Hi,
The 200pcs of assembled JTAG daughter boards is done.
1. smt factory assembled 195 full parts mounted excepts lost 5pcs 12 MHz
Xtal parts. This is quite bad. :(
After i checked first panel with 6 pieces on polarity checks, pin1 checks,
first two piece with functionality work on JTAG and serial connections. I
came back to wait. :-)
later i removed last rc1's part to rc2 5 boards, they are all
well-assembled.
2. after followed test procedures Step1: current test firstly,
http://en.qi-hardware.com/wiki/JTAG_Serial_Cable_run2_for_Milkymist_One#Test_Procedures
got one obvious soldering short by my visual inspection but pass in current
test. well...while i tested each jtag board, i still visually inspected on
mounted status. this way didn't completely auto though. U5 pin 31 is 3.3V
which shorted to pin30, It's ACBUS4 that it's useless.
I have no smart way that can detect out if short pins still there, since the
current won't be a big difference. if someone have new idea, you can share
it. :-)
3, then follow step 2 ~ 4 for each test:
I got all 100% yield rate of PASS.
This way means that i used a Milkymist One board to do the tests about
plug-in/plug-out actions on J5 and J6 at least 400 times completely in two
days. :-)
Next time i may do an extend parallel cable to connect my J6 to be as
fixture tool.
for whole steps of this run, check
http://en.qi-hardware.com/wiki/JTAG_Serial_Cable_run2_for_Milkymist_One
Comparing to previous runs, my observations on this new smt factory:
1, poor import parts control
2, very good mounted smt process results
3, not bad internal quality human checks
4, each returned parts has been re-counted, this is good for preparation for
next.
5, not openly to take videos
It's good that we have this factory as other source now.
Thanks,
adam
_______________________________________________
http://lists.milkymist.org/listinfo.cgi/devel-milkymist.org
IRC: #milkymist@Freenode
Twitter: www.twitter.com/milkymistvj
Ideas? http://milkymist.uservoice.com
