Hacking the CP-2

[George Schmermund]

The hammer sensor is holding up fine so far, but the real test will be with the powerplant charged up. I've included a few DSO images to show what's happening in these first runs sans CO2.

The images include some shots with the DSO's digital low pass filter set to 5 KHz. It can be turned on and off using the same data set. There's an extra hammer shot to illustrate the variance that can occur from shot to shot that makes it impossible to predict the muzzle movement during these forced oscillation periods.

I'll try to get some tests done with everything under power later this afternoon. I'm very interested in seeing what the P/T curve looks like with this ensemble.

[WhatUPSbox?]

George, I was trying to understand your hammer sensor data. Have you taken any datasets with a longer timeframe? I was looking to see if the low speed motions of the hammer were present. When I was monitoring the hammer motion on the 2240 there was about 10ms of hammer flight time. The CP-2 should have something similar. During that time, starting with sear release, there should be a low frequency acceleration that starts high and may or may not hit zero before the valve is hit. Then a deceleration until the hammer bounces off of the valve. Without CO₂, there may be several bounces. Noise from bouncing on the tube wall would be on top of that.

[George Schmermund]

Stan - I'm sure that the smaller signals would be detectable if the gain on the DSO was set higher. The noise floor on the 2250A is less than .002 g RMS. The shock limit is 2000 g peak. The stated calibration range of the device is +/- 500 g. The scale I setting the DSO to is 800 g full scale. Some of these hammer impacts are approaching full scale. Without the filter in place the signal has high frequency content that swamps the sensor's amplifier and causes massive clipping at 15 Volts!

[George Schmermund]

Here's a cornucopia of images with the powerplant in operation. These were all done with the 5 kHz filter on duty. I'm surprised that there was no indication of hammer bounce extending past the P/T peak.

Channel 1 is the hammer sensor.

Channel 2 is the muzzle sensor (only one of the sensors is being used for these tests).

Channel 3 is the pressure sensor.

Channel 4 is off duty.

The key to deciphering most of this info is at the bottom of the grid frame. The channel voltage/grid segment values are listed along with the time/grid segment. The channel being used as the trigger is in the lower far right. Stan was asking about the hammer sensor's sensitivity. It's output is 10 mV/g. Actually, all of the 2250A accelerometers I'm using have the same nominal output rating.

These measurements are all up for interpretation at this point so have your way with them.

[WhatUPSbox?]

George, These are great. Definitely lots to chew on. Have you made a shot without a pellet? It would be interesting to see how the muzzle responds without the pellet excitation.

[George Schmermund]

Stan - Here are a couple of shots w/o a pellet.

[WhatUPSbox?]

George has collected and posted a wealth of data on the shot cycle and vibration response in the CP-2. To help me try to understand  what is going on, I took his graphics (from posts 363 and 365) and overlayed them onto a couple of charts. Since these are his results, I checked with him for his OK before posting. The charts are just an overlay of several runs indexed to roughly the same starting point on the hammer accelerometer curves. These are just graphical overlays, no numerical processing was done. I post them because they helped me understand a bit what was going on. Note, sorry for the artifacts left over from setting the transparency.

The first image overlays the three pellet shots (1-DS0004, 3-DS0010, 4-DS0011) I also added the dashed blue line to approximate the period when the pellet is moving in the barrel (from George's previous data). To me the repeatability of the traces, including the hammer impact event was something I didn't expect.

In the second image I added an overlay of the shot without the pellet (2-DS0003). For this I switched the colors of the hammer and muzzle traces to provide some contrast.

In reading the resulting tea leaves, It looks like the initial muzzle response is very similar whether there is a pellet there or not, and then after about 1 ms the pellet causes additional response at the muzzle. I also expected the pellet shots to have higher amplitude than the 2-3X seen. In fact the no-CO2 shot (5-DS0011) have similar muzzle response amplitude to the pellet shots.

No clue if my interpretations tie to reality but it is fun trying to figure out what's going on.

Thanks George

[George Schmermund]

Stan - That's some interesting work that you've done. Thank you for the effort. From the look of the overlays there seems to be something forming as a possibly repeatable pattern in the shots that you chose. The illusion is due to the 2 dimensional data that I used to test the hammer sensor's signal and just one of the muzzle signals at the same time.

The DSO images shown in this post are using both of the muzzle's 2250's being recorded simultaneously and with the 5 kHz filter in. As can be seen, they present a pretty wild ride after the first 1.5 ms. These shots were taken without the hammer signal included because the sensor got hit one too many times while working the bolt handle. That caused the lead wires to break. I'll confess to not having thought out the experiment well enough to realize that the access slot was milled on the wrong side of the powerplant tube. That can be corrected.

Another interesting part of these images is that the 2 muzzle sensors are producing signals that are 180º out of phase on some of the major peaks. This is probably do to the muzzle rotating and the fact that they're frequencies are different as shown in the Lissajous patterns. For curiosity I've inverted channel 4 in the 2nd image. Also,as a check, I tapped the top of each sensor to assure that the direction of displacement and output signals were both in phase.

