All:
Thanks for looking, I'm trying to determine the most likely path to go down.

Problem statement: "Patterning" a shotgun involves shooting at paper targets (or a steel plate with wet paint) and counting the number of pellets (~350 per shot) in a 30' circle, the center of the pattern is also determined by estimation. Since a series of several shots over multiple distances is required, this is a very time consuming task. It is also a 2 dimensional view of the shot. Shot actually strings out and impacts the paper at various times.

Proposed solution: Have the shot impact a plate and determine the location and time of the impacts. Calculate the center and other metrics.

Unknown: How to capture the precise time and impact location on a 3' by 3' surface without the sensor(s) being destroyed by the impact.

Research so far: It looks like Piezo Vibration sensors are used by the space station to measure impacts. However it looks like there would have to be a lot of them since the ones I've seen so far are fairly small.

thanks for any direction you can provide.

 

How about an optical sensor, aimed perpendicular to the shot-stream path, and calibrated to count the number of pellets and their time sequence, as they interrupt a vertical slot-aperture infra-red beam located opposite the sensor ?

I'll duck away now, before someone asks me to be more specific about this concept.

 

Nothing beats being able to see what you're doing. How about high-speed video?

I think if you want a mechanical sensor (eg. piezo), you're going to need to define the impact energy it must survive, as well as the impact energy that is typical to an event to be counted. You'll also need to define your 2D resolution, like pixels per inch.

 

I doubt that a vertical slot would be of much help, you would get only the "density" through time, plus optipcal sensors are quite slow.

opfats: do you have some time range in mind? Like how much time is between the first pellet hitting the target and the last?

 

Here is some reading that may help you decide what you need to do and use to determine how to approach the problem.

http://www.labsmith.com/applications/LabSmith_Timing_a_Ballistics_Experiment.pdf

 

Thanks for the replies everyone. The time is very short in the 1/100 of a second to almost instantaneous. The time perspective is a like to have, not a must have.
For the high-speed camera to work, it would require image processing to be able to do the counting and location. Is there something out there that would do that or would I have to write that kind of software?

 

There are probably image processing libraries that could be used to do most of the work. You can probably help it along by using high contrast targets, such as bright orange so that the holes appear very dark by comparison.

There are a number of acoustic solutions for single-impact events that have been put together. But if they have to deal with multiple hits within their sensing window, then the processing may become extremely difficult if not outright impossible. The same is likely true for optical methods. If you have enough sensors, you can almost certainly resolve things, but it wouldn't be fun.

With the image processing approach, you could avoid having to change targets each time by having the image processor work with a differencing algorithm. Of course, any pellets that go through prior holes will get missed. But you could play a number of games by having two sheets of paper on top of each other and, after each shot, having a mechanism that slides one of them slightly relative to the other. This would make most of the holes vanish and would let you get quite a few shots before you had to change targets.

 

Just to think outside the box here, what would be wrong with putting a big piece of carbon paper over a steel plate, and then scanning the image after a shot? Then software could easily do your image analysis.

Scanning a paper target with holes in it might be difficult, so that's why I think the carbon paper and steel plate might give you a more precise image when scanned.

Obviously this low-tech approach loses any time information and will also lose data when multiple pellets hit in the same spot.

 

A target made from different plates in the shape of the standard patterning shape with individual pressure sensors for each segment of the target. Image of the standard type target for testing; http://www.shotgun-insight.com/ShotgunInsight_OnePageSummary.html

While this wouldn't tell the number of actual pellets, it would, by analysing the impact force give the concentration of the pellets. And that is what is being done by counting the pellets.

The time of hits and actual pellets per shot is a ever changing thing, from shot to shot. As no two shells are exactly the same. So what will that tell you in measuring those things?

 

A target made from different plates in the shape of the standard patterning shape with individual pressure sensors for each segment of the target. Image of the standard type target for testing; http://www.shotgun-insight.com/ShotgunInsight_OnePageSummary.html

While this wouldn't tell the number of actual pellets, it would, by analysing the impact force give the concentration of the pellets. And that is what is being done by counting the pellets.

The time of hits and actual pellets per shot is a ever changing thing, from shot to shot. As no two shells are exactly the same. So what will that tell you in measuring those things?

This is probably your best option. I would put some thought into the location, shape, and number of impact sensing regions, as that will be the limitation of your resolution. If you use piezo, be sure to mechanically isolate the plates from each other.

Some creative software engineering can tell you shot location by percentage of total shot.

Determining the length of the shot string, while achievable, will at least quadruple price and complexity. A separate high-speed camera and a scale is one way.

 

Assuming a shot is on the order of 1/10" how about an array of photo emitters and detectors? You could put a Y array just ahead of an X array to get a close approximation of the spread on a completely reuseable target.

To cover 30" you would just need 600 of these, 300 for X and 300 for Y.

I would imagine several FPGAs (several to get enough pins) would be needed to capture the data, plus one more FPGA or a (dirt slow by comparison) micro controller to process the results.

 

Wow, you folks are great! These suggestions give me several good approaches. The reason I was looking for the time was that some after-market chokes claim to change shot string length.

 

I'm with Ernie. The very same solution occurred to me while I was doing laps at the pool. No moving parts, nothing actually getting hit by pellets, essentially instant feedback of results, arbitrary precision (albeit with increasing cost for higher precision). Time data if you need it. What more could you want?