Feedback request for upcoming FDM Quality Guidelines

Georgei · 2017-01-03T10:15:07.000Z

Dimensional accuracy: +/- 1mm or 1% depending on which is greater

So

20mm object = +/- 1mm tolerance

200mm object = +/- 2mm tolerance

cobnut · 2017-01-03T10:17:35.000Z

@Vienna3DPrint the guidelines say “whichever is greater”, so for a 20mm cube, 1% is 0.2mm, yes, but 1mm is greater so you could print anywhere between a 19.01mm and a 20.99mm cube and still be (just) within the 1mm tolerance for a print that should be 20mm.

Vienna3DPrint · 2017-01-03T11:35:52.000Z

The whole sentence says: “FDM parts need to be produced within an accuracy of +/- 1mm or 1% depending on which is greater.”

greater=accuracy

Maybe a native English speaker could formulate a less contradictory sentence.

bye, Tibor

Vienna3DPrint · 2017-01-03T11:41:59.000Z

As I mentioned before: greater is misleading, as I don´t think the dimension is meant, but the accuracy.

Robin must reformulate the sentence, as it leaves room for interpretation.

bye, Tibor

Georgei · 2017-01-03T11:43:49.000Z

The greater refers to the dimension (i.e. +/- 1mm or 1%) and not the accuracy. It is common statement in engineering at least.

The below text is from the original quidelines draft text from 3DHUBS

FDM parts need to be produced within an accuracy of +/- 1mm or 1% depending on which is greater.

For example, in the case of a 40mm cube, a 1% accuracy requirement would result in a max of 0.4mm deviation. As this is below 1mm, the 1mm is leading and determines the maximum. A printed part of 40.6 x 39.4 x 39.3mm would be acceptable as the deviations are within this 1mm. Printing a cube that is 38.8 x 39.4 x 39.4mm would not pass the dimensional accuracy requirements as the width deviates more than 1 mm.

cobnut · 2017-01-03T11:56:46.000Z

@Vienna3DPrint hmm, good point, I’m now wondering if I’ve read it the right way (as a native English speaker!). A better (but not necessarily ideal) sentence would be either of:

“FDM parts need to be produced within an accuracy of +/- 1mm or +/- 1%, whichever is the lesser inaccuracy.”

(so the 20mm cube should be produced to within 0.2mm, because 1% is less inaccurate)

or

“FDM parts need to be produced within an accuracy of +/- 1mm or +/- 1%, whichever is the greater inaccuracy.”

(so the 20mm cube should be produced to within 1mm, because that is the greater inaccuracy)

depending upon which meaning 3DHubs are actually looking for, and the 1% should be +/- as well. @Robin3D ?

edit (just noticed the doc does include examples)

Vienna3DPrint · 2017-01-03T12:49:20.000Z

Also with example it makes no sense.

40mm=2,5% (1mm), 400mm=1% (4mm) Why should the tolerance be bigger at smaller objects ?

40mm=1% (0.4mm), 400mm=0,25% (1mm) would be more accurate.

Maybe more understandable would be: Accuracy can be expected within 1mm deviation in any direction for objects below 100mm expansion and 1% deviation in any direction for objects above 100mm expansion.

But ok. Shall it be. I´ll pass on any questions to 3DHubs. It´s their guidelines, it´s their explanation.

bye, Tibor

MadBuggyofDoom · 2017-01-03T16:07:45.000Z

Until the 19.01mm cools a couple of degrees and then it’s out of spec.

CWCrawlers · 2017-01-03T18:06:57.000Z

Ok since it is easier than typing out my opinions on the new “rules” I thought i’d throw them into a quick little video… 3D Hubs FDM Rules? - YouTube

MadBuggyofDoom · 2017-01-03T22:16:15.000Z

Anybody can get a part off the machine in that tolerance, even the inept, the issue is keeping it in tolerance. I could take nylon parts off my machine accurate to 0.01mm and I still wouldn’t guarantee a general 1% accuracy in nylon to a client unless it was to very specific measurement/temperature/moisture levels.

