectangular part with a custom resin setting that keeps the laser in one spot. A piece of paper placed in the printer without a resin tank or build platform would show what the spot looks like, and I was supposed to take a photo of it and send it to them for analysis. Simple enough, except another problem arose before I could even try it.

As luck would have it, the very moment I was going to do the laser spot test, the printer wouldn’t power up again. Like last time, I tried leaving the power supply unplugged for 30 minutes, but it still wouldn’t power up. It would power on for a moment, but wouldn’t finish showing the logo. I tried letting the power supply reset a couple more times before I gave up and told support what was happening. The response was what I didn’t quite want to see: they wanted DynamicCoyote sent back. It was decided there was something deeper going on that I probably couldn’t fix on my own, and they wanted to see what was wrong with it.

With DynamicCoyote packed up and the prepaid (yay!) FedEx label attached, I drove my printer to the FedEx depot and went home. The next day I got confirmation of the replacement being shipped, and it was at my doorstep in less than a week. YoungPlanula had arrived.

The first thing I decided to print on the new printer was the last thing I tried printing on my last printer: the same set of barrels that had been failing. I figured it’d be a good test to compare the old and the new. Immediately, I could tell that the new printer was working because this is what I saw a few hours later:

20130830_StoppedForming-CompletedBarrels1

It worked! Happily, I proceeded to pull the parts off the platform to wash and dry them. But there was something odd about one of them: it was really “wavy.” While one printed perfectly fine (and actually very nicely, for once), it was very apparent that the other had problems. It appeared as though several layers were printed with severe shifts in relatively large groups of layers. You can see in the photo below just how bad the shifts were.

20130830_StoppedForming-CompletedBarrels2

Those lines horizontally situated in the photo are the shifts. You can see that they’re quite apparent. On a large model, this may not be that big a deal, but on a small, finely-detailed piece, this would be a huge problem. These obviously won’t buff out. It also apparently only affects one axis, the one that travels front and back (Y, I think).

I had to test further to see what, exactly, was causing this. My first hunch was that it had something to do with more than one part being printed at the same time, so I removed one of the barrels from the print (the one that printed nicely), and left the other. Printing it in exactly the same spot, the barrel that printed badly actually came out way nicer than before.

20130830_StoppedForming-CompletedBarrels2

The two parts above are exactly the same model, printed in exactly the same spot. The only difference is the one on the left was printed by itself, while the one on the right was printed with another part. Strange..!

I printed a bunch more with tiny changes to accommodate the darts in the gun. The first several versions were a bit too small in the inner diameter, so I had to widen it successively before I got a good size. The part fit in the gun perfectly, and mated with all the other parts great as well. I could only print the barrels one at a time, or they’d suffer the fate of the wave.

A little over a week ago, I finished setting up a character I had built a while back for printing. It took a little while to get her set up, but it was a necessary step toward my ultimate goal of designing and printing my own figures. The goal was 1/8 scale or larger, but since the printer’s maximum print height is just a tad too small for her, I had to split the model into smaller parts.

20130830_StoppedForming-Girl

It was during this print that I noticed something odd about the way the printer was operating. I sat and watched the printer for a few minutes as the laser danced across the build platform’s surface, filling in the area where the raft was. Nothing unusual there. Then the motor that operates the resin tank tilting started whirring, but the tank didn’t budge.

"Oh crap,” I thought. “Is it broken..??” A few more minutes of watching this and I began to wonder if the newest version of the PreForm software I was using introduced some new process that I wasn’t aware of, so I let the print go for a while more while it finished building the raft. When the last layer of the raft finished, I watched the peel process complete, and I let out a sigh of relief. The peel was still working.

A few minutes later I watched the printer work further on the support structures. Peering into the sides of the tank, I could see that the structures were there, but they weren’t attached to the raft, they were stuck to the bottom of the resin tank... all of them. I immediately cancelled the print and observed all the little failed structures at the bottom of the tank.

