PDA

View Full Version : Cool technology



garagemark
03-11-2009, 09:19 AM
Just another cool technology you guys and gals might be interested in. Although there would be issues with the current form, I thought it was just cool as hell. That this gizmo can "grow" a WORKING model is very impressive.


http://widgets.nbc.com/o/47f1317f105123ad/498ebd00a62edaa0/47fe70d4555df05a/9e46bd46/-cpid/ba4377d3bfd6c81

Whadda you all think?

davidh
03-11-2009, 10:09 AM
link appears broken

Circlip
03-11-2009, 10:45 AM
Tain't, it's working well over here.

Evan
03-11-2009, 11:40 AM
I gave up on it.

garagemark
03-11-2009, 11:50 AM
I dunno. I just clicked on it and it worked fine. Sorry guys. I can't repair it if it ain't broke on this end.

Berniep
03-11-2009, 11:53 AM
It is a link to a flash/shockwave video application. If you don't have the appropriate software/players installed it wont work.

Spin Doctor
03-11-2009, 12:06 PM
Stereolithography, aka 3D Printing. Been around for a while in various forms (the earliest used rolls of paper IIRC). The cost will come down to the hobbiest level eventually (and most likely be made in China). There are plans out on the internet for a do it yourself version http://fabathome.org/wiki/index.php?title=Main_Page cost estimate is arund $2000

Evan
03-11-2009, 12:16 PM
$2000? I bet I can build one for much less. I'll make it from beetle killed pine. :D I wonder if you can make one to work with concrete?

Liger Zero
03-11-2009, 12:20 PM
I worked with selective laser sintering, that was really cool. I understand the range of materials it can work has really expanded in the last few years and cycle-times/accuracy are improving.

mochinist
03-11-2009, 12:48 PM
Im taking a SolidWorks class at a local college right now and in the second half we get to model something within reason and use the 3-d printer to make it. They have a display of other stuff students have made and it is pretty impressive.

I was thinking of doing something like a mobius strip or a klein bottle, but the klein bottle wouldnt be see thru so I dont know. That and I dont think my math skills or drawing skills in solidworks is good enough yet:p


Someone posted a page a long time ago that had all kinds of cool models of mathmatical shapes, anyone got a link for that?

Liger Zero
03-11-2009, 01:11 PM
Im taking a SolidWorks class at a local college right now and in the second half we get to model something within reason and use the 3-d printer to make it. They have a display of other stuff students have made and it is pretty impressive.

I was thinking of doing something like a mobius strip or a klein bottle, but the klein bottle wouldnt be see thru so I dont know. That and I dont think my math skills or drawing skills in solidworks is good enough yet:p


Someone posted a page a long time ago that had all kinds of cool models of mathmatical shapes, anyone got a link for that?

Can you put a cross-cap on a Klein-bottle or does that break the universe? :D

Wiki has some intresting stuff about Klein bottles, cross-caps and other interesting mathematical stuff.

mochinist
03-11-2009, 01:16 PM
Can you put a cross-cap on a Klein-bottle or does that break the universe? :D

Wiki has some intresting stuff about Klein bottles, cross-caps and other interesting mathematical stuff.reminds me of that time I tried to divide by zero...:D



http://lexlibertas.com/wordpress/wp-content/uploads/2008/04/i-divided-by-zero.jpg

Liger Zero
03-11-2009, 01:20 PM
Just being a smart-arse one day I casually mentioned the "square root of negative zero" in the presence of a mathematically inclined programmer, you could actually hear his brain go "pop." :D

We both had a good laugh about that.

dp
03-11-2009, 01:36 PM
It's the Jay Leno rapid prototype video - it's been presented before. Fun stuff out of my price range.

ckelloug
03-11-2009, 01:38 PM
There's absolutely no problem with the concept of the square root of negative zero. IIRC, it should be negative zero according to IEEE-754. Yes, Dorothy, negative zero is something you have to deal with on computers.

Liger Zero
03-11-2009, 01:39 PM
There's absolutely no problem with the concept of the square root of negative zero. IIRC, it should be negative zero according to IEEE-754. Yes, Dorothy, negative zero is something you have to deal with on computers.

...learned something new today. :o

lazlo
03-11-2009, 01:42 PM
Agreed Cameron, but an IEEE 754 result doesn't mean that it's mathematical meaningful. NaN's are legal IEEE 754 results, and there's no analog in "real" math.

lazlo
03-11-2009, 01:46 PM
Yes, Dorothy, negative zero is something you have to deal with on computers.

Cameron, you made me check ;) -- square root of a negative number throws an exception in IEEE 754, and returns a NaN.

