Shop OBEX P1 Docs P2 Docs Learn Events
HDMI-Output-Board for BeMicro CV — Parallax Forums

HDMI-Output-Board for BeMicro CV

65816581 Posts: 132
edited 2015-01-25 16:47 in Propeller 1
Hi there,

I just made this HDMI-Output-Board which uses 4 LVDS lanes at J2 for
HDMI video output. Code has still to be written, but it should work just fine!

iMeouzOhAgWA.png

Comments

  • Cluso99Cluso99 Posts: 18,071
    edited 2014-10-21 17:00
    Seeing the circuit (never thought about it before) it would be simpler to just buy a HDMI cable and cut off the end and crimp the wires to a 2x5 header.
  • 65816581 Posts: 132
    edited 2014-10-22 06:49
    Cluso99 wrote: »
    Seeing the circuit (never thought about it before) it would be simpler to just buy a HDMI cable and cut off the end and crimp the wires to a 2x5 header.

    The traces on the PCB are all routed differentially and with matched lenghts. I think just soldering wires
    to a 2x5 header could get a problem with this kind of high-speed signals.
  • jac_goudsmitjac_goudsmit Posts: 418
    edited 2014-10-22 07:29
    6581 wrote: »
    The traces on the PCB are all routed differentially and with matched lenghts. I think just soldering wires
    to a 2x5 header could get a problem with this kind of high-speed signals.

    I wonder, when your differentially routed the traces, did you measure the trace lengths on the BeMicro PCB and take them into account?

    I'm not a high speed PCB design expert or HDMI expert, but it seems to me that for such short traces (probably 2" / 5cm total max or so), the frequencies for which mismatched trace length would cause problems are probably in a range that the FPGA can't generate anyway. I seem to remember something about 180 picoseconds per inch, so if one trace is 1mm shorter, the difference in arrival time is going to be less than 10 picoseconds off; I would say that only matters for signals at close to 20 picoseconds time period, i.e. 50GHz or so. Or am I wrong?

    I guess it doesn't hurt to to have matching trace lengths but I'm just genuinely curious as to how much it matters in this case.

    ===Jac
  • Cluso99Cluso99 Posts: 18,071
    edited 2014-10-22 12:06
    6581 wrote: »
    The traces on the PCB are all routed differentially and with matched lenghts. I think just soldering wires
    to a 2x5 header could get a problem with this kind of high-speed signals.
    If the 2x5 header was crimped then the lengths could be held within 0.1" but by being careful the lengths could be matched. If the header was soldered vertically (or using crimped pins in a header block) the lengths could be held very close. However, I believe none of this really matters and would save requiring a pcb to transition to an HDMI socket. In fact, IMHO, it would not only be simpler, but more stable too.
  • 65816581 Posts: 132
    edited 2014-10-24 07:26
    I just ordere a batch of 3 PCBs from OSHPark including stencils. Will let you know if they
    arrive here. Then I will start with the HDL code.
  • TubularTubular Posts: 4,740
    edited 2014-10-26 19:39
    This is great, 6581. I hope it works ok for you.

    I'm interested in building a micro hdmi into a board I'm working on. I don't have room for a full HDMI connector, but Micro HDMI would fit. I believe the pin numbers may be slightly different.

    I must try the oshpark stencil sometime too.
  • porcupineporcupine Posts: 80
    edited 2015-01-25 16:47
    Any update on this? finally got my BeMicro CV working again, want to do something more interesting than blink LEDs :-)
Sign In or Register to comment.