I am actually even suggesting that we are near the end of life on our ability to build these kinds of devices in our basements/shops and get them reliable enough with a high enough yield to be useful. I am suggesting that if a cool head takes a long hard look at where we are, where the parts are, and what resources we have, that we are likely to decide we are better off having our people do the designs and then having the boards manufactured for us. The price has fallen so far on set up charge and building that this probably makes sense. I am not mandating this, I am suggesting that we should investigate this.
I could not agree more. There is no more need to solder the boards ourselves than there is to etch them in our bathtubs.
I have suggested that if we fly our own core in a fully qualified rad hard FPGA, there are OPEN CORES for processors. We can easily modify these cores to have fully EDAC registers and memory on chip for cache and off chip memory can be the slow EDAC we know about. In other words, time has marched on, and there is no reason we should not investigate whether we should not move on as well. These cores are available TODAY in the opencores.org group. We should look.
Again agreed. Also look at:
http://www.gaisler.com/cms/index.php?option=com_content&task=section&...
Which is company that has released a couple of SPARC compliant cores. They have been put into real products and into space. The rad hardened versions (ECC, etc.) are not free, but they may allow us to use them in return for space performance data.
Matt