How Much Precision in a Scale is Required for Precise Weight-Sorting?

[Rocketvapor]

more comments
Linearity is the error between the zero and cal points not attributed to calibration error, zero error, or repeatability.
Cal at 10, 20 and 50 grams on a 100 gram scale and the max linearity error would "probably" be right in the middle of two adjacent cal points. Powder charges of 2 or 3 grams are close to zero on a 100 gram scale.
Linearity error is what you would expect by calibrating @ 50 grams and weighting 10, 20, 30 grams.

With 0.02 readability digital roll over could indicate 29.98 grains with 29.97 grains on the pan.
29.99 grains in the pan could also indicate 29.98 grains. That's with NO actual scale error.
Just a feature of digital devices. One half of the least indicated interval, +/-
Note that this rollover applies to a zero, a weighed value and the closest calibration point.


Here's a test in progress with my EJ-54D2
I bought my EJ-54D2 for $355 shipped, it is now $426 plus shipping.
I'm going to go anal on a few cases for my next 600 F-Class shoot.
More anal than normal, just to evaluate scale error.
I'll track the cases, new, sized and trimmed, primers, primed cases, charges, charged cases, bullets, and loaded rounds. Probably take a few days in my spare time
6mm HAGAR CASE PREP
10G (ASTM Class 1) WITH 20G CAL
10.0000 9.9998 9.9998 10.0000 9.9998

NEW CASES weight in grams
1 8.1826 1826 1826 1826 1824 avg 8.18256 grams
2 8.1826 1828 1828 1826 1826 avg 8.18268 grams
3 8.1828 1830 1830 1830 1830 avg 8.18296 grams
----------------------------
4 8.1838 1840 1840 1840 1840 avg 8.18396 grams
5 8.1840 1840 1840 1842 1842 avg 8.18408 grams
6 8.1844 1844 1842 1844 1842 avg 8.18432 grams

10G
9.9998 9.9998 10.0000 9.9998 10.0000

I now have them sized to 22 NOSGAR and trimmed to 1.765".
Enough for tonight.



Well, here's case #1, after trim and chamfer. I'll let it rest overnight.

10G
9.9998 9.9998 10.0000 9.9998 10.0000
SIZED AND TRIMMED TO 22 NOSGAR 1.765"
1 8.1672 .1674 .1674 .1672 .1672 avg 8.16728
2 8.1694* .1692 .1692 .1690* .1692 .1692 avg 8.16920
3 8.1702 .1704 .1702 .1704 .1704 avg 8.17032
----------------------------
4 8.1704 .1702 .1704 .1702 .1706* .1702 avg 8.17033
5 8.1722* .1720 .1720 .1718 .1718 .1720 avg 8.17197
6 8.1720* .1718 .1716* .1718 .1718 .1718 avg 8.17180
I took an extra reading when range exceeded 0.0002g
10G
10.0000 10.0000 10.0000 9.9998 10.0000
Total range of the 6 cases is 126.040grains to 126.113 grains (0.073 grains)
Case #1 will be a sighter. Avg of #2 thru #6 is 126.092 grains (a range of 0.043 grains)


Just don't bother weighing CCI primers

 

[RegionRat]

Here are some definitions:

 

[cdgaydos]

So with a scale with .02 gr. readability, we can expect--from the repeatability spec--that any reading of 30.0 grains could in actuality be as low as 29.98 gr. or as high as 30.02 gr. If this is true, it would seem to suggest that, in, say, 20 different powder weighings, the range or maximum difference between heaviest and lightest of the 20 separate charges weighed might be .04 gr. For some reason that I can't explain, this seems too low.

Yes, BUT… The expected range or maximum difference of 0.04 would be for measuring the SAME charge 20 times… the same part. That is repeatability… or precision.

The accuracy of the scale is how close to the true value your readings are. This is why we calibrate scales and why check weights are necessary to verify calibration. Linearity refers to the accuracy of the scale over its range. This is important when the calibration is a single point calibration… one check weight or 0 and a weight. If it were not linear calibration would be more complicated.

A scale can be extremely precise with a very fine readability and still be “wrong”.

I have a 20g check weight that I use for calibration and to check that my scale is still in calibration (and accurate). If the check weight is off my scale will be off. So, I’ve checked my weight on a laboratory balance… I actually added a little bit of tape to the weight to make it exactly 20g (based on weighing on a certified analytical balance).

