Online since 2002. Over 3300 puzzles, 2600 worldwide members, and 270,000 messages.

TwistyPuzzles.com Forum
 It is currently Fri Apr 18, 2014 4:59 am

 All times are UTC - 5 hours

 Page 1 of 1 [ 35 posts ]
 Print view Previous topic | Next topic

How much tolerence (gap between internal surfaces) gives the best result for 3D printed puzzles?
 <0.05 mm 29% [ 6 ] 0.05 - <0.1 mm 5% [ 1 ] 0.1 - <0.2 mm 38% [ 8 ] 0.2 - <0.3 mm 29% [ 6 ] 0.4 - <0.5 mm 0% [ 0 ] 0.5 - <0.6 mm 0% [ 0 ] 0.6 - <0.7 mm 0% [ 0 ] 0.7 - <0.8 mm 0% [ 0 ] 0.8 - <0.9 mm 0% [ 0 ] 0.9 - <1.0 mm 0% [ 0 ]
Author Message
 Post subject: POLL: Ideal tolerence for 3D-printed designsPosted: Wed Feb 16, 2011 7:54 am

Joined: Mon Mar 30, 2009 5:13 pm
For those of you who have had experience with 3D-printing your puzzle designs, how much tolerence (gap between internal surfaces) gives the best result?

Many thanks!

_________________
If you want something you’ve never had, you’ve got to do something you’ve never done - Thomas Jefferson

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Wed Feb 16, 2011 8:08 am

Joined: Tue Mar 10, 2009 7:06 pm
Location: Nowhere in particular.
I've seen everywhere from .1 to .3. Some designers put a little tolerance on parts of the piece that don't necessarily need it, to make the puzzle fit together nicer. This is usually .05, or some other value <.10.

_________________
~Kapusta

PB: At home (In Competition)
2x2 1.xx (2.88)
3x3 11.xx (15.81)
4x4 1:18.26 (1:24.63)
5x5 (3:00.02)
6x6 4:26.05 (6:34.68)
7x7 6:38.74 (9:48.81)
OH (35.63)

Current Goals:
7x7 sub 6:30
4x4 sub 1:10

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Wed Feb 16, 2011 8:22 am

Joined: Fri Jul 16, 2010 8:25 pm
Location: Israel
I dont know much about puzzle designing, but I think that the tolerence should/is different from puzzle to puzzle - some need more to work and some need less.

It also depends on the printing method you use (FDM/SLS and more)

Sharon

_________________
Some of my collection for sale

The puzzles on the picture are not for sale.

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Wed Feb 16, 2011 8:24 am

Joined: Sat Jan 16, 2010 11:48 am
Location: In Front Of My Teraminx (saying WTF?)
Yeah depends. All my splits have 0 tolerance except on the splindle, where there is 0.1mm on each face. It depends where and what puzzle. Vmech would use 0.3 on some faces and 0 on others

- Greg

_________________
My Shapeways Shop!
My Designs
My Official Results
My Website on Twisty Puzzles with Gears

Grégoire Pfennig

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Wed Feb 16, 2011 8:28 am

Joined: Wed Oct 01, 2008 2:39 pm
Location: Marquette, MI, USA
Indeed as Sharon said tolerance is different for every puzzle, and indeed in different areas of different puzzles.

I generally use anywhere from .05 to .21 mm .

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Wed Feb 16, 2011 8:32 am

Joined: Tue Jan 01, 2008 7:30 pm
Location: Texas, USA
Sharon wrote:
I dont know much about puzzle designing, but I think that the tolerence should/is different from puzzle to puzzle - some need more to work and some need less.

It also depends on the printing method you use (FDM/SLS and more)

Sharon

I agree completly. I voted .1-.2, however that is for most of my puzzles and using shapeways wsf.

_________________
"I discovered the triangle one day while shaving. I trimmed my beard like the intersection of three circles and noticed how I could unfog a square in the bathroom mirror by rubbing my beard circularly against the glass."-Franz Reuleaux

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Wed Feb 16, 2011 10:21 am

Joined: Mon Nov 30, 2009 1:03 pm
Usually, I am a quite mild guy. However, twisty puzzles need zero tolerance to keep them well-behaved.

Eric Vergo explained the concept to me:
1: Pull all pieces outward, away from the center.
2: All surfaces that feel the force need zero tolerance.
3: All gaps that are pulled open need 0.4 mm clearance.
4: All gaps that are pulled laterally (conical cuts) need zero tolerance as well.
5: Pin-in-hole constructions need 0.2 mm gap at all side to assure smooth sliding/turning.
Smooth turning is achieved by proper breaking in and good lubrication.

Oskar

_________________
.

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Wed Feb 16, 2011 3:18 pm

Joined: Mon Feb 06, 2006 12:52 am
Quote:
Usually, I am a quite mild guy. However, twisty puzzles need zero tolerance to keep them well-behaved.

This line is SO good. First I thought you would state some dislike of this topic or anything. Your expressions are unique Oskar.

That "recipe" of Eric sounds reasonable and really interesting.

_________________
Life is simple.
eat. sleep. train. repeat.

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Wed Feb 16, 2011 3:32 pm

Joined: Sat Feb 03, 2007 10:03 pm
Location: Mississippi
Oskar wrote:
Usually, I am a quite mild guy. However, twisty puzzles need zero tolerance to keep them well-behaved.

