Well, since this has degraded into a comparison of automatic loaders and scales and the fact that I am both an engineer and scientist I think I will chime in on the subject.
First off, I have no experience with the Autotrickler V3. But I will accept it is a good device. That assessment is based on reviews and forum threads. I will comment on the Fx-120i. This scale is excellent. But it is also misunderstood in relation to accuracy. The manual for the 120i gives a linearity of +/- 0.002grams and a repeatability standard deviation of 0.001 gram and a minimum weighing capability of 0.001gram. None of these are an "accuracy" statement in the classical sense. The linearity is what is referred to as trueness in ISO standards and is commonly known as a bias error. It represents how far from the true value a weight may be indicated. The repeatability is a precision error meaning what variation may exist when the same weight is measure multiple times. The minimum weighing capability is the amount of weight required to increment the display. It can only resolve .001 grams. To convert these specifications to something akin to "accuracy" as nominally stated in specification requires analysis. First is the combination of the trueness and precision to give a 95% accuracy. This is calculated as the square root of the trueness^2 plus (2xrepeatability)^2. This is ~.0028grams. To this value the minimum weighing capacity of 0.001 grams must be added for a final accuracy of ~0.0038grams. This is equal to 0.058 grains. Call it .06 grains. But since we are not concerned with trueness error per se because we are only concerned with precision the 95% confidence level is 2x standard deviation of precision plus the minimum weighing capability which is 2x.001g+.001gram or .003 grams. This is .046 grains or about 0.05 grains. Bottom line, the classical accuracy of the Fx-120i is 0.06 grains.
The ChargeMaster 1500 quotes an accuracy of 0.1grain. When stated in such a manner this means that if a given weight is weighed a large number of times 95% of the measurements will fall within 0.1gr of the true value. This includes linearity (bias), precision and resolution accuracy. Since the display is 0.1 grain the effects of bias and repeatability are much less (less than 0.05 grains) and the electronics of the scale work on a more accurate level not affected by the scale resolution. I have personally verified the 1500 accuracy on multiple occasions.
Now the accuracy of the dispenser! When the ChargeMaster 1500 first came out there were too major complaints about its operation. One was that it was too slow and the other was it overthrew charges. To understand the issues it helps to understand the operation of the dispenser. It uses a single feed tube that runs at multiple speeds to dispense the charge. The speeds run from high to trickle. The dispenser slows its speed as it approaches certain load points. When in trickle mode the feed tube rotates a portion of a turn and stops to allow the scale to stabilize and then turns again until the target is met.
The problem of overthrow was minimized by the now famous straw modification that used a plastic straw to provide a smooth outlet of the dispensing tube to allow less material to drop per trickle rotation. It was later found that a bushing would also work. One of the more ingenious methods was to use the neck end of a 243 case (my method).
The speed issue was addressed by programming changes to speed up the loading. Unfortunately, doing so also increases the potential for overthrows. However, the same parameters can be used to optimize the trickling.
As it turns out, when the two methods (programming and bushing) are combined and optimized to the powder being thrown test have been published that show the 1500 is capable of throwing charges to the 0.05 grain level (verified by Fx-120i level scales).
This should not be taken to mean that the 1500 or other 0.1gr auto throwers will equal a V3 but it does mean that they are capable of 0.1 grain or better.
All good info. Great info really. I don't know if you were referring to me or not, and am by no means picking a beef with you. I'm just posting what's below to clarify what I was getting at in my post about the video.
My response to the video had nothing to do with the accuracy of scales or the ChargeMaster vs AutoTrickler. It had everything to do with the velocity variation via loading to a precision of approximately +- .1gr compared to a precision of approximately +-.03g(those two numbers were the load variation given by the poster of the video as measured by the FX-120i). It just happens the video comparing a ChargeMaster and AutoTrickler contained what I considered to be excellent data regarding a comparison that I was discussing. I personally believe that loading to +- .1gr to be good enough, while another person is claiming that his results could only be achieved by loading to the nearest 02gr. Those two stances are reasonably close to the loadings compared in the video, so I thought it was highly relevant.
I only posted the P-value of the comparison of velocity
variation in his reported shot strings according to his reported velocities. I did not report in my earlier post the P-value of the absolute difference between both data sets, only the variation. The P-value of the absolute difference between both data sets was .0636 IIRC, which is close to what would commonly be considered significant, but A) that's not what we're testing, and B) we know why that difference exists, and would expect it to be significant. The difference existed because charges thrown by the AutoTrickler V3 were heavier than the charges thrown by the ChargeMaster and that resulted in an average velocity 4fps higher for the AutoTrickler. Again, I was not attempting to address the difference between the two powder measurement methods, but rather the difference between +-.1gr precision and +-.03gr precision. As such, only the velocity
variation matters, which gives a P-value of .405, which means that there was no discernible difference between the two data sets. Said another way, loading to +-.1gr and to +-.03gr yielded statistically identical SD and DS.
There is likely some measurement error in his charge weights(all of which he claimed that measured on the FX-120i, so really what thrower he used to get there is immaterial). There is also likely some measurement error in his chronograph. I only mention the measurement error to support the stance that loading to .02gr is unlikely to be substantially important. I'm not convinced that anyone is actually accurately measuring to such precision, I'm not convinced that anyone is accurately measuring velocity to a precision that would show such differences, and I'm not convinced that if such precise powder measurement did in fact occur that the velocity spread would reflect the improvement enough to justify it. Load data suggests that for a 6.5CM .1gr of powder at actual charge weights is responsible for roughly 5fps, which is 70% of the 7fps that dividing velocity by charge weight suggests. If you go from +-.1gr to +-.02gr(which I'm not convinced anyone is doing regularly). That's 1/10th as much variation is powder charge. That would suggest that you could cut your ES from 10fps to 1fps, but that's not what happens. In the video, if we're to believe that the velocity spreads he experienced were due to powder charge variation, then when he cut his powder charge spread to 1/3 of what the ChargeMaster through, you would expect a reduction in velocity ES to 1/3 of the original and you would expect a P-value to be .05 or less, but instead he only cut it from 16fps to 13fps. He cut the velocity spread to 4/5 of the original instead of 1/3, and had a p-value of .405, which means that the reduction in velocity spread was impossible to differentiate from random effects.
In guns for which I do intense brass prep, and load development I usually get single digit SD with near single digit ES. I wish I saved more of my printouts, but I came across one that had an ES of 9 snd SD of 4. That's not the only time I've had such results. All load loading of my ammo is done by throwing about .4gr under with a Redding powder thrower, and trickling up to weight with a home made trickler my grandpa made. I weigh the charges on a $20 Frankford Arsenal scale from Cabela's. If the scale is .1 under, right on, or .1 over, I dump the charge in the case and move on with life. If someone wanted to loan me an AutoTrickler and see if my SD/ES would drop to near zero, I'd be happy to try, but I currently have to reason to believe that measuring powder charge to any more precision than I currently do would result in meaningfully improved results. At the price point of a ChargeMaster, I'm definitely interested in an upgrade at some point.
for us not well educated, what is "edjucation"?
