Repricing Advice: Puzzles & Multi-part files

Hi Everyone,

Starting October 7th, Shapeways will be charging per part, a pricing change that most negatively impacts the multi-part file community. Pete, our CEO, discussed the reasons this change was necessary on our blog, and I wanted to create a place for us all to get through this together, here. There are a few ideas for how this could be resolved, and I'll have our Materials and Factory teams chiming in here as much as possible.

We want to be transparent about how we're managing this transition and let you know we're building special tools to make ordering many parts more cost effective after repricing.

First of all, why are high part count files so expensive?
High part count files are expensive because at every step in our production process, each piece is manually handled and accounted for. This means that check each part, orient each part, breakout each part, sort each part, polish each part, dye each part, and finally sort one last time before shipping you all of the parts. Each product at Shapeways is touched an average of 10 times before making its way to you, if you multiply that by the many parts in a file, that's 500+ direct interactions with one model file. This costs considerably more to produce than if that same material volume was one product.

How will you help me cost effectively adjust the models in my shop?
We knew this was going to be a problem we had to address inevitably, so we've been building tools for this transition. These tools aren't yet ready to be released to the public, but we're going to invite you specifically to our printability center pilot preview where we'll be able to work with you and your model parts specifically to optimize for this new pricing structure. We've also put together a tutorial specifically for optimizing the cost of high part count files that will be coming out soon.

We'll be checking regularly, so let us know how you're feeling

Best Always,
Savannah

Hi Savannah,

The twisty-puzzle community is discussing the Puzzle & Multiparts files issue also at the Twisty Puzzles Forum. As you understand, the sudden huge price increase has stirred a lot of emotion, as puzzle designers have invested tens of thousands dollars into Shapeways models, and tens of thousands of man hours on the designs. Unchecked, we would lose all our customers, and our investment in Shapeways would become sunk cost. We hope you will not let that happen.

As promised, here is an initial list of questions on this issue. Expect more to follow.
-How does the new pricing model work? What are the optimizations? Do we still need to hollow our parts to reduce cost?
-Do we still need to make those silly powder holes that only aggravate the powder issue at both Shapeways and the buyer?
-What are the design parameters for sinterboxes? How does the Shapeways "machine" know whether a model has been properly sinter boxed?
-What are the design parameters for sprues? Is Shapeways willing to polish sprued models? Would sprues be cheaper than sinterboxes?
-How about our legacy model into which we have invested tens of thousands of manhours?
-What solutions can Shapeways provide that does not require us to re-model, remesh, re-upload re-prove each of these already proven models?
-What toolings can Shapeways provide for packing optimization, automatic sinterbox generation and automatic sprue generation?
-What solutions can Shapeways provide for multi-part models that is cheaper and more customer-friendly than ordering each sub-model separately?
-What solutions can Shapeways provide to handle the 1M-triangle limit? Many models go beyond this limit if we need to add triangles for sinterboxes or sprues.
-Would Shapeways be willing to offer the "lease a tray" option? That is, I arrange a full 3D-printer tray, optimally filled with puzzle pieces, and Shapeways prints the whole tray. Shapeways could shake out the recyclable powder, but you keep the pieces dusty. I would take care of the cleaning, sorting, shipping, etcetera. Would you be willing to discuss such option?

Thanks.

Oskar

And what about enclosed parts? They don't require so much work as completely separated ones.

I'm confused about how machine space is calculated. It seems like it is counting a bounding box for each shell and adding them together with a little extra gap around the edges, not just a direct multiplication of x*y*z of the file? Or is my math wrong? It seems crazy that you would double count the bounding boxes if they are sharing space.

The tutorial mentions necklacing/looping as a way to reduce costs, but as far as I can tell, it just adds costs - an extra $1.50+ for the shell + material. If that's true, I'd be better off removing the necklaces in my existing multi part files, not adding them? Perhaps you are adding a tool to do this, but what do we do in the meantime? Are my manually placed, time tested loops not acceptable? And how long till these beta tools are available? I have a handful of large jobs at the moment and it's very difficult to continue doing business if I don't know how to estimate costs effectively.

The problem I find here is that you are giving us tools to manage the changes but not yet.

The slight problem is the the changes are happening now - 7 days from now anyway. We need time to use the tools that we haven't got yet to handle these changes. This is making life very difficult for those of us that earn our living from this!

I at the moment am in the unenviable position of having no idea if I still have a viable business in 7 days.

