FDM + SLS = PCB?
Brandon Fosdick
It seems that the sticky problem is printing both the board and the traces at the same time. There are methods for printing metal and methods for printing plastic, but I haven't seen one that can print both in the same model. It occurred to me that some sort of frankenstein machine that can employ two different methods would do the trick. Off the top of my head, FDM and SLS seem like a good match.
Mechanically, the two methods seem compatible: the FDM head can be moved out of the way to accommodate the powder roller, and vice versa. To make a circuit board, the FDM head would put down a layer of plastic everywhere that the traces aren't. Then the roller would push a layer of metal powder over the board to fill in the traces. Sinter with the laser, then repeat for the next layer.
Any thoughts? I can't be the first to think of this, so why hasn't it been done? What am I missing?
°|° Walt Perko °|°
Making 3D plastic boards is pretty easy 'cept for the tiny holes ... the 3D
printers don't like holes less than about 3mm diameter. OTOH, printing
plastic boards with groves for the conductive runs is possible ... for
simple boards but then filling in the groves with conductive material and
NOT soldering become the next fence to hurdle in the process ...
I know I would be very interested in ideas on how to make a circuitboard
work starting with a 3D Printed board with grooves and maybe small dents for
the bores.
This seems like a very useful project to explore ...
===============================================================================================
C U L8r, �|� Walt Perko �|� "Kids ... teach them the good stuff,
and they still learn the bad stuff."
A robot that doesn't think ... but just teaches ... the BallBot1!
http://www.Brainless.org/R2Pv1.net/BallBot1.html
RoboToons weekly Robot Cartoons; http://www.brainless.org/RoboToons/
"The World Needs a New Economic Model"
--
You received this message because you are subscribed to the Google Groups
"HomeBrew Robotics Club" group.
To post to this group, send email to hbrob...@googlegroups.com.
To unsubscribe from this group, send email to
hbrobotics+...@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/hbrobotics?hl=en.
Dave Curtis
-dave
Brandon Fosdick
> Making 3D plastic boards is pretty easy 'cept for the tiny holes ... the 3D printers don't like holes less than about 3mm diameter. OTOH, printing plastic boards with groves for the conductive runs is possible ... for simple boards but then filling in the groves with conductive material and NOT soldering become the next fence to hurdle in the process ...
Hole size might not be a problem because through holes would need a metal cylinder inside of them anyway. That should close the hole enough to be used for through hole parts. Even if it is a problem, surface mount parts are becoming popular enough that I can see this being an acceptable limitation.
The thin walls between the leads of surface mount parts are worrisome too, but it looks like the MakerBot can do 0.5mm walls. That shouldn't be a problem for the larger packages. It's a bit tight for something like SOIC though.
Brandon Fosdick
> Well, one of the things people are working on w.r.t. the RepRap is coming up with a good dual-head design, primarily so that you can print support material. But some people have thought about using the second head to simply deposit solder or some other low melting point metal.
It seems like the material used for the traces would need to have a melting point significantly higher than the temperature of the soldering iron used to attach components to the board. How do they plan to handle that?
°|° Walt Perko °|°
The $3,000 PP3DP can print a .2mm thick resolution ... the $10,000
Dimensions prints at a 1.75mm resolution.
===============================================================================================
C U L8r, �|� Walt Perko �|� "Kids ... teach them the good stuff,
and they still learn the bad stuff."
A robot that doesn't think ... but just teaches ... the BallBot1!
http://www.Brainless.org/R2Pv1.net/BallBot1.html
RoboToons weekly Robot Cartoons; http://www.brainless.org/RoboToons/
"The World Needs a New Economic Model"
----- Original Message -----
From: "Brandon Fosdick" <bfos...@gmail.com>
To: <hbrob...@googlegroups.com>
--
°|° Walt Perko °|°
Hence the idea that experimentation is needed to develop this concept into a
viable solution ... sans heat soldering ... perhaps developing a conductive
glue?
===============================================================================================
C U L8r, �|� Walt Perko �|� "Kids ... teach them the good stuff,
and they still learn the bad stuff."
A robot that doesn't think ... but just teaches ... the BallBot1!
http://www.Brainless.org/R2Pv1.net/BallBot1.html
RoboToons weekly Robot Cartoons; http://www.brainless.org/RoboToons/
"The World Needs a New Economic Model"
----- Original Message -----
From: "Brandon Fosdick" <bfos...@gmail.com>
To: <hbrob...@googlegroups.com>
Sent: Monday, September 05, 2011 5:15 PM
Subject: Re: [HBRobotics] FDM + SLS = PCB?
