170203-2018 EST
I don't understand parts of this thread.
What I think is being ask is:
1. An inexpensive time measuring device is needed to measure up to a full scale value of 50 mS, with a resolution better than 1 mS. What resolution is really required? This somewhat relates to what accuracy is required. Is accuracy to be based on reading or full scale?
2. Time measurement is to be initiated by a contact closure, or possibly by some millivolt trigger signal that is used to initiate the trip mechanism of a breaker.
3. Time measurement is to be terminated by opening of the breaker.
4. My preference would be an oscilloscope. I did this with an ordinary oscilloscope and a camera back in the early 60s. Today I would use a digital scope.
5. A simple way to measure short times is with a current source, capacitor, and high impedance meter. The early part of an RC charge curve is approximately linear with respect to time. A high voltage source and resistor is an approximate current source to a capacitor when the voltage to the capacitor is small compared to the source voltage.
Probably a better approach is to use an operational amplifier with a resistor and capacitor as an integrator with a gated voltage source for the time period and a digital voltmeter on the output. This can be a relatively low cost approach.
Contact bounce mentioned above will cause some small problems.
6. Knowledge of your real required accuracy is important to any selection of a method for your measurement, and how the start and stop signals are created. I have used a Simpson 260 as a ballistic galvanometer to make some millisecond measurements of a Potter and Brumfield timer relay. Not highly accurate, but it worked.
7. Also note that an instrument called a counter/timer could work. But it is way overkill. And a jungle out there to figure what is available. Technical descriptions are very poor. Too much clutter and too little information.
I went looking for BK Instruments, they were in the past a low cost supplier, did not find a counter/timer, only frequency meters. An old model would be a possibility.
Found a company GW Instek with a model GFC 8270 H that is in the $ 400 range. Could be tricky to use because of its high frequency sensitivity. Would require filtering at its inputs.
There are a lot of PC based counter/timers. I did not look at any specs.
.