The Custom Integrator Show Installment 023

This Installment focuses on HDMI Extenders.

[mp3]http://media.libsyn.com/media/iandixon/TDL_Custom_20100104_023.mp3[/mp3]

Direct Download – Subscribe via RSS – Subscribe via iTunes – Subscribe with Zune

I decided to put the topic of the Podcast as the first line (as part of the New Year) so that it is easier to find information and so that it shows up as part of the Podcast description on small-screen devices.  Wow, it has been over a year since my first post on this blog site.  It seems like forever ago.  Thanks for sticking it out with me Ian!

We start to get back on track with our series on HDMI by diving into HDMI Extenders, one of my favorite topics.  We started the discussion at the end of the last Installment and pick up with the details this time.  We review some of the wiring considerations, which I think are an integral part of a successful HDMI Extender implementation.  Once again, I will use the Honeywell CURxE line for my examples because I think they are a great example of how a product should be designed and I understand the engineering internals fairly well.  I will add, however, that their current implementation still does not support full Deep Color bit depths though.  Keep an eye out in the April time frame for some products that should overcome this limitation.  Also note that I emphasize (once again) using dual CAT-6 cables over CAT-5s, which increases your headroom and reduces the possibilities of DDC corruption by reducing crosstalk, the influence of capacitance, and both the inter and intra pair skew.  It also allows for slightly longer cable runs at a little higher bandwidth.  Your mileage may vary more when using CAT-5 cables.  Even at 200 feet, the mask margin on the eye is better than 14 percent for CAT-6, so 1080P definitely is not a problem at that length.  remember that there really are two functions performed within HDMI.  One is the video itself (and the bandwidth to support the resolutions, etc.) via the TMDS lines.  The other is the handshaking (the Display Data Channel is the main one, with the CEC line being another).  The video transmissions use one of the CAT-6 cables between the units.  The DDC and CEC lines use the other.

Following is a list of what I tell integrators to look for in an HDMI Extender.  It is not all-encompassing, but covers the purchasing and installation criteria pretty well.

Note that over 60% of all HDMI failures are because of corruption of the DDC.  The focus of any good Extender is to minimize this corruption.  The CURxE Light technology essentially is an I2C repeater.  It drives the level of the DDC lines out to about 10 mA between the two boxes and uses higher current clamping gates at the receiver side, giving it extreme stability.  For instance, the pull up resistors usually would be 4.7 K.  With 5 volts, that works out to be about 1 mA.  However with the driver they use, it increases that by using a 470 ohm resister, which increases the current to 10 mA.  That is 10 times the current required under the HDMI spec.  Since this connection between the sender and receiver units is being done “outside” the HDMI world, they can drive the gates at this higher level and then restore it back to normal before it goes back into the HDMI world.  This drastically minimizes the noise and errors on these lines.

They also have incorporated adaptive EQ that matches the wire lengths and keeps the rise time under 260 nS, no matter what the length (up to about 200 feet).  You know why this is important if you have listened to the previous Installments.  This allows the rate of change of the EQ to shift depending on the length of the cable.  Some poorer implementations actually extend this rise time out to 1000 nS, which is why they do not work correctly or are intermittent at best.

A nice feature that they incorporate is a set of three diagnostic LEDs to provide a visual indication of the link status for voltage, data, and clock.  There are 3 LEDs on the receiver board and one on the transmitter.  The one on the transmitter is the Hot Plug Detect.  That goes on when you have a connection between the sink and the source.  They even retransmit the Hot Plug Detect back to the source to ensure it shows the correct handshaking is occurring.  There is a supply indicator that is green on the receiver board.  This tells you that the phantom power supply is active and running.  There are a yellow and blue LED on the receiver board.  The yellow is the data line and the blue is the clock line.

When the unit starts up, these two LEDs flicker quite a bit.  Then they will stop.  They start again for about 2 seconds.  These are the “wiggle” data lines for HDCP (you will remember that 2 second handshake for the HDCP key exchanges from Installment 020).  When you see that refresh come in, the handshaking has worked correctly and the unit should be operating.  If for some reason you still see no picture, the fact that the I2C lines are “wiggling” every two seconds indicates that you have an HDCP issue.

I also am a big fan of running that extra pair of 20 or so gauge wire to power the sink side without having to put a wall wart at the display.  Not using the transmitter’s power for any of the Extender’s electronics is important because it does not support the amount of current drain it takes to drive any additional logic over what the receiver’s basic electronics need to stay alive for completing the DDC handshaking when the device is off (these units require about 500 MA at 5 Volts DC to remain stable).  These particular ones do not support using what normally would be considered the extra “PoE lines” on a CAT-6 cable for powering it because they are used for additional grounds (they use a differential signal between the Extender transmitter and receiver).  It does require that you run an extra pair of wires (20 gauge works fine) in parallel with your twisted pairs to support it. 

I also will mention that the wiring for these particular units is “straight across” within the CAT-6 cables (not EIA/TIA 568 A or B).  This is because of using those ground lines in with the signals.  The twists are run between the transmitter and receiver as differential pairs. There are some products that specify wiring with an EIA/TIA 568-B configuration, but I still do not understand how the differential pairs can be “spot on” wiring that way.  Suffice it to say, HDMI Extenders do not use “standard” Ethernet, so your installers need to be aware of this.  With CAT-6 runs, you can get about 200 feet reliably out of them.  The DPL Labs test results show a drastic performance increase when wiring them this way (these particular products include a wiring diagram for this with the devices).

My only other advice is to look at using some of the good thinner HDMI cables (those with a higher DPL rating) for the short runs between the Extender plates and the devices.  This reduces failures because heavier cables tend to come loose from the connectors.  Panasonic (as well as others) also makes a right-angle HDMI swivel adapter for areas that do not allow a “normal” turn radius into the equipment.

HDMI Extenders are great.  Having a distributed HDMI infrastructure opens up a lot more possibilities for distributed and synchronized media throughout the home.  They are not cheap, but I firmly believe they are worth the additional investment.

If all goes well, the next Installment will take a look at the Consumer Electronic Control (CEC) features of HDMI.

  =D-