[George Schmermund]

Stan has pointed out something very interesting with his overlays. The overlays showed what might be a pattern in the first part of the shot cycle time window when the forced oscillations are running the show. My point of difference was based upon the fact that only 1/2 of the information was available in the images that I posted. Subsequently, a DSO image of both muzzle sensors recording simultaneously was posted to illustrate the real complexity of these vibrations.

My assumption up until this point was that the situation was approaching chaotic randomness until the barrel's natural frequencies took over. This might have been true if we consider that most systems which are very sensitive to the initial conditions will lean towards chaos. After some thought about the overlays it was clear that the actual initial conditions were very close to reproducible because the action to produce the forced oscillations is pretty much the same force applied repeatedly to the valve's stem and face every time.

This new notion got me going onto another experimental set up. Since the DSO wouldn't do any filtering when using the XY mode of acquisition, I decided to employ a Krohn-Hite 3202 Dual Channel Tunable Variable Filter. This instrument allows both signals from the sensors to be filtered identically before they're passed on to the DSO. I also decided to lower the low-pass upper limit to 2 kHz. This cleans up the images and uses a more realistic frequency range where the displacement is concerned. Using the dual filter also eliminates any phase shift between the signals.

The results were enlightening and I want to thank Stan for his sizable contribution to my understanding of what these measurements can reveal.

I'll try to get some images posted tomorrow. The new Lissajous patterns contain much more information now that they're filtered. 

[George Schmermund]

Here are some XY plots of the pellet's potential position after the shot cycle begins. The first image is from the naive days of using pure sine waves to test barrel harmonics. This image was posted already some time ago. The markers are positioned selectively at a crossover point that would represent the face of the muzzle at rest. The second marker is placed at 2.86 ms from the first marker to illustrate where the pellet might be at that time if the barrel was truly in resonance at both it's vertical and horizontal frequencies.

The next 3 images are the real patterns of the barrel's vibrations during the full shot cycle at the muzzle. Each is marked with overlapping points at 100 µs intervals on either side of the time when most pellets leave the barrel. The lead or lag timing of the marker in these images could easily be real by this amount of time for several quantifiable reasons.

The time position of both cursors can be seen in the YT trace at the top of the screen. The Krohn-Hite filter was in play for these last images.

The next interesting experiments would be to delay the trigger time in order to capture just the window on either side of when the pellet leaves the muzzle. This could be done by using the already reliable muzzle contact wire.

[George Schmermund]

There was a thread somewhere along the way where someone was talking about testing the hardness of the lead that they were using to cast pellets/bullets with. I got to thinking that this would be a good thing to distract myself with while I work on the design and construction of an accelerometer calibrator.

Being able to test the hardness of the alloys being used by different manufacturers would be interesting along with the alloy's composition. This alloy notion was brought up back in the days of the Vigilante hacking thread. ( https://www.gatewaytoairguns.org/GTA/index.php?topic=118339.60 ) It was post #70 and the experiment showed that a piece of solder was easy to analyze with some simple stuff found around the shop. Pellet alloys would be just as easy to analyze, but I got distracted back then, as usual, and never got around to reassembling the experiment.

Anyhow, now that I've wrestled this sliding microtome out of the attic I might as well put it to good use. I'll have to say, I was a younger man when I put this beast up there. Somehow it acquired significantly more mass during it's stay.

The first testing setup to confirm that it still works was done on a .22 wadcutter that was just clamped unsupported in the mounting vise. I wanted to make sure that the blade was up to the task. Well, it didn't even know that the pellet was there. The slicing thickness can be set from 1 to 40 microns per slice and automatically advances the sample up by the assigned thickness on each stroke. It's really an amazing instrument to operate. The blades are as sharp as you can get steel and feels like you can almost get cut just staring at them! This instrument uses a 10" blade which is quite large as bench microtomes go.

The real cross sectioning will be done with the pellets mounted in a hard wax and then sectioned properly. These sections can then be etched to see if the cooling rate or other factors involved in the manufacturing process have any significant effects. This will allow me to use some of my other metallurgical toys. Idle hands are the devil's tools!

The photos are for those who may still be following my "Mr. Toad's Wild Ride" down into the rabbit hole. Disregard any cat hair that she somehow manages to sneak into my photos. 

[WhatUPSbox?]

Wow George, I think I'd keep that in a display cabinet. Looks great...waiting to see how you'll put it to use.
Won't ask what tests you were running on Fluffy that needed a 1 micron shave off the top.

[George Schmermund]

I've decided that the sliding microtome was not the best match for slicing up pellets. The next choice was to go with a rotary microtome. It's scaled much closer to the job. Fortunately the collection of these instruments has grown over the years and fulfilled an assortment of applications. There are some things that you might want to see the inside of without having to use x-ray imaging. Of coarse the x-ray imaging is nondestructive in use, but for pellets, physically slicing them open affords the best view of the interior and they can be sacrificed.

