Powder Measure Statistics

[idntwn2]

Anyone done any statistical workups on their powder drop? I just did a little tiny one:

Sample size 30
Powder HP-38
Lee Powder drop, setting 252
Brass Case
Tare scale after every measurement

Target 5.2 grains

Results

Average 5.3
Max 5.6
Min 5.1
Range 0.5
SD 0.2
Mode 5.4
Median 5.2

I was surprised by how much the drop varied each time. With Standard deviation this high, rounds at +/-3s (99.7%) are going to be between 5.9 and 4.7 grains. Quite a bit in my book.

[TimRB]

What kind of scale did you use? Try taking thirty measurements of a check weight and see how much that varies. Sounds like the data was a bit skewed to the low side, so SD might not be as meaningful as it normally would.

Tim

[Whiterabbit]

my question exactly, about the scale. Someone using a drug-dealer hand-me-down will likely get different results than the guy who is using a lab grade .01 grain capable scale-in-a-box, even if they are using the same powder drop in the same room next to each other alternating drops.

[Fordtrucks]

I think ur taking the numbers and jumping to conclusions. Keeping ur powder level relatively constant, ur powder measure has show it will throw 5.3 ave. 5.6 max, 5.1 min. So what's with ur 5.9 to 4.7? Ur extrapolating numbers from assumptions. Ur powder measure will most likely NEVER throw 4.7 or 5.9 keeping everything constant.
Don't use ur max and min to apply ur correction. It should be applied to ur true value.

Now that being said, I see a variance of ~0.25 grn. I would call that good enough for most of my pistol rnds and surely in the realm of plinking rifle rounds. Match rifle should most likely be weighed individually for best accuracy.
Just my veiw from here.

[idntwn2]

Scale is a cheap digital from Midway (Frankford). But... it is very consistent on check weights.

Will repeat the experiment tonight.

[idntwn2]

I think ur taking the numbers and jumping to conclusions. Keeping ur powder level relatively constant, ur powder measure has show it will throw 5.3 ave. 5.6 max, 5.1 min. So what's with ur 5.9 to 4.7? Ur extrapolating numbers from assumptions. Ur powder measure will most likely NEVER throw 4.7 or 5.9 keeping everything constant.
Don't use ur max and min to apply ur correction. It should be applied to ur true value.

Now that being said, I see a variance of ~0.25 grn. I would call that good enough for most of my pistol rnds and surely in the realm of plinking rifle rounds. Match rifle should most likely be weighed individually for best accuracy.
Just my veiw from here.

5.9-4.7???

[FromTheGrave]

For reference

5.9-4.7???

With Standard deviation this high, rounds at +/-3s (99.7%) are going to be between 5.9 and 4.7 grains. Quite a bit in my book.

[chevy95]

Had similar issues with Hornady 1500 and now my Franklin

Load with Hornady single stage and normally measure each load i drop. Have kept powder amount pretty constant and tap on side of hopper once in awhile to make sure powder is settled. Stopped using the Hornady scale last year; it wandered between 5.2 to 5.9 grains of Unique going under a 230 grain berry bullet (loading for .45 acp). Bought the F/A scale. At first it seemed to be very consistent and then it too began to wander. It is now in the "wandering area" of between 5.4 to 6.0 grns.

I have tried to keep my motion constant; the power in the hopper constant. Some folks have suggested getting a model 505 or 10-10 beam scale, but I'm on a VERY tight budget.

Anyway, because it takes me SO DERN LONG to weigh my charges, it takes 2 hours to do 100 completed rounds....TOO DANG SLOW.

Suggestions are welcome folks.

[idntwn2]

For reference

Ah yes, I see, that is at 6 sigma. My bad.. I thought you were talking about observed range.

No extrapolation in any way. It is a statistical reality, albeit one based on a very small sample size.

[idntwn2]

Have discovered a scale drift issue. Quite possibly a major contributor.

[hntnnut]

Are you using the scale close to a fluorescent light or is it plugged into the same circuit as a fluorescent light?

Richard

[Dark Mod]

I did a similar experiment comparing the lee, Hornady, and Dillon powder measures. Used an extruded powder, a ball powder, and a flake powder. 10 drops and weighed on the same scale.

Extruded powder: all 3 had about a .3 grn deviance
Flake powder: lee and Hornady were about .3, Dillon varied by .5
Ball powder: Hornsby had a .2 deviance, lee had .5, and Dillon had .4 deviance.

This was done with a jeweler a Scale meant to measure carats but it's a very accurate scale. Take my findings with a grain of salt though, this was hardly a scientific experiment

[ar15barrels]

Scale is a cheap digital from Midway (Frankford). But... it is very consistent on check weights.

Will repeat the experiment tonight.

The check weights are usually WAY heavier than typical powder charges.
Keep a small and a larger paper clip on your bench.
Weigh them from time to time and see if they show different weights at different times.

[stevec223]

I have found my Hornady scale will start to wander when the batteries get low .. A change of batteries brings back the consistancy... Cheers...

[idntwn2]

The check weights are usually WAY heavier than typical powder charges.
Keep a small and a larger paper clip on your bench.
Weigh them from time to time and see if they show different weights at different times.

Good idea!