--
Ytai Ben-Tsvi
David Wyland
is exactly this idea: 3D printing of an insulating substrate + 3D
printing of wires between the parts. The obvious next step is abandon
the PCB concept.
In the newer approach, you stick the components upside down (dead bug
style) on some substrate, plastic or metal, Then you deposit plastic
around them up to the leads, Next you deposit wiring metal between the
leads of the different components. You will need more than one layer of
insulation and wiring? No problem. Use as many as you like! Then top it
off with the final covering layer.
The result is a 3D style, multi-layer "PCB" that uses a minimal amount
of metal. Harder to do rework, though. ... I think.
Kind of blows the mind,
Dave Wyland
Dave Curtis
On Sep 5, 2011, at 7:43 PM, David Wyland wrote:
> This leads into a new fab technology under development: 3D circuits. It is exactly this idea: 3D printing of an insulating substrate + 3D printing of wires between the parts. The obvious next step is abandon the PCB concept.
>
> In the newer approach, you stick the components upside down (dead bug style)
Well, as long as everything is a through-hole part in a DIP package, but otherwise "dead bug" is kind of hard to manage, mechanically. With surface mount parts, it isn't so much dead-bug as dead-no-see-um.
> on some substrate, plastic or metal, Then you deposit plastic around them up to the leads, Next you deposit wiring metal between the leads of the different components. You will need more than one layer of insulation and wiring? No problem. Use as many as you like! Then top it off with the final covering layer.
>
> The result is a 3D style, multi-layer "PCB" that uses a minimal amount of metal. Harder to do rework, though. ... I think.
This is somewhat reminiscent of "MultiWire" technology from the late 70's-early 80's. Kollmorgen was the vendor when I ran into it, but I believe they had bought out the company that invented it. Anyway... it was a small volume production technology that somehow started with component pin pads, maybe with a thin PCB or such, but the main part of the technology was fine, insulated wire, something well under 30ga wire, like 38 or such IIRC, and they basically used an X/Y robot to lay down wire in an epoxy resin. When it cured you got something like a regular PCB, but about 2-3 times as thick and surprisingly weighty. The board thickness was sometimes an issue with things like edge connectors.
I saw them used at two different companies that I worked for, but never on my project so I never had to rework one myself. I'm told it is about the same as reworking a PCB. But bear in mind that the equivalent PCB would have had 6 or 8 signal layers, so cutting a trace on a buried layer of the PCB required getting out the board plots and finding a place you could drill with a small bit in a pin vise and not cause too much collateral damage. MultiWire was the same way -- you needed to find a relatively benign place to drill through all layers to cut a buried wire.
To a certain extent, MultiWire was a solution in search of a problem. In those days, we were in the habit of sending out prototypes to a CNC wire-wrap shop. These were boards with about 200 DIP packages. Production PCBs were 8 layer boards, 6 signal and 2 power layers. MultiWire had a pretty skinny niche between those two. It did offer much better impedance control than wire wrap, though, so if you needed to do a complex prototype and required reasonable impedance control, MultiWire could it for you -- you could use it with 100K ECL logic.
I actually looked at using MultiWire prototypes for one of my projects, but it didn't make sense for two reasons; 1) the design had multiples of the same board so even for prototyping our PCB volumes got big quickly, 2) my secret weapons: Jim and Lucy. Jim was brilliant PCB designer -- didn't understand beans about electronics but he could quickly route the worst spaghetti that ever was and enjoyed every minute of it, and Lucy was a rework operator with a photographic memory and the workmanship of a jeweler -- with those two, PCBs were the obvious way to go.
-dave
>
> Kind of blows the mind,
>
> Dave Wyland
>
>
> On 9/5/2011 5:15 PM, Brandon Fosdick wrote:
>> On Sep 5, 2011, at 16:35 , Dave Curtis wrote:
>>> Well, one of the things people are working on w.r.t. the RepRap is coming up with a good dual-head design, primarily so that you can print support material. But some people have thought about using the second head to simply deposit solder or some other low melting point metal.
>> It seems like the material used for the traces would need to have a melting point significantly higher than the temperature of the soldering iron used to attach components to the board. How do they plan to handle that?
