Occam
Posted by nophead
No not the computer language of the Transputer but a new way of building circuits with encapsulation, laser drilling and copper plating :- [www.verdantelectronics.com]
[www.hydraraptor.blogspot.com]
[www.hydraraptor.blogspot.com]
|
Re: Occam February 05, 2008 03:46AM |
Registered: 16 years ago Posts: 58 |
It not really very "new" way, it's just reusing of "Gate array" approach, widely used in IC manufacturing. Read more at [en.wikipedia.org]
... we had some discussions about extruding conducting paste as vias and/or conducting trays and embedding components in single- or multilayer-compounds.
I'm playing around with my CNC-system to get a combination between PCB-milling and -drilling and extruding glues, globetops and conducting trays ...
When inserting IC's and passive and active components in freshly extruded and/or drilled sheets and 'painting' the trays with paste, we'll have the first step to 'selfreplicating electronics'
Viktor
I'm playing around with my CNC-system to get a combination between PCB-milling and -drilling and extruding glues, globetops and conducting trays ...
When inserting IC's and passive and active components in freshly extruded and/or drilled sheets and 'painting' the trays with paste, we'll have the first step to 'selfreplicating electronics'
Viktor
|
Re: Occam February 05, 2008 01:03PM |
Registered: 16 years ago Posts: 161 |
I had a quick read of the white paper, i didn't read it all but most of it seemed to be presenting a business case for it, rather than explaining much technical detail. Most of the technical info is just about concepts rather than actual processes, but hopefully they will do some good work to develop this technology, they may just be a bit paranoid at the moment.
This is definitely something that it would be great to see reprap doing in the future, presumably circuity can be made much more compact too when making 3d circuits in this way.
This is definitely something that it would be great to see reprap doing in the future, presumably circuity can be made much more compact too when making 3d circuits in this way.
|
getting it to work with reprap February 05, 2008 05:15PM |
This should work very well with reprap. It might be possible to extrude a thermoplastic base with holes where the components go and put down low melting point alloy trenches connecting them on an upper layer. It also might be possible to put uninsulated copper wire in the trenches and use low melting point allow to solder them to the components.
This definitely has some great potential, especially considering that circuitry made through this method would be naturally water proof. The only problem is getting this to work with surface mount components(which reprap doesn't have) and components that need to radiate heat.
This definitely has some great potential, especially considering that circuitry made through this method would be naturally water proof. The only problem is getting this to work with surface mount components(which reprap doesn't have) and components that need to radiate heat.
Hi gene hacker,
with 0,1 accuracy you can handle and embed SMD's in normal position or turned upside-down too, so this is no real limitation.
For heat-dissipation you can dispense conducting (or only heat-conducting) paste from the radiation-surfaces to the outside of the compound and attach (or reprap) maze-structured heatsinks ...
Viktor
with 0,1 accuracy you can handle and embed SMD's in normal position or turned upside-down too, so this is no real limitation.
For heat-dissipation you can dispense conducting (or only heat-conducting) paste from the radiation-surfaces to the outside of the compound and attach (or reprap) maze-structured heatsinks ...
Viktor
|
Re: Occam February 06, 2008 07:26AM |
Admin Registered: 17 years ago Posts: 1,915 |
Using solder for the interconnect will only work for a subset of circuits because it is ten times less conductive than copper.
The advantage I saw in the Occam process is that it uses copper plating directly onto the component leg so its performance is actually better than a PCB rather than worse.
It also seems on the face of it very simple:-
Place the components accurately legs down in a tray with something sticky in the base.
Pour a layer of some plastic that will flow and then set hard. It only needs to be thick enough to hold the parts, it doesn't necessarily need to cover the parts as far as I can see.
Remove it from the tray and support it accurately upside down.
Drill down to the component legs to reach the copper. A laser is suggested but why not just use an NC drill for our purposes?
Copper plate, I don't know exactly how but presumably not electroplating.
Add more connection layers by pouring on more plastic, drilling and plating.
It seems easier than some of the other suggestions and it doesn't compromise electrical performance.
[www.hydraraptor.blogspot.com]
The advantage I saw in the Occam process is that it uses copper plating directly onto the component leg so its performance is actually better than a PCB rather than worse.
