Design Notes: BJC Series-1 HDMI Cable
As with any of our custom cable products, there were a variety of design considerations that went into the production of our Bonded-Pair HDMI cables from Belden, and our purpose here is to run through a few of those so that you can see a bit of our own thought process and what goes into cable design.
History and Background:
At the outset, something which entered into this design project very heavily was our discovery, from discussions with a variety of wire and cable manufacturers and distributors, that no HDMI cable was being manufactured in the U.S., and that, in fact, few wire and cable engineers in the U.S. knew anything much about the product. What that meant, of course, was that this cable would need to be truly designed "from the ground up." There were no existing products to base our spec upon, at least not directly, but Belden had an extensive background in data and communications cable generally, and there was every reason to believe that Belden's expertise could be applied to HDMI cable.
As with any heavily-engineered product, of course, a key ingredient is having a good engineer. Back in 2005, when we first pressed the subject of HDMI cable with Belden, we met David DeSmidt, a Senior Product Development Engineer. David DeSmidt had worked on some related product types in the past, including doing some work on a DVI cable prototype for another Belden client, and so was well-suited to the task. While this product introduction and design has been a collaborative effort, of course, the "heavy lifting," engineering-wise, was always done by DeSmidt, whose practical aid and advice have sometimes been as valuable as his engineering skills.
What motivated us to go to Belden for HDMI cable was, of course, Belden's exceptional quality and process controls and its patented bonded-pair technology. It's very easy to source HDMI cable out of China, and it's much cheaper, too. We could easily enough have simply gone on, like every other American supplier, having our brand applied to Chinese-made product. But for precision work and advanced technology, there's no country like the United States, and we didn't just want to have a cable with the words "Blue Jeans Cable" on it; we wanted a cable with the words "Blue Jeans Cable" on it that we could say, without a hint of hyperbole, was the best HDMI cable in the world.
The most important element to our design was, from the beginning, that the cable needed to be based upon Belden's patented bonded pair technology. As we've written elsewhere on our site, HDMI cable is not a very well-conceived application. The cable is susceptible to failure due to excessive return loss resulting from poor impedance stability, and unfortunately, the HDMI standard's use of twisted-pair cable to carry high-bitrate signals imposes severe limitations on impedance stability. Related to impedance stability is skew; the less consistent the impedance of a pair, the less consistent the velocity of the signals traveling through it is, and the less consistent the velocity is, the harder it is to maintain tight timing from one data pair to the next. Bonded pairs address impedance stability and inter-pair and intra-pair skew very effectively. To understand how, let's look at what happens when we do the simple act of twisting wires together.
In a conventional twisted pair, two wires are drawn and separately insulated, and then these two spools of wire are mounted on a wire twister and are twisted together. At low frequencies, this works perfectly well; as frequencies increase, and the requirements of impedance and skew become more exacting, it works less and less well. A twisted pair is held together by the tension of the wires upon each other; if the tension imposed by the twister one one wire is very slightly different from that on the other (which it always is), one wire becomes a bit straighter than the other, while the other wire becomes a bit longer, producing intrapair skew. Multiple pairs are then cabled together; some pairs will wind up a bit physically longer than others, and some may wind up electrically longer than others because of minor variations in wire spacing and the amount and consistency of the plastic insulation extruded onto the wires--the result is interpair skew. The impedance of each pair varies from inch to inch and foot to foot because of these tiny variations in spacing, and when the bundle flexes, some pairs are compressed axially and "open up" while others are stretched and become tighter, introducing still more variations in impedance and timing.
These considerations probably sound like nit-picking. How much difference, after all, can tiny variations in the thickness of plastic insulation, or wire spacing, make? If the subject were telephone wire, carring an analogue audio signal a few kilohertz wide, the answer, for all practical purposes, would be "none at all." But with HDMI cable, the timing between bits at 1080p is less than a billionth of a second, and suddenly, differences which are meaningless at low frequencies become very important. In a billionth of a second (one "nanosecond," or ns), in a wire wrapped in polyethylene insulation, electrical current travels only about six inches, and so a difference in electrical length need not be large to wreak havoc. Impedance stability, similarly, can be a bear at these frequencies. For example, simply passing a wire over a wheel during production, where the wheel is microscopically out of round, can create a periodic change in the cable's characteristics which a person could not see with the naked eye, but which cause huge return loss spikes when signals run into the Gigahertz range. Factors that do not matter, and cannot even be detected, at low frequencies, become critical at high frequencies.
In a bonded pair, the two insulated wires are bonded to one another (if you'd like to know exactly how, you're out of luck; the process for doing this, and for controlling dimensions while doing so, is so secret that even if you buy hundreds of thousands of feet of HDMI cable, Belden won't let you see it being done). The physical lengths of the two wires are very strictly controlled to minimize intrapair skew, and the resulting consistency of twist rate, consistency of spacing, and stability of the spacing even under flexion, produces the best impedance stability available in twisted pair cable of any kind. Tight control over impedance necessarily means tight control over timing, which depends largely on the same factors which govern impedance, and so interpair skew, too, is minimized. Bonded pairs were developed for high-speed data, and that's what HDMI is.