Eric Vergo explained the concept to me:
1: Pull all pieces outward, away from the center.
2: All surfaces that feel the force need zero tolerance.
3: All gaps that are pulled open need 0.4 mm clearance.
4: All gaps that are pulled laterally (conical cuts) need zero tolerance as well.
5: Pin-in-hole constructions need 0.2 mm gap at all side to assure smooth sliding/turning.
Smooth turning is achieved by proper breaking in and good lubrication.

Oskar

Interesting - but confusing to me. Can someone provide a picture of this idea with a standard 3x3 or something? I'm a visual learner...

_________________
Space for rent

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Wed Feb 16, 2011 4:02 pm

Joined: Sat Jan 16, 2010 11:48 am
Location: In Front Of My Teraminx (saying WTF?)
Oskar wrote:
Usually, I am a quite mild guy. However, twisty puzzles need zero tolerance to keep them well-behaved.

Eric Vergo explained the concept to me:
1: Pull all pieces outward, away from the center.
2: All surfaces that feel the force need zero tolerance.
3: All gaps that are pulled open need 0.4 mm clearance.
4: All gaps that are pulled laterally (conical cuts) need zero tolerance as well.
5: Pin-in-hole constructions need 0.2 mm gap at all side to assure smooth sliding/turning.
Smooth turning is achieved by proper breaking in and good lubrication.

Oskar

When Oskar says 0 you say ok, no argument, he rules

_________________
My Shapeways Shop!
My Designs
My Official Results
My Website on Twisty Puzzles with Gears

Grégoire Pfennig

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Wed Feb 16, 2011 4:07 pm

Joined: Wed Jan 07, 2009 6:46 pm
Location: Evanston, IL
Oskar wrote:
Usually, I am a quite mild guy. However, twisty puzzles need zero tolerance to keep them well-behaved.

Eric Vergo explained the concept to me:
1: Pull all pieces outward, away from the center.
2: All surfaces that feel the force need zero tolerance.
3: All gaps that are pulled open need 0.4 mm clearance.
4: All gaps that are pulled laterally (conical cuts) need zero tolerance as well.
5: Pin-in-hole constructions need 0.2 mm gap at all side to assure smooth sliding/turning.
Smooth turning is achieved by proper breaking in and good lubrication.

Oskar

I agree with these rules, except for the 0.4mm tolerance. That is WAY too big. I use 0.1mm tolerance on those surfaces. And that means 0.1mm TOTAL, not on both surfaces.

I know that Eric has a really good picture to explain this, so I'll wait for him to post it.

-Eitan

_________________
Eitan = "EIGHT-ahn"
Check out my video: Twisty Puzzles a la Vi

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Wed Feb 16, 2011 4:10 pm

Joined: Fri Jul 16, 2010 8:25 pm
Location: Israel
Just to clerify, you are all talking about SLS (Shapeways) printing if anyone is wondering (its different in FDM and some other methods).

Sharon

_________________
Some of my collection for sale

The puzzles on the picture are not for sale.

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Wed Feb 16, 2011 4:14 pm

Joined: Fri Feb 08, 2008 1:47 am
Location: near Utrecht, Netherlands
Etian - I will not buy that. I've nearly always used 0.3mm where Eric/Oskar uses 0.4mm and it has been perfect. When I tried 0.2mm sometimes it didn't work. From my experience, 0.1mm could never be enough.
Perhaps we're not talking about the same situation?

_________________
Tom's Shapeways Puzzle Shop - your order from my shop includes free stickers!
Tom's Puzzle Website

Buy my mass produced puzzles at Mefferts:

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Wed Feb 16, 2011 4:20 pm

Joined: Wed Jan 07, 2009 6:46 pm
Location: Evanston, IL
Interesting. I've been doing 0.1 on all of my recent puzzles, and it's worked great. It does take a while to break in the puzzles, though.

The thing is, as long as you stick to those rules that Oskar posted, it shouldn't matter how much you take off of those surfaces. Thanks for the advice, Tom. I'll give 0.3 a try on my next puzzle.

(And yes, we are all talking about Shapeways WSF material, as that seems to be the most common printing method on the forum at the moment.)

-Eitan

_________________
Eitan = "EIGHT-ahn"
Check out my video: Twisty Puzzles a la Vi

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Wed Feb 16, 2011 4:24 pm

Joined: Fri Mar 06, 2009 9:23 pm
see post below

_________________
--Eric Vergo

My Shapeways shop

Last edited by gingervergo on Thu Feb 17, 2011 5:56 pm, edited 1 time in total.

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Wed Feb 16, 2011 4:38 pm

Joined: Wed Aug 01, 2007 3:14 pm
Location: Orange County, CA
If using the method above I stick with .2 - <.3 but it really depends on the situation. Some puzzles don't require tolerances at all.

_________________
-Garrett

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Thu Feb 17, 2011 6:37 am

Joined: Sun Mar 15, 2009 12:00 am
Location: Jarrow, England
I hope you guys will accept my (non-expert) view on this.

As an (electrical) engineer, I know that tolerances are a necessary evil which must be used to calculate worst case scenarios e.g. allowable manufacturing errors in resistors, capacitors, timing errors in digital circuits etc. etc. They are a pain in the behind, but I work in the real world, so I accept them.

