Home »  » Suggestions » Preventing Rejections - Update on Project Caterpillar
Search Search  
 () 4 Vote(s)
Show: Today's Messages    Show Polls    Message Navigator
Re: Preventing Rejections [message #77309 is a reply to message #77308 ] Tue, 22 October 2013 15:36 UTC Go to previous messageGo to next message
avatar Vidalcris  is currently offline Vidalcris
Messages: 168
Registered: March 2012
Go to my shop
Senior Member
Rejected for a spike as always.
it's easy to measure a spike and say that this is not thick enough...


Re: Preventing Rejections [message #77310 is a reply to message #77309 ] Tue, 22 October 2013 15:47 UTC Go to previous messageGo to next message
avatar mkroeker  is currently offline mkroeker
Messages: 1572
Registered: June 2012
Go to all my models
Senior Member
Huh ? There goes my polished theory... Did they share any information about the previous print success rate of this model
(as stonysmith suggested they would do now), or did you make any changes to the "successful" model that would invalidate
its previous history ?
Re: Preventing Rejections [message #77312 is a reply to message #62565 ] Tue, 22 October 2013 17:08 UTC Go to previous messageGo to next message
avatar Youknowwho4eva  is currently offline Youknowwho4eva
Messages: 5300
Registered: September 2008
Go to my shop
Senior Member
I work here
Just popping it to let you guys know we are still listening and still working diligently on tools, tutorials, and guidelines to assist in making perfect products every time.

For all individual issues, please email service and feel free to post about your issue separately so that we can work on individual issues outside the mass of this thread. We have members dedicated to helping you solve any and all print-ability issues. Please keep the mass discussion here, and any individual posts please try to keep on topic of the original post.

As for discussion of competitors, we do appreciate insight into what they can and can't do, but please refrain from suggesting other services.

Thank you again for your input, and I hope to have news for you in the near future.


Your friendly neighborhood Moderman
michael@shapeways.com
Re: Preventing Rejections [message #77315 is a reply to message #77312 ] Tue, 22 October 2013 18:34 UTC Go to previous messageGo to next message
avatar dcyale  is currently offline dcyale
Messages: 106
Registered: February 2013
Go to my shop
Senior Member
Actually, discussing specific rejections is helpful so designers know what to avoid. If I can learn from someone else's rejection it may prevent me from making the same mistake. It is from discussing specific examples that this thread actually has a lot of learning in it, along with some messages that might have crossed the respectful disagreement line into something less helpful.

Also, when I started getting rejections on models that previously test printed OK, it helped me understand that it was a system wide issue and many designers were facing the same situation, not something that was only effecting me and my designs. That was actually important and somewhat reassuring.

Almost everything I have learned has come from Shapeways' forums and tutorials, or some other public forums, and I feel I have progressed from a rank amateur up to a novice in my design skills.

As to discussions of other services, I do see your point in Shapeway not wanting to host ads for competitors.

Dave Yale

Re: Preventing Rejections [message #77318 is a reply to message #62565 ] Tue, 22 October 2013 18:41 UTC Go to previous messageGo to next message
avatar Youknowwho4eva  is currently offline Youknowwho4eva
Messages: 5300
Registered: September 2008
Go to my shop
Senior Member
I work here
Dale,

I'm glad that we have been helpful in advancing your knowledge. This thread certainly has a place, I just wanted to point out that if someone wants one on one guidance, this thread probably isn't the place to post. I don't want someone to post here and not feel heard and helped, and I don't want someone to post elsewhere looking for help and be sucked into a wider ranging discussion such as this one.


Your friendly neighborhood Moderman
michael@shapeways.com
Re: Preventing Rejections [message #77348 is a reply to message #77318 ] Wed, 23 October 2013 06:53 UTC Go to previous messageGo to next message
avatar MitchellJetten  is currently offline MitchellJetten
Messages: 592
Registered: June 2011
Go to all my models
Senior Member
I work here
Hey guys,

Please, please keep in mind that it's only a printed before if THAT model ID has been printed before (not including updated version).

Vidalcris, your model was rejected for all wallthickness of 0.54mm this is around 5 layers thick in SLS:
[img]http://i.imgur.com/pztb866.jpg[/url]
Polishing will remove around 0.1mm thickness of your model which is why we ask people to please make sure that the model meets the minimum required thickness.

mkroeker wrote on Tue, 22 October 2013 15:47


(as stonysmith suggested they would do now), or did you make any changes to the "successful" model that would invalidate
its previous history ?


If printed before and rejected it will indeed show the success rate.
A different model or similar model will not have this information in the rejection mail.


Kind regards,

Mitchell Jetten
Customer Service Coordinator
Shapeways
Re: Preventing Rejections [message #77400 is a reply to message #77348 ] Wed, 23 October 2013 18:43 UTC Go to previous messageGo to next message
avatar railNscale  is currently offline railNscale
Messages: 48
Registered: June 2013
Go to my shop
Member
On the printed before issue, I'd like to add something.
The Printed before flag will be gone as soon as a model was changed (even if the model is actually improved in respect to producebility). The same is happening if you're creating a 'set', being a cluster of already printable models. To me as a designer, this is a little awkward, since in both cases I'm pretty sure the models can be produced, but I cannot guarantee the model will not be rejected by SW.

Now, IF SW would just consult the designer before any rejection, I really wouldn't be bothered so much. But this still is not he case. Communication with SW-team goes pretty good I must say, but the damage was already done. So, when is the the 'consult-first, reject later (if needed)' policy being adopted?

