Speaking of telephony experience, I just used a kick meter to estimate the amount of cat 6 cable left in a box.
TDRs are for the weak.
"What's a kick meter?", you ask?
"Kick meter" is the common name for the Bell System KS-8455 loop tester, a simple analog volt-ohm meter introduced in the 1940's and still made to this day. (They're also called "brownie meters", for the brown leather case they came in).
What's so special about this meter? Glad you asked.
On the surface, not much. It's 0-100V DC voltmeter plus a simple Ohm meter function. But it has a special feature....
Read on...
... In addition to the usual switch to measure Volts or Ohms and resistance zero pot, there's a "REV" switch that reverses the polarity of the input leads. That switch, along with the known ballistics of the meter movement, lets you make approximate capacitance measurements, if you know the trick.
When you measure ohms on an open circuit, you're putting DC voltage on the circuit. If the circuit is open, you won't see the needle move (which is why Ohm meters have the high value first).
But...
... If there's any capacitance on the input, what the ohm meter was actually doing was charging the "capacitor". When you reverse the input leads (by flipping the switch), the Ohm meter will briefly deflect ("kick") forward as the the capacitance discharges back into the meter. The amount of kick is proportional to the amount of capacitance.
Why is this useful for measuring the length of twisted pair cables?
Well, you see...
... a long twisted pair cable (such as a phone local loop or Ethernet cable) builds up capacitance as it gets longer (the two wires act as the plates, and the insulation the dielectric). The longer the cable, the more capacitance, proportional to the length.
A long time ago, someone noticed this, and figured out that this particular meter kicks a particular percentage of the scale for a given number of feet on open circuit twisted pair. And so for about 70 years...
@mattblaze This is slightly inaccurate: The line also contains inductance (it can be modelled well as a long sequence of segments with serial resistance and inductance in each wire in each segment and parallel capacitance and resistance in each segment), which will affect the kick: it will slow down the discharge and thus prolong the kick while reducing its amplitude (and will add a smaller counter-kick after the main kick).
Capacitance is affected by the dielectric constant of the material used for insulation of individual wires in the pair (and also by wire thickness), while inductance isn't by either. Do you know if the constants they had to use changed whenever the wire material changed?
@robryk The inductance is tiny compared with the capacitance here.
@mattblaze The only way I can think of comparing the two (comparing impedance) is frequency-dependent. What kinds of frequencies are you talking about (or is the difference so huge that it would hold well into tens or hundreds of MHz?).
@robryk perhaps you misunderstand the measurement technique here. You hook up an analog ohm meter (a DC source) to an open circuit (a cable) and manually flip a switch to reverse the polarity, observing the deflection of the needle. The maximum deflection increases with the capacitance of the cable, and is proportional to the length.
@mattblaze Sure. So at a very crude approximation, the circuit we have is a series connection of voltage source, internal resistance of the meter, inductance of the cable, capacitance of the cable. (A less crude approximation would assume that the capacitance is spread across the cable, behind various amounts of inductance.) (I mentioned frequencies, because impedances at frequency f are TTBOMK a very crude approximation of what matters after a transient + 1/f delay.)
One thing is that the total charge (summed with sign) transferred during the transient does not depend on anything other than capacitance. So, the interesting question is how long does it take to charge the capacitor to, say, half of the input voltage, and how that time is affected by inductance. (As long as that time stays smaller than something related to ammeter's needle's inertia, we will be getting an ammeter reading that is roughly proportional to the total charge that moves through it.) I suspect that we'll end up finding that this time is a sum of something internal-resistance-derived and inductance-derived (well, in first order approximation for small inductances it must be something like that).
If we want to look at transients, then
V_res + V_ind + V_cap = V_source
V_ind = I'*L
V_res = I*R
V_cap' = I/C
I' = 1/L*(V_source - V_cap - IR)
I'' = -1/L*(V_cap' + I'R)
I'' = -1/L*(I/C-I'R)
... and I'll write out how that affects the time when the integral of current reaches sth like 1/2*CV_source tomorrow
@robryk I'm really not interested in debating transmission line theory with you.
@mattblaze In that case would you mind measuring (or reading off its documentation) the internal resistance of the kick meter you have? (I suspect it's raised by an explicit resistor, to reduce the impact of changes in battery's internal resistance as it discharges on the measured value.)
@robryk E.g., you can use the RC time constant to calculate how long the meter will take to discharge the capacitance. But that's not sufficient to tell you how far the needle will travel in that amount of time, which depends on the ballistic properties of the movement.
@robryk Analog Ohm meters always have a pot you adjust to zero the meter before taking a reading so there’s no constant resistance. But the meter is spec’s for 450microamps @ 45V, implying nominal 100k ohms. But as I noted, the capacitive kick effect depends heavily on the specific ballistics of the needle movement, which are known for this particular meter.