Reprap style SMD options

nophead Wrote:
-------------------------------------------------------
> I think we have to aim for BGA and QFN because
> that is the way the market is rapidly going.

And the only reason to pursue this in the context of reprap is planning for the future.

I do disagree, though, unless we're building something outside the realm of microcontrollers, QFP has a lot of life left in it.

This thread seems to have run through the prickle bushes, but it raises some good points.

I think that unless you are only printing other people's parts to get your enjoyment, then a builder needs to be unusually competent to be able to take advantage of the technology.

Where the confusion creeps in, is that some of those unusually competent people are gifted with electronics and/or software as well. Some, are gifted in mechanical aptitude as well.

The issue, is that Adrian Bowyers self replication goal (of the machines) is really muddied by the typical absence of some of those skills, in most of the population, and even in many of the users.

Circuitboard production is a specialised task, and for economy, most outsource at least some of the job.

It's an admirable goal to have a machine that can replicate 100% of itself, but in reality, most of the "vitamins" are outside the scope of most people to personally make, or even to assemble.

Threaded rods, silicon chips, acrylic sheet, all very exotic. Some you get from the hardware store, some from an electronics distributor.

Within the project itself, it's nice to know how the parts function, but an outsider would just want to flick the switch, and select product.

-Self reproduction is not viable in a simple machine
-A complex machine may be able to reproduce a wider variety of it's own components, but at that cost of complexity. Ultimately, it would need material refinement capabilities, acquisition capabilities (hunter-gatherer) etc, etc.

~The reprap machine is a simple machine.

Availability goals are reached vis cost reduction and knowledge.
Performance goals are achieved via sophistication.

For a "product" to deliver on all points, standardisation on a cheap and yet sophisticated design allows the benefits of mass production to kick in. That's not a task reprap was ever designed for. It's a build machine in essence, and takes too much of an economy hit if it's used in ways other than additive or subtractive (deposition/milling) manufacture. That's not to say that a reprap PCB couldn't be reprogrammed to drive a different type of machine though !

grael Wrote:
-------------------------------------------------------
>
> -Self reproduction is not viable in a simple
> machine
> -A complex machine may be able to reproduce a
> wider variety of it's own components, but at that
> cost of complexity. Ultimately, it would need
> material refinement capabilities, acquisition
> capabilities (hunter-gatherer) etc, etc.
>
> ~The reprap machine is a simple machine.
>
So basically, you appear to reject the core premise of the reprap project as impractical. Mind, many share that view.

---

-------------------------------------------------------

Hell, there are no rules here - we're trying to accomplish something.

Opportunity is missed by most people because it is dressed in overalls and looks like work.

Thomas A. Edison

I accept the core premise of reprap. I accept that many people will not want to build circuit boards. I simply believe that the platform should be built with extensibility in mind. I think one very important extension is electronics manufacturing.

That's it. That's all there is to it.

LOL! I know what you think, Annirak. I was actually wondering if Grael felt the same way.

---

-------------------------------------------------------

Hell, there are no rules here - we're trying to accomplish something.

Opportunity is missed by most people because it is dressed in overalls and looks like work.

Thomas A. Edison

A lot has already been achieved with reprap, but yes, if you put it like that Forest ("core premise as being impractical"), then I have to agree. We(humans) are not technically advanced enough (yet) to be to making Von Neuman machines.

As such, there is a decision to me made about how much of a reprap should be self build, and how much should be mass produced. I think the term used in relation to the self built parts was "critical mass", in relation to other, already built units.

Then, you have to look at product lifetimes, and how many others are built for each starter machine. It's still very much a niche item, as the level of complexity of a start to finish operation would be beyond most people's grasp.

As far as I know, Forest is getting the RRF PCBs mass produced.
Why not have them also mass produced with the SMD components already fitted ?
(assuming that some of us here like SMD functionality )
One of our local electronic retailers does that for boards with fine pitch SMD, the kitset has those parts pre soldered.

I also think there is a place for totally through hole component machines, for those who want to make soldering part of the "reprap adventure". Through hole components are still around, but they are no longer a priority for suppliers or manufacturers these days.

And, if you do see reprap as needing to incorporate the soldering adventure as well as the other challenges, then it's certainly easier for a constructor to use distributed intelligence, as the individual package pin count requirement is reduced.

I have the luxury of a fresh start, compared to many of you, so of course I want the ideal entry level machine for my own purposes. I can understand for those with existing investment though, that leveraging existing boards seems like an attractive option... hence talk on the forums about Micros with more RAM, code space, same pin counts, etc. For them, maybe a CPLD with function transparency on some existing pins, special pin combination, and you select the CPLD's extra ports. Special combination again, and you are back to the transparent (normal functions. CPLDS are available in PLCC, so can be through hole socketed.

I want the luxuries afforded by a big chip though, so I'm going SMD.

Graham.

grael Wrote:
-------------------------------------------------------
>
> As far as I know, Forest is getting the RRF PCBs
> mass produced.
>
I think you meant to say Zach, not me. I'm committed to DIP self-assemble with printed circuit boards that Tommelise 2.0 can mill itself. That's an interim solution, mind. I'm all for the idea of repraps printing their own circuit boards. I just think that it is a few years off still.