Or the other way around, when is SW providing the true model check/guarantee that once a model was printed successfully, this will be respected (even if a model or parts of the model was updated/improved by the designer)?

regards,
Maurice
RAIL N SCALE


rail N scale 'Een verrijking voor uw miniatuurwereld' railNscale.wordpress.com
Re: Preventing Rejections [message #77420 is a reply to message #62565 ] Wed, 23 October 2013 22:59 UTC Go to previous messageGo to next message
avatar uncommented  is currently offline uncommented
Messages: 9
Registered: December 2011
Go to my shop
Junior Member
Has the policy of requiring a 50% yield changed? I recently received an email telling me that Shapeways took my model down for having a "only" a 96% success rate.
Re: Preventing Rejections [message #77421 is a reply to message #77400 ] Wed, 23 October 2013 23:00 UTC Go to previous messageGo to next message
avatar FabMeJewelry  is currently offline FabMeJewelry
Messages: 247
Registered: December 2012
Go to my shop
Senior Member
railNscale wrote on Wed, 23 October 2013 18:43


Or the other way around, when is SW providing the true model check/guarantee that once a model was printed successfully, this will be respected (even if a model or parts of the model was updated/improved by the designer)?

regards,
Maurice
RAIL N SCALE


Sometimes a designer can think he/she improved a model but infact made it worse, every human makes mistakes from time to time. The same for when a designer first ordered a set and then decides to make it one piece, even if it's the same model shapeways has no guaranty that the designer didn't make a mistake.

A successfully printed flag should mean that the geometry of a model is printable, if the geometry changes the flag should dissapear until it has been sucessfully printed again.

/Quote

It's better for all involved to buzz the designer about design flaws before a model is ordered by customers. Customers who have no technical knowledge and can't do anything about the problem probably are more disappointed when they receive a rejection than we are ?

A long time ago i too received a few rejections for models that were ordered by customers from /FabMe and my other shop /GAD, it really gave me a bad feeling not being able to contact the customer. From then i decided i had to make most parts of my designs 10/15% thicker just to be sure they will be printed & shipped in perfect condition.

[Updated on: Wed, 23 October 2013 23:16 UTC]


Karen & Wesley - FabMeJewelry.com - Set a course, for the future !
Re: Preventing Rejections [message #77432 is a reply to message #77420 ] Thu, 24 October 2013 06:13 UTC Go to previous messageGo to next message
avatar mkroeker  is currently offline mkroeker
Messages: 1572
Registered: June 2012
Go to all my models
Senior Member
uncommented wrote on Wed, 23 October 2013 22:59

Has the policy of requiring a 50% yield changed? I recently received an email telling me that Shapeways took my model down for having a "only" a 96% success rate.

Best contact service directly for clarification - I guess this is about one of your "Discord" models (that seem to work surprisingly well along their own
storyline) ? (OTOH, the material overview just told me that today it will be possible to print in "Oops", so you may want to try that... Twisted Evil )
Re: Preventing Rejections [message #77457 is a reply to message #77432 ] Thu, 24 October 2013 14:11 UTC Go to previous messageGo to next message
avatar Youknowwho4eva  is currently offline Youknowwho4eva
Messages: 5300
Registered: September 2008
Go to my shop
Senior Member
I work here
mkroeker wrote on Thu, 24 October 2013 06:13

uncommented wrote on Wed, 23 October 2013 22:59

Has the policy of requiring a 50% yield changed? I recently received an email telling me that Shapeways took my model down for having a "only" a 96% success rate.

Best contact service directly for clarification - I guess this is about one of your "Discord" models (that seem to work surprisingly well along their own
storyline) ? (OTOH, the material overview just told me that today it will be possible to print in "Oops", so you may want to try that... Twisted Evil )


I investigated, and it appears this was a mistake on our end and corrected.


Your friendly neighborhood Moderman
michael@shapeways.com
Re: Preventing Rejections [message #77498 is a reply to message #62565 ] Thu, 24 October 2013 17:12 UTC Go to previous messageGo to next message
avatar TrainThingz  is currently offline TrainThingz
Messages: 106
Registered: March 2012
Go to my shop
Senior Member
I am rapidly coming to the conclusion that the issues with rejections for some items may not be the fault of the modeler, but of SW's checking software.

I have had, on a number of occasions now, models rejected for this reason:
Quote:

The file contains shells which aren't attached correctly, printing it will result in them not being attached and getting lost in the printing process.

This is rather odd... it's always 3D raised text, and there's accompanying graphics that show the individual letters highlighted.

Now, I'm using Sketchup 2013 for these models, and it generates the text as a group, not individual letters, and attaches them as a unit.

However, the graphic from SW shows some letters highlighted, and some not, implying that some letters are properly attached, and some are free. It also never shows all instances of text with this problem.

index.php?t=getfile&id=43643&private=0

Now, I'm using the "Solid Inspector" plugin in Sketchup, as well as CleanUp3. None of these show any issues at all with the text. I also ran it through MeshLab, and it finds no errors. (Don't EVER try to FIX a model with MeshLab, though... it absolutely destroys them) I'd post a picture of the model being checked with Solid Inspector or Meshlab, but there's little point - they don't show anything.

So I'm getting claims of errors from SW that I cannot find, replicate, or identify. I create the text exactly the same way every time, and yet sometimes it bounces, and sometimes it is accepted.

I think it's time to at least entertain the idea that the software SW is using to check the mesh might not be properly interpreting the model.

Re: Preventing Rejections [message #77500 is a reply to message #77498 ] Thu, 24 October 2013 17:16 UTC Go to previous messageGo to next message
avatar Youknowwho4eva  is currently offline Youknowwho4eva
Messages: 5300
Registered: September 2008
Go to my shop
Senior Member
I work here
Can you email me the file?

A few things to point out, In sketchup the letters may be grouped, but we use netfabb for checking, and netfabb looks at individual shells not groups.

But I can take a quick look to see if I see the issues they are communicating.


Your friendly neighborhood Moderman
michael@shapeways.com
Re: Preventing Rejections [message #77503 is a reply to message #77457 ] Thu, 24 October 2013 17:55 UTC Go to previous messageGo to next message
avatar stop4stuff  is currently offline stop4stuff
Messages: 3109
Registered: June 2010
Go to my shop
Senior Member
Youknowwho4eva wrote on Thu, 24 October 2013 14:11

mkroeker wrote on Thu, 24 October 2013 06:13

uncommented wrote on Wed, 23 October 2013 22:59

Has the policy of requiring a 50% yield changed? I recently received an email telling me that Shapeways took my model down for having a "only" a 96% success rate.