As a fledgling puzzle designer, as I understand it, the word "tolerance" should be applied to the manufactured product pieces. This tolerance will be different for different pieces; 3D printed by SLS or FDM, or injection molded, or for different types of materials (WSF, nylon, ABS). You have to allow for printing errors, mold errors, mold wear, material shrinkage/expansion and so on.

The aim of adding "gaps" in the puzzle design is to ensure that no matter what size (within tolerance) the piece is made, the puzzle will still turn. I am guessing here, but if you allowed for all errors of the manufacturing process, and calculated the worst-case gaps required, and designed the puzzle pieces to these values, the puzzle would most likely be very loose and lock or pop easily.

So, it seems that the puzzle design process very trial-end-error, using experience of the particular manufacturing process to guide the designer on the "gaps" needed. I suspect that you must err on the "tight" side, for as long as it turns, then it can be broken in; this essentially means that the puzzle surfaces rub against each other and change shape to conform to the errors inherently produced by that particular manufacturing process.

Even this rubbing process will be different for different materials. Thus, for a puzzle which works very well in Shapeways WSF, the same design would not be suitable for mass production in injection molded ABS, which explains why Mefferts (and others) puzzles are so expensive to bring to the mass production stage.

Anyway, I think that in this context, the idea that people can decide on an ideal tolerance for the design of twisty puzzles is really not possible. The only answer is "... well on this puzzle design, using this material, I used +/- 0.xxx mm and it worked this well after being broken in for this long ..." is the only valid answer.

I suppose that this is why puzzle design is so frustrating, and so much fun

_________________
My Shapeways Shop: http://www.shapeways.com/shops/gus_shop

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Thu Feb 17, 2011 7:19 am

Joined: Wed Jul 21, 2010 4:01 am
By "tolerance", is the post actually refering to Engineering fit? That is, spacing that you want in the final puzzle, with the hope that production produces very little change in scale, where engineering tolerance is considered. Or are we actually discussing tolerances due to the printing process? In which case, I'd change my vote because of casting shrinkage when I cast prints.

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Thu Feb 17, 2011 7:38 am

Joined: Mon Mar 30, 2009 5:13 pm
Agalloch wrote:
By "tolerance", is the post actually refering to Engineering fit? That is, spacing that you want in the final puzzle, with the hope that production produces very little change in scale, where engineering tolerance is considered. Or are we actually discussing tolerances due to the printing process? In which case, I'd change my vote because of casting shrinkage when I cast prints.

I was referring only to "3D-printed puzzles", particularly via Shapeways as that seems to be the most popular source for TP designers at present. I think the discussion and insights have been really helpful. Of course the gap/tolerence should depend on the puzzle, material, manufacturing process, type of cut, etc., but it has been very useful to hear all these different experiences, and get a good idea of the required tolerence, at least as a good starting point for further optimization.

It seems that 0 tolerence is OK for interfaces where any necessary gap can be provided by springs, etc., but greater tolerence is required for pieces that slot into constrained grooves or holes of other pieces, as shown in Eric's picture. Of course the ideal tolerence depends on all these other factors mentioned above, but where tolerence IS required it should be based on the accuracy/resolution of the manufacturing process (which happens to be 0.2mm for Shapeways-printed components). I hope this makes sense.

PS. Gus, despite claiming to be a "non-expert", your explanation and insight were very accurate and helpful.

_________________
If you want something you’ve never had, you’ve got to do something you’ve never done - Thomas Jefferson

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Thu Feb 17, 2011 4:53 pm

Joined: Fri Mar 06, 2009 9:23 pm
Tolerances basically serve one main purpose in puzzles: make the turning of the puzzle easier. They reduce the total surface area contact between the two turning part groups during a turn (reducing the total friction, thus making it easier to turn)

I am going to go through an explain using pictures an words explaining my tolerancing style, and the reasoning behind it.

This is a 2d analogue of an abstract shelled mechanism puzzle. The red part is screwed into the core (it is the "arm" piece) and the rest of the parts are alternating colors to make viewing easier. The tolerances on all of the shell mechanism puzzles I make are all done following one rule: add tolerance between any two spherical faces that do not keep parts "held in" This will be explained in greater detail later on.

Here is the 2 dimensional sketch used to create the mechanism for my master pentultimate. When this is sketch is rotated around the turning axis (and subsequently applied to each face of the puzzle) the interactions of these surfaces create the basic shapes of the parts of the puzzle. As you can see, the total length of this sketch (think distance, not displacement) is pretty high. For the master pentultimate, as well as almost all other puzzles, adding tolerances are necessary to make a puzzle that works well.

In the next photo I have colored certain lines of the sketch red (please excuse the poor ms paint job, I don't have better tools/skills) The parts colored red are areas where tolerance is added. This means that instead of the parts that would be touching the revolved surface (remember that this line is being revolved into 3 dimensions) they wont touch each other anymore, eliminating all friction between them. You can notice a pattern evolving, which tends to happen on all "standard" shelled mechanism puzzles.

Going back into "the world of the abstract" I have made a small part chain here showing how parts would interact ( here you need to imaging that the red part on the left is anchored to an imaginary lower layer, and is responsible for holding the parts in the layer that is shown in) notice how in this photo the parts are based around spherical shells and conical cuts. I have included the tolerances I would normally add to a part like this. Notice how the parts are kept from moving away from the core of the puzzle.

