Preventing Rejections - Update on Project Caterpillar

Discussion in 'Suggestions & Feedback' started by Roy_Stevens, Feb 21, 2013.

Thread Status:
Not open for further replies.
  1. Roy_Stevens
    Roy_Stevens Member
    So a typical product cycle goes like this: I design and refine a new product, print it, occasionally refine it more, print it again, post-process the print including paint, take good photos, upload them, spend some time getting the word out on my new product on message boards, and then the customers start purchasing it and Shapeways REJECTS THE PRINTS!. So all my time is wasted, my name is mud, and I look like a fool. And this has happened more than once. I'm well aware of the design rules and follow them, (thus the I get the prints first time around) but it doesn't seem to matter. So what can I do to prevent this? Are there other services available that don't seem bent on making their designers look like idiots?
     
    Last edited by a moderator: Jul 9, 2014
  2. Dragoman
    Dragoman Member
    Unfortunately, this happens to a lot of designers - including me :( . Major nuisance.
    But Shapeways are working to reduce this problem by marking models as "Successfully printed" to make sure they ae not rejected later.

    This usually happens when items are designed close to the limits of the design rules. For each print, the item is checked by an operator and some are a bit stricter than others.

    Keep complaining about this! Hopefully, Shapeways will end up developing consistent processes.

    In the meantime, there's nothing left but repairing the problem item quickly, put up an apology note as comment for the model and explain things to the customer (if you know who she/he is).

    Greetings
    Dragoman

     
  3. CGNScale
    CGNScale Member
    I sincerely hope this is true. My process is similar to Roy Stevens and it gets very frustrating when, after several dozen successful prints you receive the "Could not be printed" notice.

    I've not been able to offer anything new because all my time is spent maintaining what has already been proven printable, yet keeps getting rejected.

    Come on Shapeways! Let's move forward, not backward! :(
     
  4. Youknowwho4eva
    Youknowwho4eva Well-Known Member
    I can tell you that we are progressing to a solution. I can't tell you when or what to expect. From what I have seen, it'll be up to your expectations and maybe even beyond.
     
  5. Roy_Stevens
    Roy_Stevens Member
    I hope that 'when' is soon. It would be nice to be able to request that a new or existing design be subjected to a validation process, and then not have to worry about it ever again.
     
  6. stop4stuff
    stop4stuff Well-Known Member
    Currently, the prduction team willnot/cannot print my Twin Rail Mobius Pendant - small in silver, although before the Christams rush, many had been printed with no reported issues.
    Since then every one has been rejected for a variety of reasons, it was only last week that Joost conveyed to me the real reason as to why - the production team want me to add bars to the model to aid the silver flow... hang on isn't that what part of the handling fee is for? I have no knowledge about adding sprues or vents to aid the silver flow :(

    Anyhow since then another rejection has come through
    Reason: Can not be cleaned
    Additional information: to fragile, will bend


    What can't be cleaned? The wax print or the silver model?
    And if the silver wires bend, it's no big deal as they are 1.3mm diameter and will bend back easily.

    So, in addition to a 'printed before' or whatever flag, the communication about a reason for rejection needs to be clear and not change every time (the model hasn't changed since the day I uploaded the original),

    Cheers,
    Paul
     
    Last edited: Mar 1, 2013
  7. I have also had a similar problem to everyone else here. A customer who bought 2 prints of a model ordered a third, but received the message that it could not be printed, even though he received the first two no problem. I then received an email saying the model could not be printed because of thin walls. the diagram attached to the email showing the problem parts points to a section of the model, highlighting it as 0.8 mm. This is incorrect. It measures 1mm on all faces, so it is Shapeways who have made a mistake on a part which is perfectly printable.
     

    Attached Files:

  8. stop4stuff
    stop4stuff Well-Known Member
    Nathan2012, do you the rejection image you can share?

    I have had a similar rejecion in the past, however the operator used the wrong tool for the measurement, they measured point to point rather than wall thickness. That was a simple mistake, which was rectified and the model printed.

    Paul


     
  9. I am having problems uploading the image. Looking at a larger version of the image I can now see the highlighted figure is actually 1mm. What is wrong with this? This is the minimum thickness they go on to suggest in the email!
     

    Attached Files:

    Last edited: Mar 2, 2013
  10. CGNScale
    CGNScale Member
    I had another model rejected yesterday. Again, a model that has printed successfully at least 30 times.

    The rejection is based on wire minimums. According to the design guidelines:
    How to design thin, unsupported wires:
    "0.3mm-0.6mm wire thickness: keep under 6mm wire length"
    https://www.shapeways.com/materials/frosted-detail-design-gui delines

    [​IMG]

    Here I have .44mm wires supported at multiple points. True, the supporting points are closer to 6.6mm apart, however this should still be well within the acceptable tolerance, ESPECIALLY for a model with a numerous printable history.
    (At least I would consider about 30 prints to be numerous :rolleyes: )

    Here are some photos of one of my successful prints. The same photos that accompany the model on the product page:
    https://www.shapeways.com/model/651286/n-scale-42-dry-bulk-tr ailer-2-pack-kit.html

    [​IMG]

    Careful observers will notice the railing for the ladder is broken. More careful observers will notice the thickness for the railing parts are thinner than the ladder. And further yet, clever observers will notice that in the rejection photo above I had thickened the railings to match the ladder. Finally, I'll admit that it was me that broke those railings during cleaning :( But notice that the overall ladder assembly is printed PERFECTLY.
    [​IMG]

    [​IMG]


    I think the best test should be, does the product have accompanying photos of a successful print?
    Then why perform the manual check? Shouldn't that be sufficient proof?