I hope I’m not rambling and confusing things more.

 

[cdgaydos]

more comments
Linearity is the error between the zero and cal points not attributed to calibration error, zero error, or repeatability.
Cal at 10, 20 and 50 grams on a 100 gram scale and the max linearity error would "probably" be right in the middle of two adjacent cal points. Powder charges of 2 or 3 grams are close to zero on a 100 gram scale.
Linearity error is what you would expect by calibrating @ 50 grams and weighting 10, 20, 30 grams.

With 0.02 readability digital roll over could indicate 29.98 grains with 29.97 grains on the pan.
29.99 grains in the pan could also indicate 29.98 grains. That's with NO actual scale error.
Just a feature of digital devices. One half of the least indicated interval, +/-
Note that this rollover applies to a zero, a weighed value and the closest calibration point.


Here's a test in progress with my EJ-54D2
I bought my EJ-54D2 for $355 shipped, it is now $426 plus shipping.
I'm going to go anal on a few cases for my next 600 F-Class shoot.
More anal than normal, just to evaluate scale error.
I'll track the cases, new, sized and trimmed, primers, primed cases, charges, charged cases, bullets, and loaded rounds. Probably take a few days in my spare time
6mm HAGAR CASE PREP
10G (ASTM Class 1) WITH 20G CAL
10.0000 9.9998 9.9998 10.0000 9.9998

NEW CASES weight in grams
1 8.1826 1826 1826 1826 1824 avg 8.18256 grams
2 8.1826 1828 1828 1826 1826 avg 8.18268 grams
3 8.1828 1830 1830 1830 1830 avg 8.18296 grams
----------------------------
4 8.1838 1840 1840 1840 1840 avg 8.18396 grams
5 8.1840 1840 1840 1842 1842 avg 8.18408 grams
6 8.1844 1844 1842 1844 1842 avg 8.18432 grams

10G
9.9998 9.9998 10.0000 9.9998 10.0000

I now have them sized to 22 NOSGAR and trimmed to 1.765".
Enough for tonight.

View attachment 1338453

Well, here's case #1, after trim and chamfer. I'll let it rest overnight.
View attachment 1338455
10G
9.9998 9.9998 10.0000 9.9998 10.0000
SIZED AND TRIMMED TO 22 NOSGAR 1.765"
1 8.1672 .1674 .1674 .1672 .1672 avg 8.16728
2 8.1694* .1692 .1692 .1690* .1692 .1692 avg 8.16920
3 8.1702 .1704 .1702 .1704 .1704 avg 8.17032
----------------------------
4 8.1704 .1702 .1704 .1702 .1706* .1702 avg 8.17033
5 8.1722* .1720 .1720 .1718 .1718 .1720 avg 8.17197
6 8.1720* .1718 .1716* .1718 .1718 .1718 avg 8.17180
I took an extra reading when range exceeded 0.0002g
10G
10.0000 10.0000 10.0000 9.9998 10.0000
Total range of the 6 cases is 126.040grains to 126.113 grains (0.073 grains)
Case #1 will be a sighter. Avg of #2 thru #6 is 126.092 grains (a range of 0.043 grains)


Just don't bother weighing CCI primers

How do you like that scale? I have a GemPro 250 I was considering replacing… because, “how can it be accurate?” It was inexpensive and it’s old and I need a new toy… but, I don’t want to spend $700… on a whim. I looked at the one you have (I believe)… then I decided to recheck my GemPro vs an analytic balance and it was spot on (within its repeatability and readability)… but, I’m still curious about other scales…

 

[Rocketvapor]

I like it, but the price has gone up.
Maybe I bought on a sale?
Technique and environment can make a big difference in how a scale performs.
I use several GOOD check weights (no tape ) to verify results.
I use a technique I call a Fake Zero to monitor zero drift for small weights close to zero,
like powder charges and now primers.
The picture shows a 1 gram weight, and one gram plus a primer.
Looks like I have enough resolution to sort primers.


Another primer from the same box.

Without the extra digit and counting by 2's, the 0.2348g primer would show 0.234 or 0.236 and the 0.2394g primer would show 0.238g or 0.240g. That might not be good enough.