This next photo is the same as the last, except that it has tolerances added to the conical cuts. its pretty easy to see here that this part would be very unstable now. The green part (and ultimately the red part on the right) has the ability to move in multiple directions, and can rotate. This would be a serious problem for any puzzle. The multiple degrees of freedom would cause lock ups and pops, which is something that obviously needs to be avoided.

You may be asking yourself " how come parts are allowed to move towards the center of the puzzle, wont that cause the same problems as described above?" The answer to that question is "no". By looking at the photo of the arch below you can see that even though there is ample amounts of room for any one of the blocks to move in toward the center, they are all prevented from moving down because the conical cuts force all of the parts to be shaped like wedges, disallowing them from moving in the only direction that the tolerances allow! In a puzzle that utilizes the shells mechanism, this sort of situation happens in every layer, and is a 3d analogy of the 2d arch.

Hopefully my guide isn't too confusing, and I'm hoping people will tell me where I'm not being clear enough, so I can update it, and make it easier for everyone to read.

_________________
--Eric Vergo

My Shapeways shop

Last edited by gingervergo on Mon Feb 21, 2011 1:12 pm, edited 2 times in total.

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Fri Feb 18, 2011 5:41 am

Joined: Wed Jul 21, 2010 4:01 am
gingervergo wrote:
Tolerances basically serve one main purpose in puzzles: make the turning of the puzzle easier. They reduce the total surface area contact between the two turning part groups during a turn (reducing the total friction, thus making it easier to turn)

I am going to go through an explain using pictures an words explaining my tolerancing style, and the reasoning behind it.

Again, this is engineering fit as opposed to and at odds with engineering tolerances mentioned in the previous two posts before it, including by the OP. Fit seems to be what most people are refering to when they talk about tolerance on the forums, however tolerance as described in this post is distinct from considerations of material shrinkage and inaccuracies. Which one is actually being discussed?

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Fri Feb 18, 2011 6:01 am

Joined: Mon Nov 30, 2009 1:03 pm
Agalloch wrote:
considerations of material shrinkage and inaccuracies.
3D-printing does not have shrinkage. The discussion is about the gaps between pieces to make the puzzles turn smoothly. As you know, a pin gets stuck if you push it into a hole without gap/clearance.

Oskar

_________________
.

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Fri Feb 18, 2011 6:10 am

Joined: Wed Jul 21, 2010 4:01 am
Oskar wrote:
Agalloch wrote:
considerations of material shrinkage and inaccuracies.
3D-printing does not have shrinkage. The discussion is about the gaps between pieces to make the puzzles turn smoothly. As you know, a pin gets stuck if you push it into a hole without gap/clearance.

Oskar

Thankyou very much Oskar, that is what I thought, I shall steer away from Gus and Kelvins post directly above and leave my vote as is - the process being similar but not nearly as well thought out as your own.

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Fri Feb 18, 2011 7:14 am

Joined: Mon Mar 30, 2009 5:13 pm
Agalloch wrote:
Oskar wrote:
Agalloch wrote:
considerations of material shrinkage and inaccuracies.
3D-printing does not have shrinkage. The discussion is about the gaps between pieces to make the puzzles turn smoothly. As you know, a pin gets stuck if you push it into a hole without gap/clearance.

Oskar

Thankyou very much Oskar, that is what I thought, I shall steer away from Gus and Kelvins post directly above and leave my vote as is - the process being similar but not nearly as well thought out as your own.

@Allagoch, I would be interested to know where you think there is any discrepancy or contradiction between what Gus and I were saying, and what Oskar said:

Oskar, Gus and I are all referring to the tolerence/gaps required in 3D-printed puzzles, which has absolutely nothing to do with shrinkage of cast puzzles that you keep mentioning for some reason. This is clearly stated in the main title of this thread:
Kelvin Stott wrote:
POLL: Ideal tolerence for 3D-printed designs

... in the main poll question:
Kelvin Stott wrote:
How much tolerence (gap between internal surfaces) gives the best result for 3D printed puzzles?

... in my original post:
Kelvin Stott wrote:
For those of you who have had experience with 3D-printing your puzzle designs, how much tolerence (gap between internal surfaces) gives the best result?

... AND was later confirmed directly to you in my subsequent post after you couldn't read all these "clues":
Kelvin Stott wrote:
I was referring only to "3D-printed puzzles"

And yet you are still talking about shrinkage??

Furthermore, both Oskar and I are in clear agreement that zero tolerence is generally OK (or even preferred) in most cases, especially where springs can provide any necessary tolerence. However, we also agree that some tolerence is required where one piece slides into another, as per Oskar's pin-in-the-hole example, and my reference to Eric's diagram above:
Oskar wrote:
Usually, I am a quite mild guy. However, twisty puzzles need zero tolerance to keep them well-behaved. ... Pin-in-hole constructions need 0.2 mm gap at all side to assure smooth sliding/turning.

Kelvin Stott wrote:
It seems that 0 tolerence is OK for interfaces where any necessary gap can be provided by springs, etc., but greater tolerence is required for pieces that slot into constrained grooves or holes of other pieces, as shown in Eric's picture.

Oskar wrote:
As you know, a pin gets stuck if you push it into a hole without gap/clearance.

As Gus and I clearly explained, much of this tolerence is required to allow for errors in the manufacturing process (0.2mm inaccuracy in the dimensions of 3D printing):
Gus wrote:
As a fledgling puzzle designer, as I understand it, the word "tolerance" should be applied to the manufactured product pieces. This tolerance will be different for different pieces; 3D printed by SLS or FDM, or injection molded, or for different types of materials (WSF, nylon, ABS). You have to allow for printing errors, mold errors, mold wear, material shrinkage/expansion and so on.