>>
>
°|° Walt Perko °|°
Thinking about this a PP3DP has a feeder system ... but why not add another
feeder to the side of the print head to feed thin wire down into a hole,
then along the printed ditch to create an electrical run? Another head on
the other side of the plastic print head could push a rivet down into the
holes after the wire has been inserted ... then the plastic print head would
cover everything up creating a layer.
The PP3DP can print a 140mm x 140mm area ... so that's a pretty decent size
PCB. Making the PCB multi-layer would also help to create strength in the
plastic.
Another idea is that being able to print 3D ... now the circuitboard could
be printed inside the 3D shape of the device it runs!
===============================================================================================
C U L8r, �|� Walt Perko �|� "Kids ... teach them the good stuff,
and they still learn the bad stuff."
A robot that doesn't think ... but just teaches ... the BallBot1!
http://www.Brainless.org/R2Pv1.net/BallBot1.html
RoboToons weekly Robot Cartoons; http://www.brainless.org/RoboToons/
"The World Needs a New Economic Model"
----- Original Message -----
From: "Dave Curtis" <da...@dave-curtis.com>
To: <hbrob...@googlegroups.com>
Sent: Monday, September 05, 2011 9:50 PM
Subject: Re: [HBRobotics] FDM + SLS = PCB?
Brandon Fosdick
Chris Palmer
printing circuits using inkjet cartidges. Here's an old link to an
artcle about Epson experimenting with it back in 2004:
http://arstechnica.com/old/content/2004/11/4366.ars
I thought I'd seen some other variants on this since then, where folks
were trying to even print simple components (or maybe that was the
goal).
Another similar story from 2007:
http://m.zdnet.com/blog/btl/printing-circuits-on-your-ink-jet-printer/4893
This kind of stuff looks like it has great potential for both flexible
circuits and "print at home" circuit boards.
I also remember seeing an article in one of the hobby magazines
(Servo, Circuit Cellar, Nuts & Volts) where they were printing a
simple wing-flapping bird robot (more like a desk sculpture).
Anyone see anything newer in these areas?
-Chris
Dave Curtis
I'm going off-grid for a couple of days, but when I get back I'll try to find the information on the company. The prices were actually pretty reasonable. I thought it would be a great way to make a robot skin with simple tactile sensing switches in it.
-dave
Dave Curtis
-dave
Chuck McManis
A fairly creative technique might be to attach what are essentially crimp on pins on to the leads of the components, lay them down dead bug style, fill resin leaving just the pins exposed, and then wire wrap the result.
An extruded conductor though which could withsand soldering later would be killer. You might need a really low temp solder though.
--Chuck
Alan
For plated-through holes, I understand they use a carbon coating of some sort first, then plate with copper. Maybe find out how they do the plated-through holes, and apply the same process to lay out traces? Should be able to lay down the carbon powder; then later plate. Wouldn’t that be sharp!
Alan KM6VV
On Behalf Of Chuck McManis
Brandon Fosdick
> This is an area I've wondered about as well, sintering is not a good candidate if you are going to later solder on components, but if you build in the components this isn't an issue. I've seen systems which plate copper on to an insulator once the insulator has been patterned by a mask but I'm not sure what chemistry they use in that process.
Why is sintering incompatible with soldering? I'm assuming that if the resulting traces are copper or aluminum they should be just as solderable as etched traces. Am I missing something?
> A fairly creative technique might be to attach what are essentially crimp on pins on to the leads of the components, lay them down dead bug style, fill resin leaving just the pins exposed, and then wire wrap the result.
>
> An extruded conductor though which could withsand soldering later would be killer. You might need a really low temp solder though.
My grand scheme here is to produce a machine that pops out a completed board at the press of a button (sans components of course). Plating would be hard to integrate into a single machine that's small and cheap enough for home use.
Chuck McManis
On Sep 6, 2011, at 14:20 , Chuck McManis wrote:Why is sintering incompatible with soldering? I'm assuming that if the resulting traces are copper or aluminum they should be just as solderable as etched traces. Am I missing something?
> This is an area I've wondered about as well, sintering is not a good candidate if you are going to later solder on components, but if you build in the components this isn't an issue. I've seen systems which plate copper on to an insulator once the insulator has been patterned by a mask but I'm not sure what chemistry they use in that process.
> A fairly creative technique might be to attach what are essentially crimp on pins on to the leads of the components, lay them down dead bug style, fill resin leaving just the pins exposed, and then wire wrap the result.My grand scheme here is to produce a machine that pops out a completed board at the press of a button (sans components of course). Plating would be hard to integrate into a single machine that's small and cheap enough for home use.
