Networking & Cables

Boccaccio: A Better Way to Power Active HDMI Cables

HDMI cables, power supplies and DC currents don't always play well together.

Boccaccio: A Better Way to Power Active HDMI Cables
Jeff Boccaccio on the age-old question: "Is there an actual HDMI failure or is it related to a power supply issue?"

Jeff Boccaccio · July 3, 2018

Like any active device — whether it be HDMI cable extenders, baluns, HDBaseT, or fiber — some form of supply voltage is required. It’s always been a big no-no to tap into the lifeblood supply voltage that resides within the HDMI bus. This portion of the interface was intended for the system to function under normal HDMI applications and not to be used to power up any other ancillary devices. Even HDCP relies on this supply voltage by way of the HDMI’s internal DDC instruction channel.

As we have noted in the past, the spec for this voltage can only support small amounts of DC current. This is why you see some cable companies provide additional power to run their devices.

But we’re now experiencing an increase of reported failures with active cable products due to power issues even though a good majority of them offer external power. We’ve requested and received many of these products to determine the origin of the problems, and it gets a bit more complicated than just offering external power.

Some manufacturers recommend using any USB port to power their devices ... but what form of USB?

The first thing you should do is check the power rating of each device if it is published. Some manufacturers recommend using any USB port to power their devices, which typically ends up being one in a display or A/V receiver. But what form of USB?

Since, in many cases, the device’s power or current requirement is not published one would assume that the device can work on any USB format. However, feedback from the field reports these products are still not working very well in many applications. There are different USB power standards and problems can be as simple as the type of standard used.

A Better Way to Secure Power 

A better way to secure power for each cable device is to provide its own power source via some kind of AC converter similar to a wall wart. This can be a very good move since now each device can be matched to a known power source. As the failing samples arrive into DPL Labs they are categorized according to if a device is internally powered, requiring external power or supplying its own external power source.

Let’s say a cable device incorporates chipsets having an operating voltage of 5 volts. Let’s also assume the current demand is 500ma. Well, that is easy, right? Just get a power supply that has a rating of 5 volts output with at least 500ma and you are done. The rule of thumb is to buy more power than you need just to be safe.

But how are these power supplies rated? If it claims 500ma does that mean it can provide 500ma and hold its regulated 5-volt output? We’ve had power supplies come in that when pushed to their max current the output voltage fell to less than 4.2 volts DC.

Now go one step further and consider what would happen if an active cable device operated at voltages less than 5 volts, perhaps 3.3. These voltages continue to drop, some down to as low as 1.2 volts VCC. So are you really safe since your supply is 5 volts, and even though you are down to 4.2 under load you have plenty of headroom to still operate the device? Well, yes and no.

To provide these lower voltages some form of voltage regulation is built inside each device like with a 3.3-volt TMDS driver, for example. Depending on the type of regulator its quiescent current can be proportional or inversely proportional to the incoming voltage.

A regulator designed to accept 5 volts can in fact draw more current if the input voltage goes low or high. The current demand will also be affected by the cable’s internal electronics. The more current, the more heat and noise, which are a semiconductor’s worst enemy. No matter what the case may be, it can set up a supply failure over time — whether hours, days, months or years later.

Now consider what happens if the supply finally gives up and dies. The HDMI device will shut down killing the video, but what’s worse is if its internal current protection mechanism automatically turns the supply back on after some cooldown period or if somebody removes the device from the AC, creating a hard restart.

And the question then looms, is there an actual HDMI failure or is it related to a power supply issue?



  About the Author

Jeff Boccaccio is president of DPL Labs. Jeff can be reached at info@dpllabs.com. Have a suggestion or a topic you want to read more about? Email Jeff at jeff@dpllabs.com

Follow Jeff on social media:

Jeff also participates in these groups:
LinkedIn · Google+

View Jeff Boccaccio's complete profile.



  Article Topics


Networking & Cables · HDMI · News · HDMI · All Topics
CE Pro Magazine

Read More Articles Like This… With A Free Subscription

CE Pro magazine is the resource you need to keep up-to-date on the latest products, techniques, designs and business practices. Subscribe today!

Subscribe Today!

Comments

Posted by Jeff Boccaccio on July 4, 2018

Sorry folks, my last comment from Targa 100 was with my old CEpro account.
My new one is boccaccio

Posted by Targa100 on July 4, 2018

We have tested a bunch of these from all walks of life and most did not come up with the correct numbers. One big reason for this was in fact the USB cable used with it. So in order to actually do a physical test on the supply we had to tap into the 3 power pins on the USB connector directly so there would be no other form of resistance distorting the test. Even doing that we still could not get most of the devices to meet spec. So you can see that this issue can manifest themselves by using long USB cables and relatively small gauge.

We also ran some numbers on “average current demand” from a sampling of products that need external power where David is spot on. They came in anywhere from 160ma to as much as 300ma.  You must also take into account that these numbers must be measure with the device fully operational with all three TMDS channels operating at rail into a differential load of 100 ohms with the necessary pull ups, any additional features like LED’s and such, and all DDC Data operating at specified rail (5volts) and running at 100Khz.

