Search CE Pro

Print  |  Email  |  Comments (9)  |  Share  |  News  |  Follow on Twitter, Facebook, Google+ or RSS

Can Category Cable Replace HDMI?

Jeff Boccaccio says if you’re going to use Cat 5 and 6, seek out products built more along HDMI’s performance limits and not Ethernet’s.


An inter-skew test on an oscilloscope reveals the gap in time accuracy among three sets of 25-foot twisted pairs representing three video channels (red, blue and green lines) vs. the clock channel (yellow). In this test, the blue cable performed the worst. Click here to enlarge image.

As we ramp up to start testing ATDs (baluns), there is much to do documenting and organizing the testing process, performance limits, and the kinds of tests that would expose the true operational characteristics of these heavily used extension devices.

Since there is no said standard for testing ATDs and the cable they use, DPL had to develop a new set of standards that would emulate the same types of testing limits as HDMI’s current cable specification. There are a series of tests to guarantee that HDMI cable transmission lines function in harmony with all HDMI products. Some of the key evaluations are:

Video integrity: Determined by sending high-speed data through the cable and examining what it looks like as it exits the transmission line.

Supply voltage: There are certain limits of loss over the length of an HDMI cable. These voltage limits are very small and must be measured down to a tenth of a volt.

Near-field and far-field crosstalk: This can be the silent HDMI killer that many don’t know exists and in many cases gets passed over as part of the operational specification. There are well-defined tests to determine the integrity of each video channel’s crosstalk values.

HotPlug detect: This function is also mandatory for the interface to work and supplies the necessary trigger to start HDCP and EDID. These tests are much lower in frequency and in many ways stay under the radar from many firms.

This free White Paper reviews common HDMI field problems and their troubleshooting solutions, details HDMI performance signature, and provides answers to the question of why HDMI cables perform differently. Download White Paper

But the one we will focus on here is for Category 5/6 HDMI cable substitutes: timing.

Timing is one of the mandatory attributes needed for HDMI to work. There are two types of timing functions that need to be within certain performance limits: inter-pair skew and intra-pair skew, and they’re taken from the red, blue and green video channels and the clock channel.

Inter-pair skew is the time accuracy between two separate twisted pairs. In this case, the clock channel and all three video channels. Each video channel within the HDMI interface is differential, which means it uses two wires instead of one: a non-inverting wire matched with an inverting wire. In most cases these two wires are twisted pairs, and in an effort to reduce crosstalk these twists must be perfect in symmetry for each wire within the pair to be equal in length. The length differences between each pair cannot be more than 1.7ns (nanoseconds).

Intra-pair skew is similar in that it is the measurement of lengths between the two wires that make up a twisted pair. There are a total of four video channels that use this configuration and all of them must be below 112ps (picoseconds).

Has anyone taken the time to measure Cat 5/6 using the HDMI specification as the base? We did, and the findings were pretty astonishing. We used four brands of Cat 5/6 and cut each brand to 25, 50 and 100 feet. They were then terminated with RJ45s to achieve accurate in-field results.

There was no question that the longer we went, the more skew errors there were. Fig. 1 demonstrates what the interpair skew test looks like on our oscilloscope. The yellow is the clock trace from which we reference. Then there is a trace for red, blue and green. The difference between the clock channel from each of the other color channels is the measurement we are examining. When measured for 25 feet, the inter-pair skew was reasonable at 920ps, or .920ns, from the worst channel (in this case blue). This was at least below the spec limit of 1.7ns.

Then we moved to the 50-foot category cable. With the longer cable, the inter-pair skew climbed from .920ns to 1.8ns, breaching the HDMI maximum skew specification.

As you would imagine, it only got worse when we sampled our 100-foot cable. This sample produced an inter-pair skew of a whopping 2.5ns, almost twice as much over the rated minimum. And these samples were the best of the bunch.

Lesson learned? If you’re going to use Cat 5 and 6 you best seek out these products that are built more along HDMI’s performance limits and not Ethernet’s. Then, of course, you have to take into consideration the RJ45 installation and the reality of wiring a home through studs and walls. There is no question that a well-manufactured HDMI cable is the way to go. Add the necessary EQ to the cable and you’re off and running.

Subscribe to the CE Pro Newsletter

Article Topics

News · Wire and Cable · HDMI · All topics

About the Author

Jeff Boccaccio, President, DPL Labs
Jeff Boccaccio, president of DPL Labs, can be reached at

9 Comments (displayed in order by date/time)

Posted by Steve  on  11/13  at  04:53 PM

Category5, 6 etc. are Not cable types.  They are actually specifications within the EIA/TIA standards realm.  Cables constructed to meet these EIA/TIA specifications are commonly referred to as “Cat_” cables.  Many AV signal types - via modern electronics - can now be transmitted effectively on cable constructed for Category_ applications.