I’ve just taken a part off my cheap home machine which is supposed to be 130 x 50 x 25mm - it measures 130.05 x 50.02 x 24.95mm. You couldn’t really want any better, it’s full of 13mm holes which are 12.94-13.00mm in diameter, there’s 6 x 3mm holes in there which are all 3.00mm. The ovality and coning are all way within 1% of true too.

The issue? They won’t be tommorow when it’s absorbed some moisture and the temperature changes. Or next week when the stresses have relieved after spending some time at working temperature.

Filemon · 2017-01-04T16:17:34.000Z

Hi all,

Allow me to check-in on this thread as well with an attempt to summarize and hopefully shed some light on the most asked questions (forgive me if it’s a bit lengthy).

Why these guidelines?

As mentioned in the original thread, the goal of the FDM quality guidelines is to improve FDM printing standards globally.

One of the top customer complaints is about print quality being below expectations. Phrased differently, there’s a huge opportunity for both Hubs and 3D Hubs to grow their business if a cross platform standard is set. This can be used to set the right expectations up front, as well as setting a minimum requirement for FDM printing, which enables us to cater to more professional audiences.

To be clear, the main message is to manage the customer’s expectation. These guidelines serve as a standard reference (e.g. backup) whenever there’s a dispute between customer and Hub. If a Hub and customer agree to deviate from the guidelines, that’s completely fine.

I hope this also explains that this is not about “policing” Hubs, but rather a tool to help set clear expectations.

Implementation of Guidelines

As many of you highlighted, after we’ve asked for feedback over the holidays (and you delivered) we failed to respond. This was a mistake on our side and I understand the negative impression it has left. I do want to stress that we’ve discussed all feedback elaborately and made changes accordingly. Below, more information on the main topics.

On dimensional accuracy

After checking with a significant sample of top Hubs as well as 3D printing experts from the industry, the current suggestion is considered to be reasonable.

To illustrate the variety of opinions, while the feedback here has been that the requirement might be too harsh, the opinion on Reddit has been that the requirements are too loose. My personal opinion is that with a well calibrated printer these requirements should be met without difficulty in most cases.

Clearly, there are scenarios in which an accuracy of 1mm might be tough to achieve - say, a FDM printed sphere. In those cases, be sure to set expectations with your customer before accepting the order. Again, the guidelines will be the point of reference whenever there’s a dispute between customer and Hub. If you both agree, then all is good.

On support removal

The majority of the customers expect to receive a finished product when ordering a print and we think that’s reasonable. Delivering a print with the support still on is considered below the minimum required level of service. Another indicator is that a large part of Hubs already delivers support removal in their service.

That being said, if a Hub and customer both agree that support can stay on the part - for faster delivery for example - then that’s their decision and 3D Hubs will not get involved.

We are aware that at this point there’s no automated system in place for calculating (and thus pricing) support. These tools are planned for Q2 / Q3 2017. Currently, we communicate to customers that possibly cost will be added for models that need a lot of support. For example, in the checkout; “Models with overhangs steeper than 45 degrees, needs support material possibly resulting in additional costs”.

Ok, hope that explains a bit. Please let me know if any questions remain.

Best,

Filemon

Enza3D · 2017-01-04T16:44:42.000Z

@Filemon

What about policing the dimensional accuracy? I really do not feel comfortable with only pictures, and what happens if I measure the parts before I send them, and so does the customer, but we have different measurements? This is really my main point of concern, so if this could be clarified, that would help quite a bit.

PepCo_Parker · 2017-01-04T16:50:10.000Z

Agreed. No offense @filemon but you’re just saying the same argument that has been countered time and time again on here. Our concerns weren’t hitting the targets, as a few of us have stashed several times, it’s enforcing them. It’s starting to feel a bit like beating a dead horse here, plenty of people have come onto this forum saying that these tolerances are easy to hit, which in my opinion they are, excluding post processing expansion or manipulation. Our concern was people manipulating the system to their advantage, or at least a fair few of us had this concern and I do feel it must be addressed.