Tom

I can see that we are going to have to print more plastic to deal with the new pricing structure. I think that Shapeways also need to look at their equipment and systems to come up with a more economic multi-part costing. The simple model signboards in my example below would be very easy to clean, yet the new pricing structure has $60+ of labor in the price. I could attach them all to a printed spru and save $58.50 of labour. It would take just as long to clean post print. I understand that maybe it is harder to track the multiple parts versus the single parts, but this is surely a solvable problem with a better tracking system.

After some experimenting, I'll revise my question some. It seems like interlocking/necklaced parts are recognized by the software. And that if objects are necklaced, there can be savings by nesting objects within one another. It would be exceedingly helpful if the staff at shapeways could clarify the underpinnings of the updated rules. With one week to update markups, the deadlines for testing holiday designs coming fast, it would be helpful to make educated decisions rather than guessing.

I appreciate that many other services have expressly forbidden multi part files, and while this pricing change at Shapeways reflects a hike in my prices in some items, is necessary for the new industry of 3D printing. Just get the info out so we can get our work done!

Additionally, are there any provisions for grandfathering in printability for files that have been simply rearranged?

Thankfully, I'm not affected, since I do not have any designs that incorporate multiple parts in polyamide.

I see the problem you're having with multiple parts, but is this really the solution? The way I see it Shapeways is damaging the foundation stones of its business by making this move. I don't know what the actual numbers are and I'll bet the numbers look good on paper, but I feel this is not going to be good for Shapeways' overall image that has been built up over the years.

What else can be done? Couldn't Shapeways write code that automatically encapsulates each multipart file in a mesh container and then charge for the extra material and build space volume?

Or what about trying to automate the sorting process. In this regard I'd look into the way diamonds are sorted from ore concentrates via x-ray fluorescence separator equipment. The advantage you have is that you know exactly what each part looks like so you could use machine vision in place of the x-ray emitter detector combination.

Make a game out of it! HAHAHA! No seriously though, make it a game. Partner up with MMO game developers and figure out a way that the players can sort the parts to get some kind of in game reward.

Also along the lines of the game aspect, you could partner up with people that develop captchas. Captcha solving is being used to do all kinds of interesting things with all those billions of brains at work solving captchas. You could simply ask people solving captchas, "Do these two parts match, yes or no?" This along with standard captchas would make for a very powerful captcha while getting the part sorting done at minimal costs.

Outsource it. There billions of people around the world who would love to help out for a price far below what workers in Holland and America want to be paid. The sorting can be completely implement overseas via online.

Use a SCARAs to slice up the powder cake into sections containing multiple part orders. I'm seeing two robot arms with something like steel spatulas that are cutting up the block according to what's inside. They are then placing the sections into baskets for automated depowdering equipment to process.

Stop leasing those monstrous machines and build your own compact machines in various sizes. The patents are expired. That way you can, not only offer next day service more readily, but you can also dedicate multiple part orders to each machine.

That's all I can think of for now.

What if there were two different pricings ?

One without sinter box, the new pricing which depends on the number of parts.

One with sinter box where shapeways doesn't post-process and clean in depth, the user does, and that doesn't depend on the number of parts because the number of parts doesn't make the job harder !
I have small models whose prices have been multiplied by 10 because of this...

You're obviously selling the volume on the machine (powder being only 1200/metric ton at a very low volumic mass) so why not sell models with sinter boxes by bounding box size and not by number of parts ? It's only logical...

Just wanted to add to this last idea, which I think is probably the best idea of the ones I mentioned. That is, these could be designed in such a way that they operate by not heating the entire layer to just below the melting point, thus, nearly 100% of unused powder could be reused. There would need to be some heating to eliminate uneven shrinkage, but not so much heating that the physical properties of the unused powder are degraded to the point where there would be problems with reuse.

The disadvantage to this is speed. You could increase laser power to compensate for this to a degree, but you would need to keep the laser power down price wise, such that the idea of creating numerous inexpensive machines would be practical. I'm thinking that small machines, with a build volume of 150 mm cubed could be built for under $1000 and up depending on the laser power. Also, in regards to speed is the cool down time. This would be greatly reduce given the small size.

If I were in charge there I'd have some of my engineers working on this. It's a way to get the advantage over the competitors.

@UniverseBecoming:

If you heat parts of they layer to below the melting point, that will cause the same problem, with powder being in a transition state and either ruining the print, having the same amount of 'damaged powder' or both (which is why SLS powder can only be recycled max 5 times, otherwise printing errors are bound to happen).