I'll continue with the scale test but I need to make a range trip to finalize a load first.
The 22 Nosgar with SB6.5, and 88's will be new.
If all goes well maybe I'll be allowed back in

 

[BOB LEE SWAGGER]

I'm of the view - I could be wrong - that so long as I have decent but basic reloading equipment, I can improve my scores much more by learning to read conditions. And until I do that, spending $5,000 on whizz-bang reloading gear is cart b4 the horse.

It is not a bad thing to have good equipment. Usually you come to the place where you wonder if you are better than what you have.

 

[cdgaydos]

I like it, but the price has gone up.
Maybe I bought on a sale?
Technique and environment can make a big difference in how a scale performs.
I use several GOOD check weights (no tape ) to verify results.
I use a technique I call a Fake Zero to monitor zero drift for small weights close to zero,
like powder charges and now primers.
The picture shows a 1 gram weight, and one gram plus a primer.
Looks like I have enough resolution to sort primers.
View attachment 1338483

Another primer from the same box.
View attachment 1338486
Without the extra digit and counting by 2's, the 0.2348g primer would show 0.234 or 0.236 and the 0.2394g primer would show 0.238g or 0.240g. That might not be good enough.

I'll continue with the scale test but I need to make a range trip to finalize a load first.
The 22 Nosgar with SB6.5, and 88's will be new.
If all goes well maybe I'll be allowed back in

To be clear I adjusted the check weight with a tiny piece of tape to agree to 0.0001g on a lab grade analytical balance… if I recall it was off 0.0001-2g and I was not going to do anything as it was within the acceptable error… but, why not fix it.

Another technique for low mass items is to add a constant weight to the pan or tray (a washer for example… keep it centered)… this gets to scale away from its minimum weight region and closer to the calibration point….

A scale with higher resolution than 0.001g (0.0001g) gets a bit frustrating to use unless you have very accurate calibration weights, a hood to protect from drafts (we use one in addition to the enclosure on the balance… and a very sturdy table. Ours is about 6” thick marble on marble legs… heavy). So, I think you are doing about all you can do…. I hope it pays off and gives you an improvement worth the effort… or it shows nothing so you don’t have to worry about it… I guess a drastic improvement would be worth it.

Good luck!

 

[South Pender]

Yes, BUT… The expected range or maximum difference of 0.04 would be for measuring the SAME charge 20 times… the same part. That is repeatability… or precision.

Yes, I realize that repeatability refers to the same charge being weighed 20 times. But what I was wondering is whether this might also provide the range that you'd expect from weighing 20 different charges. That's the question I'd like to get an answer to. In other words, if I weighed 20 different charges each to exactly 30.0 gr., would the range of weight of those 20 different charges also be 29.98 to 30.02?

 

[johnfred1965]

Ok, just to be silly, gravity pulls harder the closer you are to the center of the earth. If you are in say Denver as opposed to say San Diego, your primers will be lighter. Do any of you have a scale that can measure that?

Or...does the additional mass of the mountains offset your increased distance from the center of gravity. Also at different times of the year the earth is millions of miles closer to the sun, does that effect your measurement? Or how about high tide vs low tide?

How far are you going to go? I've heard of (never met) benchresters dropping loads at the range rather than weighing them and still winning, can that be explained? Please don't consider this mocking, I don't want to be like some people.

 

[Rocketvapor]

@South Pender
I'm going to say
"I doubt it".
A stable object be it a case, bullet, check weight (even with tape) has a property of Sameness.
Precision weights properly cared for, in addition to being accurate have the property of sameness.
Unless you can prove scale performance you will always have some unknown.
You could toss the charge, lift pan and reweigh. One problem is the autozero feature of some scales will find a slightly different zero each time you toss a charge. Trickle with a slow approach to final value can cause an undershoot or even to hunt for a final value.