It also seems on the face of it very simple:-
Place the components accurately legs down in a tray with something sticky in the base.
Pour a layer of some plastic that will flow and then set hard. It only needs to be thick enough to hold the parts, it doesn't necessarily need to cover the parts as far as I can see.
Remove it from the tray and support it accurately upside down.
Drill down to the component legs to reach the copper. A laser is suggested but why not just use an NC drill for our purposes?
Copper plate, I don't know exactly how but presumably not electroplating.
Add more connection layers by pouring on more plastic, drilling and plating.
It seems easier than some of the other suggestions and it doesn't compromise electrical performance.
[www.hydraraptor.blogspot.com]
|
Re: Occam February 06, 2008 10:16AM |
Registered: 16 years ago Posts: 622 |
I think its OK to do it at home as long as you don't exploit it commercially. Same with FDM which is also patented but due to run out soon.
[www.hydraraptor.blogspot.com]
[www.hydraraptor.blogspot.com]
|
mccoyn
Re: Occam February 06, 2008 07:52PM |
Nah. You can use chemistry to deposit a layer of metal on something.
An alternate approach to consider.
Instead of using Surface Mount devices, use the more traditional leaded variety.
Mount them back down, with the legs up, on flypaper, or a fitted tray, then pour your epoxy to cover the components, but not the leads. Shear off the leads where they protrude from the epoxy, wash, then plate the surface of the epoxy, (and the surface of the leads still visible as dots, or even very short risers.) Before plating, you could draw on the surface with a crayon equipped plotter to prevent those areas from being plated, or grind away the surface in much the way McMaster did with his original mill.
To add a second layer, do either of the following, (or both, but not recommended, for the sake of simplicity.)
A. Don't clip all the leads so short they'd be covered by a second layer of epoxy. Use the ones you don't clip short AS your risers. The drawback is now you need to be able to control how high an individual lead gets trimmed; you can't just mow them down to a common height.
B. Stick on risers between the plating stage and the second epoxy stage. The drawback here is now you need a separate tool to stick the risers in. Basically a wirefeed mechanism that can clip off a 1/4" length of wire after sticking it in mostly cured epoxy.
This way, you get the benefits of the described system, without having to actually drill so accurately as to touch a component without modifying either its resistance or its capacitance. You would get slightly larger finished products, however.
One thing to consider. If you can use pick-n-place capabilities to place your components, you can either use the same mechanism, with the addition of a wire-feed spool, to add risers, or you can use the same mechanism, with the addition of a pair of end cutting pliers beside the gripper tool.
Pouring the epoxy should be done from the side. Pour it so it covers the surface of the electronics like heavy objects left in a pool or tub when the water is turned on. The exposed surface should stay dry right up until the epoxy covers it.
Just an idea.
P.S. Depending on how soft the epoxy is, and how pliable the wire is, it should be possible to stick the protruding end of a wire off a spool into the soft epoxy, pull it out, (unwinding it, but not using a separate wire-feed motor,) and clip off the length that protrudes past the end of the clippers, (using the same gripper motor that places the circuits.)
The drawback of using the same motor for both operations is you eventually have to throw in another motor to raise the unused tool, so it doesn't interfere, or you're limited on how close you can pack the circuits; you wouldn't be able to place a circuit in a location that would interfere with depositing a riser, or you wouldn't be able to leave certain risers, as they intersected with the location of the grippers when the clippers were cutting a lead short elsewhere.
It might be possible to use your clippers as grippers, but I'd prefer to use a plunger mechanism for through-leaded components, as it'd orient the wires for us.
Take a piece of L-shaped channel, and cut a slot large enough to push a resistor through sideways, (if its leads were cut.) Attach your spool so it lines the components up with this hole. Now drop the mechanism to the board, and push a plunger through the slot, forcing the component out onto the board, (flypaper,) while also forcing the leads to bend upward to clear the channel. You'd still need some other mechanism for can-type capacitors, transistors, and DIP type components. Perhaps pliers that can be controlled to grip these, while still being strong enough to cut leads.
Then add a "wrist" actuator so you can remove one from play while using the other, (or a tool-change mechanism, which would actually be better.)
Edit 2, typo.
Edited 2 time(s). Last edit at 02/08/2008 06:12PM by Sean Roach.
