Just curious: why isn't BeagleBoard HDMI shell grounded?
1,392 views
Skip to first unread message
John Beetem
Feb 7, 2012, 3:47:00 PM2/7/12
to Beagle Board
I have a BeagleBoard B4 which has always worked fine over an HDMI to
DVI-D cable. I recently acquired a cheap "HDMI Switcher" so that I
could easily switch between multiple HDMI sources, including the
BeagleBoard. However, when I tried to connect a cheap 3' HDMI-to-HDMI
cable between BeagleBoard and the HDMI switch it didn't work at all.
It turns out the cheap 3' cable didn't bother using the individual
ground pins and relied on the shell to provide a common ground, which
I guess must be normal for HDMI products or they wouldn't be able to
get away with selling cables like this.
So I'm curious why BeagleBoard (at least classic BeagleBoard) doesn't
connect the HDMI shell to ground. I went ahead and soldered the shell
to a good ground and now the cheap HDMI cable and HDMI switch work
fine.
DVI-D cable. I recently acquired a cheap "HDMI Switcher" so that I
could easily switch between multiple HDMI sources, including the
BeagleBoard. However, when I tried to connect a cheap 3' HDMI-to-HDMI
cable between BeagleBoard and the HDMI switch it didn't work at all.
It turns out the cheap 3' cable didn't bother using the individual
ground pins and relied on the shell to provide a common ground, which
I guess must be normal for HDMI products or they wouldn't be able to
get away with selling cables like this.
So I'm curious why BeagleBoard (at least classic BeagleBoard) doesn't
connect the HDMI shell to ground. I went ahead and soldered the shell
to a good ground and now the cheap HDMI cable and HDMI switch work
fine.
Gerald Coley
Feb 7, 2012, 4:25:39 PM2/7/12
to beagl...@googlegroups.com
The BeagleBoard-xM has the shield grounded to frame ground which connects to the system ground through a resistor, R119. The original BeagleBoard version does not have the connector shield grounded. Only the signals in the cable are grounded..
Gerald
--
To join: http://beagleboard.org/discuss
To unsubscribe from this group, send email to:
beagleboard...@googlegroups.com
Frequently asked questions: http://beagleboard.org/faq
Dave Higton
Feb 7, 2012, 5:26:31 PM2/7/12
to beagl...@googlegroups.com
In message
<CAHK_S+c=BSW=LvfoHYKpqAN7hD89H_r...@mail.gmail.com>
Gerald Coley <ger...@beagleboard.org> wrote:
<CAHK_S+c=BSW=LvfoHYKpqAN7hD89H_r...@mail.gmail.com>
Gerald Coley <ger...@beagleboard.org> wrote:
> The BeagleBoard-xM has the shield grounded to frame ground which connects
> to the system ground through a resistor, R119. The original BeagleBoard
> version does not have the connector shield grounded. Only the signals in
> the cable are grounded..
Interesting. Best EMC practice suggests that connector shells should
be grounded at as many points as possible, as directly as possible,
which means all the shield pins/pads should go directly to the ground
plane. You really don't want the inductance of a resistor in the way,
low as it might be in absolute terms; it's high in EMC terms.
Dave
rickman
Feb 7, 2012, 6:02:21 PM2/7/12
to Beagle Board
On Feb 7, 5:26 pm, Dave Higton <davehig...@dsl.pipex.com> wrote:
> In message
> <CAHK_S+c=BSW=LvfoHYKpqAN7hD89H_rcJsAx_N62noOVhu8...@mail.gmail.com>
EMI is one of those things that people often don't really understand
or sometimes they just don't believe the facts. People use terms
like, "noise getting into my circuit" and other meaningless things.
So they will leave pins unconnected to leave the noise "out" or use
odd connections which reduce the effectiveness.
But shields are not typically used for connecting grounds of two
boards or chassis. It is more common to connect a cable shield only
at one end. I believe the reason for this is to prevent ground
loops. Two units really should be connected by a single low impedance
path.
I think the cheap HDMI box that isn't using the ground wire(s) in the
cable is really the true culprit. I expect the "standard" way of
doing this would be to connect the shield pins of both boards to
ground and let the cable determine which end the shield is not
connected. Anyone know the details about this?
Rick
> In message
> <CAHK_S+c=BSW=LvfoHYKpqAN7hD89H_rcJsAx_N62noOVhu8...@mail.gmail.com>
or sometimes they just don't believe the facts. People use terms
like, "noise getting into my circuit" and other meaningless things.
So they will leave pins unconnected to leave the noise "out" or use
odd connections which reduce the effectiveness.
But shields are not typically used for connecting grounds of two
boards or chassis. It is more common to connect a cable shield only
at one end. I believe the reason for this is to prevent ground
loops. Two units really should be connected by a single low impedance
path.