Fasttrack
03-11-2009, 02:01 PM
There's nothing wrong with dividing by zero either. We have an ongoing joke in the physics department about divide by zero's. They come up alot in physics and you'll find that they are often meaningful. Divide by zero either produces inifinity or some finite result if you take a limit as the argument approaches 0. Shows up all the time in just about every field theory.

davidh
03-11-2009, 02:07 PM
it looks like a real neat thing to play with. . . im betting leno has one,

whatta suppose a super gadget like that would really cost to buy ? ? ?

oh, that right, i think you need to be very computer literate to operate it and the economy is a little bit in the toilet so i should just watch and envy. . .

MickeyD
03-11-2009, 02:38 PM
I played with the previous version of the scanner and it is pretty impressive. With software to make it useful and a support contract, it is about $5300. It was not as easy to use as they show and the software (at least the version from 2 years ago) was clunky, but it was on the right path.

I looked at one of the 3D rapid prototype machines like Leno has and they start around $25K and go up from there. The consumables are also expensive as well as the service agreements.

johnnyd
03-11-2009, 02:49 PM
There's a convention coming up in May 12,13,14th in Shaumburg,Ill. that is specifically oriented to rapid manufacturing.
They can even make you a finished metal part. Parts can be made from cad files or scanned image files.
I posted a link over on the Practical Machinist website.
I booked my trip thru expedia & got a 2 day motel,car rental & roundtrip airfare to/from Florida for $424.00.
I've seen the 3D printers in action & they are awsome! Now I have to know more about that technology.

ckelloug
03-11-2009, 03:36 PM
As a complete unabashed trolling pedant:

Wikipedia's entry on sqrt of negative zero says it should be negative zero. I believe the result returned depends on the trap mode the processor is in based on my experience doing mathematical software. My paper copy of IEEE-754 is missing so I couldn't check. I've never encountered a need for negative zero but I think there are cases in numerical arithmetic where it could be important otherwise, IEEE-754 would not call out specifics for handling it.

Back to our regularly scheduled OT conversation!

Liger Zero
03-11-2009, 03:44 PM
As a complete unabashed trolling pedant:

Wikipedia's entry on sqrt of negative zero says it should be negative zero. I believe the result returned depends on the trap mode the processor is in based on my experience doing mathematical software. My paper copy of IEEE-754 is missing so I couldn't check. I've never encountered a need for negative zero but I think there are cases in numerical arithmetic where it could be important otherwise, IEEE-754 would not call out specifics for handling it.


In other words its meaningless when programming a mill in G-code... useless for anything other than making the programmer blink. :D




Back to our regularly scheduled OT conversation!

mmmm rapid prototype machine mmmmmmaarrrrgggggllllhhhh :p

lazlo
03-11-2009, 04:15 PM
There's nothing wrong with dividing by zero either.

You deal with divide by zero in signal processing all the time.

In the IEEE 754 floating-point gospel, dividing a positive value or +infinity by zero results in +infinity and dividing a negative value or -infinity by zero results in -infinity.

lazlo
03-11-2009, 04:19 PM
As a complete unabashed trolling pedant:

Wikipedia's entry on sqrt of negative zero says it should be negative zero.

I looked it up in the official IEEE 754 spec when I quoted that Cameron -- real square root of a negative number returns a NaN. If Wikipedia says otherwise, they're wrong, especially since the floating point designers literally test against the IEEE spec. That's why the revised 754 R spec, which includes MAC-based floating-point, has been in committee for 10 years -- everyone (IBM, Intel, AMD, ...) argues over the smallest details, because that's what 100's of millions of floating-point units are going to be designed to...

Spin Doctor
03-11-2009, 05:03 PM
$2000? I bet I can build one for much less. I'll make it from beetle killed pine. :D I wonder if you can make one to work with concrete?

Or Jatoba? :D

ckelloug
03-11-2009, 05:12 PM
I get -0.0000 when I take the sqrt(-0.0) on a linux machine with IEEE-754 math on an AMD processor:




#include<stdio.h>
#include<math.h>

int main(int arc, char**argv)
{
fprintf(stderr,"%f\n",sqrt(copysign(0,-1)));
fprintf(stderr,"%f\n",sqrt(copysign(1,-1)));
}

Output:
-0.000000
nan



Note what may be the first actual legitimate use of the vB code tag on a machining forum :p

Liger Zero
03-11-2009, 05:14 PM
Ok, BACK ON TOPIC... how can we use this square-root-of-negative-zero in home-built rapid-prototyping 3D printers?

ckelloug
03-11-2009, 05:17 PM
sqrt(-0) could be used as a flag to tell the printer that the height it has received is invalid and it should not try to make the plastic part infinity units high at that point :)

I think that the mini discussion on the square root of -0 is however proof that there are at least two exceptionally bored electrical engineers posting here today!

