Texas Tightwad Electronics

I started working on this because I think it's unreasonable to pay $200 for what is effectively an 8 bit computer running at 16mhz. Using Tramflug's Gen7 design as a starting point, I put together a schematic that uses 100% smd components. I had intended for the entire board to cost about $40 in raw materials, but it's starting to look like 50 or 60 is more achievable for batches of one. Right now I could hit the $37.5 price point in batches of 100, and with a 50% profit, a retail price of $75 for a finished board seems practical. In theory the reprap should be able to make the pcb, and either spread the solder paste directly or print a stencil. Then a person populates the board with components and pops the whole thing into the toaster oven... 10 minutes later the board is completely finished.

I came up with a basic BOM from mouser:
[www.mouser.com]

And a schematic in gschem:
[github.com]

It's been awhile since I took those circuits classes in college, so if I missed anything please point it out. I think the data bus and I/O part of the circuit is good, but I may need some help sizing on the passive components. Also my only direct experience with a reprap is seeing one at the Atlanta, GA hacker space for about 10 mintues, so if anyone sees anything blatantly wrong with the interface feel free to point that out too. Because of the components available for smd mounting, I can't find a wire connector that can handle more than 3A for the heated bed. Any suggestions?

If the user is going to want to pare down the expense to the minimum, then it might not be unreasonable to just have a set of largish surface pads to solder wires to. Or maybe a few through hole pads just for the high current connections. It might still be a good idea to have a few standard ATX power connectors as used by most for the high power supply, and maybe even some through hole stuff for motor connections. Not sure how that would affect the cost considering higher board space etc. , but it would help a lot when the power supply fails, and for troubleshooting motor problems. Cool project though. The electronics are becoming the highest cost item in the build now that RP parts are coming down in price.

I'm thinking direct solder pads are the best way to connect the power supply if I'm trying to go 100% SMD. If I end up having to drill some holes anyway, you're right going with the 4-pin molex connector from a PC power supply would be best. Even though it can only supply a max of 16 amps a 12V, that should still be plenty for the application. I've never had a computer power supply fail that was unrepairable. The biggest issue I've ever had with one is bad caps or a bad fan.

I finished the pcb layout and uploaded everything to github, so I'm ready for beta testers. It may be awhile before I build one myself and test it since I still have to get the rest of my reprap parts together. If there is a more windows friendly format I could convert the files too, I would be happy to do so.

Being a tightwad myself I would offer to test them, but I have never soldered surface mount stuff before, and I just built a couple gen7 boards, so my electronics budget is shot for a while.
Plus my motors are currently 2A so I don't know if your electronics would drive them with sufficient torque, especially the two ganged Z steppers on the Prusa. I'm having trouble even with the pololus (with heatsink) on Z, I think I'll have to buy lower current motors to get my Z axis to work well.
Great idea though. I'm thinking of doing a version of Gen7 and/or Sanguinololu with the through hole Toshiba drivers AKA47 is using. That should bring the cost down by a good $20 or $30 vs pololu driver modules. That way the SMT averse crowd will have a similar solution. That is way down the road though at the rate I'm going

If you are struggling for current on Z you can wire the two motors in series instead of parallel, but you will sacrifice some speed.

---

[www.hydraraptor.blogspot.com]

OK, I really should have thought of that. Thanks