Best contact service directly for clarification - I guess this is about one of your "Discord" models (that seem to work surprisingly well along their own
storyline) ? (OTOH, the material overview just told me that today it will be possible to print in "Oops", so you may want to try that... Twisted Evil )


I investigated, and it appears this was a mistake on our end and corrected.


Mike, along the same lines about 50% success rate.

I had v1 of a model which successfully printed a few times without any reported issues, then the model was rejected. I adjusted the model as per recommendations and uploaded the replacement model, v2 (which to be honest now looks amaturish, the original v1 was bang on scale)

My question, is the success rate data for v1 available?
I'd like to revert to v1 if possible.

Cheers,
Paul

[edit] sorry about the caps Embarassed

[Updated on: Thu, 24 October 2013 17:56 UTC]

Re: Preventing Rejections [message #77504 is a reply to message #77503 ] Thu, 24 October 2013 18:01 UTC Go to previous messageGo to next message
avatar Youknowwho4eva  is currently offline Youknowwho4eva
Messages: 5300
Registered: September 2008
Go to my shop
Senior Member
I work here
stop4stuff wrote on Thu, 24 October 2013 17:55


Mike, along the same lines about 50% success rate.

I had v1 of a model which successfully printed a few times without any reported issues, then the model was rejected. I adjusted the model as per recommendations and uploaded the replacement model, v2 (which to be honest now looks amaturish, the original v1 was bang on scale)

My question, is the success rate data for v1 available?
I'd like to revert to v1 if possible.

Cheers,
Paul



I don't have access to that data. I'll point Mitchell in this direction to see what he can tell you.

Reverting is not possible, and success rate of v1 is lost after updating

[Updated on: Thu, 24 October 2013 18:39 UTC]


Your friendly neighborhood Moderman
michael@shapeways.com
Re: Preventing Rejections [message #77510 is a reply to message #62565 ] Thu, 24 October 2013 18:55 UTC Go to previous messageGo to next message
avatar MrNib  is currently offline MrNib
Messages: 459
Registered: November 2012
Go to my shop
Senior Member
I've had mysterious shell type errors once in a while if I didn't slightly overlap items prior to merging, although that's in the Tinkercad world. Sometimes you have this problem sometimes you don't. The best way to check this in Tinkercad is to define the entire finished object as a "hole" and the planes where outside faces butt up to each other inside the object look darker.
Re: Preventing Rejections [message #77511 is a reply to message #77504 ] Thu, 24 October 2013 18:58 UTC Go to previous messageGo to next message
avatar stop4stuff  is currently offline stop4stuff
Messages: 3109
Registered: June 2010
Go to my shop
Senior Member
Thanks for the info Mike.

Back to TrainThingz shells issue - run the stl through NetFabb Cloud service - the service will boolean merge any overlapping shells. Any shells that are face to face will mess up the normals and fail the process or make a complete mess of the model. If there are shells that do not overlap, the resulting 'fixed' model will have the number of shells on the fixed model download page.

Unless your model is chainmaille Wink

Paul
Re: Preventing Rejections [message #77513 is a reply to message #77511 ] Thu, 24 October 2013 19:09 UTC Go to previous messageGo to next message
avatar Youknowwho4eva  is currently offline Youknowwho4eva
Messages: 5300
Registered: September 2008
Go to my shop
Senior Member
I work here
stop4stuff wrote on Thu, 24 October 2013 18:58

Back to TrainThingz shells issue - run the stl through NetFabb Cloud service - the service will boolean merge any overlapping shells. Any shells that are face to face will mess up the normals and fail the process or make a complete mess of the model. If there are shells that do not overlap, the resulting 'fixed' model will have the number of shells on the fixed model download page.


The automated software on upload should do this. If it doesn't, then my guess there is a gap. If there is no gap, then I'd think the software isn't working properly.


Your friendly neighborhood Moderman
michael@shapeways.com
Re: Preventing Rejections [message #77514 is a reply to message #77513 ] Thu, 24 October 2013 19:15 UTC Go to previous messageGo to next message
avatar MitchellJetten  is currently offline MitchellJetten
Messages: 592
Registered: June 2011
Go to all my models
Senior Member
I work here
Hi Trainthingz,

One side is fine, the other has a gap between the parts
index.php?t=getfile&id=43661&private=0

  • Attachment: model.JPG
    (Size: 91.17KB, Downloaded 389 time(s))


Kind regards,

Mitchell Jetten
Customer Service Coordinator
Shapeways
Re: Preventing Rejections [message #77515 is a reply to message #77513 ] Thu, 24 October 2013 19:21 UTC Go to previous messageGo to next message
avatar stop4stuff  is currently offline stop4stuff
Messages: 3109
Registered: June 2010
Go to my shop
Senior Member
Youknowwho4eva wrote on Thu, 24 October 2013 19:09

stop4stuff wrote on Thu, 24 October 2013 18:58

Back to TrainThingz shells issue - run the stl through NetFabb Cloud service - the service will boolean merge any overlapping shells. Any shells that are face to face will mess up the normals and fail the process or make a complete mess of the model. If there are shells that do not overlap, the resulting 'fixed' model will have the number of shells on the fixed model download page.


The automated software on upload should do this. If it doesn't, then my guess there is a gap. If there is no gap, then I'd think the software isn't working properly.


Not sure if much has changed on the automated side, however there were circumstances whereby a model would pass the automated checks, yet have shells too small to be printed/recovered.

Paul
Re: Preventing Rejections [message #77531 is a reply to message #77514 ] Thu, 24 October 2013 22:40 UTC Go to previous messageGo to next message
avatar TrainThingz  is currently offline TrainThingz
Messages: 106
Registered: March 2012
Go to my shop
Senior Member
Well, at least I know what the actual issue is now...

