Most wont w/o being disturbed. That's where weighing the charge on a better scale comes in, trust me, yours does under throw.
One guy in this post was claiming "to the kernel", well bs, the scale doesn't have the resolution to do so. It's a 1500 gram capable scale weighing 40+ grain charges for most of us.
What I was suggesting is that you can CHARGE to the kernel, as verified on a more accurate scale.
You cannot weigh individual kernels with a CM scale.
So how can you charge more accurately than your scale can directly measure?
What is a 'feel for' that allows this with a CM, and other scenarios including electronic alignments, process control tuning, and various load cell calibrations?
It's static -vs- dynamic
Hysteresis of the system (electronic sampling and mechanical sensing combined).
Consider hysteresis the added energy needed to cause change in a system.
It takes more energy to cause change from a static condition, and this is accounted for with the CM's scale rating.
While static, you could add or subtract a kernel and this will not overcome hysteresis in a CM.
But once the system is already in motion (dynamic state), hysteresis drops to near nothing, and the scale can then
sense individual kernel changes(even though it cannot measure them).
Drop kernels at a rate that causes the reading to eventually increment, but not lock on as stable, stop and compare this charging with a better scale, and you'll soon get a feel for the kernel rate it takes between indicated increments -to fall within a kernel of desired.
Now, we're not measuring kernels, but we're charging to desired with their weight. That's any scale + your engaged brain.
This means; most accurate charging from a CM follows a best trickle rate for the powder.
I was an early adopter of CM, probably a decade+ ago. The forums were all a buzz then with CM mods for better accuracy. Program changes and 'the straw' were employed within a week of it's release, so this thread isn't introducing something new -except, for the focus on speed instead of accuracy. This, as though accuracy in charging with a CM has been preset by RCBS, and somehow out of our control.
I've always employed program settings and a straw myself. The straw is merely to reduce occurrence of kernel clump drops, programming is to set efficient transitions to each trickle speed. I can't tune charging with either. For this, I adjust the final trickle jogs with a 10-turn potentiometer setting for each powder in use. I've tested & verified this so much that i don't need to anymore. I can now watch and see the best for myself, with no extra scale needed.
My CM seems neither faster nor slower than original until the final trickle speed. At this point things slow way down, adding around 20sec before locking on. Then I disturb the pan and watch how the reading settles. Regardless of final reading I can tell if I'm off a kernel right here. With that feeling, I dump the powder back and recharge another. Happens here & there, I don't miss it.
I personally would not want a more accurate scale on my CM. Once I learned all I could using a far more accurate Acculab, I sold it off, as it was truly horrible for a reloading scale. The CM scale is practical for it's purpose, and the charging system overall can be very accurate where operated well. Maybe you can get better accuracy than a thrower at wide open operation? I don't know, but I'm sure charging is not a race, and that a CM is not accurate at higher speeds than mine.