What happened was that the peel process apparently hadn’t operated correctly while it was printing the raft. Since the resin tank wasn’t pulling away from the print as the raft formed, the build platform may not have been moving upward at the right heights. This meant that the raft wasn’t printing as tall as it should. So as the final layer of the raft completed and the resin tank was finally able to pull free, the build platform was already set too far away from the bottom of the resin tank to allow any more layers to adhere. Nuts.

Time to clean up the bottom of the resin tank. Knowing the silicone on the bottom of the tank is prone to pitting if I use tweezers or the spatula just a bit too roughly, I started using a 120x120x1mm part that does the “cleaning” for me. If the little bits aren’t too tall (no more than 2 or 3 layers, max), printing this for a few layers or so allows the bits to become encased in cured resin while the peel process pulls it off the silicone for me. So far, this process has proven to work well for me, so I loaded it up in PreForm and sent the print on its way.

20130830_StoppedForming-Cleaner

The first layer already proved problematic. The peel process still wasn’t operating. I cancelled the print, but this time, the build platform refused to return to the top of the tank. The printer made a loud mechanical grinding sound as it tried to pull the build platform away from the resin tank, but since it was unable to, it gave up after three attempts. I wasn’t sure what to do at this point; I had never seen the printer do this, and with the build platform still firmly stuck to the resin tank, I wasn’t sure how I was going to get the two separated.

By the time I had figured out that I could remove both the build platform and the resin tank together, the resin tank slowly dropped away from the build platform, tilting on its own. At this point, I began to wonder if the motor operating the tilt just wasn’t strong enough to pull the tank away from the platform if the area of solidified resin was too large. Another experiment was necessary.

I loaded up the character file again, but this time I sized her to 1/16 scale. I sent the print on its way, and several hours later I had my first printed figure:

20130830_StoppedForming-GirlPrint1-16

Not bad, I thought, considering the small size. A lot of details were lost, such as the small creases along her mouth and eyes and some of the subtle folds in the leotard, but I was surprised that the loincloths printed so well. Unfortunately, the shoulder decoration (dunno what it’s actually called) failed to print since the scale made it a bit too thin, and there are some heavy waves along her left arm. I don’t exactly know why the waves formed this time. I figured if there were waves on her arm, other parts should have experienced similar waviness. The support structures cleaned up pretty nicely, except for where they intersected the model along her left collarbone and her back. Since they were deep gouges, they weren’t going to clean up at all. Some sanding and polishing was attempted, but not much.

“Ok, so small scale works... maybe the motor just isn’t strong enough for the peel.” I decided to try another go at the 1/8 scale, but this time I’d pause the print between peels while it printed the raft. It didn’t work. Apparently, something changed between firmwares or my memory is different than what was actually happening, because I could have sworn pausing prints left the resin tank in its dropped position. Instead, it was pausing at the upright position. I decided to try something else: offering the tank some assistance.

I positioned a magnet where the lid sensor was, and pressed down on the resin tank while it was peeling. The tank dropped, but I didn’t feel any of the forces I was expecting. It was at this moment that I realized that there was nothing operating the tank at all. I cancelled the print and inspected the resin tank carriage. It was still in its dropped position, so I had to lift it to get the tank off. A flashlight revealed this:

20130830_StoppedForming-Disconnected

The tilt motor screw that drives the resin tank carriage had somehow unscrewed itself from the clevis. The clevis is that rectangular block situated directly above the screw. I turned the printer on and off a few times with the cover off to watch how the motor interacted with the screw. The screw would go up and down, but it wouldn’t do much with the resin tank in terms of pulling. It would push it up, but that’s about all. I tried screwing it back in by hand, but this didn’t do much. I was certain I would need some kind of tool to screw it back in, but I had no idea if this would void my warranty (if I had any left at this point).

I started another support ticket, writing about the waves in the prints and the screw that had undone itself. I asked if this was something I could repair myself, and if I’d void any warranty I had left.