An alternate approach to consider.
Instead of using Surface Mount devices, use the more traditional leaded variety.
Mount them back down, with the legs up, on flypaper, or a fitted tray, then pour your epoxy to cover the components, but not the leads. Shear off the leads where they protrude from the epoxy, wash, then plate the surface of the epoxy, (and the surface of the leads still visible as dots, or even very short risers.) Before plating, you could draw on the surface with a crayon equipped plotter to prevent those areas from being plated, or grind away the surface in much the way McMaster did with his original mill.
To add a second layer, do either of the following, (or both, but not recommended, for the sake of simplicity.)
A. Don't clip all the leads so short they'd be covered by a second layer of epoxy. Use the ones you don't clip short AS your risers. The drawback is now you need to be able to control how high an individual lead gets trimmed; you can't just mow them down to a common height.
B. Stick on risers between the plating stage and the second epoxy stage. The drawback here is now you need a separate tool to stick the risers in. Basically a wirefeed mechanism that can clip off a 1/4" length of wire after sticking it in mostly cured epoxy.
This way, you get the benefits of the described system, without having to actually drill so accurately as to touch a component without modifying either its resistance or its capacitance. You would get slightly larger finished products, however.
One thing to consider. If you can use pick-n-place capabilities to place your components, you can either use the same mechanism, with the addition of a wire-feed spool, to add risers, or you can use the same mechanism, with the addition of a pair of end cutting pliers beside the gripper tool.
Pouring the epoxy should be done from the side. Pour it so it covers the surface of the electronics like heavy objects left in a pool or tub when the water is turned on. The exposed surface should stay dry right up until the epoxy covers it.
Just an idea.
P.S. Depending on how soft the epoxy is, and how pliable the wire is, it should be possible to stick the protruding end of a wire off a spool into the soft epoxy, pull it out, (unwinding it, but not using a separate wire-feed motor,) and clip off the length that protrudes past the end of the clippers, (using the same gripper motor that places the circuits.)
The drawback of using the same motor for both operations is you eventually have to throw in another motor to raise the unused tool, so it doesn't interfere, or you're limited on how close you can pack the circuits; you wouldn't be able to place a circuit in a location that would interfere with depositing a riser, or you wouldn't be able to leave certain risers, as they intersected with the location of the grippers when the clippers were cutting a lead short elsewhere.
It might be possible to use your clippers as grippers, but I'd prefer to use a plunger mechanism for through-leaded components, as it'd orient the wires for us.
Take a piece of L-shaped channel, and cut a slot large enough to push a resistor through sideways, (if its leads were cut.) Attach your spool so it lines the components up with this hole. Now drop the mechanism to the board, and push a plunger through the slot, forcing the component out onto the board, (flypaper,) while also forcing the leads to bend upward to clear the channel. You'd still need some other mechanism for can-type capacitors, transistors, and DIP type components. Perhaps pliers that can be controlled to grip these, while still being strong enough to cut leads.
Then add a "wrist" actuator so you can remove one from play while using the other, (or a tool-change mechanism, which would actually be better.)
Edit 2, typo.
Edited 2 time(s). Last edit at 02/08/2008 06:12PM by Sean Roach.
Just a few semi-skeptical notes:
First, here's a page about creating PCBs via copper plating:
[www.thinktink.com]
and here is the recipe for the plating solution, which is a fairly standard version:
[www.thinktink.com]
Now my reading suggests this is a not uncomplicated process involving conductive inks and large finicky designed tanks and that all the paths must be temporarily shorted together for plating to occur. More importantly the plating solution is a nasty little bugger that you wouldn't want around the house. This solution will burn you in the time it takes to find a rag to wipe it off. And it's poisonous besides. This particular solution contains some unspecified organic conditioners of the hazmat variety. This is not a process for the masses.
Second, embedding the components in plastic will certainly improve durability and perhaps even keep the nasty acid off, it will also have a drawback which is rather cavalierly dismissed in the whitepaper, namely heat. While heat dissipater can certainly be built in there will be a cost in simplicity; also many components that do not normally require cooling, such as resistors and capacitors may well begin to build up heat in amounts that effect circuit operations when buried in insulation. Supplying every resistor with a heat sink could become bothersome.