Kelvin Stott wrote:
Of course the ideal tolerence depends on all these other factors mentioned above, but where tolerence IS required it should be based on the accuracy/resolution of the manufacturing process (which happens to be 0.2mm for Shapeways-printed components).

You should read people's posts (as well as the title/subject of the thread) properly before discrediting them.

_________________
If you want something you’ve never had, you’ve got to do something you’ve never done - Thomas Jefferson

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Fri Feb 18, 2011 10:57 am

Joined: Wed Jul 21, 2010 4:01 am
Kelvin Stott wrote:
@Allagoch, I would be interested to know where you think there is any discrepancy or contradiction between what Gus and I were saying, and what Oskar said:

Oskar, Gus and I are all referring to the tolerence/gaps required in 3D-printed puzzles, which has absolutely nothing to do with shrinkage of cast puzzles that you keep mentioning for some reason. This is clearly stated in the main title of this thread:

I don't want to argue, but I'm not suprised you failed to read my posts when you couldn't even spell a name you quoted only lnes above correctly. My tag is clearly "Agalloch", and I did not mention cast puzzles a single time in my last post (could you quote me if I'm just being very forgetful, and apolgies if I am?) I have therefore mentioned shrinkage due to casting once, and would personally consider it a stretch to say that I therefore "keep mentioning" it!

In my opinion - and again massive apologies if I am wrong - I think I made it very clear that I was discussing the difference between tolerance and fit (which is often called tolerance), and even in my first post this point was not dependant on casting 3d printed puzzles.

This distinction can easily be shown by quoting (or if you had simply read) Oskar and Gus' originally posts.

These two indicate the disctintion between tolerance and fit respectively.

Gus wrote:
tolerances are a necessary evil which must be used to calculate worst case scenarios

Oskar wrote:
The discussion is about the gaps between pieces to make the puzzles turn smoothly

The first of these refers to design tolerances used to ensure quality when there can be mistakes in production, while the second refers to measurable differences you want in all cases (known also as engineering fit) in order to ensure the functioning of the puzzle. If you can guarantee that all produced puzzles were 100% identical then you do not need the form of tolerance which Gus was refering to, however, you still need tolerance as Oskar stated very clearly.

In case you are confused by my mention of shrinkage in my second post - as you incorrectly seem to think it had something todo with casting - i took it directly from this quote by Gus (as an example problem in place of an manufacturing errors)

Gus wrote:
You have to allow for printing errors, mold errors, mold wear, material shrinkage/expansion and so on.

I was not refering to casting, however this post which you agreed with explained tolerances include those in casting. This is very peculiar, and I'm not sure why a criticism of my post which has caused you to ignore my valid points doesn't invalidate this post.

Finally, you seem to misunderstand Oskar.

Oskar wrote:
As you know, a pin gets stuck if you push it into a hole without gap/clearance.

Kelvin Stott wrote:
As Gus and I clearly explained, much of this tolerence is required to allow for errors in the manufacturing process (0.2mm inaccuracy in the dimensions of 3D printing):

He seems to be clearly refering to the difficulty of inserting a pin into a hole without clearence. This has nothing to do with errors in the manufacturing process, as clearence is engineering fit, not engineering tolerance.

I hope this helps you to understand your mistakes and better understand the difference between intentional clearence and worst case measurements.

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Fri Feb 18, 2011 11:25 am

Joined: Mon Mar 30, 2009 5:13 pm
Clearly you do not understand that "tolerence" and the "gap" you need to ensure that a puzzle turns are effectively one and the same thing (or at least that's how I chose to define it in my original post), and in the *real* world (which is where Oskar, Gus and I live, I'm not sure about you) they DO depend on things like friction and manufacturing imprecision, which is what this thread/poll is all about - reality.

In theory, a pin *could* slide in a hole without any gap/tolerence/clearance (whatever you want to call it), provided there is no friction and/or manufacturing defects, but this does not happen in the real world. Even the temperature of the room can have an effect (either allow or disallow the pin to slide freely) as the parts may expand/contract at different rates. So Gus's post is spot on.

I'm sure Oskar (who has actually printed all of his puzzles) can confirm that he is talking in practical terms here, from real experience (i.e., taking into account things like friction and printing resolution), like Gus and I, not pure theory like you. In fact I had already discussed the pin-in-the-hole case with Oskar by PM before he even mentioned it in his first post, so I *know* we are all fully aligned here.

I suggest that if you have no experience of printing your puzzles then you have nothing to add (and no right to vote), which is precisely what I wanted to avoid in my original post:

Kelvin Stott wrote:
For those of you who have had experience with 3D-printing your puzzle designs, how much tolerence (gap between internal surfaces) gives the best result?

By all means, please feel free to start your own thread discussing theoretical definitions of gaps/tolerence/clearance and theoretical scenarios/hypotheses as much as you want. This thread/poll is about reality, and what people have actually found to work best in practice, through real experience. That is what people will find useful to know, not your theoretical analysis or nitpicking over what the gap/clearance/tolerance (whatever you want to call it) should be called.