This time, the ticket took a few days to get a response. Understandable, as I’m sure they’re very busy with other customers now that they’re getting more of the printers out. Jory again handled my ticket, courteous and informative as always :). He put my mind at ease when he mentioned that I shouldn’t worry about the warranty... they wouldn’t leave me hanging with a defective machine. (Formlabs, you guys are AWESOME.)

Jory sent me a few photos of the parts of the screw (tilt motor screw, clevis, screw nub), and said that I should be able to twist the tilt motor screw back into place by hand, securing it with super glue or, preferably, some kind of permanent thread locking glue.

Try as I might, I couldn’t get the screw back into the clevis. I twisted as hard as I could in the confined space, but it just wouldn’t go in. I ran the printer a few times after I’d noticed this problem, so I began to wonder if I had damaged the thread. The only other option I could think of was to remove the clevis and tilt motor screw to see if I could get a better look at it.

The clevis was secured to the resin tank carriage by a rod with a C clip on each end. Some needle nose pliers helped pull one of them off so I could slip the rod out, freeing the clevis, then I unscrewed the nuts on both sides of the stepper motor and took the tilt motor screw out. I tried screwing the clevis onto the screw again, but I still had no luck.

Peering into the screw hole of the clevis, I noticed that there was some odd irregularities in the threads. I used a scribe to clear some of it out, stuff that looked like black scum. I then tried twisting the two parts together again with the help of some pliers to hold the tilt motor screw, with some cloth to protect it from the pliers’ teeth. SUCCESS! It was a tight fit, but it worked.

As it turns out, the clue was in what Jory said about thread locking glue. No doubt this was some factory thread locker, and for some reason this application just didn’t hold up. I cleaned off the threads on both the tilt motor screw and the inside of the clevis with some alcohol, and the two parts now screwed together very cleanly and easily. Unfortunately, the hardware store didn’t have any permanent thread locker (what the heck?) so I went with the super glue. A drop on the threads and the two parts were screwed back together, hopefully more permanently than before.

20130830_StoppedForming-Reconnected

After the glue felt solid enough, I put the printer back together, turned it on and off a few times to see if the screw was working, and fired up a print. The printer worked like it should, and a decent print formed into existence. It wasn’t perfect—there were some surface anomalies thanks to some residue on the bottom of the resin tank from the protective film—but the printer works now, which is the important part. Now I can get back to forming! :)

While the issues with my printer may not be over, I really want to give a shout out to Formlabs and their incredible support. I’ve had some really bad experiences with support from companies, but Formlabs has consistently met or exceeded my expectations. Formlabs, thank you, thank you, for being such an amazing company to deal with thus far, and I really hope this level of support continues. ^-^)/

27 comments:

  1. Good to hear you're on your way to printing again! I think you might have inspired me to actually deal with my not-currently-working Form1. I've had a support ticket going for a long time but I haven't had the time/motivation to perform *my side* of the support.

    ReplyDelete
    Replies
    1. To be honest, there have been times when I've thought about throwing in the towel, but these printers are still new, and I like to think that all the problems we're experiencing, reporting, and getting fixed will hopefully help Formlabs produce higher quality machines for everyone in the long run.

      I'd say definitely keep trying to get your printer working right. :) What kind of problems have you been running into, and what has Formlabs asked you to do to try to fix it?

      Delete
    2. I printed several Yoda busts in a row using various settings. I was using a brand new resin tank and all new/fresh resin. The first one came out nearly perfect; I call it "95% awesome" or something like that. Unfortunately, the 5% that isn't awesome basically makes the print worthless.

      I messed with the settings and tried three more times. All three of those times the bust basically cut off at the neck. An issue I had with the first print was the sides of the supports sticking to the sides of the model. I tried tweaking the support settings to rectify that problem - the only way to do it was to basically randomly change the settings until something looked good. Unfortunately, all the busts (after the first one) printed headless.

      On one of my bust prints I tried doubling-up the model. I placed two Yodas on top of eachother as perfectly as I could by zooming in a whole lot. That cured the resin too much I guess and the peel motor wasn't able to get it off the build platform (similar to your issue, but there was a good explanation for this).

