On Tue, Mar 21, 2006 at 05:23:32AM -0700, James Richard Tyrer wrote:
> Dieter wrote:
>
> <SNIP>
>
> >Hopefully the DVI will turn out to be good enough. I like coax as much
> >as the next guy, but I assume that whatever chip the signals come out
> >of doesn't have a coax connector on the side.
>
> Answers.com states that the bandwidth is 400 MHz without any citation:
>
> http://www.answers.com/topic/digital-visual-interface
>
> AMP data sheets indicate that the rise time would be 840 ps which would
> be 416.6 MHz. That is 700 ps rating for the connector and 140 ps
> degradation for a cable.
>
> So the 400 MHz figure may have validity.
This sounds like the DVI analog pins are good enough for now, though
the dotclock rate is approaching its limits. It should be just about able
to handle the 330 MHz rate, and that would meet the requirement for 75 ohm
lines on OGD1. As far as hydra cables go, SI carries them.
In summary: since OGD1 can't accommodate multiple 75 ohm connectors,
DVI still likes the choice that will suit the most users directly, and
adapter cables are commercially available for everybody else.
Since the 500 MHz signals are going to 50 ohm coax anyway, DVI isn't
an issue there. The 13W3 probably isn't suitable either, because it's 75
ohms. That leaves either 5 50 ohm BNCs, 5 smaller-size 50 ohm coax
connectors such as SMAs, or a multi-pin 50 ohm connector in a D shell.
Unless there is some issue with additional pins to handle plug-and-play
configuration, coax connectors look like the simple choice. Adapters to
interface different type 50 ohm connectors are stock items, so if panel
space is insufficient for 5 BNCs, that's not a problem.
Somebody mentioned running coax through holes in the case. DON'T!!
The outer conductor should enter the case through a flange that gives it 360
degree grounding to the case wall. Otherwise you let ESD in and RF
interference out. Also, a low-reflection stripline-to-coax transition is a
lot easier to make with a right angle connector designed for the purpose.
_______________________________________________
Open-graphics mailing list
[email protected]
http://lists.duskglow.com/mailman/listinfo/open-graphics
List service provided by Duskglow Consulting, LLC (www.duskglow.com)