I think the cheap HDMI box that isn't using the ground wire(s) in the
cable is really the true culprit. I expect the "standard" way of
doing this would be to connect the shield pins of both boards to
ground and let the cable determine which end the shield is not
connected. Anyone know the details about this?
Rick
John Beetem
Feb 7, 2012, 6:36:43 PM2/7/12
to Beagle Board
Gerald, thank you for the quick reply!
On Feb 7, 3:02 pm, rickman <gnu...@gmail.com> wrote:
> I think the cheap HDMI box that isn't using the ground wire(s) in the
> cable is really the true culprit. I expect the "standard" way of
> doing this would be to connect the shield pins of both boards to
> ground and let the cable determine which end the shield is not
> connected. Anyone know the details about this?
It's definitely the cheap cable that's not connecting any of the HDMI
ground pins. It's just using the shell for all the grounds. HDMI
calls 4 of the ground pins "shields": they're associated with the 4
high-speed differential pairs. Only the CEC/DDC ground is actually
called "ground". The cable actually has a rather elaborate foil
shield which isolates the pairs from each other. It's connected at
each end of the cable to the HDMI shell. Yep, I cut one open as part
of "WTF analysis". Other than the missing wires, it's well made and
performs just fine for 3 feet.
The HDMI switch connects every ground together, along with the
shells. Yep, I had that baby open also. Fortunately, it has screws.
It uses a Pericom PI3HDMI series HDMI switch.
On Feb 7, 3:02 pm, rickman <gnu...@gmail.com> wrote:
> I think the cheap HDMI box that isn't using the ground wire(s) in the
> cable is really the true culprit. I expect the "standard" way of
> doing this would be to connect the shield pins of both boards to
> ground and let the cable determine which end the shield is not
> connected. Anyone know the details about this?
ground pins. It's just using the shell for all the grounds. HDMI
calls 4 of the ground pins "shields": they're associated with the 4
high-speed differential pairs. Only the CEC/DDC ground is actually
called "ground". The cable actually has a rather elaborate foil
shield which isolates the pairs from each other. It's connected at
each end of the cable to the HDMI shell. Yep, I cut one open as part
of "WTF analysis". Other than the missing wires, it's well made and
performs just fine for 3 feet.
The HDMI switch connects every ground together, along with the
shells. Yep, I had that baby open also. Fortunately, it has screws.
It uses a Pericom PI3HDMI series HDMI switch.
rickman
Feb 7, 2012, 6:59:14 PM2/7/12
to Beagle Board
four of the grounds "shield" is because that is what they are! But a
pin called ground should be connected at each end to make the
connection. It sounds like the cable is using the shell connection
for the individual pair shields in addition to an overall cable
shield. As I said, this only needs to be connected at one end to work
properly. Is the ground pin connected in the cable at both ends? I
think you are saying it is not which is not the fault of any board.
The cable shields should not be used as the ground connection between
chassis as they can often be very thin layers of Aluminum with
noticeable resistance on longer runs, but I suppose the HDMI cable is
not very long really. Is the outer cable shield the braided type?
That would have a pretty low resistance.
Rick
Gerald Coley
Feb 7, 2012, 7:20:59 PM2/7/12
to beagl...@googlegroups.com
As I said. It is connected on the BeagleBoard-xM version. There was a reason we didn't do it, but that was four years ago and I don't recall the exact reason..
Gerald
John Beetem
Feb 7, 2012, 7:33:00 PM2/7/12
to Beagle Board
On Feb 7, 3:59 pm, rickman <gnu...@gmail.com> wrote:
> Is the ground pin connected in the cable at both ends? I
> think you are saying it is not which is not the fault of any board.
> The cable shields should not be used as the ground connection between
> chassis as they can often be very thin layers of Aluminum with
> noticeable resistance on longer runs, but I suppose the HDMI cable is
> not very long really. Is the outer cable shield the braided type?
> That would have a pretty low resistance.
The HDMI CEC/DDC ground pin isn't connected in the cable. It looks
> Is the ground pin connected in the cable at both ends? I
> think you are saying it is not which is not the fault of any board.
> The cable shields should not be used as the ground connection between
> chassis as they can often be very thin layers of Aluminum with
> noticeable resistance on longer runs, but I suppose the HDMI cable is
> not very long really. Is the outer cable shield the braided type?
> That would have a pretty low resistance.
like there is a single bare copper wire which connects the HDMI shells
end to end, and shorts to the aluminum foil shield built into the
cable which separates the differential pairs. Not braided, but it's
only a 3 foot (1 meter) cable.
I also have two PCs with HDMI outputs. Both of these connect all the
HDMI grounds and shell together, so the cheap 3' cables work fine for
those. I suppose if I did a lot of flexing the cables wouldn't last
long, but the whole point of the HDMI switch is to avoid moving cables
around.
Reply all
Reply to author
Forward
0 new messages