@johnfred1965
Most scales measure gravity behind your back. You must calibrate at your local gravity.
The EJ-54D2 allows you to enter a local gravity constant (or just recalibrate at the new location.
Weighing scales are calibrated to Apparent Mass. A 10 gram cal weight will have the same mass In Denver as it does on the moon. Force calibrations need corrections to mass due to local gravity (and bouyancy for precise force measurements). For example, calibrating a torque wrench with dead weight in Denver, or on the moon


Notice that the calibration uncertainty of the Mass calibration, and the Mass value is less than the 0.05mg class tolerance. This is a small fraction of the 0.0002g (0.2mg) resolution of the EJ-54D2.
This mass standard is adequate Using multiple weighings I can use it to calibrate less accurate weights still suitable for use with this scale (avg of a count) to save wear and tear on the more expensive standard.
I have two 10 gram working standards that I can not detect a weight difference from the Standard.
Good enough for me.

Some of you older guys might remember two events that had an impact on precision mass standards.
One when NBS (NIST) changed reported Apparent Mass in air from A.M. vs Brass (8.4g/cc) to A.M. vs Stainless Steel (8.0g/cc). Most lab standards were S.S. anyway. This was a buoyancy correction.

The other was when a deviation of the U.S. standard Kilogram from the International Kilogram required a mathematical adjustment to all certified U.S. standards.

 

[cdgaydos]

Yes, I realize that repeatability refers to the same charge being weighed 20 times. But what I was wondering is whether this might also provide the range that you'd expect from weighing 20 different charges. That's the question I'd like to get an answer to. In other words, if I weighed 20 different charges each to exactly 30.0 gr., would the range of weight of those 20 different charges also be 29.98 to 30.02?

Oh! If you DISPENSED charges to 30.0 on the scale…. Then yes…

If you trickled up you would see 29.96, 29.98, 30.0…. If you stop at 30.0 your ACTUAL charge would be between 29.98 and 30.02… you would not know. If you weighed each of those many times they may show 29.98 to 30.02 (assuming the repeatability to be 0.02).

Another aspect of a scale is how small of a change in weight it can detect once it has “settled”. Similarly how quickly it settles. I always pick the pan (lift it slightly) up and put it back down when trickling in the last kernel to verify… a “better” scale may not require this step.

 

[RegionRat]

Some of you older guys might remember....

 

[FNSafari]

Not my game. (except for rimfire BR) I kinda been scared off by all the internet experts who mock my 0.30" groups with their one hole, 0.19" groups. Figger I'd just embarass myself. Working toward F-class tho.

GetReal,
They aren't mocking you, they are just telling you the simple truth of where a group or score benchrest rifle needs to agg.

 

[Rset]

Oh! If you DISPENSED charges to 30.0 on the scale…. Then yes…

If you trickled up you would see 29.96, 29.98, 30.0…. If you stop at 30.0 your ACTUAL charge would be between 29.98 and 30.02… you would not know. If you weighed each of those many times they may show 29.98 to 30.02 (assuming the repeatability to be 0.02).

Another aspect of a scale is how small of a change in weight it can detect once it has “settled”. Similarly how quickly it settles. I always pick the pan (lift it slightly) up and put it back down when trickling in the last kernel to verify… a “better” scale may not require this step.

Haven't read this thread, but my question with all this talk is simple. Can YOU see the difference on the target?

 

[DocCarbon]

I'm mostly in agreement, but will just add a parting thought and leave the rest to the forum and readers, while begging your indulgence to expand on your point if you are saying a typical 0.1 gr scale isn't adequate for brass sorting.

There tends to be controversy surrounding concepts such as the tolerance of brass weight (or volume) to have a significant difference on target. After all, that is the heart of the matter with this forum so it is the right place to spend time on it.

The controversy happens when it takes significant resources or barrel life to prove a point, and since most reading here are not the Fortune 500 or trust fund babies, we are stuck with certain controversies due to physics, the difficulty of setting up definitive tests, and the budgets it takes to run them. That said, I think we can do a better job of being clear for the sake of the beginners.

When it comes to records and state of the art shooting, beginners reading here are not always BR shooters or familiar so it is up to folks like you and I to steer their development concepts. We need to keep in mind we have to cater to all of them, the rookies with their first box stock RPR, and the experienced Master who is trying to make it to the podium at a national or international level. Good ballistics discussions can often serve both types at the same time, or at least build the reference frame for state of the art and what that takes, even if the answer cannot be applied universally.

When it comes to primers, the total weight is so much lower than typical case weights, that I tried to steer the thread away to avoid boiling the ocean over too many conflicting tolerance requirements.

