Mold sdm - way better than any 3d printing process

[Anthony Douglas]

I want to risk for other people interested in this. I think this prices called Mold sdm can frankly beat the pants of every other 3d process in existence, *on the bread and butter things* like dimensional tolerance, scalability to larger sizes, surface finish, and material properties.

Not printing tissues, or functional gradients, I know.

Here are some links below, but I actually think that the authors themselves font see the potential it has. Although granted, I like it because of the potential for *grassroots* technological development, like the reprap. They are thinking in a very different context.

A lot of these documents are super old, yet no progress has been delivered... If this potential is to be delivered there probably has to be a project to get it going like the reprap did for filament deposition manufacturing.

Btw, the executive summary is that you produce a highly accurate mold of almost any desired shape, of a destroyable mold material, then fill and destroy the mold material.

There are a few limitations on geometry, but less even than powder bed laser sintering. Like tubular objects or hollow tanks have to have some holes in the side, which you might have to patch up later. And minimum feature sizes, but they could be five thousands of an inch easily.

When you are actually involved in manufacturing, you see what small fries that is, in comparison...

articles:

https://www.sciencedirect.com/science/article/pii/S0261306999000138
Automated fabrication of complex molded parts using Mold Shape Deposition Manufacturing

https://www.researchgate.net/profile/Jurgen_Stampfl/publication/2554741_Application_Of_The_Mold_Sdm_Process_To_The_Fabrication_Of_Ceramic_Parts_For_A_Micro_Gas_Turbine_Engine/links/09e41505833780f08a000000.pdf

https://www.researchgate.net/profile/Jurgen_Stampfl/publication/250678938_Fabrication_of_High_Quality_Ceramic_Parts_Using_Mold_SDM/links/02e7e539776e00ed5d000000/Fabrication-of-High-Quality-Ceramic-Parts-Using-Mold-SDM.pdf

The stuff below I c cannot remove, my browser is malfunctioning.
https://patents.google.com/patent/US6375880B1/en

10.1016/S0921-5093(01)01800-7 we

10.1002/9780470294628.ch3

[Anthony Douglas]

No comment at all? How else are we going to improve over 3d printing.

I don't think we will ever or at least in the near future, be able to get a combination of tight tolerances like utc-5, good surface finish, normal engineering materials with the range of properties we need, no limits on general geometry, and speed, our of any of the printing processes currently under development in the public sphere.

Does anyone?

[John Griessen]

On 07/15/2018 05:11 PM, Anthony Douglas wrote:
> I don't think we will ever or at least in the near future, be able to get a combination of tight tolerances like utc-5, good surface finish, normal engineering materials with the range of properties we need, no limits on general geometry, and speed, our of any of the printing processes currently under development in the public sphere.

Can't decipher the above sentence.

"no limits on general geometry, and speed, " Can first comma be left out? Probably.

then this shows up as last element in a list:

"our of any of the printing processes currently under development in the public sphere."

What is that?


"under development in the public sphere" That suggests fiction of the fantasy category to me.

[Anthony Douglas]

Second try:

 I don't think we will ever, or at least in the near future, be able to get a combination of: tight tolerances (like utc-5), good surface finish, normal engineering materials with the range of properties we need, no limits on general geometry, good speed ( a few hours for a fully dense part 10 cm by 10 cm, perhaps), out of any of the printing processes currently under development in the public sphere.

Sorry, I didn't know I was that bad sometimes.

By public sphere, I mean the stuff you might read about in the news or blogs.  There is stuff that is just not being discussed due to patent concerns, I assume. Or it is just buried.

On Sun, Jul 15, 2018, 7:04 PM John Griessen <jo...@industromatic.com> wrote:

On 07/15/2018 05:11 PM, Anthony Douglas wrote:

> I don't think we will ever or at least in the near future, be able to get a combination of: tight tolerances (like utc-5), good surface finish, normal engineering materials with the range of properties we need, no limits on general geometry, good speed ( a few hours for a fully dense part 10 cm by 10 cm, perhaps), out of any of the printing processes currently under development in the public sphere.

Can't decipher the above sentence.

"no limits on general geometry, and speed, "  Can first comma be left out?  Probably.

then this shows up as last element in a list:

"our of any of the printing processes currently under development in the public sphere."