    Additionally, this model was ordered and printed more than a dozen times! THAT, should be definitive proof.

     
  11. CGNScale
    CGNScale Member
    Great to hear Youknowwho.

    This is a rough outline, but it closely models my suggestion for this process.
    [​IMG]
    The standing question would be what is considered "Tolerable". I would suggest to have the checkers error on the side of pass, and deliver the resulting model. As a serious designer, once I see and hold the model in hand and can physically observe the tolerance areas identified by the checker, I would be MUCH more motivated to refine appropriately and ultimately produce a quality design that would profit both me and Shapeways. (Win-Win)

    When the design is rejected based on virtual dimensions, I am given the ability to doubt, scoff, and never return. We usually call this a Lose-Lose situation.
     
  12. Youknowwho4eva
    Youknowwho4eva Well-Known Member
    The problem with your chart is the mildly successful prints. A failure can cause a whole tray of production to be lost, can cost hours, and can damage the printer. Problem 2, which is part of our thinking on the solution to some issues, is what you might find acceptable, another customer may not. So if you have a piece with thin rails that ships to you ok, we also have to be able to ship it across the world to arrive ok. And you may accept a broken rail, but a customer may want a reprint.
     
  13. stop4stuff
    stop4stuff Well-Known Member
    And today's rejection is...

    Twin Rail Mobius Pendant - small in Silver
    Reason: Can not be cleaned
    Additional information: has to be more connected

    Same model different reason... Additional information: has to be more connected ... The model is a single shell, it cannot be more connected!!!

    Paul
     
  14. CactusBones
    CactusBones Member
    I unfortunately have also had this same issue with customers having prints rejected after I successfully printed and took pictures. It is a bit embarrassing and hope that there is at least a message being sent to the customer relieving designers of responsibility if the print had been successfully completed before and suddenly becomes unprintable. I even went as far as to re-upload files....but now that seems like that was completely unnecessary and that this issue happens to many people and is unpredictable.
     
    Last edited: Mar 4, 2013
  15. Roy_Stevens
    Roy_Stevens Member
    OK, I'm going to have to call BS on the idea that a broken wire can cause an entire tray of prints to fail. Gravity, at least where I live, only works in the down direction. The problem that I have found is not a failure of the wire, but a failure of the support material. I have one model where I had a number of wires, which were designed to be printed in the X-Y axis, but the operator in his infinite wisdom oriented the model vertically, causing some wires attempting to be supported at the top of 6 cm of 0.6mm thick wax walls. I added 'plates' near the wires to cause the support material to be solid throughout the print and the next print came out perfect. But this wouldn't be an issue if orientation could be specified.
     
  16. stannum
    stannum Well-Known Member
    The mobius pendant should try to connect with its inner self.

    The tray reference is probably to the box used in SLS, full of nylon dust. The box is filled a bit each time, with a blade moving over it, so if a block of material becomes lose, it could disturb the smoothness of the dust. Imagine a small rock falls while flattening a pile of very fine sand.

    In any case, the recurrent issue is that problems are reported with new orders, and rarely along the fullfilment of current order, and with cryptic descriptions, no photos (that's different than renders), or just contradicting past prints experience. Last one was a 0.02mm difference, and after fixing that, claims that final customers would prefer it even thicker (final customers depend on the flexibility and "invisibility") for a part that has printed hundreds of times (it repeats many times per item, just in case anybody wants to guess which one it is).

    To make it worse, as you have no way to contact customers of failed prints, you & SW lose face and sales, specially with impulse buys.
     
  17. Innovo
    Innovo Member
    I agree. At least 90% of my clients with a failed print never re-order it.
     
  18. Youknowwho4eva
    Youknowwho4eva Well-Known Member
    I take your call of BS and tell you that for WSF, the pieces are built by putting a layer of powder that is synthered with a laser. If the powder sinks or shifts, it can alter the entire layer, and once the powder isn't level, it won't be level the rest of the print. And I think if FUD sinks, and the printer tries to add another layer, that material has to go somewhere, and it's not going where it's supposed to. I don't know if FUD can be as catastrophic as WSF.
     
  19. stonysmith
    stonysmith Well-Known Member Moderator
    It took me a long time to understand the "destroying a batch" comment.

    I finally came to understand: A small piece of material has significantly less mass than the printhead (or the wiper) does. As the printhead passes over it, it is possible that some friction is generated between the small peice and the printhead. That small piece of material could then be shifted laterally by the print head (or wiper).. causing that peice to "surf" for some distance, possibly even landing on somebody else's model. Then, as small as it is, that sliver can cause the print head to "bump" upwards, causing a gap or irregularity on other people's models... "destroying the batch"
     
  20. stannum
    stannum Well-Known Member
    If the mass of the wiper is big enough to push things, why would it bump next time instead of pushing them again? And the laser is above the box, plastic melts "itself" with nothing moving near it (invisible laser beam).

    Based in one print with a defect (broken and bent tube, like if slices had been pushed) the issue seems to be that it makes a mess of dust instead of perfectly flat surface, layers don't match with ones above and below, and soft parts deform. Given enough deformations, the items are useless.
     
Thread Status:
Not open for further replies.