>
> An extruded conductor though which could withsand soldering later would be killer. You might need a really low temp solder though.
Brandon Fosdick
> Three things I would investigate fully about sintering if I were taking that path;
> 1) Is the grain structure sufficiently large to maintain the conductivity that is expected or needed for traces?
> 2) If the sintered trace/feature is re-heated to say 400 degrees how does it react? (or not)
> 3) Can copper be sintered in an atmosphere with oxygen or does it require a non-oxygen environment like Al sputtering does?
This is going to require a lot of experimentation. Needing a nitrogen atmosphere wouldn't be too bad, but requiring something like argon could be a problem. It would be interesting to see how bad the sputtering is. Traces don't really need to be smooth or pretty, just conductive.
The layer thicknesses for FDM are a lot bigger than what appears to be typical for sintering, so there may only be one material layer for each circuit layer. If only the top of the trace oxides then it may still be conductive underneath. Hopefully the laser will be able to entirely melt the thicker layers. Without overheating the surrounding plastic of course.
I need to think about how to start experimenting without having to jump straight to a full prototype. Maybe if I could get ahold of some aluminum or copper powder I could take it to some place that has a laser engraver to see how well it sinters. I know a 20W laser engraver will cut aluminum foil, so melting should be possible at a lower setting.
> "cheap enough" will always be a challenge. One can robo-mill a PCB with a $3,500 CNC mill, or a $1,500 xy table and a dremel, but neither is low cost enough to make it into wide 'home' use. Most 3D printers are around $1000 and that keeps them out of most homes too :-). Not to be discouraging here but shooting for 'most cost effective' is always a win, evaluating based on whether or not you've achieved a mass market price point is kind of a set up for disappointment.
Maybe "home use" was the wrong terminology. Targeting the same market as the $1000 3D printers seems like a reasonable goal.
> I agree with you that plating is the hard step. I wonder if there is some sort of 'lost wax' equivalent where your 'conductor' material could be a chemical pre-cursor material that you later 'develop' by immersing the result in a bath of the reagent that reacts with the pre-cursor to become the conductor.
The whole "immersing in a bath" part is problematic. Vats of chemicals are exactly what I'm trying to avoid.
> Al's comment about using carbon is pretty spot on. If you could figure out how to deposit graphene as the the conductor you would do well. I always wonder given all the cool physics paper on various carbon structures if we won't soon read about carbon 'chips' which are fabricated by a 3D printing process rather than etched like the current generation. It would be slower (well at least electron beam lithography is slower than current mask based processes) but its only one material to work with rather than several.
I agree, graphene would be an excellent way to do this. If only I knew how. But, I imagine if I did know how, I wouldn't be sitting here right now. I'd be on some tropical beach enjoying my riches. :)
Brandon Fosdick
> I need to think about how to start experimenting without having to jump straight to a full prototype. Maybe if I could get ahold of some aluminum or copper powder I could take it to some place that has a laser engraver to see how well it sinters.
Amazon really does have everything...
http://www.amazon.com/Aluminum-Powder-micron-uncoated-ounces/dp/B004R80FW4
Peter N. Glaskowsky
I need to think about how to start experimenting without having to jump straight to a full prototype. Maybe if I could get ahold of some aluminum or copper powder I could take it to some place that has a laser engraver to see how well it sinters. I know a 20W laser engraver will cut aluminum foil, so melting should be possible at a lower setting.
Brandon Fosdick
> I really think you'll want to limit yourself to copper. Fine aluminum powder is already largely aluminum oxide, and even if you can sinter it, it will be very hard and brittle. It's a ceramic, after all.
That's a good point, I had forgotten about that. Although, if sintered aluminum oxide doesn't conduct, maybe it can be used to make very strong yet brittle circuit boards.
How hot does aluminum oxide need to get to drive off the oxygen? Maybe it can be sintered into pure aluminum? I imagine that would melt the surrounding plastic though.
I think both copper and aluminum need to be tried, just to find out exactly what happens to each. So far my limited googling hasn't turned up a convenient source of copper powder. I did find copper oxide at Alpha Chemicals, which probably won't work, but I imagine I need to try a wide variety of materials before settling on any one of them.
> Plus you will want to limit your exposure to aluminum powder for health reasons. Copper isn't exactly health food either, but it probably isn't as bad as aluminum.
