Audiogeek Cables | Nitrogen Interconnects

Introduction

Effectively, the ProSink termination ensures that the signal travels only one way down the cable. In an normal cable, the signal travels down to the end, and reflects back -- a product of the fact that the impedance at the end of the cable is more than a thousand times higher than the impedance of the cable. The ProSink termination, developed by Dr. Kevin Gilmore, ensures that no such reflection occurs.

Figure 1. TDR with no cable attached,
pulse magnified 2x.

Figure 2. TDR with Belden 89259 as load,
75 ohm output, pulse magnified 2x.

Figure 3. TDR with expensive audio cable as load,
50 ohm output, pulse magnified 2x.

The Basics of Time Domain Reflectometry

A time domain reflectometer is a remarkably simple device. It sends a pulse down the cable under test, and displays the results. If the cable changes in impedance at any point on its length, that point will show a reflected pulse corresponding to the change in impedance. Effectively, the time domain reflectometer -- also known as a "cable radar" -- is showing the impedance of the cable as a function of time.

If you look at a normal cable on a time domain reflectometer, the display will show two significant peaks and a few little ones. For example, compare the two pictures to the left. Figure 1 is a TDR display with no cable connected; Figure 2, a TDR display with a normal audio interconnect connected.

The first peak shown on the display is the output of the TDR. It's a sharp spike that travels down the cable, and is reflected back in direct proportion to the impedance of the cable at that point. In essence, the TDR gives us a display of impedance as a function of distance.
Therefore, the second big peak we see is the end of the cable. Why? At the end of the cable, the impedance reaches infinity -- and since the highest peak that can show on the display, we see a peak equal in height to the initial peak.

But what of the "other stuff" on the screen? In the case of the Belden 89259 stock under test, there's a second little peak to the right of the "end" of the cable, as if there was still a phantom cable there. Some of the shoulder arises as an artifact of the TDR's pulse-generating circuitry. But if you compare Figures 2 and 3 to Figure 6, it becomes clear that not all of the "phantom cable" comes from the test instrument -- the secondary spikes are simply too large and defined. What's happened is one of two things. Since the cable has capacitative, inductive, and resistive elements, it's possible that the initial "test spike" set up oscillations in the cable. (This occurs very commonly with certain cable geometries, such as the various braided/twisted designs. In those cases, you can see multiple "phantom cable" spikes, and the distance between them changes as you bend and flex the cable -- changing the inter-conductor capacitance and inductance.)

Also possible in some cases is that the signal has reflected off the non-ProSinked end, traveled back down the cable, reflected off the source end (which happens with sources that have an output impedance not equal to the impedance of the cable and cannot handle the fast [1nS] risetime of the reflections), reflected off the non-ProSinked end again, and hit the detector in the TDR precisely one cable-length-time after the main reflection. However, we can prove this is not the case here -- in Figure 2, the cable is driven with the TDR set to a 75 ohm source impedance (which would eliminate the second reflection), and yet the second peak still exists. Therefore, in this cable, the TDR is showing us that the cable has a tendency to "ring" when a signal is applied.

An example of an extreme case of cable-induced distortion can be seen in Figure 3. This is a certain, very expensive audio interconnect that was shipped to our laboratories for testing by an audiophile that wanted to see how it would fare on the TDR. Figure 3 suggests that the cable is less than 50 ohms (comparing the relative height of the pulses) and the spreading of the reflected pulse suggests that the cable has a limited high-frequency response -- though a spectrum analyzer would be required to properly determine the exact frequency response. It's quite possible that the resistance and capacitance of the cable are high enough to have formed a low-pass filter.

The Effect of the ProSink Termination

Consider first figures 4 and 5. In figure 4, we see the same spike we did in figure 1 -- like figure 1, there is no cable connected.

Now, let us connect a ProSink-terminated Nitrogen cable to the TDR -- figure 5. You'll note that with the ProSink termination, there is no second peak. The ProSink termination is absorbing the whole of the initial pulse, resulting in no back-reflecting signal down the cable. (In this case, the "shoulder" on the right side of the pulse is an abberation induced by the pulse-generator in the TDR. It exists to a smaller degree even when the TDR is unloaded, as seen in figure 6.)

As a result of the ProSink termination, any cable of constant, controlled, and measurable impedance that is hooked up to the TDR appears almost identical to any other cable. Silver, copper, foil shielded, braid shielded, twinline or coax -- they all become exactly the same. Now, you'll note I said two things -- any cable of constant impedance, and almost identical. This is because many common cable geometries, including the vast majority of hand-made cables, do not in fact have a constant impedance. Twisted- or braided- cables, for example, cannot be ProSinked because their impedance varies often quite significantly down the length of the cable. The greater the precision to which the impedance of the cable is controlled, the greater the effect that the ProSink termination has.

Figure 4. TDR with no cable connected, pulse magnified 2x.

Figure 5. TDR with Nitrogen cable connected, pulse magnified 2x.

Figure 6. TDR with no cable connected, pulse magnified 1x.