Want to bet, though, that if I was having a model printed with surfaces I WANTED to be separate parts that were that close together, it would be rejected because they would fuse? Those letters are 0.5mm thick/tall. That makes that gap... something on the order of 0 .005mm? Smile So just how thin is a layer of FUD, anyway (since it's the highest detail material...)

I've tried NetFabb, BTW... the downloadable version completely fails to load on my Mac. Trying again just wants me to register... again. It accepts the registration, then never opens. (as in "waited 20 minutes for it to do something while I watched the spinning beach ball" never opens... I can create a YouTube video for upload faster than that!) This on a quad-core iMac running OSX 9 Mavericks.

I still don't understand why the two sides are different though - the text was attached in EXACTLY the same way on each side.

I've tried again, running it through the cloud version of Netfabb... which I REALLY don't like, because you get no practical feedback at all... So now I guess I have to wait for another rejection to know whether it's really fixed or not.
Re: Preventing Rejections [message #77534 is a reply to message #77531 ] Thu, 24 October 2013 23:38 UTC Go to previous messageGo to next message
avatar stonysmith  is currently offline stonysmith
Messages: 1854
Registered: August 2008
Go to my shop
Senior Member
moderator
In truth, the actual problem is probably the evil computer gremlin called "Rounding Error". Whatever drawing tool you used rounded UP the size/distance one way on one side of the surface, and rounded DOWN for the same object on object on the other side.

====

The automated software (MeshMedic) that Shapeways uses is a very close version to Netfabb Cloud.
Both of them go thru your model, welding any shells that overlap into a single shell. But.. they must actually overlap, no "tolerance" is accepted.

The difference is that MeshMedic (the Shapeways version) then looks for any small loose shells.
One example is that letter "e" shown on the picture.. it's floating away from the rest of the model.

This is one of those cases where automated software does a worse job than a human would. In this case, the human operator MIGHT decide that 0.005 is "close enough" and allow the model to pass. (the actual clearance delta is 0.05mm) But.. MeshMedic is being very literal. It is not sophisticated enough to measure the distance between too shells.. it simply determines that there are two separate shells, and that one of them has a net volume below a certain threshold.

===
I happen to use Windows machines, and Netfabb works great for me. If you would like to share the model with me, I'd be happy to take a look and let you know if there's any problems I can spot - to try to save you a rejection.


Patience, Persistance, Politeness - the 3Ps will help us get us to Perfect Printed Products
Re: Preventing Rejections [message #77565 is a reply to message #77531 ] Fri, 25 October 2013 13:52 UTC Go to previous messageGo to next message
avatar TrainThingz  is currently offline TrainThingz
Messages: 106
Registered: March 2012
Go to my shop
Senior Member
Quote:

I've tried NetFabb, BTW... the downloadable version completely fails to load on my Mac. Trying again just wants me to register... again. It accepts the registration, then never opens. (as in "waited 20 minutes for it to do something while I watched the spinning beach ball" never opens... I can create a YouTube video for upload faster than that!) This on a quad-core iMac running OSX 9 Mavericks.


Well, we got THAT issue solved, anyway... I e-mail the NetFabb people yesterday about this, they responded asking for my system details. I essentially dumped everything in "About This Mac" in their laps.

This morning, I got an e-mail from them that they had found the problem in their software and fixed it. Sure enough, it now actually loads on my Mac. Now if I can just figure out how to use it...
Re: Preventing Rejections [message #77566 is a reply to message #77565 ] Fri, 25 October 2013 14:02 UTC Go to previous messageGo to next message
avatar mkroeker  is currently offline mkroeker
Messages: 1572
Registered: June 2012
Go to all my models
Senior Member
Check the Blog - Part 13 of the "Shop Owner Challenge" series has a micro-tutorial on using netfabb...
Re: Preventing Rejections [message #77579 is a reply to message #77566 ] Sat, 26 October 2013 01:02 UTC Go to previous messageGo to next message
avatar TrainThingz  is currently offline TrainThingz
Messages: 106
Registered: March 2012
Go to my shop
Senior Member
Yes, it does... but there's no mention of how to fix detached shells in the tutorial, and I can't find any way to do it in the program either. Anyone have any ideas?
Re: Preventing Rejections [message #77581 is a reply to message #77579 ] Sat, 26 October 2013 01:11 UTC Go to previous messageGo to next message
avatar stonysmith  is currently offline stonysmith
Messages: 1854
Registered: August 2008
Go to my shop
Senior Member
moderator
TrainThingz wrote on Sat, 26 October 2013 01:02

Yes, it does... but there's no mention of how to fix detached shells in the tutorial, and I can't find any way to do it in the program either. Anyone have any ideas?

Let's move this particular discussion outside of this forum thread.

Rather than deal with the limited formatting choices of the forum, I saved my notes here: http://stonysmith.com/wired/Netfabb_Tutorial_1
I'll work with Shapeways to try and get it moved over to the Tutorials.

[Updated on: Sat, 26 October 2013 03:09 UTC]


Patience, Persistance, Politeness - the 3Ps will help us get us to Perfect Printed Products
Re: Preventing Rejections [message #77582 is a reply to message #77581 ] Sat, 26 October 2013 01:24 UTC Go to previous messageGo to next message
avatar TrainThingz  is currently offline TrainThingz
Messages: 106
Registered: March 2012
Go to my shop
Senior Member
I'd love to see it have it's own sticky thread so we can help others, too
Re: Preventing Rejections [message #78409 is a reply to message #62565 ] Thu, 07 November 2013 17:46 UTC Go to previous messageGo to next message
avatar natalia  is currently offline natalia
Messages: 786
Registered: September 2008
Go to all my models
Senior Member
I work here
Hey guys,

We know rejections are a huge issue that has been affecting our community for some time now and we want you to know that we recognize that it's a complex issue that affects your entire experience of Shapeways. What we also want you to know is that it's a complex issue for us, and that the entire team at Shapeways is working to fix.

