White paper on trigger press measurement.

[JCN]

https://www.nist.gov/system/files/do...12-16-2019.pdf

Came across this as I was trying to learn methodology of trigger weight measurement and it makes sense but didn’t occur to me that each cylinder of a revolver would be measured separately.

Thought I would share.

[OlongJohnson]

That's because the fitting of the ratchet teeth can have a significant effect on DA trigger press. If the teeth are too long, they squeeze the hand between the tooth and the frame window as the hand rises up along the side of the tooth. The extent of this squeezing and resulting excess drag (it may even be severe enough to be called "binding") will vary tooth to tooth, as the lengths of the teeth vary. I fitted the teeth on my M&P 340 until this binding/excess drag was just eliminated. Three of the teeth required work, the other two were just a hair shorter than I'd like due to the defective repair done by S&W customer service, but still serviceable.

The only other factor that changes by cylinder position is the cylinder stop rubbing against the cylinder and the notches. The greatest variation I've seen in that is the finish of the ramps leading into the notches and whether there is a burr at the edge of the ramp where it meets the notch. Not likely to lead to much difference compared to that caused by the ratchet.

In single action, the only parts that move are the same, and have the same movement, regardless of which chamber is up.


I found this interesting:

4.6.2 The standard deviation of the ten average values should be calculated for each firearm. For revolver(s) operating in double action, the standard deviation of the ten average values should be calculated at each cylinder position.

The average values are average of three. Assuming a normal distribution, n = 30 is generally accepted in the manufacturing quality systems I've been trained in as an appropriate sample size to reasonably accurately characterize variation. The outlined procedure starts with 30 samples, converts that to 10 samples by averaging them in groups of three, and then takes the standard deviation of the 10 average values. The standard deviation associated with a smaller number of samples is larger given equal/similar dispersion of values (by the way that SD is calculated), but the averaging will tend to damp out variation before it is fed into the SD calculation. It's beyond my expertise to predict how those differences will offset, but it strikes me that the averaging could damp out really quite a lot of variation if the variation is truly random.

I guess it is defensible if one takes the position that the methodology for reporting in "production" (i.e., case work) is to take three samples from a firearm and average them, and then apply the uncertainty value for the lab's measurement system derived from the outlined process to that single average of three measurements. It must be assumed that the variation is likely or primarily due to measurement process variation, rather than to variation in performance of the firearm. It also strikes me that much depends on the characteristics of the firearms chosen for the process capability study, and there is a built-in assumption that firearms of a particular general category will perform equivalently. My default assumption is that the latter is not dependable.

[octagon]

Interesting paper. I wish there was a source that used a trigger pull machine that graphed the entire pull weight from start to finish and shared the results across various firearms. Seeing how the pull actually measures shows if it is a rolling break, light take up and solid break, over travel etc... It would be interesting to compare TDA, various striker fired guns and unusual designs like the LEM to each other. Why no Youtuber, Gun mag or other hasn't filled the gap is surprising. The machines aren't cheap but not crazy expensive either.