_________________
If you want something you’ve never had, you’ve got to do something you’ve never done - Thomas Jefferson

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Fri Feb 18, 2011 1:17 pm

Joined: Thu Dec 02, 2004 12:09 pm
Location: Missouri
jabeck wrote:
Oskar wrote:
Usually, I am a quite mild guy. However, twisty puzzles need zero tolerance to keep them well-behaved.

Eric Vergo explained the concept to me:
1: Pull all pieces outward, away from the center.
2: All surfaces that feel the force need zero tolerance.
3: All gaps that are pulled open need 0.4 mm clearance.
4: All gaps that are pulled laterally (conical cuts) need zero tolerance as well.
5: Pin-in-hole constructions need 0.2 mm gap at all side to assure smooth sliding/turning.
Smooth turning is achieved by proper breaking in and good lubrication.

Oskar

Interesting - but confusing to me.
Same here...
1: One at a time or simultaneously?
2: What force? Thoughs of Star Wars come to mind. I think we are talking about the surfaces that come into contact with the pieces that hold the pulled piece in place. Correct?
3: If all pieces are pulled don't all gaps open? If just one piece is pulled and there is no tolerence with the pieces that hold it in then nothing moves and no gaps open. I guess we are talking about those surfaces that would be pulled away from the other parts IF the piece was allowed to move out.
4: I think I follow... but a 3x3x3 edge can be pulled parallel to a face or at 45 degrees to 2 neighboring faces. In one case you have surface cuts moving laterally and in the other you don't. I'm looking at the surface cuts seen on the outside regardless of the 3x3x3 mech that is used on the inside. Am I just supposed to pull in a strait line with the center of the puzzle? Even there the edge is pulled away from the face centers but there is laterally movement along the cut between the corners and the edge.
5: This is the only statement that is totally clear to me.

I've got to read the rest of the post here a few more times yet.

Carl

_________________
-

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Fri Feb 18, 2011 1:40 pm

Joined: Wed Jul 21, 2010 4:01 am
Kelvin Stott wrote:
Clearly you do not understand that "tolerence" and the "gap" you need to ensure that a puzzle turns are effectively one and the same thing (or at least that's how I chose to define it in my original post), and in the *real* world (which is where Oskar, Gus and I live, I'm not sure about you) they DO depend on things like friction and manufacturing imprecision, which is what this thread/poll is all about - reality.

In theory, a pin *could* slide in a hole without any gap/tolerence/clearance (whatever you want to call it), provided there is no friction and/or manufacturing defects, but this does not happen in the real world. Even the temperature of the room can have an effect (either allow or disallow the pin to slide freely) as the parts may expand/contract at different rates. So Gus's post is spot on.

I'm sure Oskar (who has actually printed all of his puzzles) can confirm that he is talking in practical terms here, from real experience (i.e., taking into account things like friction and printing resolution), like Gus and I, not pure theory like you. In fact I had already discussed the pin-in-the-hole case with Oskar by PM before he even mentioned it in his first post, so I *know* we are all fully aligned here.

I suggest that if you have no experience of printing your puzzles then you have nothing to add (and no right to vote), which is precisely what I wanted to avoid in my original post:

Kelvin Stott wrote:
For those of you who have had experience with 3D-printing your puzzle designs, how much tolerence (gap between internal surfaces) gives the best result?

By all means, please feel free to start your own thread discussing theoretical definitions of gaps/tolerence/clearance and theoretical scenarios/hypotheses as much as you want. This thread/poll is about reality, and what people have actually found to work best in practice, through real experience. That is what people will find useful to know, not your theoretical analysis or nitpicking over what the gap/clearance/tolerance (whatever you want to call it) should be called.

I'm hoping this is just a language barrier and you're not being intentionally obtuse. Tolerance and Fit are not only "theoretically" different, they 100% are and Oskar and Gus have both used different definitions. I understand that the confusing use of tolerance in English may confuse you, but I wasn't making suggestions, only posting facts.

I do have experience printing puzzles, and this is an odd attempt to deflect your own misunderstanding. I know being wrong can be annoying, but theres no need to get rude when I'm trying to educate you about the difference between tolerance and fit and how they are very very different things.

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Fri Feb 18, 2011 1:47 pm

Joined: Mon Mar 30, 2009 5:13 pm
Agalloch wrote:
I'm hoping this is just a language barrier ... I understand that the confusing use of tolerance in English may confuse you...
FYI, I happen to be British, native English speaker, and lived in England for 40 years (including PhD at Cambridge), so I don't think it's a language issue. Maybe just a misunderstanding about whether this thread was supposed to be about technical definitions, or just a simple question that everyone else seems to have understood very clearly:

"How much tolerance (gap between internal surfaces) gives the best result for 3D printed puzzles?"

But just to clarify, according to Wikipedia:

Quote:
Engineering tolerance is the permissible limit or limits of variation in

1. a physical dimension,
2. a measured value or physical property of a material, manufactured object, system, or service,
3. other measured values (such as temperature, humidity, etc).
4. in engineering and safety, a physical distance or space (tolerance), as in a truck (lorry), train or boat under a bridge as well as a train in a tunnel (see structure gauge and loading gauge).
5. in mechanical engineering the space between a bolt and a nut or a hole, etcetera.
I think it is clear that my original question is based on the last definition (no. 5), as I specifically refer to "gaps between internal surfaces" of mechanical parts. However, Gus and I (and others) then explained that this gap should depend on physical variability of the components - which just happens to be an alternative definition of the word tolerence (see no. 2 above) - so maybe that's what has caused the confusion: one word with two alternative definitions that happen to physically relate to each other.