Behind the scenes we have been making a lot of measurable progress and I want to share that with you. Below you'll find details on the recent changes we've made and how we plan to continue to improve your experience over the coming months.

BUT FIRST:
There is one thing that we all need to understand: 3D printing is a process that we are all learning together. We either embrace the technology and accept its inherent risks or we walk on the safe side and throttle creativity for the sake of consistency. Shapeways, at its core, is about enabling. So we err on the risky side so YOU can create whatever you imagine.

There is a choice we had to make between being conservative and having consistency in prints, and taking risks and allowing you to push the envelope of what is possible, which can lead to printed before rejections and inconsistencies. We choose to walk on the edge because we want to let you to make whatever you want. If we want to experiment and really push the envelope, then we need you to embrace risk, inconsistency and rejections with us. If we want to eliminate rejections entirely we would need to increase all the design guidelines into a very safe zone. That zone would rule out the fine detail and intricacies that 3D printers allow in the first place. Shapeways has chosen the edge, and much of the excitement that brings us to work everyday is what YOU create with this design freedom.

Your models help us learn what our machines are capable of, and what you WANT them to be capable of. Your designs push us to push the machine manufacturers to improve their machines. Everyone wins. It's a long process, but growth is not smooth sailing, you must take risks to gain rewards. As Pete put it, "if you want to live on the edge, you have to sometimes hurt yourself"

That said, there is a lot we can do to mitigate risk and make this process easier. So we are taking the top rejection reasons and eradicating them one at a time by updating software and our processes.

In short, we are approaching rejections from three ways:
Automated software to help you see and help you fix your printability issues
Improving our internal checking processes even further
Improving the way you communicate with us about rejections

Here's a summary of the recent projects we've been working on and how they'll impact you:

1. Upload
We recently improved the upload experience - now you can see your model online almost instantly, it is processed in real time and the prices are instantly displayed. This new interface lays the groundwork for more improvements and we are building even more automated software to help you see, improve and fix your printability issues.

2. Process Change: Checking before assignment
Previously, it could take days or even a week for you to receive a rejection. Now, we check your model BEFORE we send it to manufacturing so if there is an issue, we inform you faster, giving you a chance to fix it. Checking also happens continuously during all business hours in real time, so your order is checked by a 3D printing engineer much faster.

3. Process Improvement: More rejection reasons
We want to give you as much information as possible to help you improve your model. Previously, we had a limited number of rejection reasons and now we have almost double. This increase helps us give you more accurate feedback about the issues in your model and how to resolve them. We want to make sure you are receiving accurate and actionable information so you are well equipped to improve your model for printing. Permanent rejections are a part of this too, as soon as we know there is an issue with your model, it's not visible for sale until you have had a chance to improve it for printing. To help with some of the jargon, we are trying to give better explanations. We check up to 700 models per day, so our 3D printing engineers are becoming experts, but they are not infallible.


4. Model Resolution Project
Currently, when you upload a model, it first goes through an automated check then it is manually reviewed by a 3D printing expert before it's sent to a printer. In an effort to try to eliminate some of the printed-before rejections, we launched the model resolution project with Mitchell, our awesome customer service coordinator, where he did a SECOND manual review on each rejection. This was a critical step for making sure printed before models were being printed again instead of being rejected and getting rejections feedback to you faster. Mitchell was offering customers the choice to print the item even if it was going to fail. Most people said yes, so we know that is something you wanted, and in certain situations, we are able to offer this as a choice going forward. Note this is not something you can chose when you order, but only on certain models that are already queued for rejection, and we think there is a good chance they might survive regardless. This is the edge we're talking about!


5. Process Improvement: Consistency in Checking
The last, and perhaps most important piece is our steps to improve consistency. This is about consistency across our factories, consistency between 3D printing experts and consistency of your model being printed again and again. Unlike a regular manufacturer, almost everything we print is unique. That means, of the million models we have printed to date, 99% were different. Right now, there is not a guaranteed way to know that a specific model will print. We can only make educated estimates on the probability that it will print based on similar products we have printed previously, and on knowing the limitations of our machines. This makes it incredibly difficult to guarantee certainty. What we CAN do however, try to ensure consistency. So whether you order today or next week, or your order is produced in either factory, it will be consistently checked for markers of printability. So in the coming months as we roll this out, you can look forward to a little more consistency in your rejections - including the printed before rejections.

I know printed before rejections are the absolute worst and as long as we allow you to push the limits, there is no way that we can guarantee these will be totally eliminated. However, since we've implemented all these changes, printed before rejections have fallen by a third! I'm sure you'll agree that's a huge improvement and you should see fewer and fewer of these going forward.

All of these improvements so far are the building blocks, we are continuously improving the process, our software and the way we communicate our leanings with you so we can increase printability for more models.

As I said at the start, we either embrace the technology and accept its inherent risks to learn or we walk on the safe side and throttle creativity for the sake of consistency. We want you to experiment and really push the envelope, so we ask you to embrace the risk with us and together we can make push the boundaries of manufacturing.

I hope we can continue this journey together.

It's certainly why we come to work everyday

-Natalia, Bart, Michael and Savannah - your community team.


Shapeways Community Manager
Re: Preventing Rejections [message #78423 is a reply to message #78409 ] Thu, 07 November 2013 19:51 UTC Go to previous messageGo to next message
avatar UniverseBecoming  is currently offline UniverseBecoming
Messages: 975
Registered: March 2012
Go to my shop
Senior Member
Nice write up Natalia!

Good to see Shapeways is shaping up! Very Happy

Perhaps I'll try printing some of my models that Shapeways refuses to print, that is, if they can also be put into production. Smile

By the way, not related to rejections, but one of your competitors made a huge announcement a few days ago. Looks like they're going to actually start working with designers to actually help them with a HUGE discount and also professional photography to help them develop their products and their stores. If you're not aware of these developments you can contact me (Shapeways employees only Smile ) behind the scenes and I'll tell you more about what's going on. This is one of the things I've been saying is going to happen eventually and that Shapeways should be the first to do it since Shapeways is the leader. I've been warning about what will happen if a competitor steals all the designers away by making deals so tempting that every designer would be a fool not to take them up on it.