My interest at the moment is to see what the cross sections look like. There are lots of pellet photos out there, but very few of them are inside views with much detail. A good cross section will allow the sample to be chemically etched to give high contrast to the grain structure for comparison to other samples. If nothing else they should be interesting to look at. It doesn't take much to get me chasing my tail with another experiment!

The .22 cal. wadcutter in the image is only tacked onto the sample holder for a sense of scale compared to the sliding microtome. To do a proper sectioning will require that the pellet be embedded in the appropriate mounting medium.

[George Schmermund]

The accelerometer used in the hammer experiment was short lived because of my lack of foresight about which side of the main tube to mill out the access slot on. Since I'll have to take the gun apart to move the slot onto the other side of the bolt, The idea of replacing the accelerometer was worth considering. After rummaging through a few drawers I found an Endevco model 22 device.

I thought the 2250A was small. This on is claimed to be the world's SMALLEST! It weighs in at 0.14 gm! There won't be any mass loading issues with this flyweight. The next thing that needs doing is to see if it still works. The 2250A has a built-in preamplifier that requires a current source signal conditioner to get the signal out to the measurement system. The model 22 needs a charge coupled preamplifer to get the signal out. Fortunately I'm well stocked in both requirement.

The specs on the 22 are excellent. The amplitude linearity is 1% out to 4,000 g. The frequency response is +/- 1dB from 3 to 12,000 Hz. I'm looking forward to getting the gun back on the test bench.

The image is for comparison of the 2 sensors.

[George Schmermund]

I got a chance to check out the new micro accelerometer and it works to spec. This pushed me to start the machining for the hammer to be instrumented on the other side of the bolt. The photo shows the new slot and hammer in place. I'll still have to machine a new shelf for the sensor on this side of the hammer.

The rifle barrel will probably be the next project for new vibration testing now that the pistol arrangement has be overly measured at this point. The rifle barrel will need to be machined to accommodate the pressure sensor.

[George Schmermund]

The new accelerometer's ledge was milled into the hammer and fit tested today. Progress is slow, but little steps will get this project back onto the test bench. Tomorrow I may be able to mount the accelerometer and also machine the pressure sensor port into the .22 cal rifle barrel. 

[George Schmermund]

The .22 cal rifle barrel was drilled and tapped today so that the pressure sensor could be mounted. The photos show the barrel in the lathe getting the front sight turned down after the blade was bandsawed off. The breach was then deburred after the drilling and tapping operations and polished using the Dremel tool.

Next is to reassemble the gun and make some measurements with the new smaller accelerometer installed on the hammer and with the longer barrel.

[George Schmermund]

I finally got around to finishing the machine work on the carbine version of the CP2. This included the milling on the pistol grip to accommodate the new hammer and accelerometer position on the side opposite the bolt lever.

The gun was reassembled without using the barrel band and tested with some RWS wadcutters at the 15" target in the machine shop area. These guns are really impressive for the price.

Installation of the new micro accelerometer for the hammer was abandoned because of the precarious mounting necessary to install it safely for such hazardous duty. The previous 2250A device is better suited for this job and is safer now that it's terminal wiring won't get clobbered by the bolt action.

The next tests will include vibration testing with and w/o the barrel band along with timing info for the full shot cycle. Another pressure sensor at the muzzle is in the works and should provide some interesting numbers along with the exit timing.

[George Schmermund]

I shot 10 pellets with the new mounting and everything stayed in place. Next is to attach the signal wires and make some measurements with this longer barrel.  I'm trying to weave this project into something less protracted, but Ive got a lot of irons in the fire at the moment.

I've become preoccupied with designing and building new versions of the earlier accelerometer and instrumented hammer boondoggle. That and a bunch of other things.

The ballistics experiments are still fascinating, though.

[George Schmermund]

Well, It's the end of another year. I decided to use today as a reminder that I'm somewhere in the middle of a barrel test on the CP2 carbine. I just need to attach the new hammer/accelerometer arrangement up to the preamp and get a few DSO screen shots posted. It's amazing how time flies by when there are other distractions.

Speaking of which, the instrumented hammer and accelerometer project has made considerable headway lately. The parts and methods of assembly are closing in on something resembling standardization. This includes a small shaker for calibration and small structure excitation table. There may be a market for these devices as an entry level front-end for a modal analysis systems. If not at least I'll have it all for my own experiments.

Part of the above system includes a force transducer that gets built into the impulse hammer. I've posted several variants of these sensors in the "hacking" threads and how they can be used to make an assortment of measurements. Now that I've arrived at a somewhat standardized force transducing device it would be interesting to use one as a "ballistic pendulum" sans the pendulum part. The output signal from these sensors is very reproducible and could be useful for measuring a pellet's energy at the muzzle and at a target some distance away. There could be some very interesting info relating to energy loss vs distance if it's done right. There would be some challenges in the design and construction, but that's what a New Year's project list is all about!

This is my second full year of experiments in the airgun world and I want to say "Thank You" to all of the people who have been keeping track of these experiments for all this time. I know that the replies have been few and far between this year, but the thread doesn't seem to be withering on the vine, so I'll keep posting as long as folks stay interested in reading about my perpetual follies!