What is that?


"under development in the public sphere"  That suggests fiction of the fantasy category to me.

--
You received this message because you are subscribed to a topic in the Google Groups "Open Manufacturing" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/openmanufacturing/Hr0YjbC1_bw/unsubscribe.
To unsubscribe from this group and all its topics, send an email to openmanufactur...@googlegroups.com.
To post to this group, send email to openmanu...@googlegroups.com.
Visit this group at https://groups.google.com/group/openmanufacturing.
For more options, visit https://groups.google.com/d/optout.

[John Griessen]

Will we see in the news a combination of: tight tolerances (like utc-5), good surface finish,
normal engineering metals, no limits on general geometry,
good speed a few hours for a fully dense part 10 cm by 10 cm?

The good surface finish is not happening any time soon.
Making molds by fancy SD manufacturing might help get a part,
but none of these has a price you want -- all are very high.
SDM sounds very high priced.

For niche market low volume things, I like that tactic of lowering
standards to make slightly imperfect shapes, then go to normal injection
mold tooling if the volume gets high. That's why I described how I am working
on methods to use slightly inaccurate molds and allow some parting lines and
allow that to work at all by molding at lower pressure than
the usual injection mold technique. The mold can be 3D printed by imaterialize
as sintered steel tiny spheres held together by bronze alloy. They start out as
steel spheres glued together, then shrink a little in an oven, then get infused with
bronze. I'll be reporting here how it turns out as a mold.

[Anthony Douglas]

Why does mold sdm sound expensive?  It looks like a router plus some deposition stuff, to me.  No lasers as the primary energy source or anything.

It is more expensive and complicated than fdm, certainly, I grant.  It might be $5000 in parts and materials to get started. Still, a basic 300 pound class 3 axis mill goes for about that.

Also the cost to performance ratio matters a lot. Cheap but almost useless is not worth much, either.

Like the reprap, I would hope that mold adm can build a large fraction of its own parts, so that tells with the cost. You could perhaps make a lot of the router by using polymer concrete and hydrostatic bearings, including hydrostatic leadscrews.

[John Griessen]

On 07/16/2018 09:04 AM, Anthony Douglas wrote:
> It is more expensive and complicated than fdm, certainly, I grant.  It might be $5000 in parts and materials to get started.

That's capital intensive for this audience. So, next question is which service companies
are offering tooling made by that method, since DIY is a hurdle?

imaterialize.com offers bronze sintered steel sphere prints now.

3D Printed Steel: printed in 420 stainless steel and infused with bronze. The composition will be approximately
60% steel and 40% bronze (Cu 90%/Sn 10%)

Part of the pricing formula is $6 US per cc as of 3 yrs ago:
https://i.materialise.com/blog/en/3d-printing-in-steel-finishes-and-prices/

$6 US per cc of volume of the 3D model. So, a thin walled, (.25cc per sq cm area), mold with air channels built in could be
under $288 for a 8cm x 12cm x 2cm mold.

Maybe that could be a starting point for more machining of the mold surface
to be smooth? Is there some SDM way of adding a thin layer of metal?

You mentioned dissolving away parts of molds to get a result, but that won't go over big for $1k molds...

[Anthony Douglas]

Now it's my turn to have some difficulty deciphering... I will reply to some approximate interpretation, because it's an interesting discussion :).

 I agree $5k is a lot to an individual.  Especially if they don't have much design expertise.  Plus actually making stuff has more to it than making the parts and assembling them. A cloud of designs would help a lot.

To be honest, the machine in the basement of someone's house or garage is a concept I regard as not very interesting, I know it's not necessarily what you are suggesting, but it seems to be the default.... 

It's got to be a small group, some ten to couple hundred people, collaborating like with a hacklab or fab lab. Otherwise the machine isn't giving good value, it would sit there a lot of the time.

Right now, making parts to tight tolerances etc. Requires a lot more than $5k (4k usd).  And has high labor input and material costs.  A fab lab is about $50,000.  And they can't even do a lot of this stuff, at all.

A solution for parts that cost $5k would be great progress over what we have right now.  A $1k machine would be nice, but I think the path to that we must travel  includes through a period when the machine is $5k.  We can only leapfrog so much over that intervening space...