It's so exciting to see how the story of 3D printing is playing out for the world. I can't wait to see what's going to happen next! Very Happy






It always was. It always was because somethingness cannot spawn from nothingness. And in the was of the past there is the forever of the now. Only now. Only now and nothing new, for anything new would add to the infinite, yet there can only be one infinite. Only one. The universe is only becoming something new in the delusion of our minds. This delusion that makes life worth living in our perceived universe becoming.
Re: Preventing Rejections [message #78508 is a reply to message #78409 ] Fri, 08 November 2013 16:51 UTC Go to previous messageGo to next message
avatar SnowyRiver  is currently offline SnowyRiver
Messages: 10
Registered: September 2012
Go to my shop
Junior Member
natalia wrote on Thu, 07 November 2013 17:46

Hey guys...

4. Model Resolution Project
Currently, when you upload a model, it first goes through an automated check then it is manually reviewed by a 3D printing expert before it's sent to a printer. In an effort to try to eliminate some of the printed-before rejections, we launched the model resolution project with Mitchell, our awesome customer service coordinator, where he did a SECOND manual review on each rejection. This was a critical step for making sure printed before models were being printed again instead of being rejected and getting rejections feedback to you faster. Mitchell was offering customers the choice to print the item even if it was going to fail. Most people said yes, so we know that is something you wanted, and in certain situations, we are able to offer this as a choice going forward. Note this is not something you can chose when you order, but only on certain models that are already queued for rejection, and we think there is a good chance they might survive regardless. This is the edge we're talking about!


5. Process Improvement: Consistency in Checking
The last, and perhaps most important piece is our steps to improve consistency. This is about consistency across our factories, consistency between 3D printing experts and consistency of your model being printed again and again. Unlike a regular manufacturer, almost everything we print is unique. That means, of the million models we have printed to date, 99% were different. Right now, there is not a guaranteed way to know that a specific model will print. We can only make educated estimates on the probability that it will print based on similar products we have printed previously, and on knowing the limitations of our machines. This makes it incredibly difficult to guarantee certainty. What we CAN do however, try to ensure consistency. So whether you order today or next week, or your order is produced in either factory, it will be consistently checked for markers of printability. So in the coming months as we roll this out, you can look forward to a little more consistency in your rejections - including the printed before rejections.

I know printed before rejections are the absolute worst and as long as we allow you to push the limits, there is no way that we can guarantee these will be totally eliminated. However, since we've implemented all these changes, printed before rejections have fallen by a third! I'm sure you'll agree that's a huge improvement and you should see fewer and fewer of these going forward.

...

-Natalia, Bart, Michael and Savannah - your community team.


Hi Natalia,

Thanks for the great explanation. I do have a question, however. If a model has been successfully printed before, and even been printed many times before, why does it need to be reviewed at all? I have heard from customer service reps via email that the reviewers are always very busy, and time for reviewing models is a premium. If this is the case, why not relieve them of the job of reviewing already printed models entirely? If a model has been reviewed and printed, why is there a need to review it again and again and again? Isn't this a waste of your time, energy and money? Isn't it leading to the issues that we, as customers, are having?

If a model is, as you say, right on the edge, then perhaps you could have the reviewers give it a once-over, but from the perspective of knowing that it has already been successfully printed, to help take them out of the frame of mind of "find a reason to reject this model" and put them into the frame of mind of "just look for anything that is glaringly bad that might have been missed before". Also, if there are reasons that the model is "right on the edge", if it is printed anyway, some form of communication should be sent to the designer to let them know that their model is considered to be on the edge and why.

I have some models that were printed even though they have some areas (and not small areas, but significant areas of the model) with a 0.56mm wall thickness. This wasn't caught until a customer ordered one of these models, despite the fact that I had already ordered several of these models. My evaluation is that the material handled this wall thickness quite well, though I was happy to fix the problem. This was a case where there WAS something glaringly wrong that had not been caught before, and I'm fine with the rejection (though I would have preferred the rejection had come when I was ordering the models, not when a customer was ordering the models).

While I really do appreciate that you are working to improve consistency, I think that it has been mine and other's experience that there is still a significant problem with consistency. This problem makes it hard for us to have confidence in the SW service as a means of dealing with people who buy our designs. I don't know about anyone else, but every time I sell a product, I find that I'm holding my breath for days waiting for the model to clear so that I know that I've REALLY made a sale, and not had a customer get ticked off at me because of a rejection.

I realize that fixing complex systems is not easy, but there are a lot of suggestions around here that could REALLY help mitigate some of the issues (including in the thread that I have about my own list of suggestions), and some of them should be pretty easy to implement.

Thanks for the great service, and I really do hope that you keep the improvements coming.

Best,
Geoff.
Re: Preventing Rejections [message #78511 is a reply to message #78508 ] Fri, 08 November 2013 17:09 UTC Go to previous messageGo to next message
avatar mkroeker  is currently offline mkroeker
Messages: 1572
Registered: June 2012
Go to all my models
Senior Member
Quote:

take them out of the frame of mind of "find a reason to reject this model"

This accusation is IMHO both unnecessary and unfair - remember that rejection means that shapeways loses money, as we do not pay for the time spent checking if the model turns out to be unprintable.
BTW I am not sure if you expect or encourage comments on your private monologue thread ? Most of what you wrote there has been discussed at length elsewhere (viz. this thread), and there may be some misconceptions e.g. about accessibility of superseded versions of a model, feasibility of comparisons between previous and "improved" versions during checking etc.
Re: Preventing Rejections [message #78526 is a reply to message #78511 ] Fri, 08 November 2013 19:38 UTC Go to previous messageGo to next message
avatar Roy_Stevens  is currently offline Roy_Stevens
Messages: 143
Registered: November 2009
Go to my shop
Senior Member
mkroeker wrote on Fri, 08 November 2013 17:09

Quote:

take them out of the frame of mind of "find a reason to reject this model"

This accusation is IMHO both unnecessary and unfair - remember that rejection means that shapeways loses money, as we do not pay for the time spent checking if the model turns out to be unprintable.