Power supply’s have always been critical and will continue to be as we move into Rev 2.1.

Posted by David Meyer on July 3, 2018

Julie, that is actually a very good question, but sadly the answer is no. The reason is that power is never sent, only drawn. Voltage could be measured easily enough within the 4.7-5.3V range in question, but to measure a source’s current output capacity would entail a tester drawing current, starting low and ramping up to see where the fail point is. The problem is that it doesn’t really indicate safe or stable limits, and it’s also possible that damage could result to the source (rare, but possible nonetheless).

The other issue is that active cable/extender manufacturers don’t usually disclose their operating power requirements either. So we’re in the dark on both power needed and power available. But Jeff’s absolutely right about going to the only predictable external power source. USB 2.0 only defines 100mA per unit load, and 150mA for USB 3.0, but many products need >200mA. Many USB ports can handle multiple unit loads, but they’re not always labelled! A wall wart or USB power adapter main be a pain, but it’s the only way to be sure.

Posted by Julie Jacobson on July 3, 2018

David, thanks so much for chiming in. Consider that I’m pretty ignorant on the video side of things, but is there a simple tool to test the power coming out of an hdmi port? The beauty about being a site moderator is that if this is deemed a stupid question, I can delete it. jj

Posted by David Meyer on July 3, 2018

Thank you Jeff! I totally agree about the absolute need for integrators to take command of an active transmission line by using external power wherever available. Some active cable vendors infer that a source not providing enough power to drive their product is somehow inferior. But in fact a source need only provide 55mA to be complaint, of which the sink needs to be able to draw 50mA. Providing anything more is purely optional. However, this is typically not specified by manufacturers, so how do you know? You don’t. Well, you do Jeff, but integrators don’t wink.

Further to this, HDMI 2.1 effectively outlaws the practice of in-line devices drawing anything more than 5mA through the HDMI port, and mandates the inclusion of external power for such active cables or extenders.

Relying on the unknown quantity of HDMI internal power is a recipe for a callback. I’ve long advised that if the active device has a power socket, use it.

Posted by David Meyer on July 3, 2018

Thank you Jeff! I totally agree about the absolute need for integrators to take command of an active transmission line by using external power wherever available. Some active cable vendors infer that a source not providing enough power to drive their product is somehow inferior. But in fact a source need only provide 55mA to be complaint, of which the sink needs to be able to draw 50mA. Providing anything more is purely optional. However, this is typically not specified by manufacturers, so how do you know? You don’t. Well, you do Jeff, but integrators don’t wink.

Further to this, HDMI 2.1 effectively outlaws the practice of in-line devices drawing anything more than 5mA through the HDMI port, and mandates the inclusion of external power for such active cables or extenders.

Relying on the unknown quantity of HDMI internal power is a recipe for a callback. I’ve long advised that if the active device has a power socket, use it.

Posted by Julie Jacobson on July 3, 2018

David, thanks so much for chiming in. Consider that I’m pretty ignorant on the video side of things, but is there a simple tool to test the power coming out of an hdmi port? The beauty about being a site moderator is that if this is deemed a stupid question, I can delete it. jj

Posted by David Meyer on July 3, 2018

Julie, that is actually a very good question, but sadly the answer is no. The reason is that power is never sent, only drawn. Voltage could be measured easily enough within the 4.7-5.3V range in question, but to measure a source’s current output capacity would entail a tester drawing current, starting low and ramping up to see where the fail point is. The problem is that it doesn’t really indicate safe or stable limits, and it’s also possible that damage could result to the source (rare, but possible nonetheless).

The other issue is that active cable/extender manufacturers don’t usually disclose their operating power requirements either. So we’re in the dark on both power needed and power available. But Jeff’s absolutely right about going to the only predictable external power source. USB 2.0 only defines 100mA per unit load, and 150mA for USB 3.0, but many products need >200mA. Many USB ports can handle multiple unit loads, but they’re not always labelled! A wall wart or USB power adapter main be a pain, but it’s the only way to be sure.

Posted by Targa100 on July 4, 2018

We have tested a bunch of these from all walks of life and most did not come up with the correct numbers. One big reason for this was in fact the USB cable used with it. So in order to actually do a physical test on the supply we had to tap into the 3 power pins on the USB connector directly so there would be no other form of resistance distorting the test. Even doing that we still could not get most of the devices to meet spec. So you can see that this issue can manifest themselves by using long USB cables and relatively small gauge.

We also ran some numbers on “average current demand” from a sampling of products that need external power where David is spot on. They came in anywhere from 160ma to as much as 300ma.  You must also take into account that these numbers must be measure with the device fully operational with all three TMDS channels operating at rail into a differential load of 100 ohms with the necessary pull ups, any additional features like LED’s and such, and all DDC Data operating at specified rail (5volts) and running at 100Khz.

Power supply’s have always been critical and will continue to be as we move into Rev 2.1.

Posted by Jeff Boccaccio on July 4, 2018

Sorry folks, my last comment from Targa 100 was with my old CEpro account.
My new one is boccaccio