I am going to answer my own question. Feel free to offer alternatives. I won't accept my answer.
When I was a junior engineer, a senior engineer demonstrated transmission lines in the lab with a pulse generator, a scope, and a few BNC cables. I learned more playing with real circuits than I learned in school. Later, I would repeat the demonstration for other engineers.
A typical home lab won't have a pulse generator with a fast rise time (my function generator has a rise time of 20 ns). Fortunately, it is not hard to build one. This circuit has a measured rise time of 3 ns (could be limited by the scope, it may be a little faster). And if you build a pulse generator that can create pulses shorter than the propagation time of your cable, the behavior is easier to understand.
For a demonstration setup, there are three important parameters: pulse rise time, scope bandwidth, and cable length. With this circuit and a 50 MHz scope, you need about 50 ft of cable to get nice waveforms. With 20 ft of cable, the short pulse are more like humps.
I built the circuit on a PCB with a solderless breadboard layout.
I only had top mount connectors, but I wanted them mounted on the edge. I improvised (hacked) an edge mounting.
Wire to connector should be just long enough to connect a scope probe.
Terminators:
RF TV Coax is cheaper than BNC cable, so I bought 100 feet of TV Coax.
If you want pretty plots, you need to fine-tune the termination resistances. Cheap cables may be off by 10%.
I have saved plots of 5 simple configurations, each with short and long pulses. Short being defined as a width less than the propagation time of the cable.
Note that there are people here who understand this better than me, and can explain it better than me. The Q&A is the pulse circuit, the scope plots and descriptions are a bonus.
1. Source Impedance = 0, Load Impedance = Infinite:
Short pulse: The pulse is reflected at the load non-inverted, and at the source inverted. At the load, the incident voltage and the reflected voltage are superimposed (double). The cable has losses, so the pulse amplitude is reduced each time it makes a round trip through the cable. The pulse also suffers from dispersion, i.e., the shape changes from rectangular to trapazoid because the higher frequencies travel more slowly.
Long pulse: Time magnification of the dreaded ringing that we want to avoid. At the load, the incident voltage and the reflected voltage are superimposed (double). This voltage can damage ICs or cause upsets in the receiving IC. When the signal is reflected off the source and arrives at the load a second time, it is the opposite polarity. This dip can cause double clocking.
2. Source Impedance = 0, Load Impedance = 75 Ohms:
Short pulse: The pulse is absorbed by the load, no reflection. However a lot of energy is wasted, and a typical source can't drive an impedance this low.
Long pulse: same.
3. Source Impedance = 75 Ohms, Load Impedance = Infinite:
Short pulse: The transmission line initially looks like 75 ohms, so you have a voltage divider. A pulse of amplitude Vs/2 travels down the line. At the end, the reflection causes the pulse to double. The non-inverted reflection returns to the source and is absorbed.
Long pulse: The transmission line initially looks like 75 ohms, so you have a voltage divider. After the source resistor there is a step to about 1/2 Vs. Only when the reflection returns does the source voltage step a second time to Vs. This is a very popular termination method for digital circuits. But it is very important to note that the signal only looks good at the end. If you branch off a second load, the signal will not look as good.
4. Source Impedance = 75 Ohms, Load Impedance = 75 Ohms:
Similar to #3, but the voltage at the load is only Vs/2 and it is absorbed, no reflection.
5. Source Impedance = 75 Ohms, Load Impedance = 0 Ohms:
Only useful for demonstrations.
Short pulse: The transmission line initially looks like 75 ohms, so you have a voltage divider. After the source resistor there is a step to about 1/2 Vs. The reflection is inverted, returns to the source, and absorbed by the source impedance.
Long pulse: The transmission line initially looks like 75 ohms, so you have a voltage divider. After the source resistor there is a step to about 1/2 Vs. The returning reflection cancels the pulse.
I have documented 5 simple cases. There are many other possibilities, non-matching impedances, stubs, etc.
These plots are from my 200 MHz scope. They look similar with my 50 MHz scope, maybe a little better since the ringing is reduced.
Note that in long pulse mode, with a source resistor equal to the line impedance, you have a poor man's Time-Domain Reflectometer (TDR). If you want to see how to use a pulse generator as a TDR and/or build a simpler version that only creates long pulses, see this video (not mine). https://www.youtube.com/watch?v=9cP6w2odGUc