I'm hoping a respirator and gloves are sufficient. Anything more elaborate could be a product killer.
> It occurs to me that you don't need very good results from sintering copper if you just run the board through a tin plating operation-- however porous the copper is to start with, it'll be filled in and reinforced by the tin, and you'll probably get decent results.
>
> If it turns out nobody has done this before, remember that I invented it. :-)
Well, I had thought of that too :)
But it looks like neither of us is the first. SLS uses powder coated in a binder that gets burned off to leave a porous material that's then infused with some other material. Infusing it with solder would just be another case of the same idea. Still not a bad idea though.
Despite the subject of my original email I've actually been focusing on DMLS instead of SLS because I want the finished traces to be solid. But SLS with solder infusion might be an excellent fall back solution if DMLS isn't feasible.
It would be interesting to see if the solder-infused traces would wick solder away while soldering components, or if the component pad/pin would draw the solder out of the trace.
Peter N. Glaskowsky
On Sep 7, 2011, at 2:22 PM, Brandon Fosdick wrote:
> On Sep 7, 2011, at 14:01 , Peter N. Glaskowsky wrote:
>> I really think you'll want to limit yourself to copper. Fine aluminum powder is already largely aluminum oxide, and even if you can sinter it, it will be very hard and brittle. It's a ceramic, after all.
>
> That's a good point, I had forgotten about that. Although, if sintered aluminum oxide doesn't conduct, maybe it can be used to make very strong yet brittle circuit boards.
Aluminum oxide in the conventional sense has long been used to make electronic substrates. I suppose if sintering produces comparable results, it can be used to make comparable substrates.
> How hot does aluminum oxide need to get to drive off the oxygen? Maybe it can be sintered into pure aluminum? I imagine that would melt the surrounding plastic though.
As you probably know, this is the basic problem in aluminum refining, which is not exactly energy-efficient. :-)
With pulsed lasers, you can heat the surface of the target so briefly that the heat doesn't penetrate any further than you want it to.
> I think both copper and aluminum need to be tried, just to find out exactly what happens to each. So far my limited googling hasn't turned up a convenient source of copper powder. I did find copper oxide at Alpha Chemicals, which probably won't work, but I imagine I need to try a wide variety of materials before settling on any one of them.
Alfa Aesar offers pure copper powder, but it has to be packed under argon so that it isn't all copper oxide by the time it arrives. I don't know how long you'd have to pattern and sinter it, but I don't think copper is as greedy for oxygen as aluminum is, so here again copper is more likely to be the better choice.
. png
°|° Walt Perko °|°
Maybe we need a penny grinder since a real copper penny is worth a nickle
now days ???
===============================================================================================
C U L8r, �|� Walt Perko �|� "Kids ... teach them the good stuff,
and they still learn the bad stuff."
A robot that doesn't think ... but just teaches ... the BallBot1!
http://www.Brainless.org/R2Pv1.net/BallBot1.html
RoboToons weekly Robot Cartoons; http://www.brainless.org/RoboToons/
"The World Needs a New Economic Model"
----- Original Message -----
From: "Brandon Fosdick" <bfos...@gmail.com>
To: <hbrob...@googlegroups.com>
Sent: Wednesday, September 07, 2011 1:49 PM
Subject: Re: [HBRobotics] FDM + SLS = PCB?
George Warner
>
> I agree, graphene would be an excellent way to do this. If only I knew how. But, I imagine if I did know how, I wouldn't be sitting here right now. I'd be on some tropical beach enjoying my riches. :)
Note: grapheme is a non-conductor; you'd have to dope it with Nitrogen or Boron to have extra or missing electrons.
As long as we have to assemble these one atom at a time we could go 3D and make (doped) diamonds. ;-)
--
Enjoy,
George Warner,
eMail/iChat: <geo...@mac.com>
Randy M. Dumse
> As long as we have to assemble these one atom at a time we
> could go 3D and make (doped) diamonds. ;-)
Stayed out of the discussion to this point. Just had a thought.
You know the amalgem dentists use? Silver, mercry and something?
Sure sounds like a moldable conductor that sets after placement
to me. Some chemist will probably get rich when the need and the
existing metalurgical technology get together.
Randy
Chris Palmer
Might have some legal issues with that grinder. ;-)
-Chris
Hi,
Maybe we need a penny grinder since a real copper penny is worth a nickle
now days ???