Proper SDI interconnect paths II

The last Transition to Digital newsletter covered traditional wiring for NTSC plants and how it does not fit into today’s digital facilities as well as the bandwidth requirements for SDI. In this issue, we continue the discussion with how to avoid some SDI wiring pitfalls.

What to look for

How do you know if your system is error free? Testing has always been the best answer to this question, but what are you looking for in these digital signals? The SD-SDI signal can carry an error detection signal as defined by SMPTE RP 165, called EDH (Error Detection and Handling). EDH is a simple CRC (Cyclic Redundancy Check) that counts the bits in one frame and adds this number to the beginning of the following frame. If what the receiver counts and the EDH states do not match, there is an error. The EDH standard also incorporates flags that specify where the error was detected.

Most digital waveform monitors, rasterizers as well as more sophisticated monitoring equipment today will display information about the EDH signal. Some will log the errors and let you store them. Sometimes improperly configured equipment will trigger EDH, causing a rash of errors that can cause engineers to ignore them or turn off error detection. This would be a mistake, because the source of the error should be found and corrected as soon as possible.

You might be able to visually detect errors in the digital signal by watching for dropouts in the picture, but this would be time consuming and no guarantee that the error would be found. Monitoring the EDH would detect these errors much sooner.

Another way to monitor your SDI signal is with test equipment that displays the eye diagram and jitter measurements. This equipment will give you an analog display of the characteristics of your SDI signal. Using the eye diagram, the voltage level (0.80V) and DC offset of the digital signal can be measured. The rise and fall times as well as overshoots can also be observed, giving you an indication of the relative health of your transmission path. Jitter is an important parameter of SDI; SMTPE standards call for no more than 3.7nS for SD and 673pS for HD. If jitter exceeds these values, errors can occur in the recovery of the data contained in the SDI signal.

Of course, the transmission path can be directly measured with a network analyzer, which will ensure that the impedance of the complete path is correct and alert you to any other problems. But this is only for out-of-service testing, and would only be performed at the time of installation and on large, complex systems.

SDI inputs

The SDI input that receives the signal performs several functions. One of the first differences between NTSC and SDI inputs is the lack of loop through; most SDI inputs terminate to reduce the effects of RL. Another difference is that all SDI inputs use an automatic equalizer that is used to restore the rise times of the edges in the SDI signal to a required level.

On some devices, the SDI signal is also re-clocked at the output before passing on to the next device in the transmission path. There are two ways to re-clock a signal, the more common being to extract the clock signal from the SDI and use that to create new pluses with specified rise times. But if the input signal is suffering from jitter, so will the re-clocked output. A better way is to feed the house clock to the equipment and use it to regenerate the pulses of the output SDI. Of course, this can only be accomplished if the SDI is synchronized with the house sync, otherwise, you will need an SD frame synchronizer.

How to avoid problems

As with all projects, you should start with quality components. Starting with the coax cable, they are not all made the same. From the consistency of the thickness of the center conductor to the composition of the foam core, these all contribute its impedance and RL. Care must be taken with cable length, because it will directly affect the error rate in SDI. Even for SD-SDI, lengths beyond acceptable limits will bring about a total loss of picture. Always check with the cable manufacture for recommendations on maximum distances for a particular data rate.

Here is a list of Do Nots for digital cable —

Do not step on the cables.
Do not lay equipment on the cables.
Do not kink the cables.
Cable pulls should be done in a slow, steady fashion — no jerking. Do not exceed the cable’s maximum pulling tension (call the manufacturer for this information).
Do not exceed the minimum bend radius of the cable — 10 times the diameter of the cable.
Do not cinch cable ties too tightly. If you cannot move any cable inside a tied bundle, the cable tie is too tight. Slip an extra cable in when tightening and then remove it to leave room.
Do not put cable ties or J hooks at identical distances apart. This can lead to deformation at a given wavelength, which can cause RL. Place cable ties at random distances for the same reason.
• Maintain the original physical shape of the cable.

In the past, BNCs were mostly all 50ohm in impedance without any trouble, but 75ohm impedance must be maintained throughout the digital signal path. These BNCs can be recognized by the lack of a dielectric in the mating area surrounding the center pin. Also, the center pin is the same thickness throughout its length and does not taper at its point, like 50ohm BNC pins do. Always use patch bays that are rated for the data rate you are going to be using, usually this means 270Mb for SD and 1.5Gb for HD.


The new world of digital video brings its own set of challenges, whether designing a new facility or upgrading an existing one. Most of us have an exceptional working knowledge of how to wire NTSC plants, and it would seem that digital is just an extension of that knowledge, but there is much more to it. By knowing more about the signals that will pass through your plant and understanding the parameters under which they operate, you can build a reliable, error-free plant.