TDR Messungen im Heimlabor
TDR Measurements in the Home Lab
Die Längenmessung eines Kabels wird in meinem Beispiel durch die recht langsamen Flanken meines Funktionsgenerators beschränkt. Es ist daher naheliegend sich nach einer Signalquelle mit kürzerer Steigzeit umzusehen. Danach muss man aber untersuchen ob die Messmethode mit dem Tastkopf noch gut genug ist?
Kennt jemand einen schnellen digitalen Treiber neben einem 74Fxx?
Bitte Vorschläge an mich: alfred_rosenkraenzer@gmx.de
Fortsetzung folgt.
English Text:
TDR Measurements in the Home Lab
The TDR (Time-Domain Reflectometry) measurement method allows one to determine a cable's length and characteristic impedance. Plenty of literature on this subject can be found online. Specialized measuring instruments exist, though they are rarely found in home labs.
Here, I want to investigate what can be achieved using the test equipment typically found in a home lab, what the limitations are, and how those limitations might be circumvented.
Figure 1 shows the block diagram of a TDR setup.
The output of a square-wave generator feeds the cable under test, and measurements are taken with an oscilloscope directly at the injection point.
Let’s look at the oscilloscope first. I have a Hameg HMO1524 with a 150 MHz bandwidth.
Its rise and fall times are specified in the datasheet as ≤ 2.4 ns (10–90%).
A measurement using a borrowed LVDS generator showed a rise time of approx. 1.5 ns (10–90%) and about 1 ns (20–80%). In the following, the rise time is measured over the 20–80% range.
When measured with the oscilloscope—using a direct connection via a 50-ohm cable and a 50-ohm termination within the scope—my function generator shows a rise time (20–80%) of approximately 4 ns for a 100 kHz square-wave signal with 1 Vpp amplitude (into open circuit). Figure 2 shows the result.
For TDR measurements, however, it is necessary to measure as close as possible to the generator's BNC jack while minimizing signal disturbance.
It therefore makes sense to start the first experiment using a high-quality oscilloscope probe.
A BNC T-connector is attached to the function generator's output; the cable under test will be connected to one end, while the probe is attached to the other. Its tip is carefully held against the signal contact, and its ground lead is clipped to the metal, as shown in Figure 3.
Figure 4 shows the measurement result. The rise time is barely changed, but a ripple is visible after the signal rise. Experience suggests this might be related to the ground connection. An adapter is available for my probe to allow direct measurement in a BNC jack; see Figure 5. This eliminates the ripple; Figure 6 shows the resulting waveform.
For the first measurement, a 2-meter-long, unterminated 50-ohm cable is connected; its propagation delay is in the range of 5 ns per meter. Figure 7 shows the result. The "plateau" in the middle corresponds to twice the cable's propagation delay. However, measuring the delay is not straightforward because the rise time is sluggish relative to the propagation delay itself. The oscilloscope's automatic rise-time measurement function (20–80%) can help here. First, measure the rise time without the cable connected, then subtract this from the rise-time measurement taken with the cable attached; the result is twice the cable's propagation delay.
You can also measure the cable's characteristic impedance. Adjust the generator's amplitude so that the oscilloscope displays a level of 1 Vpp after the reflections have settled.
Then, measure the voltage of the step. If it is 500 mV, you can assume the cable has a characteristic impedance of 50 ohms (based on the valid assumption that the generator's output impedance is 50 ohms).
For comparison, a 3.1-meter-long 75-ohm video cable is now connected.
Multiple reflections are visible. The step in Figure 8 has a level of 600 mV, corresponding to the voltage divider formed by the 50-ohm and 75-ohm impedances. In my example, the length measurement of a cable is limited by the relatively slow edges of my function generator. It therefore makes sense to look for a signal source with a faster rise time. However, one must then investigate whether the measurement method using the probe remains adequate.
To be continued.

Discussion (2 comments)