Re: "Everybody knows that you should measure battery voltage under load, but exactly what load"

The best load conditions to measure are those that apply to your loads. Different for a speedlight vs a remote control vs a battery powered LED strip vs a remote temp/humidity sensor.

Also I am doubtful about the usefulness of constant current load testing these days. The most common loads for NiMH batteries these days are "constant power" loads: boost regulators. As their input voltage falls, they draw more current to supply the same voltage and current to their own load.

I put "constant power" in quotes because the efficiency of boost regulators usually goes down with decreasing input voltage, so they are usually "increasing power" loads. Constant current testing would overpredict.

I have a project on my list to make a NiMH capacity tester using as the load a boost regulator loaded with a range of resistors. But at my current rate of progress through the list, I'll be dead first.

Re: "Everybody knows that you should measure battery voltage under load, but exactly what load"

The best load conditions to measure are those that apply to your loads. Different for a speedlight vs a remote control vs a battery powered LED strip vs a remote temp/humidity sensor.

Also I am doubtful about the usefulness of constant current load testing these days. The most common loads for NiMH batteries these days are "constant power" loads: boost regulators. As their input voltage falls, they draw more current to supply the same voltage and current to their own load.

I put "constant power" in quotes because the efficiency of boost regulators usually goes down with decreasing input voltage, so they are usually "increasing power" loads. Constant current testing would overpredict.

I have a project on my list to make a NiMH capacity tester using as the load a boost regulator loaded with a range of resistors. But at my current rate of progress through the list, I'll be dead first.

@alec44 We need a little more information.

Is it flexing or axial movement - rotation, or translation? How fast, how much, and how often? How long is the wire? Is it moving along its entire length? Is it a single wire or a multi-wire cable? Is it in an inaccessible environment, or can it be moved for testing? Are there temperature, pressure, optical or chemical environmental considerations? What is it connecting? Can the circuit be altered at all? Is it possible to inject signals at one end and detect them at the other? What signal (voltage, current, frequency) does it carry now, and does it need to carry them for the life cycle test? How thin is the wire? What structures are near the wire, and can they be modified? How precisely do you need to know the point of failure (e.g. "101 273.27 flexions" vs "more than 100 000, fewer than 110 000")?