I have read about infiltrating sintered parts.  It has many advantages over full melting with lasers, especially but the tolerance is still quite poor.

[John Griessen]

On 07/16/2018 01:37 PM, Anthony Douglas wrote:
> I have read about infiltrating sintered parts.  It has many advantages over full melting with lasers, especially but the tolerance
> is still quite poor.

Does SDM offer a good way to put a surface layer of metal like bronze on a substrate of bronze infiltrated sintered
metal?

CMU shows a diagram here: http://www.cs.cmu.edu/~sdm/methodolgy.htm I think I just got something
about SDM that I missed before -- You could deposit support and part material 1cm deep, then mill some smooth,
then deposit more, then work on a zone between 1cm and 2cm deep, and repeat until tall enough.

I read "remove sacrificial support material" as dissolve before, which was because I was missing
that a milling machine carves in layers that can be thick.

This starts to get more specific finally:

http://www.cs.cmu.edu/~rapidproto/students.99/myh/project2/materials.html

The metal parts are mostly made from steel, using a copper support/sacrificial structure. The support structure is there to
stabilize the part while it’s being manufactured. During the production of metal parts, shot peening is necessary to force the
metal together while it cools and solidifies.

Wow! Now my old bronze foundry experience is coming alive and I want to combine oxy torch steel blob dripping/welding with CNC
milling and shot peening and make some smoke! Funny that they say copper as support above, since it sucks heat so well
it's hard to weld, easy to get a pile of disconnected drips instead of a single hunk of metal...

Tin bronze just like imaterialize infiltrates sinterings with is low toxic and could be a good support material
for oxy-torch dripped steel -- and it's 90% copper. Maybe that's what they meant...

[Anthony Douglas]

They use stainless steel as the build material and copper as the support because copper can be etched away with nitric acid, which has negligible effect on the alloy of ss they were using.

Elsewhere they describe how you can get perfect welding, so perfect strength between layers, to occur, and zero remelting, despite the differences in melting point.  It's because the heat gets conducted away fast enough to prevent remelting. It's a pretty cool revelation. 

However I trend strongly away from sdm of that kind.  I much prefer mold sdm, where you make and then fill a mold.  The mold could be a  material like colloid bonded zirconia powder, used in investment casting, so you can mold metals.  The process to get tolerance would not be easy, of course, as there is a lot of shrinkage when casting most engineering metals... But the coefficient of expansion of the zirconia or blend of ceramics can match the metal.  You can do somethong like rheocasting, where the metal is 80 percent solid particles, reducing shrinkage.  Lastly, directional solidification as is done with turbine blade casting, could eliminate some types of distortion.  Forcing the metal in under pressure perhaps after solidification might be useful. 

Phase transitions of the metal are the biggest problem I would foresee, especially steel is bad for that. If the ceramic could elastically distort and yet still provide plenty of constraining force, things could be cooled with highly predictable distortion and zero final stress, and desirable crystalline phase properties... And if the distortions are sufficiently well understood, the cast shape can be compensated ahead of time to give the desired final shape, perhaps for utc 5 or higher precision maybe.

 The shot peening is not to bond the layers, it is to remove the residual stress left by the deposition process.

The residual stress thing is mission critical for additive manufacturing.

If you drop melted metal on a plate, it shrinks as it solidifies and cools. In the end it is in a state of being slightly stretched, like an elastic band that was stretched and them glued down.  The shot peening squishes the material, eliminating that state of being stretched. If done too much of course it will leave the material in a state of compression instead.

The big problem it's that the stress of each layer adds up, leading to distortion, curling, and other problems.

[John Griessen]

On 07/16/2018 09:38 PM, Anthony Douglas wrote:
> they describe how you can get perfect welding, so perfect strength between layers, to occur, and zero remelting, despite the
> differences in melting point.  It's because the heat gets conducted away fast enough to prevent remelting. It's a pretty cool
> revelation.
>
> However I trend strongly away from sdm of that kind.

I like the possibility of reducing mold tooling costs for thermoplastic molding.
I thought the original topic, Mold sdm, meant SDM for making molds.

Did Mold sdm mean something else?