The title of the thread and the later details within the OP's first post were not lined up well, so the aim point of the question started with the question of case sorting and before we came to any good conclusions, we quickly digressed into all sorting and the significance or insignificance of primer sorting and single kernel loading.

While I highly value the opinions of highly accomplished competitive shooters, we run the risk of turning off the beginners when threads wander and go down rabbit holes where evidence is scant because it requires state of the art science and big budgets to be definitive.

When some of us are still solving the case sort question and others are off into the definition of calibration and tolerance of primers, things get drawn out and muddy with conclusions on none of those.

We may or may not all agree on this, but I will back up a level higher to before the cause for the question in the first place... and state that the reason this comes up is that things like bullets, powder, brass, and primers are all made with tolerances that can and do cause problems from time to time when the topic is state of the art accuracy performance.

If someone isn't shooting for state of the art in BR or highpower competition, then they should frame their answers in context, and the opposite is also true in that some advice is based on extreme accuracy requirements. Sometimes, we just don't have the budgets or enough data to answer their questions or the data is proprietary and not to be published due to legalities.

If we focus on just the brass, we all (I should say some of us) know that even Lapua isn't perfect for all purposes or from time to time. Many batches of Lapua have checked out good, yet others seem to be escapes. It is a damn sight better to start with Lapua as an example, but the reason we might sort it and when we say it needs to be better is getting muddy here.

I will go on the record and tell you primers have been just as vulnerable to escapes as bullets and brass, but thank the stars it is rare.

Taking a good batch and using that to prove the bad ones are also good isn't good advice. To be fair, not everyone has tracked primer weights (or bullets or cases either), so if you have been blessed to never have seen a bad batch good for you. My advice to beginners is to be prepared to establish your own criteria, but that starts with knowing when you are just lucky versus screening for the eventual escapes.

That circles us back to the age old question of advice. I suggested early that a scale to detect if brass even needs sorting can be a typical inexpensive 0.1 gr scale. If I am wrong then I would assume someone with even better data should be able to say that a sort finer than the typical uncertainty of a 0.1 grain scale needs to be better because of XYZ, etc. and point out a topic or context where a sort bin of less than 0.3 grains makes a difference on target, and therefore the answer to the scale question becomes clear.

I am not calling you or anyone else out and apologize if it comes off that way cause it is certainly not my intention, just suggesting we get the thread back on tracks to at least solve the first question the OP asked without drift into powder kernels and primers.

Those other components are important enough to warrant their own thread in my opinion. And, because I am still interested in hearing your point on if a lab grade scale with better than 0.1 grains is needed for brass sort for something I am unaware of. YMMV.

I'm in agreement 100% . Scales were a sticky point with ME personally ,as I never mentioned What scale or degree of accuracy was needed for any purpose ,let alone case sorting .
My emphasis was on ACCURATE SCALES ,as being a Well Trusted manufacturer and being certified as for legal trade . ( This ensures accuracy ) I have perhaps #7-10 older cheapo electronic scales ,people have given Me over the years . I'm sure You know the Name brands of shooting reloading equipment and accessories manufacturers that I'm referencing . Chinese Junk ,each and everyone of them , I tested them all and repeatability is not in the realm of their frame work . I matched up two scales which supposedly were same Chinese manufacturer ,with different brand name . I used a small medium weight 2Gram ; One scale DIDN'T register ,the other read 3.4 Gram . I switched them off and repeated and the previous failed to register ,showed 1.3 Gram ,the other 2.7 Gram .

Let Me be absolutely crystal clear and Drive this point home for everyone .

When You buy a scale any brand or price category ,just be damn sure it's ACCURATE and will weigh a Known weight or #3 actually small medium and larger weight and is repeatable .
This ensures YOUR SAFETY !.

I shutter to think someone may actually have relied on one of those cheapo Elec. scales and Blown their gun and hurt themselves or another shooter on the line ,because the scale showed !!!!!!!! .

Reality is a quality manual balance beam is pretty sensitive ,especially IF one cleans and replaces pivots .

As with anything else quality cost $$$ ,I'd personally Not feel comfortable with an electronic scale which cost less than $200.00 ,because it says somewhere ,something was undermined to reach that price bracket .

And electronic scale with the degree of accuracy required by nearly all reloaders ,is going to be in the $200-500 range . Main thing CERTIFIED LEGAL TRADE ,the degree of weigh accuracy is an individual decision IMO .