Posted by DrFlick  on  11/14  at  05:31 AM

1) I assume these tests and specs are for HDMI Extenders that essentially run the normal HDMI cable wires themselves over longer distances (requiring two CAT cables). I wonder what the results are for HDMI Extenders that encode the data to send it across a single CAT cable and then decode it at the other end where the cable itself probably does not cause the timing errors (like HDBaseT). I look forward to seeing your results for these types of products.
2) I notice you lump CAT-5 and CAT-6 classes of cables together. It would seem that the skew factors for CAT-6 level cables would be tighter than the CAT-5 ones. Is that what you found?
3) How does one decide which actual cable type/brand to use for minimizing these types of errors? Do the cable specifications detail skew factors over x number of meters/feet or can this be deduced from the specific electrical characteristics like capacitance over length or tolerances in the way they are twisted?



Derek R. Flickinger
Interactive Homes, Inc.

Posted by Steve  on  11/14  at  04:31 PM

...left out an important detail; cable construction deviations that result in non-compliance with the EIA/TIA standards, does mean the cable can no longer be accurately referred to as “Category_” cable.  Composite video, line-level audio, mic-level audio and many other common AV signal types have been successfully altered (via electronics; ie “baluns”) to be effectively transmitted down Category_ cable; HDMI signals - not as successfully.  The important difference to grasp is changing the signal type to fit the cable construction vs. changing the cable construction to fit the signal type.

Posted by Brad Temple  on  11/14  at  08:16 PM

Understood.  If we are simply looking at the integrity of the CAT cable, what standards specifically should we be looking for that would register along HDMI standards.  Is there a number or something.

Posted by Ernie Gilman  on  11/15  at  04:51 AM

Another very informative and totally unhelpful article.  You do not name names and offer no suggestions as to HOW we can obtain cables good for the job.
Now, what do the spec numbers mean if translated into things we can comprehend? A bit of arithmetic gives us the speed of light as .982 feet per nanosecond.  A permissible delay of 1.7 nanosecond means two wires of a pair can differ by as much as 20” and still meet spec. Really? NO! Electricity flows slower than the speed of light, on CAT cables somewhere around 60%.  That means the spec of 1.7ns would be caused by a difference between two conductors of a pair of more than 33 inches! While this sounds possible for the 100 foot cables, and is definitely in the realm of possibility (that is, it’s not an inch and it’s not 100 feet) it just seems that cable would have to be pretty horrible and visibly bad to get three feet of DIFFERENCE in length between, say, blue/white and white/blue!  Is something more than just wire length at play here?

Posted by Mark Ilpo  on  11/15  at  01:02 PM

Another completely worthless Jeff Boccaccio article.  It’s as if the point of every single article from him is “hey, check out how much I know and all these measurements I can take”.  Are we left with anything remotely useful in the end?  Of course not!  You know, stuff like what cables performed poorly.  What cables performed well.  Of course that never gets published because that information might actually be useful.

Posted by JerryS  on  11/22  at  10:52 AM

The skew difference is perfectly understandable.  To prevent inter-pair crosstalk, CAT-5 and CAT-6 cables have a different number of twists per foot for each pair.  The more twists, the longer the pair per foot of cable.  The longer the pair, the more skew.

But the author seemed surprised by this simple fact.

Posted by Steve  on  11/22  at  11:29 AM

@Brad, good question and I wish I had as good an answer.  Consider HDBaseT, which is (as best I can interpret) designed to be used with standard Cat5e cable.  No additional benefit is realized by using Cat6 or higher rated cable, and the construction differences may in fact hinder performance for HDBaseT applications.  Another example of changing the signal type to fit the established cable standard.  Development of a single, new cable construction standard that works equally well for All HDMI baluns - regardless of mfr. - hinges first on All mfrs. adopting the same transmission standards.  Unfortunately, I’m not aware of any such agreement in the works anytime soon.

Posted by DrFlick  on  11/26  at  10:04 AM


Actually, CAT-6 with HDBaseT potentially does play a role if drawing “the rated” power across the line. The temperature coefficients and the affects of temperature susceptibility of the insulation are different for CAT-5 and CAT-6. This could play into temperature rise factors, especially if bundling multiple HDBaseT cables together as they interface with a rack. CAT-6 is preferred for this scenario.

What is not documented is how the potential impact of noise affects error rates on the line. Since CAT-6 theoretically is less susceptible to noise ingress, CAT-6 potentially could allow the encoding/decoding process to run more efficiently in noisy environments.

Also note that, since HDBaseT, by design, is limited to 100 Meters, using different cable types has less of an impact on bandwidth requirements over longer distances since there is a built-in limit to the single-run architecture.


Page 1 of 1 comment pages
Post a comment
Choose smileys | View comment guidelines
Remember my personal information
Notify me of follow-up comments?

Sponsored Links

  About Us Customer Service Privacy Policy Contact Us Advertise With Us Dealer Services Subscribe Reprints ©2013 CE Pro
  EH Network: Electronic House CE Ideas Store Commercial Integrator ChannelPro ProSoundWeb Church Production Worship Facilities Electronic House Expo Worship Facilities Expo