      After those failures I decided to try the default Formlabs "butterfly clip" print. It basically failed completely. A couple mm of the base printed and that's about it. The rest was stuck to the PDMS in the resin tray.

      I opened a support ticket and explained what happened. They basically told me to print the butterfly clip again with slightly different settings. I wanted to be more willing to try what they asked, but it didn't seem like printing the same model again with slightly tweaked settings was going to tell me much. It seemed like a waste of time. I also got busy with other projects and stuff. I have basically let the printer idly sit on my desk for a month (at least).

      A new problem (which I haven't reported to them yet) is that my power plug seems to have quit somehow. I have another AC adapter that seems to work, but it's not as powerful (output: 24V 1A instead of the OEM 24V 2.5A).

      Delete
    3. Yikes, that does sound awful. Seems like you ran into a lot of the kind of problems I ran into, even with the power supply. I've started to wonder if the power supply quitting and the parts not finishing properly has something to do with the laser not getting enough power. If the laser isn't as powerful as it should be, the rafts may still print since they get multiple laser passes, but since subsequent layers only get 1 pass each it could be possible that they're not curing enough to adhere to the rest of the part. I'd definitely shoot them another support ticket, as it sounds like your printer might be experiencing problems similar to my original unit.

      The support structures, they seem very hit-or-miss. I feel your pain. :P I've been experimenting with building my own supports with some decent results, but I haven't gotten it to a point where it's a science just yet. Once I get close, I'll probably post about it. I know others have been experimenting in this area as well. I remember seeing a + and - on one of Formlabs' screenshots (or was it video?) of the PreForm software. That made me wonder if they had (or still have) plans to allow adding or subtracting of individual support columns. It would be nice if they could bring such a feature to the software.

      Delete
    4. Oh, another thing to note: my replacement printer came with a different power supply, if I'm not mistaken. It has the same power rating, but either the brand or the design is different. I'm wondering if they've found another model to be more reliable than the one they initially shipped?

      Delete
    5. I just thought I'd come in to say that I have finally tried my printer again in earnest and it seems to be printing pretty well again!

      After my last post I contacted Formlabs and they sent me a new AC adapter without delay. I'm not sure if the solution was the new AC adapter, the new software, or something else. One way or another I seem to be up and running again... I'm going to go order some of that gray resin and try it out. :)

      Delete
    6. Nice! Glad to hear it. :D Perhaps it was the AC adapter, maybe not providing enough power to the machine. Early adopter pains, I suppose. Sorry for the late reply, too, btw... my day job's been keeping me pretty busy. >_<)\

      Delete
  2. Hi thanks for your experience sharing. So how is the wave line problem, is this solved after you fixed the screw?

    ReplyDelete
    Replies
    1. I haven't had a chance to see if the wave problem's been fixed since prints have been coming off "dirty" from the residue on the bottom of the resin tank. I probably won't be testing the two barrels print again (or any other "serious" print for that matter), till Formlabs contacts me back about how to remove the residue safely. Hopefully I hear from them soon... I'm itching to print more. ^-^)

      Delete
    2. Looks like the waviness hasn't been fixed after fixing the screw, and it seems really random. Even redoing the same print in the same spot with the same settings, the print will come out wavy in a different spot, or not at all. Even weirder, it seems like it gets better with subsequent prints, almost like the printer needs a warm-up before it decides to print well. :<

      Delete
  3. Hi TJ, sorry to hear about your troubles. I am also on my second form1, and really hoping not to have any issues this time!

    Formlabs support is extremely well managed, so at we know we're in good hands!

    ReplyDelete
  4. Makelijk te verhelpen. Boor een gat met een 1mm hss boor. pen erdoor. verbuig de pen. Opgelost. Het rare geluid is weg. Ik had het probleem ook met de Form1.