 

[South Pender]

Rocketvapor, that A&D EJ-54D2 scale is very attractive. It evidently gives readability to .0002 grams, or .003 grains which makes it about 5 times more precise than the very-popular A&D FX120i that a number of reloaders are using. Here in Canada, the two seem to be pretty close to the same price, so the EJ-54D2 would definitely seem to be the better deal. It's hard to imagine that you'd need any more precision for the majority of reloading tasks involving weighing--even weighing primers.

 

[Coyotefurharvester]

Problem with most digital reloading scales, their not even calibrated in the range of normal use. 10 grams 154.324 grains,20 grams 308.647 grains, of course the 50 gram weight is 771.618. I would like to see capacity, work range and calibration synchronized. Graduation 0.05 grains ok, accuracy +,- 1 grad , 250 gr capacity. I don't shoot any centerfire competition, just like repeatable hunting accuracy.

 

[Rocketvapor]

The EJ-54D2 is not compatible with the auto trickler set ups.
They do offer a data output option that I did not buy.
The price I bought at, $355 shipped, no tax was sort of a no-brainer at the time.
(scalesgalore). I compared it to the Creedmore Sports scale for near the same price.
Still a lot for a strain gage scale.

I agree that calibration points should be coordinated with the desired target weight.
If you are weighing 30 grain charges on a 100 gram scale, calibrated at 50 grams then don't expect good results.
UNLESS you CHECK near your target weight.
Proving scale performance is on the user. Your legal for trade banana scale at the market may rely on the state scale guy to come around once a year, but the reloader needs to KNOW how to verify performance.
For personal use you are not bound by government weighing regulations.

In effect you should be using the scale as a high resolution comparitor, not a full range measuring device.
Instead of automating the charge, invest in GOOD check weights that represent your application. Refine your techniques, refine your using environment.

Some scales, the EJ-54D2 included, even allow you to specify the calibration point.
If you want, you can set the cal point in the 22 gram low range to 1.0000 grams, or 2.0000g, or 3 or 4.
OR, you can use an appropriate Fake Zero to continuously monitor scale performance.
(Note the 1 gram Fake Zero in my previous primer measurements).
I might modify the primer sorting to cal at 2 grams and use a 1 gram Fake Zero.
Just checked a 1 gram calibration. Linearity above 10 grams to 20 grams pretty much is crap.
1 gram displays 1.0000, 20 grams displays 19.9974.

ASTM Class 1 mass standards are more than adequate for most reloading applications

https://www.uniquetek.com/store/696296/uploaded/calibration_weight_tolerances.pdf

 

[tom]

In other words, if I'm loading a large number of cases and trying for charge weights of precisely 31.0 grains and have all the weights on my .02-readable balance showing exactly 31.0 grains, what does that translate into in terms of the range I might expect in true actual weights across the large number of cases processed?

Not being a very "techy" guy....actually not even very smart, lol.

But I do have an fx120 with the auto trickle thing setting on a quartz slab, and a sartorius gd503 8" left of it on the same slab. In the real world it means adding up to 2 kernels and subtracting up to 2 kernels (h4895), when double checking on the better scale. That's roughly .08 grain error, and would be fine loading to the middle of a node. Not so fine if you're looking to take it to the last kernel where she shoots a dot, right before it blows.

If my 503 went down, I would spring for whatever model replaced it. Or I would load more in the center of the node. I already load more centered for heavy (10 shot groups) than I do for light (5 shot). I'm sure a guy may have to approach 20 shot differently yet again.

My take from reading some of this thread, and admittedly skipping some. Buy the best you can justify. And Benchrest is way cooler, lol. But the F class shooters I know work just as hard on tuning, despite being uncivilized, lol.

Tom

 

[Mulligan]

But the F class shooters I know work just as hard on tuning, despite being uncivilized, lol.

Tom

This is awesome!!
Tom, I agree!!!

The local F-class shooters work darn hard at the precision game. They sort most everything, with a handful of national level guys at nearly every match, they couldn't sleep at night knowing they left something on the table. The Ninja is at the range testing about every time I show up there. As uncivilized as they are, they kick some arse regularly at our 600 yard BR match's with 284's and variants.

CW