Last, I'm not sure about the effects of charging up transistors, PICs and other ICs with fairly high positive voltages during the plating process. I understand that a uniform potential throughout a circuit is effectively zero but how will that translate over semiconductor gates and what counts as "uniform" potential on the scale inside an IC?
Anyway, I think the process has possibilities - the upside down build process has possibilities although I would embed the components in soluble support material until it was time to lay down a plastic layer with inset channels for the paths, which are then filled with some sort of solder type metal. Are there low resistance solders? Further research I guess, in my copious free time ;-)
First, here's a page about creating PCBs via copper plating:
[www.thinktink.com]
and here is the recipe for the plating solution, which is a fairly standard version:
[www.thinktink.com]
Now my reading suggests this is a not uncomplicated process involving conductive inks and large finicky designed tanks and that all the paths must be temporarily shorted together for plating to occur. More importantly the plating solution is a nasty little bugger that you wouldn't want around the house. This solution will burn you in the time it takes to find a rag to wipe it off. And it's poisonous besides. This particular solution contains some unspecified organic conditioners of the hazmat variety. This is not a process for the masses.
Second, embedding the components in plastic will certainly improve durability and perhaps even keep the nasty acid off, it will also have a drawback which is rather cavalierly dismissed in the whitepaper, namely heat. While heat dissipater can certainly be built in there will be a cost in simplicity; also many components that do not normally require cooling, such as resistors and capacitors may well begin to build up heat in amounts that effect circuit operations when buried in insulation. Supplying every resistor with a heat sink could become bothersome.
Last, I'm not sure about the effects of charging up transistors, PICs and other ICs with fairly high positive voltages during the plating process. I understand that a uniform potential throughout a circuit is effectively zero but how will that translate over semiconductor gates and what counts as "uniform" potential on the scale inside an IC?
Anyway, I think the process has possibilities - the upside down build process has possibilities although I would embed the components in soluble support material until it was time to lay down a plastic layer with inset channels for the paths, which are then filled with some sort of solder type metal. Are there low resistance solders? Further research I guess, in my copious free time ;-)
Are there any suitable epoxies? I'm thinking if you can silver epoxy in the same manner as glass, (it's important to remember glass is hydrophilic, so I don't know if you can,) you can use straight silver. As you're going to coat it immediately afterward, tarnish shouldn't be an issue.
[www.alanmacfarlane.com]
As for heat conductance. Yes, that's a likely problem. Whatever the components are place on, or in, needs to be reasonably heat-conductive, preferably while also being an electrical insulator.
I think I'll stand by my flypaper suggestion, but modify it to double stick film on aluminum.
[www.alanmacfarlane.com]
As for heat conductance. Yes, that's a likely problem. Whatever the components are place on, or in, needs to be reasonably heat-conductive, preferably while also being an electrical insulator.
I think I'll stand by my flypaper suggestion, but modify it to double stick film on aluminum.
One idea that I've been kicking around for a while is to use a two-part paste that, when combined, deposits copper. One component would be a copper compound (e.g. cupric oxide, cupric sulfate, blue vitriol). The other would be a strong reducing agent (e.g. ascorbic acid).
In theory, you should be able to just mix the two in the right proportions and get a nice copper deposit. In practice, you'd probably have to fiddle with temperature and PH to get the desired results.
In theory, you should be able to just mix the two in the right proportions and get a nice copper deposit. In practice, you'd probably have to fiddle with temperature and PH to get the desired results.
Some things to keep in mind:
Conduction is largely a surface effect. Sure the bulk conducts, but electrons like to travel the most at the surface.
Signal and power are 2 different things. Signal doesn't require the high conductivity you might want with power (to drive a motor for instance) eutectic metals are probably fine for sending signal maybe even up to driving an LED. Keeping power and signal apart during design means only needing the high conductivity in those areas required to power something.
Light, infrared, radio waves, static charge and mechanical strain can all be used to transmit signal so it isn't always necessary to use conductors for this purpose.
Duration, frequency, voltage and distance all affect conductance properties. A single high voltage high frequency pulse will travel with far less loss than a standing DC charge.
There are many new organic polymer conductors coming near production. Fullerene doped polymers are nearing the conductance range of metals.
Edited 1 time(s). Last edit at 02/09/2008 01:38PM by The Guy.