Anyway, I hope it's clear now.

_________________
If you want something you’ve never had, you’ve got to do something you’ve never done - Thomas Jefferson

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Fri Feb 18, 2011 5:01 pm

Joined: Sun Mar 11, 2007 3:11 am
Location: Oregon, USA
Folks, you seem to be ignoring the "3d printed" part of the question. 3d-printed objects suffer aliasing at two stages in the production process:

• When the model data is converted into STL.
• When the model is built.

Conversion to STL causes aliasing because STL only describes polygons. For example, a cylinder will always be represented as a polygonal prism. You can minimize this effect by rendering at high resolution, but you can never eliminate it completely. A sphere will always be aliased as some sort of polyhedron.

Building a model entails reducing the STL to quantifiable physical units that approximate the contours of the model. 3D printers like Objet ('white detail') and ProJet literally print a matrix of plastic dots at fixed positions; ZPrinter ('sandstone') prints a binder matrix into a bed of gypsum powder; EOS ('strong & flexible') raster scans a laser across a bed of plastic powder.

Here's an illustration in 2D (from a discussion of image processing) that shows the challenge of rendering pixels of a curve. The same thing happens when building solids: the built material is going to be too high in some places, too low in others. (Imagine building your part with very tiny Lego blocks.)

The practical effect of all this seems to vary from one process to the next, and it can be affected by build orientation. In my opinion:

'White strong & flexible' (SLS) seems to vary by as much as ~0.1mm so allow ~0.2mm spacing for movement between curved surfaces.

'Gray Robust' (FDM) allows tighter tolerance in the XY plane (possibly ~0.1mm) than along the Z axis (possibly ~0.3-0.4mm). Exploit this by orienting the axis of cylindrical mechanisms parallel to the Z axis. Send FDM designs to http://printo3d.com/ instead of Shapeways because Bradley allows you to specify build orientation.

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Sat Feb 19, 2011 2:03 pm

Joined: Thu Dec 02, 2004 12:09 pm
Location: Missouri
Oskar wrote:
Eric Vergo explained the concept to me:
1: Pull all pieces outward, away from the center.
2: All surfaces that feel the force need zero tolerance.
3: All gaps that are pulled open need 0.4 mm clearance.
4: All gaps that are pulled laterally (conical cuts) need zero tolerance as well.
5: Pin-in-hole constructions need 0.2 mm gap at all side to assure smooth sliding/turning.
Smooth turning is achieved by proper breaking in and good lubrication.

Ok, I think I follow what each of these means now. We are basically talking about the spherical shell mech, cut up with conical cuts. So we are talking about the tolerence of the cuts on the inside of the puzzle.
gingervergo wrote:
Hopefully my guide isn't too confusing, and I'm hoping people will tell me where I'm not being clear enough, so I can update it, and make it easier for everyone to read.

But what if the final puzzle isn't a sphere. Say the surface cuts in a 3x3x3 for example. Based on what we seem to be saying here I think all those cuts should have zero tolerance as well. Correct? I don't think these cuts are addressed directly in the above list.

Carl

_________________
-

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Sun Feb 20, 2011 9:20 am

Joined: Sun Mar 15, 2009 12:00 am
Location: Jarrow, England
Who would have thought a technical discussion about engineering would result in such a heated discussion? Which channel to watch on the TV maybe, but tolerancing?

Anyway, arguments about semantics aside, I would always bow to opinions from people like Oskar. For goodness sake, the man has designed and built a 17x17x17 cube! As I said in my post, my thoughts were from the perspective of an engineer, not a puzzle designer.

As I see it, all parties have made valid comments to this discussion. In order to make a puzzle which turns well, you have to know about and understand the process in which it is going to be made as well as the tricks of the puzzle designer. Here, experience is king.
VeryWetPaint wrote:
Folks, you seem to be ignoring the "3d printed" part of the question. <snip> Building a model entails reducing the STL to quantifiable physical units that approximate the contours of the model. <snip> The same thing happens when building solids: the built material is going to be too high in some places, too low in others. (Imagine building your part with very tiny Lego blocks.)

VeryWetPaint makes an interesting point. There are always errors in analogue to digital conversion (in electronics I know this problem very well). I just wonder if the stated manufacturing tolerances quoted by 3D printing companies already allow for these quantisation errors. And of course, these edges and corners will be quickly rubbed down during the breaking in process.

_________________
My Shapeways Shop: http://www.shapeways.com/shops/gus_shop

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Sun Feb 20, 2011 1:37 pm

Joined: Thu Dec 02, 2004 12:09 pm
Location: Missouri
gingervergo wrote:
This is a 2d analogue of an abstract shelled mechanism puzzle. The red part is screwed into the core (it is the "arm" piece) and the rest of the parts are alternating colors to make viewing easier. The tolerances on all of the shell mechanism puzzles I make are all done following one rule: add tolerance between any two spherical faces that are touching, and do so in such a fashion where no part is allowed to move away from the core of the puzzle.

Ok, to test my understand look at this pic made in the same style as yours:
Attachment:

Tolerence.png [ 11.98 KiB | Viewed 4065 times ]

None of the surfaces of the black piece should have any tolerance... correct?