    ReplyDelete
  5. Remedy. Makelijk Drill a hole with a 1mm HSS drill. pen through. bend the pin to the other part and it Solved. The weird sound is gone. I had the problem with the Form1.

    ReplyDelete
  6. Easy to remedy. get out the two parts. screw it together. Drill a hole with a 1mm HSS drill. pen through. bend the pin. Solved. The weird sound is gone. I had the problem with the Form1.

    ReplyDelete
  7. Hey!
    Here is an online community of 3d artists.Jion it and get discussed your animations.

    ReplyDelete
  8. Tj, I'm curious about your design for the 120x120x1mm cleaning piece. Why the circles?

    ReplyDelete
    Replies
    1. The circles are to reduce resin use and pull on the silicone. I think the "cleaning" may be slightly compromised in the centermost area of the circles, so the design may see a revisit in the future. :)

      Delete
    2. Would you be willing to share your .stl file for the cleaning piece? I am having some of the same problems you have experienced.

      Delete
    3. Certainly! I've uploaded it here: https://db.tt/tva541Nd

      Delete
  9. I use Form1 too!
    Your advanced report is very intersting.
    Thank you for information!
    By the way. You can use japanese character like a ' ^-^) '.
    great(。・ω・)ノシ

    ReplyDelete
    Replies
    1. Awesome! I wish I knew how to type the special characters to make faces like that. I can type some Japanese, but I don't know how to type the ones like the bear-like face or the shocked one (I have to copy and paste haha). ^^;

      Like this one lol --> (/°д°)\

      Delete
  10. Thanks for posting such details! Quite frustrating with 70% failure in prints, but I agree, formlabs support has been amazing

    ReplyDelete
  11. Dear all,
    70% failure in prints, but you agree that formlabs support has been amazing.... I guess something going wrong with human beings....someone can explain me? I was thinking to buy a Form 1, but 70% failure sounds a little depressing...

    ReplyDelete
    Replies
    1. Hi there :)

      I haven't had time to put together a new post in a long time due to work and family stuff, but my print success rate seems to have improved since I posted this article. I received my 4th printer earlier this year and haven't had any of the hardware issues I'd had with previous printers so far. Before the latest printer arrived in early March, I was told that the optical degradation that was causing so many failed prints was a common issue, and that they were working on resolving it. I haven't had any of the print problems I'd expect from a degraded optical system on my current printer, so I think they've fixed it. They also announced the Form 1+ earlier this year, which is supposed to have beefier internals, so it sounds like they've managed to work out hardware issues.

      Software-wise, things have been improving greatly (custom supports placement! yes!!!), and my prints have been coming out well when they don't fail. Sometimes the failures are my own, sometimes they're the result of the software not being able to calculate a good angle for the part, or from the software not placing supports where they were needed (I've seen it miss overhangs on multiple occasions). Occasionally, a part will just fail for what seems like no reason at all, but that doesn't seem to happen all too often. I suspect when this does happen it's caused by small imperfections in the resin, but it's just a guess. Most other times there's a perfectly logical explanation of why the part failed... you just have to know what to look for.

      I hope that helps you feel a little better about it. I was pretty disappointed by all the problems I'd been having for a long time, but I feel like my current printer has been serving me pretty well. :)

      Delete
  12. My Form 1+ on 10 print jobs, half of them are successfull. I am pissed. This is the worst deal I made in my life. It is totally unreliable. I want to sell it. But...who would buy such a piece of crap ?

    ReplyDelete
  13. The writer understand better the mind of people what they want to learn through their writing therefore this article is outstanding. Thanks!!!Open in a new window

    ReplyDelete
  14. I'm on my fourth failed printer (three form 1's and a 1+). I thought that having upgraded to a Form1+, the laser would be much better. I have gotten about six months of trouble-free use out of it, but now it, too, is starting to fail. Crap. :( Looking at moving to an LCD based system or maybe a Milkshake. I don't want to think of the $$$ I've wasted on failed prints on the Form1 (and the VFLASH before it).

    ReplyDelete