I kind of like the method where you read the current directly as it is applied to the stepper while the stepper is not stepping but is set to a position and held there because the meter will tell you exactly how much current is going to the motor vs using the voltage from the pot then doing the math to give you the current. In terms of accuracy, the meter method tells you what's going on inside the circuit and would compensate for any kind of variance between different steppers where inevitably, the impedance of the coils will vary from one stepper to the next and using the pot method only tells you generally what target current the driver is going to strive for whereas the meter method leaves nothing to guess. Though I've not compared the pot method to the meter method, which would be an interesting test and would reveal how accurate the driver is when it does its own adjustments to the output current. Though I suspect it will be fairly spot on ... the DRV8825 is no slouch in the engineering quality department. TI doesn't tend to "half-ass" anything they do.
It just so happens that in this specific application, I'm driving four electromagnetic coils for an electronic stirrer where a Teflon coated magnet (called a pill because of its shape) is spun on its center axis inside a solution in a beaker. The DRV8825 turned out to be a perfect choice since the way the coils are energized is exactly the same as they are inside a stepper motor, though I would really like to have a driver that graduates voltage during the transition from one coil pair to the next keeping the intensity of the magnetic field constant throughout the circle that the pill spins in, but I think the only way to really do that effectively would be to have two more coil pairs inside the radius of the sweep, where I could have three points of field generation for each end of the magnet where the center point is always the strongest and the leading point is stronger than the tailing point. I think it would help combat some of the problems I have spinning the pill inside a thick solution where the "jerkiness" of transitioning between only two coil pairs seems to cause enough injected resistance between the pill and the solution that it causes the pill to leave the magnetic field and it just starts jumping around inside the beaker.
As it turns out, it's not imperative that I get the current precise because as it happens to be, I need to drive the coils as hard as I can for obvious reasons. I'm driving these coils with 32 volts and they tend to consume around 400 to 800ma of current when it's spinning inside a thick solution. They do get hot ... too hot to touch for very long ... but I've had them powered up and running - one time for a straight 7 days without stopping and they hold up just fine, though removing heat would give me a stronger magnetic field ... I tried using a fan but the magnetic field from the fan interfered with the pill so I might try something at a distance - short of being able to submerge the coils inside liquid nitrogen would be perfect ... but WAY overkill for this application lol.
I step the stepper with a very simple method where I calculate RPMs into microsecond delays between steps and I use a timer library that I wrote called BlockNot - though I have a microsecond version of it for this app, which I'm going to add to the published library once I have the time to do that. But I didn't actually expect BlockNot to work in this application, I just tried it as a test but it's been working flawlessly which was a pleasant surprise. BUT, I am using it on a Teensy 4.1 which is far superior to a standard Arduino in terms of processing speed. I have my stepper code running on a loop on one CPU core while the rest of the code runs on the other core. Those Teensy's are really nice.