However you say "add tolerance between any two spherical faces that are touching". Here there is a spherical surface between the black and blue pieces yet you can't add any tolerance to either surface without allowing the blue piece to move away from the core.

Carl

_________________
-

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Mon Feb 21, 2011 1:04 pm

Joined: Fri Mar 06, 2009 9:23 pm
wwwmwww wrote:
gingervergo wrote:
Hopefully my guide isn't too confusing, and I'm hoping people will tell me where I'm not being clear enough, so I can update it, and make it easier for everyone to read.

But what if the final puzzle isn't a sphere. Say the surface cuts in a 3x3x3 for example. Based on what we seem to be saying here I think all those cuts should have zero tolerance as well. Correct? I don't think these cuts are addressed directly in the above list.

Carl

You are correct in saying that the so called "surface cuts" should not have any tolerances at all. The photo where I show where I added tolerance on the master pentultimate does not show tolerance on the "surface cut"

wwwmwww wrote:
Ok, to test my understand look at this pic made in the same style as yours:
Attachment:
Tolerence.png

None of the surfaces of the black piece should have any tolerance... correct?

However you say "add tolerance between any two spherical faces that are touching". Here there is a spherical surface between the black and blue pieces yet you can't add any tolerance to either surface without allowing the blue piece to move away from the core.
Carl

You are again correct. My wording is a Bit confusing and I will go back and change it to something (hopefully) clearer.

_________________
--Eric Vergo

My Shapeways shop

Top

 Post subject: Re: POLL: Ideal tolerence for 3D-printed designsPosted: Wed Feb 23, 2011 5:05 pm

Joined: Sun Mar 11, 2007 3:11 am
Location: Oregon, USA
These academic discussions are theoretically valid, but still ignore some important realities of 3D printed objects:

3D printed tolerances are not rounded to the best match. The same parts may build differently in different runs due to part placement in the machine. Multiple identical parts in the same build may build differently.

An example of this is my gift-cube box in which the STL depicts a stack of 6 flat, flexible sheets exactly 0.8mm thick. Of the six parts I received, four are ~0.72mm thick and two are ~0.83mm thick. That's a variation of +0.3/-0.6mm.

Your wall thicknesses can always be rounded-up and rounded-down to the machine's nearest match. The EOS supports Z-axis layer thicknesses of 60μm, 100μm, 120μm, 150μm, and 180μm; although we don't know which setting Shapeways uses for its 'strong & flexible' parts, you can easily see that most of these settings are incapable of exactly building a 1mm wall. With a layer setting of 150μm, for example, a 1mm wall may be built as 6 layers (0.9mm) or 7 layers (1.05mm) depending on where the machine just happens to 'slice' your STL file.

Errors don't average-out between parts because each part can independently be built oversized or undersized depending on how they get sliced.

To illustrate this, here's real STL for two parts that would fit together if built perfectly to scale. I've scaled the walls to exactly 1.00mm to exaggerate the effect of slicing/building, just to help illustrate. Ordinarily you wouldn't build parts this small!
Attachment:
File comment: Sample STL data for cube and cutout with 1mm walls.

Slicesample.png [ 8.83 KiB | Viewed 3961 times ]

Below I ran the STL through a representative slicing script to demonstrate how the resulting wall thicknesses can vary, just depending on where the operator positions the part within the build chamber.

Here's how this part would be built on an EOS (or similar 3D printer) using a layer thickness of 150μm if the part had been positioned on the bottom of the build chamber. The build software slices the model at 150μm intervals, starting 75μm from the bottom of the chamber. The result is that the bottom walls were built with 7 layers totalling 1.05mm thick but the opening was built as only 6 layers totalling 0.9mm. Hence the little cube is 16% taller than the opening it was supposed to fit into.
Attachment:
File comment: How the 3D printer would slice the STL if it were resting exactly at the base of the build chamber.

Sliced_plus_0_0.png [ 14.05 KiB | Viewed 3961 times ]

And in the next picture I moved the STL data 0.1mm higher and ran it through the slicer again. That's how the data might get built if the operator positioned the part in a different location in the build chamber.

This time the bottom layer was sliced into 6 layers totalling 0.9mm in height, but the opening spans 7 layers and is 1.05mm high. The little cube would fit into the hole with room to spare.
Attachment:
File comment: Same part sliced in a different position in the build chamber.

Sliced_plus_0_1.png [ 12.85 KiB | Viewed 3961 times ]

These illustrate actual variations you will encounter when building parts on any 3D process. These pictures aren't just artwork; that's an actual slicing algorithm against actual STL data. Your designs need sufficient gaps to tolerate parts up to 1 quantum too large, yet sufficient interlock to securely hold pieces up to 1 quantum too small.

Top

 Display posts from previous: All posts1 day7 days2 weeks1 month3 months6 months1 year Sort by AuthorPost timeSubject AscendingDescending
 Page 1 of 1 [ 35 posts ]

 All times are UTC - 5 hours

#### Who is online

Users browsing this forum: No registered users and 4 guests

 You cannot post new topics in this forumYou cannot reply to topics in this forumYou cannot edit your posts in this forumYou cannot delete your posts in this forumYou cannot post attachments in this forum

Search for:
 Jump to:  Select a forum ------------------ Announcements General Puzzle Topics New Puzzles Puzzle Building and Modding Puzzle Collecting Solving Puzzles Marketplace Non-Twisty Puzzles Site Comments, Suggestions & Questions Content Moderators Off Topic