Filemon · 2017-01-05T08:40:46.000Z

Ok, got it @Enza3D @PepCo_Parker.

The guidelines only serve as reference for when there’s a dispute. Platform wide this is a very low number. The situation in which the Hub claims to deliver according to specs but the customer claims the part was not is expected to be even lower.

Having said that, if this does happen we’ll take a case by case approach. If your Hub has a good reputation you’ll obviously get the benefit of the doubt and in some cases the refund to customer will be on us. I’m curious to hear if you guys have any other suggestions on how you feel guidelines should be applied in situations?

Joerg_4 · 2017-01-05T09:35:36.000Z

Hi all & @Filemon

I don’t see a big issue with the guidelines as it helps both sides to adjust expectations vs. possibilities.
But may be some fine tuning is inevitable:

1. Dimensional accuracy: May be it makes sense to add specific tolerances for specific materials, like:

PLA: +/- 1mm or 1%

Nylon: +/- 1.5mm or 1.5% (or whatever makes sense)
ABS: +/- 2mm or 2% (or whatever makes sense)

2. Dimensional accuracy: Contrary to popular belief it is still my opinion that it would make sense if 3D Hubs would provide a file to approve the dimensional accuracy (and may be other parameters).
All existing Hubs shall print this file to approve their print settings.

Instead of printing Marvin new Hubs will have to approve that they are able to print according to the guidelines.

Sure, there are plenty of models on Thingiverse, etc. for testing the printers accuracy, but if there were an official file all hubs will do the same test.

What do you think?

Cheers,

Joerg

Georgei · 2017-01-05T09:53:46.000Z

“All existing Hubs shall print this file to approve their print settings.
Instead of printing Marvin new Hubs will have to approve that they are able to print according to the guidelines.
Sure, there are plenty of models on Thingiverse, etc. for testing the printers accuracy, but if there were an official file all hubs will do the same test.
What do you think?”

And the question still remain: How do you check if it was printed accurate? I can say that it has been printed OK. How 3DHUBS check that it has? From the photos? I do not think so. Only if you physically you sent the part there and they measure it. I do not think they want their office filled with test prints. Lets be pragmatic and realistic.

cobnut · 2017-01-05T10:13:08.000Z

Sorry @Georgei but your comment has made me think of this episode of Doctor Who: I can just see all the staff at 3DHubs fighting their way past mounds of 30mm cubes (to within 1% accuracy, of course).

Georgei · 2017-01-05T10:15:51.000Z

Exactly…We try to solve a problem. Not create more.

CWCrawlers · 2017-01-05T10:22:57.000Z

Maybe to help quell the “uprising” support can pick top performing hubs, have them try the new rules out on a few orders and see how it goes… I feel like hub reputation is something people are forgetting plays a role here! If you consistently do garbage prints then that weighs different than if you do good work consistently. In the end I feel like we can sit and bicker but till we TRY the new rules we can worry ourselves bald over them… but till we begin to use them we can’t really say much.

cobnut · 2017-01-05T10:29:54.000Z

@Joerg_4 on a serious note, the problem with approval by file (apart from the one given by @Georgei ) is that it’s a single instance of proof. Marvin suffers from the same problem; I can take a photo of a lovely Marvin and upload it, but there’s no actual proof it was printed by me, not a friend with a lot more experience and/or a better printer. If we were to use a test print for dimensional accuracy, ignoring the problem of the 3DHubs offices being overrun with thousands of prints, there would still be no proof that you were responsible for that accuracy or if it was your print, that you would continue to deliver such accuracy for all your clients.

Quality standards need to be something that can be checked by the Hub, the customer and, if necessary, 3DHubs for any print carried out, not something that’s done once (without any real proof) and then never assessed again.