> I can see a need for a very cheap Milkymist device. Right now, > selling the Milkymist One is difficult to two groups of people: > 1) open source enthusiasts and developers, who do not want to spend > much money on a tech product (unless there is a Apple logo on it).
Indeed almost same feedback i had receive about this > 2) VJs. While the Milkymist One sure is competitive when taking into > account the cost of the proprietary software they use at the moment, > they have a strong culture of using this software - which they > sometimes jump into using a pirated copy. The Apple logo also seems > to help here. I actually dont get the idea about this argument > > I think we should keep promoting and improving the Milkymist One > (more media exposure, development of RC4 hardware features such as > DVI/HDMI compatibility, better software). But I also believe we > should work on lowering one of the main barriers to entry into our > project. > > How about trying to make a $99 minimal version? Here is how I see it: > > * Artix-7 XC7A30T FPGA - most probably cheaper than the one we have > at the moment, and even a tad faster. > * HDMI only video output. Direct (read "cheaper") connection to the > FPGA without DAC and people keep asking for it (even though their > projectors are generally VGA, go figure). So that would kill two > birds with one stone. > * SDRAM/Flash: to avoid development and maintainance headaches, > let's try to stay compatible with what we have on the current board, > unless big savings can be made. Software would be the same except > the FPGA design which would only undergo minor modifications to > target the Artix-7 instead of the Spartan-6. > * Memory card. Updates are via files on the memory card. > * Video-in based on the ADV7181C (like in M1 RC4). Only composite > input accessible with an external connector. > * Audio: the Artix-7 FPGAs now include ADCs, so we can get away with > at most an op-amp on the board. Only internal microphone. > * Two USB ports. If the transceivers are expensive, consider using > resistors instead. (This hack worked for me, even though I'm pretty > sure it's not USB compliant) > * Power supply: right now we use a foolproof design whose main goal > was to reduce development time and probability of verification of > the Murphy laws. It can certainly be made cheaper by using a power > management IC (like TPS75003 or such). > * That's all for the PCB! No Ethernet, DMX, etc. Hopefully we'll > have MIDI over USB support. > * Case: injection molding? Seems aluminum molds (usable up to a few > thousand pieces) can cost only a couple of thousand dollars. > * Accessories: only a power supply transformer. No camera, cables, etc. > * Packaging: the cheapest we can find. I think this device will be > pretty small so we could even look at things like > http://blog.nicolasrenous.fr/wp-content/uploads/2008/01/tv-b-gone.jpg > > Comments? May be you can achive that price if keep the board even more minimal, and just route all you want to include to an extra HDMI-like connector so (there will be two of then) Also try to stay compatible with current jtag-serial. I recall i dont see this idea very comaptible with low end M1 for VJ,so why not get it ready for developers and hold a bit VJing part for later as an accesory that also (why not?) can be compatible with future M1 rc4 and its posible coming HDMI-like port. Regards cristian paul > > S. > _______________________________________________ > http://lists.milkymist.org/listinfo.cgi/devel-milkymist.org > IRC: #milkymist@Freenode _______________________________________________ http://lists.milkymist.org/listinfo.cgi/devel-milkymist.org IRC: #milkymist@Freenode