Most HDMI cable is made with stranded wire. Stranded wire is easier to manage in some ways, and is more flexible than solid wire, but it has a couple of disadvantages. First, stranded wire of any given gage is larger than solid wire of the same gage, leading to the need to increase physical dimensions all around to maintain proper impedance. Second, impedance stability in cables using stranded wire simply isn't as good as with solid wire; the irregular outer surface, and the potential for strands to open up or tighten (just as with a twisted pair) make physical dimensions harder to control. Accordingly, we decided that solid wire was a better choice than stranded, and this allowed us to go just a bit large on wire gage, at 23.5 AWG.
The choice of solid wire has another benefit as well. As we've indicated elsewhere, at present we are having our HDMI cable terminated for us in China due to the high labor cost of installing 19-pin solder-on connectors at each end. But we do intend to offer American-terminated HDMI cables in the future, either alongside or instead of Chinese-terminated cables. Solid wire, while a bit more troublesome to attach to existing Chinese HDMI connectors, is actually easier to terminate in the connector design which we are currently prototyping, and expecting to introduce, for American terminations.
Interaction between shields can increase crosstalk, so the individual pair shields need to be kept out of contact with one another, but foil coverage needs to be complete and the shield needs to be reasonably terminable with HDMI connectors; accordingly, we went with inward-facing mylar tape/foil shields with drain wires, and to ensure complete shield isolation, the inward-facing mylar foil shields are then wrapped in a separator tape to guard against pinpricks and the like.
The outer shield on the bundle is akin to the shielding on other Belden broadcast-quality products; there is an outward-facing mylar tape/foil shield, together with a tinned copper braid shield providing about 85% coverage. In general, outer shielding is, contrary to common perception, not really all that critical on HDMI cable, but we decided to go with a heavy outer shield configuration because it improves overall cable strength, as well as providing just a bit of additional "insurance" against high-frequency noise leaking either in or out of the cable.
By the way, we get quite a few questions about this sort of thing. A vendor will say that his cable is "triple-shielded" or some such thing, and we'll be asked whether ours is. It's important, if you're trying to compare shields between products, not to just judge on the basis of the number of shields, but to know what the quality and type of shields being compared is. We've even seen people count the drain wire--which, by itself, would effectively provide about 5% shield coverage--as a separate shield in order to boost the number of shield "layers" they can count. Similarly, a 40% aluminum braid is "one" shield just like an 85% copper braid is; but the two provide substantially different coverage and quality of shielding.
Increasingly, HDMI cables are being installed not only in residential installations, but also in commercial settings as well. Rather than settle for a CL2 rating as is found on most HDMI cable, we wanted something that would be approved for a wider variety of installation situations, and we were able, after going through full UL burn testing, to obtain a CM rating for this cable. CM is higher on the ratings hierarchy than CL2 (for a more detailed treatment of this subject, see our article on in-wall installation ratings), so this cable is lawful for installation wherever CL2 cable is called for.
As odd as it might seem, even the jacket printing style on this cable is the result of engineering considerations. Cable jackets can be printed either by a non-contact inkjet printer, resulting in lettering with a sort of "dot-matrix" appearance, or by a "print wheel" style printer, where a print wheel is rolled over the cable jacket following jacket extrusion. Print wheel printing looks nicer than inkjet, and consequently, most a/v cable for the consumer market is print-wheel printed rather than inkjet-printed. But a print wheel can introduce periodicity to a cable, which can cause some strange effects at high frequencies, which show up as return loss "spikes" at frequencies having a wavelength related to the length of the print wheel legend. Consequently, we went with non-contact, inkjet printing to preclude this finishing step from having any impact upon the cable's performance.
We're very proud of the resulting cable. It shows exceptional high-frequency performance, which becomes critical at long distances. At this writing (October 2007), the Series-1 has passed HDMI's 1.3 compliance testing at a longer distance than any other cable we know of (Category 1, 45 feet; Category 2, 25 feet; and if you think someone else has a longer compliance test result, ask for a copy of his compliance testing certificate and see what kind of answer you get!). In actual use, we have found that it handles high-definition video on ordinary consumer devices at distances of 100 feet and more. Our intention was to produce a cable with unique characteristics, with sufficient "headroom" over the minimal spec requirements to allow our customers some security; the cable ought, we felt, to work so well that a customer should not have to fear that when he buys a new device that supports deep colour, or higher resolutions, or that simply has a more finicky input or output circuit, something will stop working. We think we've accomplished that goal, and we will continue to look for ways to improve the performance of our Series-1 HDMI cable.