So, we raised our eyebrows...and decided to test both chronos, his and mine, together

Last sunday, we mounted the Chrony on top of the PACT PC2 he owns, in order to take simultaneous readings and check the differences.

Surprise!! both were no more than 1 or 2 FPS between them!!

I would like to know if other colleagues have found similar experiences.

PS: we checked the hot .300 RUM load again, and it was OK...but too hot!!!!

Regards, Gustavo

quote:

---

Originally posted by denton:
I'm not at all surprised by your result. Modern electronics can do the counter/timer function very cheaply, and very precisely. The crystal clock in mine is 12 MHz, which allows the 1 foot distance (1/3 msec) between sensors to be resolved to about 1 part in 4,000. Thermal drift for a cheap, non temperature controlled crystal is about 10 ppm per degree C, so figure 350 ppm (.035%) for going from freezing cold to too hot to want to go out. Even inexpensive chronographs do a great job... as nearly as I can tell, the main difference in precision between really expensive and very economical is a slight difference in the sensors, and the amount of dough the owner parts with. If you want the stat routines, it's vastly cheaper to run them in a spreadsheet at home.

---

Denton do you have at hand these routines ? Tks!

If you need the formula for standard deviation, it's (sum(Xi-Xbar)^2)/(n-1))^.5.
Xi is all your observations, Xbar is the mean, and n is the size of the sample. So you find the mean of your sample, take the difference between it and each of your observations, square the differences, add them up, divide by n-1, and take the square root of the whole mess.

For samples of 10 or less, the range (extreme spread) is almost as good an indicator of spread. Standard deviation can easily be estimated from range. If you shot 2 cartridges, just divide the range by 1.128. For 3, divide by 1.693; for 4, 2.059; for 5, 2.326; for 6, 2.534;, 7, 2.704; 8, 2.847; 9,2.970; 10, 3.078. You can use this technique for getting the standard deviation of group size, muzzle velociy, or, if you've got strain gauges, for chamber pressure.

T tests are good for testing whether something you did really changed something, or if the change you're seeing could as easily be explained by random variation.

Regression lets you plot an input variable vs. an output variable. For example, I use it a lot to plot muzzle velocity as a function of grains of powder.

If you want the formulas for these last two, I'll have to copy some material out of a text and email it to you.