---

-------------------------------------------------------

Hell, there are no rules here - we're trying to accomplish something.

Opportunity is missed by most people because it is dressed in overalls and looks like work.

Thomas A. Edison

Whoops... Yes, I should have written "Zach" there.
sorry !

Graham.

Edited 1 time(s). Last edit at 01/30/2009 02:48PM by grael.

Here's a few random ideas I've had about the smt process while reading this thread:

1) use some printed plastic traces for board alignment. Shouldn't be too hard. The board can be cut in an asymmetrical fashion to prevent placing upside down.

2) Have a very simple single tape reel system for pick and place. This could just be a reel on one of the solarbotics motors currently used for the extruder. Then cut and splice the tapes from the different parts together in the right order so that the reprap can just go to the same spot for each part and just keep track of which part # it's ready for.

3) We could possibly make a simple suction system using a solenoid on a rubber tube. It would take some fiddling with the sealing to get it to work, but you just apply voltage, and the solenoid creates a vacuum and lifts the part.

4) Since interest has already been shown in some sort of heated table for making the plastic stick better and not warp as much, maybe a slightly more sophisticated table could be used for heating the board and melting paste. I don't know if we could make it so that only a controlled portion of the build surface heats up. i.e. a cross thatch network of nicrome coils so that you can heat up individual squares of the build surface. That would reduce the chance of burns, and also reduce wasted energy. Unfortunately, using only a heated table limits the pcbs to single side.

5) Maybe such a table could be used to pre-heat instead of actually solder. In this mode, it would heat the pcb to near-soldering temperature, and the parts would be temporarily pushed over the edge by a heat gun. (nichrome wire + small fan in tube?) This could be done at a temperature that might allow for two sided boards. Of course, the heated surface may not be necessary at all.

I personally like what this debate has shown about the reprap community. It demonstrates that most of us can handle opposition with reasonable courtesy. It also demonstrates that the community is interested in improving the reprap. I understood the purpose of the reprap's reprapability to be twofold. a) decrease costs so that people in africa can use it. b) increase it's own rate of evolution. If a reprap can build itself, people can use it to make improvements. What I see in this debate is that different people want the reprap to go in different directions. Great! That increases the likelihood of beneficial mutations, and having people who like to work on reproducibility and people who like to work on more features is a bonus. Having only one or the other would be a drag, and limit its possibilities. We should (hopefully) see the reprap as a community, not as a single device or implementation. As soon as there were more than one in the wild, there were two different people with two different opinions on how they wanted theirs to work. Yes, there will probably always be a "canonical" reprap. It should be made as easy as possible for new people to use and modify, and hopefully reproduce. But it should also have the ability to work with features that the fringe have developed for personal use. If we can get a decently modular design, then people like Forrest can focus on replacing current tech with more reproducible equivalents, and people like Annirak can focus on the cutting edge. However, it is probably more productive if we actually do that, as opposed to debating about whether or not Forrest should be working on the things Annirak wants to work on, or vice versa.

I think that reprap will probably turn out a lot like Linux: A complicated core managed by a few dedicated individuals that is supported by a vast community of people, some of which work on making it accessible to people who don't want to worry about complexities, and some who are interested in configuring it exactly as they like it or taking it the next step. They are (generally) helpful when you want to configure yours in a different way, and everyone is (mostly) free to fork the project and make their own version. That's what open source is all about.

I think I understand Samuel:

___reprap___
| |
Repbuntu Uberrap

You might want to talk to Adrian Bowyer about Cannonical though, the applications are not yet appearing in the default Ubuntu repositories yet

Samuel Wrote:
-------------------------------------------------------
> Here's a few random ideas I've had about the smt
> process while reading this thread:
>
> 1) use some printed plastic traces for board
> alignment. Shouldn't be too hard. The board can be
> cut in an asymmetrical fashion to prevent placing
> upside down.
>
> 2) Have a very simple single tape reel system for
> pick and place. This could just be a reel on one
> of the solarbotics motors currently used for the
> extruder. Then cut and splice the tapes from the
> different parts together in the right order so
> that the reprap can just go to the same spot for
> each part and just keep track of which part # it's
> ready for.
>
> 3) We could possibly make a simple suction system
> using a solenoid on a rubber tube. It would take
> some fiddling with the sealing to get it to work,
> but you just apply voltage, and the solenoid
> creates a vacuum and lifts the part.

This is somewhat beyond my skill set, so bear with me if my idea is unworkable, but couldn't the whole pick and place be avoided by the use of the reprapped plastic template you talked about on #1?

I've been thinking about this same idea. Using the Gerbers or other files, maybe one can convert the files to Gcode for a paste extruder to put down solder paste for the SMTs. Maybe the file could also be modified to create a "mask" for the SMT parts that need to be installed. The mask can be used to reprap a plastic template with indexing pins to lock it in the proper position on the board. The mask wouldn't need to be much more than a raft with indexing pins and holes that correspond with the locations of SMT components. The reprapper can then remove the board and mask template from the reprap, and on their workbench, put the template over the board. Hand install the parts through the alignment holes, and use the "toaster" method to bake the board. I would suspect that one would need to carefully remove the template just before cooking, but wouldn't the paste keep the components in place?

The template could then be reused to make multiple copies of the board.

What do you all think? Is this lunacy?