Liger Zero
03-11-2009, 05:18 PM
sqrt(-0) could be used as a flag to tell the printer that the height it has received is invalid and it should not try to make the plastic part infinity units high at that point :)

I think that the mini discussion on the square root of -0 is however proof that there are at least two exceptionally bored electrical engineers posting here today!

I could try diverting them by mentioning calibration again. :eek: (runs)


:D

lazlo
03-11-2009, 05:52 PM
I could try diverting them by mentioning calibration again. :eek: (runs) :D

Someone post some more pictures from the Mitutoyo calendar :)

Seriously, one of our English members was kind enough to send me the UK version of the calendar -- thank you so much!

lazlo
03-11-2009, 06:20 PM
Stereolithography, aka 3D Printing.

There are plans out on the internet for a do it yourself version http://fabathome.org/wiki/index.php?title=Main_Page cost estimate is arund $2000

The Fab@Home project is pretty amazing. A sprocket from the link Spin posted:

http://fabathome.org/wiki/uploads/thumb/2/21/UVG1_sprocket_finished.JPG/800px-UVG1_sprocket_finished.JPG

It looks like the major limitation on Fab@Home 1.0 is that it uses a single syringe, which limits the kinds of fluids you can pump. This is a version from Cornell that uses two syringes. I can't hear the audio at work, but I gather that's two part epoxy, although I don't understand why they're laying it down like they are (not equal parts of each):

http://fabathome.org/wiki/uploads/a/a7/2-SyringeToolDemo.wmv

lynnl
03-11-2009, 06:47 PM
Isn't i (imaginary #) the sq root of -1? ...wouldn't 0i be the sq root of negative zero?

Just asking. It's been a long time since my math days.

lazlo
03-11-2009, 08:26 PM
Isn't i (imaginary #) the sq root of -1? ...wouldn't 0i be the sq root of negative zero?

You're absolutely right. Unfortunately, the floating-point hardware doesn't understand imaginary math, or calculus for that matter. If you want to do higher-level math on a modern CPU, you need a software library like Mathematica, Maple or Mathcad. All are frequently used in engineering, science and physics...

Evan
03-11-2009, 08:59 PM
I've never encountered a need for negative zero

It is used all the time to report fine gradations of temperature. It can be zero degrees outside but if there is a slight drop then it will shift to negative zero. Negative zero is a lot colder than positive zero since it represents the difference in heat content that occurs when the latent heat of crystallization of water is released. Negative zero is also used to indicate which direction the value of zero was approached from. If the temperature drops to zero then the proper display is positive zero. If it rises to zero then negative zero is appropriate. It also brings up a very important question. What is the result returned if you divide zero by negative zero?



:D

Liger Zero
03-11-2009, 09:05 PM
It also brings up a very important question. What is the result returned if you divide zero by negative zero?
:D
http://www.entity.cc/central/atomic/mush_or_80.gif


My question involves the height gauge I just purchased. If the tolerance is +-.000 and my gauge reads -.000 is the part in spec?

lazlo
03-11-2009, 10:44 PM
I finally got to listen to Jay's voice on GarageMark's video now that I'm home -- one really cool feature of that Dimension 3D printer is that it prints the 3D object completely assembled, with all the working parts inside.

I thought that was BS at first, but they actually print spacers/standoffs for internal 3D parts that are soluble in a special solvent, and when you wash the finished part in the solvent, all the internal parts are freed.

So they printed a Crescent wrench as a single assembly, but the worm gear actually drives the movable jaw. Amazing!

macona
03-12-2009, 04:52 AM
Were supposed to be getting one of the Dimension SST's down at techshop. Pretty neat. The special solvent is water. You basically get a self contained dishwasher to finish the part. Not cheap to run and the item is still kind of grainy. Our patternmaker uses one sometimes for complex parts.

As for the laser scanner I wonder how accurate it is.

John Stevenson
03-12-2009, 06:44 AM
In other words its meaningless when programming a mill in G-code... useless for anything other than making the programmer blink. :D


Don't know about that.
My old controller made by Ahha which is an old DOS based ones requires the 4th axis to be controlled by two variables, one is degrees and the other is the sign + or -

Degrees is obviously movement and sign is direction.
So you are at +30 and want to go back to zero.
Zero is always taken as + so the command W 0.0 [ W is equal to A on the Ahha ] makes it move forward 330 degrees to 0.0

Not ideal if the tool is in a cut.