It may be your opinion that it is unnecessary and unfair, but many of us feel this way after being beat down time and time again with silly rejections, such as measuring a 0.5mm long detail connected at both ends as a wire, measuring the narrow side of a taper, and other things that have me convinced that Shapeways would like to discontinue anything that isn't a 1cm cube in WSF. If I can't be assured that my customer will receive the model when ordered, what is the point of having a shop?


Earl Grey, hot.
Re: Preventing Rejections [message #78530 is a reply to message #62565 ] Fri, 08 November 2013 20:24 UTC Go to previous messageGo to next message
avatar MrNib  is currently offline MrNib
Messages: 459
Registered: November 2012
Go to my shop
Senior Member
Working on the edge is great for prototypes and developing processes but it's not so great for general sales to other people (which leads to the consistency complaints). There needs to be an effective way for that edge to move into the realm of normal and boring with design guidelines changing over time to reflect the advances and improved capabilities of people and process. Otherwise you never have a good grasp on where that edge is and it's likely to cut you when you least expect it.

Re: Preventing Rejections [message #78555 is a reply to message #78511 ] Sat, 09 November 2013 05:14 UTC Go to previous messageGo to next message
avatar SnowyRiver  is currently offline SnowyRiver
Messages: 10
Registered: September 2012
Go to my shop
Junior Member
mkroeker wrote on Fri, 08 November 2013 17:09

- remember that rejection means that shapeways loses money, as we do not pay for the time spent checking if the model turns out to be unprintable.


But this is part of the point. If their focus is on getting things through the inspection process as quickly and efficiently as possible, then things like print history should have an immense impact on the review process. Instead, there are countless instances of models being successfully printed, sometimes many times over, and then getting rejected. Also, I have had a similar experience to the 0.5mm detail that is connect on both ends being called a wire that Roy mentioned. This stems from looking for reasons for a rejection, rather than looking to see if the particular instance doesn't really cross the line.

Further, I've had parts rejected that, upon contacting customer service, they went back and reviewed the part again only to admit that it shouldn't have been rejected, and the reason didn't really apply.

Now, don't get me wrong, I'm not saying that I harbor I'll feelings toward SW. Quite the contrary. I think they have a great service, and they're working in a difficult, nacient industry. But their current business model feels underdeveloped, and they need to move to get changes in place to put their shop owners forward as their highest priority customers. When they reject a part that I'm ordering for myself, it's a nuisance. When they reject a part of mine that someone else has bought, often despite some tremendous efforts on my part to try to ensure that the model meets the design guidelines, it hurts me as a business.

Also, on the topic of SW losing money on rejections, I'm curious how you figure that. When I order a part, SW gets my money. If the part is rejected, they do not refund me that money. Instead, they give me credit. In other words, they get my money either way, and it's up to me to order a new model to try to get my money's worth. So, they're getting money whether or not models are rejected.

Quote:

... there may be some misconceptions e.g. about accessibility of superseded versions of a model, feasibility of comparisons between previous and "improved" versions during checking etc.


During discussions of a modeling bug from the CAD program that I use that was plaguing a couple of my early models, I had occasion to want to have one of the customer service reps look at a version of the model that had been updated twice since. I assumed that they didn't keep these, so I offered to send an archived copy that I had to him. However, he promptly wrote back saying that he had pulled it up and looked at father issue we were discussing, and so on. In short, the accessibility of previous versions has been directly demonstrated to me. As for comparisons, most CAD programs offer relatively easy tools that allow you to inspect the differences between two models. If SW doesn't have the tools to allow them to do this, then they are choosing to spend time and have their inspectors reinspect every model every time, rather than investing in proper tools to accomplish their task with better efficiency.

Long and short, there are clearly many ways that SW could work to improve their service. There is no doubt that they have choices to make about which ones they want to implement. But, it is quite clear, also, that they try to be nimble and move quickly. These rejection issues have been going on for a long time. While reducing reprint rejections by a third is a pretty good chunk, the question must be asked why reprint rejections cannot be reduced by 90%-100%? And the fact that it is taking them so long to try to get these kinds of rejections reduced also shows that they're potentially leaving room for a competitor to come in and pull the rug out from under them.

I like SW. I want them to succeed. I wouldn't be putting the effort into writing posts such as this one if I didn't.
Re: Preventing Rejections [message #78558 is a reply to message #78555 ] Sat, 09 November 2013 06:43 UTC Go to previous messageGo to next message
avatar MrNib  is currently offline MrNib
Messages: 459
Registered: November 2012
Go to my shop
Senior Member
I think they are using ClearCase/ClearQuest for documentation control with regards to model updates so yes, all previous uploaded models should be available. I surmised this because I once had an error or something pop up with regards to things going wonky during an upload. Of course they may be using something else by now.

I'm very familiar with looking at changes in CAD files with a differential check but that's for the integrated circuit design world where things can be analyzed with 2 dimensional Boolean checks - including the design rule checking. For 3D stuff differential checking becomes more problematic particularly if the two models have changed significantly, scaling has occurred, or origin points have shifted. I suppose it's simpler for 3D designs with very minor localized changes. In the IC world the foundry generally sends a copy of the final fabrication designs back to you. That way you can use a differential check to make sure what you sent to them is what they will be fabricating. At this final QC step you'll catch errors like the wrong version of a design being used, things being inadvertently filtered or corrupted, layer mapping errors, etc. But unlike the 3D world there's a lot of automated checking capability for IC design rules. It's decades old established technology. You can generate a list of hundreds of small errors that need to be worked through before a design will be accepted by the foundry. A circuit like a microwave amplifier or mixer circuit may be checked on your desk top computer in 30 seconds or less for gross errors. (I'm talking simpler analog circuitry here, not complex microprocessors with millions of transistors.) Sending a design out to the foundry's server for more complete checking may take anywhere between minutes and an hour to get the full monty check depending on how many other people are sending in designs for checking. I suspect an equivalent amount of full checking for a 3D design could take many hours or even a good chunk of a day, but I still haven't found enough information on the subject to know what's possible or what design rules could be reasonably checked in today's world.