I was working with a lot of students with this (I'm a mechanical engineering student) and we were trying to build a company with SLS printers under 1000$ but the problem is that lasers have to be about 200W, the cheapest being about 10k $. We could use a lense but print time would be days. SLS printers will be cheap one day, but laser need to be cheaper, because the smallest laser itself costs in the thousands (even a 20W laser)...

If you were in charge there you would probably have some extra expenses for a project that wouldn't be practical and you'd have to make Shapeways users pay much more to backup the losses, just like they are doing right now Why not wait until the Ice1 and Ice9 come out ?

@Shapeways:

Could you easliy compute the total price of all models uploaded on shapeways and the total price before and after the change ? I'm pretty sure your 75% cheaper models are cheaper by 1$ and the 25%, more expensive by 200$ on average... The total value of all the models was probably increased by 25% or something... I even have a model that was made 3250% more expensive !

I'm not sure I understand your first paragraph, but it sounds like you are talking about preheating the entire layer to just just below the melting temperature, which is exactly the opposite of what I was saying. What I was saying is no preheating of each layer is needed. Though some heating would eliminate warping, yet on second thought, it probably wouldn't even be necessary for such small build volumes.

200 watts? I think you may be remembering that as the power levels needed for laser sintering of metals as only a very small amount of laser power is needed to do laser sintering of polyamide. One could probably get away with powers as low as one watt, though the printing speed at that power level would be very very low. The biggest most powerful commercial SLS machines for sintering polyamide use laser power levels in the 70 watt range.

One can obtain a 60 watt laser for under $450, so that is plenty of power for a small build area that doesn't use preheating of the layers. There rest of the machine is just off the shelf motion control components and some optics. Completely doable for under $1000. Slow, but it would only need to be fast enough to complete the sintering of the build volume overnight while employees are having their off time. For a large scale operation like Shapeways one would have a large number of these small machines running simultaneously.

to the readers of this thread, I just realized that I moved the thread off topic. Please feel free to comment on the topic again.

If I were faced with the multi part problem I would simply attach minute sprues to everything with attachment points that were reduced in volume so as to leave minimal surface imperfections. Customers would just have to deal with the removal of the sprues.

I realize that a number of you have too many models to go through and correct any time soon. I wonder if it would be feasible to have Shapeways develop code for an automatic sprue attaching procedure? The sprues wouldn't need to be pretty, just functional. I'm thinking of something like the automatic support generation that is used for SLA, something like that, but connecting the parts together rather than connecting to the build platform.

RubixFreakGreg, if you'd like to continue discussing SLS machines in more detail perhaps we could start a new thread somewhere in one of the forums.

disappointed that Shapeways staff has yet to respond to Oskar's post at the beginning of this thread.
You used to heavily promote his puzzles on your site, but now you're leaving Oskar (and the rest of us) without a
reasonable path forward.

The maths skills of the puzzle people probably got many of the questions above solved already, given the radio silence.

This one is interesting, at least if it can be done as a layer. Rent 20 mm thick layer, for example.

The lease a tray is similar to shared wafers in the integrated circuit business. Shared wafers are often done for small runs or protoypes before moving on to a full wafer (which would be the analog for the whole tray). I could envision shared trays where people purchase one of several standard available block sizes that would consist of an entire printed cage filled with parts. The cage bar gaps could be variable and depend on the minimum sized objects in a cage. In this scenario there would be no intrusion of parts from one block to another. It would be a pure play on the ability of a designer to pack things hyper efficiently in their block. It's somewhere between the new standard ordering and a full tray purchase. It's definitely geared towards someone who needs a bunch of parts either for puzzles or reselling. The price of a block should be easy to calculate based on the portion of volume that is shared and probably with an add on for material volume used. It's also not an on demand thing. Generally shared wafers are done on a specific schedule such as once per month so you would have to plan ahead and purchase your tray space in advance.

Hello! I made bezels for jewelry into which I put my handmade pieces. I was super excited about being able to use 3d printing in steel because I can design my own bezel into any shape I want. I can get cheap silver-plated bezel from China and Mexico for about $2-3 a piece, but they are not my own design. I wanted to opt for Made In America, and with multi-part pricing I could create my own bezels in steel for $4-5 a piece. (The steel is not perfect because it's slightly yellow from the bronze. I would love silver plating but that's not available either.) Now the cost of my bezels are doubling in price to $8-9 a piece. I charge under $50 for my pieces so now Shapeways is taking 25% of my earnings. This is going to be difficult. I really want to design my own bezels and I love to support made in America, but you're making it really difficult for small business to use your service.