Giving it the command -0.0 is ignored as it thinks zero is always plus.

The only way we got rund this was to give it a command W- 0.0000000001 which it accepted but treated as zero.

.

Rustybolt
03-12-2009, 08:06 AM
The Fab@Home project is pretty amazing. A sprocket from the link Spin posted:

http://fabathome.org/wiki/uploads/thumb/2/21/UVG1_sprocket_finished.JPG/800px-UVG1_sprocket_finished.JPG

It looks like the major limitation on Fab@Home 1.0 is that it uses a single syringe, which limits the kinds of fluids you can pump. This is a version from Cornell that uses two syringes. I can't hear the audio at work, but I gather that's two part epoxy, although I don't understand why they're laying it down like they are (not equal parts of each):

http://fabathome.org/wiki/uploads/a/a7/2-SyringeToolDemo.wmv


Now imagine the same thing only when the machine is finished making the part you can take that sproket and mount it on your bike. MIT has been working wiith the same process, only using metal. I wrote something about it here a couple of years ao. At that time I think MIT was using sintered metal and then fusing it. I may be mistaken , but I think, now they are using a vitreous metal.

A.K. Boomer
03-12-2009, 11:06 AM
I finally got to listen to Jay's voice on GarageMark's video now that I'm home -- one really cool feature of that Dimension 3D printer is that it prints the 3D object completely assembled, with all the working parts inside.

I thought that was BS at first, but they actually print spacers/standoffs for internal 3D parts that are soluble in a special solvent, and when you wash the finished part in the solvent, all the internal parts are freed.

So they printed a Crescent wrench as a single assembly, but the worm gear actually drives the movable jaw. Amazing!


My friend went to a demonstration about a year ago in denver and came back with a little plastic part with a built in spring/mechanism ---- The other neat thing about this tech. is the parts can not only be built in multi-functional pieces --- they can also be produced hollow as well, these units are capable of building things that just plain cannot be built by any other method, The only thing is they have to be done in plastic -- right now.

Liger Zero
03-12-2009, 11:16 AM
With the right knowledge you can select a polymer that can perform as well as a metal in many applications.

Not *all* applications by any stretch of the imagination, but give it time. Material-science marches onwards.

Evan
03-12-2009, 11:59 AM
The only thing is they have to be done in plastic -- right now.


Not so. Metal is available. See here:

http://www.bathsheba.com/

lazlo
03-12-2009, 12:30 PM
Were supposed to be getting one of the Dimension SST's down at techshop.

Wow, that's great! The one on Leno's video looks really expensive -- how much are they?

lazlo
03-12-2009, 12:33 PM
At that time I think MIT was using sintered metal and then fusing it. I may be mistaken , but I think, now they are using a vitreous metal.

The one that Evan linked is using powdered stainless, with some kind of heat activated binder. You have to bake the whole assembly in an oven when you're done to fuse the whole thing, but the resulting "metal" is a foam (40% air), and it's extremely fragile:

From Evan's link:

"After the whole model is built up, and the extra powder is shaken off, the piece goes into an oven, where heat drives off the binder and fuses the steel powder. There's just enough heat to make the granules weld together where they touch, without collapsing the entire piece into a puddle. This produces a porous steel part that's about 60% dense, like the one at left.

This "green" material is matte gray, feels like sandstone, and won't take a polish. It is soft enough to cut quickly with a hacksaw, but can't quite be dented with a fingernail. It's considerably lighter than steel, which is not surprising since it's 40% air. "

Evan
03-12-2009, 12:46 PM
From what I understand they aren't very fragile. The density is about the same as aluminum and the metallic particles are actually fused themselves, not by a binder.

lazlo
03-12-2009, 01:54 PM
"Can't quite be dented by a fingernail" sounds fragile to me :)

The artist in your link pours/wicks bronze into the foamy stainless to make those beautiful sculptures. Kind of a high-tech Oillite sculpture.

By the way, for those who haven't (or can't) watch the Leno video, the "Dimension" 3D printer uses a big roll of ABS plastic string -- looks like a Weed Wacker coil :) It winds through the machine, and there's a heater tip on the head mechanism that liquifies the plastic and then squirts it down, a lot like an ink jet printer.

I wonder if you could do the same with the Bismuth-based low-temperature fixturing alloys. Be expensive as hell though...

Evan
03-12-2009, 02:56 PM
There are 3D machines currently available that produce hard metal parts equivalent to those produced by compressed metal powder metallurgy. The parts are ready to use as actual production pieces.