The Guy
Conduction is largely a surface effect. Sure the bulk conducts, but electrons like to travel the most at the surface.
Signal and power are 2 different things. Signal doesn't require the high conductivity you might want with power (to drive a motor for instance) eutectic metals are probably fine for sending signal maybe even up to driving an LED. Keeping power and signal apart during design means only needing the high conductivity in those areas required to power something.
Light, infrared, radio waves, static charge and mechanical strain can all be used to transmit signal so it isn't always necessary to use conductors for this purpose.
Duration, frequency, voltage and distance all affect conductance properties. A single high voltage high frequency pulse will travel with far less loss than a standing DC charge.
There are many new organic polymer conductors coming near production. Fullerene doped polymers are nearing the conductance range of metals.
Edited 1 time(s). Last edit at 02/09/2008 01:38PM by The Guy.
The Guy
I like the signal vs power idea Guy. Here's an idea for a two mode conductor toolhead:
Signal mode: The head positions itself over a connection point (component lead); the heating element moves down and across to engage the lead, after a moment for the lead to heat up a motor feeds a solder wire onto the heating element, after a few moments the land is filled and the head begins to travel along the channel that defines the conductor path leaving a trail of solder behind as the conductor (the channel may need to be sprayed with flux for the solder to adhere). When another component lead is met the wire feed stops, the heater moves up and back down as the toolhead positions it to engage the new lead and the process resumes.
Power mode: The head positions itself over a connection point (component lead); a motor feeds a copper wire (say from the right) into position against the lead, the heating element moves down and across (say from the left) to engage the lead pressing it down onto the copper wire (or tape), after a moment for the lead and wire to heat up a motor feeds a solder wire onto the heating element, after a few moments the land is filled, the heating element (and solder) withdraw and a presser foot descends to ensure the wire sets into the conductor channel; the head begins to travel along the channel that defines the conductor path leaving a wire behind as the conductor. When another component lead is met the toolhead circles the lead, the wire feed stops, the presser foot rises, the heating element moves back down to engage the new lead. Once the land is filled the heater moves up and the wire is snipped or the presser foot returns and the process resumes. Once all the wires are connected a new layer of plastic is laid down to keep them in place.
Advantages of signal mode - faultless connections, ease of recycling, no need for cover layer, decreased material count
Disadvantages of signal mode - cost of solder vs copper? difficulties laying solder traces.
Advantages of power mode - higher watt/volt capacity, ruggedness?
Signal mode: The head positions itself over a connection point (component lead); the heating element moves down and across to engage the lead, after a moment for the lead to heat up a motor feeds a solder wire onto the heating element, after a few moments the land is filled and the head begins to travel along the channel that defines the conductor path leaving a trail of solder behind as the conductor (the channel may need to be sprayed with flux for the solder to adhere). When another component lead is met the wire feed stops, the heater moves up and back down as the toolhead positions it to engage the new lead and the process resumes.
Power mode: The head positions itself over a connection point (component lead); a motor feeds a copper wire (say from the right) into position against the lead, the heating element moves down and across (say from the left) to engage the lead pressing it down onto the copper wire (or tape), after a moment for the lead and wire to heat up a motor feeds a solder wire onto the heating element, after a few moments the land is filled, the heating element (and solder) withdraw and a presser foot descends to ensure the wire sets into the conductor channel; the head begins to travel along the channel that defines the conductor path leaving a wire behind as the conductor. When another component lead is met the toolhead circles the lead, the wire feed stops, the presser foot rises, the heating element moves back down to engage the new lead. Once the land is filled the heater moves up and the wire is snipped or the presser foot returns and the process resumes. Once all the wires are connected a new layer of plastic is laid down to keep them in place.
Advantages of signal mode - faultless connections, ease of recycling, no need for cover layer, decreased material count
Disadvantages of signal mode - cost of solder vs copper? difficulties laying solder traces.
Advantages of power mode - higher watt/volt capacity, ruggedness?
Here's another related article about "Solder made obsolete by growing copper pillars" which looks to play right into a lot of the electro-deposition stuff Adrian has been writing about too:
[www.eetimes.com]
The Guy
[www.eetimes.com]
The Guy
Sorry, only registered users may post in this forum.