Andrea

Going back to the old pricing would solve everything and this problem would probably not have happened in the first place.

If all else fails, what should I do? Should I:

A: Send an angry email to Shapeways?
B: Go on a Shapeways spending spree and buy all the puzzles I can?
C: Use this as proof the world is starting to end?
D: Email them and calmly tell them to go back to the old pricing?
E: Propose an ultimatum to them?

And if all of the choices, in what order should I do the choices?

Don't take this seriously. This seems like the best choices if all else fails.

I'll answer what I can, while Savannah is getting the rest of the answers.

-How does the new pricing model work? What are the optimizations? Do we still need to hollow our parts to reduce cost?
The optimization is a lower price per cc for models that don't use excessive machine space. Here's the blog on how it works. Hollowing parts doesn't reduce the cost like it used to, as you are still charged for the machine space, just not the material.

-Do we still need to make those silly powder holes that only aggravate the powder issue at both Shapeways and the buyer?
Right now there still must be an escape hole

-What are the design parameters for sinterboxes? How does the Shapeways "machine" know whether a model has been properly sinter boxed?
You have to have the minimum distance listed on the guidelines, and follow the wall thickness rules for the material. I do not know the magic that is used to see if a model is trapped in a box or not.

-What are the design parameters for sprues? Is Shapeways willing to polish sprued models? Would sprues be cheaper than sinterboxes?
At least the minimum wire size in the guidelines. The more models and weight you have on that sprue the thicker it needs to be. Sprue vs sinterbox will depend on layout and material used.

Hey guys,

I've asked Raphael, Alan and Greg to respond to some of your questions and they should chime in here shortly.

Cheers,
Natalia

Hi all,

Here are some answers to the questions raised in this thread:

How does the new pricing model work? What are the optimizations? Do we still need to hollow our parts to reduce cost?

I think this blog post describes it best.

Do we still need to make those powder holes that only aggravate the powder issue at both Shapeways and the buyer?

There is no longer as much incentive to hollow parts. One of the goals of the new pricing structure was to avoid designers having to make design adjustments merely for pricing.

What are the design parameters for sinterboxes? How does the Shapeways "machine" know whether a model has been properly sinter boxed?

The design parameters for best results with sinter boxes are as follows: minimum 1.2mm wire thickness. Hole size smaller that the 2nd largest dimension of your smallest part. The Shapeways 'machine' does not look for proper sinter boxing.

What are the design parameters for sprues? Is Shapeways willing to polish sprued models? Would sprues be cheaper than sinterboxes?

Design parameters for best results with sprues are as follows: minimum 2mm sprue wire thickness. Classic Design for Manufacturing priniciples such as Length/Diameter limitations and the increased streght of increased fillet radius should be observed. Material anisotropy due to orientation should be taken into account. Shapeways will polish and dye sprued models but these will be QA'd and eventually rejected if they are poorly designed and do not survive the manufacturing process. Whether sprues would be cheaper than sinterboxes depends entirely on your model, again, I would not encourage you to design merely for cost saving, though this can (and will) likely factor into it.

What solutions can Shapeways provide to handle the 1M-triangle limit? Many models go beyond this limit if we need to add triangles for sinterboxes or sprues.

The limit was raised to 2M triangles recently. We are planning on raising that limit again this year for most materials.

What solutions can Shapeways provide that does not require us to re-model, remesh, re-upload re-prove each of these already proven models? What toolings can Shapeways provide for packing optimization, automatic sinterbox generation and automatic sprue generation?

This is something we are currently working on and should have more news about it next week.

What solutions can Shapeways provide for multi-part models that is cheaper and more customer-friendly than ordering each sub-model separately?

Ordering each model separately is now the same as placing them all together in one file. They can be sinterboxed together to remove the
per-part charge but that really only works for non-polished/dyed items.

Would Shapeways be willing to offer the "lease a tray" option? That is, I arrange a full 3D-printer tray, optimally filled with puzzle pieces, and Shapeways prints the whole tray. Shapeways could shake out the recyclable powder, but you keep the pieces dusty. I would take care of the cleaning, sorting, shipping, etcetera. Would you be willing to discuss such option?

This is a discussion we are having over email with you Oskar let's keep it going there.

I recognize that for this group in particular, you are some of our most creative and intelligent designers that invest a lot of time into optimizing for 3D printing, so I appreciate that you are interested in designing for manufacturing as well.

We will keep you posted on the new tools we should be rolling out soon as we believe they will be of great value to you.

-Greg