Spin Doctor
03-12-2009, 05:38 PM
I first heard about laser scintering powder metal stereo lith oh say 15 years ago. Read about it in an article in Analog. They talked then about a screw and nut made as one operation that you could unscrew. Now as to what class thread :D

George Bulliss
03-12-2009, 06:16 PM
I stopped into my old shop in Rochester, NY last spring and got a tour of some of their new machines, one of which is a Direct Metal Laser Sintering machine. It is able to build parts out of S-7 that can then be machined on a mill or EDM, just like regular S-7. It can also be heat treated to the same hardness.

I was told that you could not polish it to as high a level as regular S-7 but it all other regards it was almost identical.

George

lazlo
03-12-2009, 06:44 PM
There are 3D machines currently available that produce hard metal parts equivalent to those produced by compressed metal powder metallurgy. The parts are ready to use as actual production pieces.

Now that truly is something out of science fiction novel :)


I stopped into my old shop in Rochester, NY last spring and got a tour of some of their new machines, one of which is a Direct Metal Laser Sintering machine. It is able to build parts out of S-7 that can then be machined on a mill or EDM, just like regular S-7. It can also be heat treated to the same hardness.

That sounds like that would allow you to just spit-out any kind of arbitrary form or cutting tool. Heck, could you "print" endmills too?

So how long before we can print an internal combustion engine? :)

Rustybolt
03-12-2009, 07:11 PM
There are 3D machines currently available that produce hard metal parts equivalent to those produced by compressed metal powder metallurgy. The parts are ready to use as actual production pieces.


The article I read mentioned they were making hollow turbine blades.

Liger Zero
03-12-2009, 07:32 PM
The article I read mentioned they were making hollow turbine blades.

The machine in question at Rochester (If George is talking about the one I am thinking of) was acquired to do rapid prototyping originally. The idea was to be able to rapidly produce a usable part from a drawing while conventional production was scaling up. You could form usable parts with this machine while the tooling and fixtures were being produced for conventional production methods.

At least that was how it was explained to me at the time. I've been gone from there for awhile.


The technology of Selective Laser Sintering as I understand it now has advanced to the part where you can produce very complex very high tolerance parts in a variety of materials, with more materials being developed ongoing. There are small prototyper scale machines and big SLS production cells out there now.

Stereo laser lithography/stereolithography (SLA) is a bit more limited as the polymers have to be photosensitive but I understand they have widened the range of materials and the properties of these materials quite a bit since I played with it.

Just like CNC mills we will eventually see home-shop scale technology within our lifetimes. Heck I forsee the day when you can pick up one at Harbor Freight for $399. 20 years from now. Mark my words.

Spin Doctor
03-12-2009, 09:38 PM
Just like CNC mills we will eventually see home-shop scale technology within our lifetimes. Heck I forsee the day when you can pick up one at Harbor Freight for $399. 20 years from now. Mark my words.

And the "inkjet" cartridges will be $249.99 :D

lazlo
03-12-2009, 10:20 PM
The machine in question at Rochester (If George is talking about the one I am thinking of) was acquired to do rapid prototyping originally. The idea was to be able to rapidly produce a usable part from a drawing while conventional production was scaling up.

It depends on your notion of "rapid". If you watch the videos of these sintered metal machines in operation, they have a tray of powdered metal, and a little squeegie runs across the top and flattens it out. Then the laser, or a resin inkjet head, selectively fuses a very thin sheet of the metal (like a 32nd). Then the whole tray drops down, the squeegie runs across again to smooth the powdered metal, and the whole process starts over. It's very, very slow:

http://www.youtube.com/v/QMbQ5dsGB-U&hl=en&fs=1

By the way, I was looking at the sintered metal process that's linked in the artist's site that Evan linked, and they have a different process that produces sand casting mold cores directly from a CAD drawing. The tray contains casting sand, and the inkjet head applies foundry resin:

http://exone.com/eng/technology/x1-prometal-rct/equipment_prometal_rct.html

lazlo
03-12-2009, 10:39 PM
Here's another video from the same company that makes the machine George and Liger are describing: EOS.

The turbine blade is at 5:56. Surface finish is pretty lousy -- all the shiny, pretty parts have a subtitle that says that the parts were electropolished :)

http://www.youtube.com/v/C9awF5te_2w&hl=en&fs=1

It looks like really neat technology for niche markets, but they've got a long way to go before it's usable for the mass market.

loose nut
03-13-2009, 12:31 AM
I always new CNC was just a fad.