Anyway, I don't know what the best immediate solution is to the rejection thing as long the set of guidelines is not large enough to define 90 to 99% of problematic printer situations and there isn't practical automated software that can correctly check models to list all guideline (aka design rule) violations in a model. I know that thoughts of automated checking leads to complaints like 1) but that would limit designers or 2) there's no way to come up with comprehensive potential printer issues for 3D models or 3) there isn't software that can do that cost-effectively or 4) you can't expect most designers to follow 30 pages of illustrated guidelines. Fine, but I still think the only realistic long term solution is automated checking based on comprehensive guidelines which leads to better overall process control, product consistency, and less impact by the human checking factors. I don't know if that will take 2 years, or 10 years, or 20 years but I think it is the inevitable and necessary solution.
Re: Preventing Rejections [message #78561 is a reply to message #78555 ] Sat, 09 November 2013 07:49 UTC Go to previous messageGo to next message
avatar AmLachDesigns  is currently offline AmLachDesigns
Messages: 1144
Registered: September 2011
Go to my shop
Senior Member
Quote:

Also, on the topic of SW losing money on rejections, I'm curious how you figure that. When I order a part, SW gets my money. If the part is rejected, they do not refund me that money. Instead, they give me credit. In other words, they get my money either way, and it's up to me to order a new model to try to get my money's worth. So, they're getting money whether or not models are rejected.


This is easy: if you ask for your money back, they will return it, end of story. SW are scrupulous in this regard.
Re: Preventing Rejections [message #78566 is a reply to message #78555 ] Sat, 09 November 2013 10:51 UTC Go to previous messageGo to next message
avatar mkroeker  is currently offline mkroeker
Messages: 1572
Registered: June 2012
Go to all my models
Senior Member
Quote:

In short, the accessibility of previous versions has been directly demonstrated to me

So the misconception was on my end,sorry. Strange, I am quite sure I saw in a recent thread someone asking if a previous version of a model could
be reinstated and I think the reply was that this was not possible. (Maybe it depends on how long ago the desired version uploaded, or I am just
imagining things.)
On the topic of automated analysis, shapeways appear to be actively fostering research - see this project that was briefly mentioned in the forum some time ago. Also keep in mind that shapeways hired many new workers for their NY factory, few of
whom we can reasonably expect to have any previous 3d printing experience. Such an event has to be disruptive, even if it is now more than a year ago
experience levels among the team will not have equalized.
Re: Preventing Rejections [message #78567 is a reply to message #78530 ] Sat, 09 November 2013 10:52 UTC Go to previous messageGo to next message
avatar dcyale  is currently offline dcyale
Messages: 106
Registered: February 2013
Go to my shop
Senior Member
MrNib wrote on Fri, 08 November 2013 20:24

Working on the edge is great for prototypes and developing processes but it's not so great for general sales to other people (which leads to the consistency complaints). There needs to be an effective way for that edge to move into the realm of normal and boring with design guidelines changing over time to reflect the advances and improved capabilities of people and process. Otherwise you never have a good grasp on where that edge is and it's likely to cut you when you least expect it.




Well said. Perhaps there need to be design guidelines for "Print almost always" for productions items, and another "We will try but make no guarantees" for innovative designs.

Dave
Re: Preventing Rejections [message #78586 is a reply to message #78561 ] Sat, 09 November 2013 16:32 UTC Go to previous messageGo to next message
avatar stonysmith  is currently offline stonysmith
Messages: 1854
Registered: August 2008
Go to my shop
Senior Member
moderator
Quote:

Also, on the topic of SW losing money on rejections, I'm curious how you figure that.


A note from basic economics:

Shapeways has to pay the human to check your models. Any rejection (with a full refund) means that they have to absorb the cost of the manhours spent checking your model. This cuts into the profit they make on the model when a copy is finally ordered. I've got eight models in my shop right now that have been rejected without subsequent patron orders. That means that Shapeways "lost" the payroll dollars paid to the humans that performed the checks, along with lost the profit they would have collected from a successful print. So, it is NOT in their interest to "deliberately try to annoy us" as has been suggested.

That's why I personally advocate a $1 option to have a model checked and validated so that it can be released to the public without my having to pay for a full print. I've got too many models where the net markup collected to date doesn't cover the cost of a single print. I'd be losing money if I paid for a test print of every model just to ensure it won't be rejected. Of all the items in my shop, 67% have not produced enough markup to cover the cost of buying a test print of the item. But again.. I'm in this for the designs, not the money. This is a HOBBY for me.


Patience, Persistance, Politeness - the 3Ps will help us get us to Perfect Printed Products
Re: Preventing Rejections [message #78599 is a reply to message #62565 ] Sat, 09 November 2013 18:07 UTC Go to previous messageGo to previous message
avatar MrNib  is currently offline MrNib
Messages: 459
Registered: November 2012
Go to my shop
Senior Member
At least we can all have a good catharsis every two weeks or so on this subject. Am I right?!!!

And on the bright side at least the healthcare.gov people aren't developing design rule checking software. Laughing


Pages (15): [ «    1  2  3  4  5  6  7  8  9  10  11  12  13  14  15    »] 
   
Previous Topic:DUPLICATING A MODEL OR EVEN ALLOWING EMPTY REFERENCES
Next Topic:Feature request: Integration with 3rd party marketplaces like Amazon and Ebay in my Shop settings

Logo

Hello.

We're sorry to inform you that we no longer support this browser and can't confirm that everything will work as expected. For the best Shapeways experience, please use one of the following browsers:

Click anywhere outside this window to continue.