Hacking the CP-2

[WhatUPSbox?]

George, getting the valve open duration would be great data. Earlier I missed the the fact that you had the brass shim on the valve body, I thought it was on the stem. I don't know if you have some insulation on the hammer side of the shim so it is forced to ground through the valve stem rather than through the hammer. Ultimately, you could have two pieces of shim, one on the stem and one on the body to get both the start and fully open points. Unfortunately the shape of the hammer face makes that a little trickier.

I've been thinking about the pellet start measurement. Other than drilling portholes into the barrel for an optical measurement, an insulated bolt with a wired tip (similar to my original skewer hammer probe) may be easier. 

Great work as always

[George Schmermund]

Things are getting more interesting with these new measurements. The photo shows the basic setup for finding the break free time on a pellet in the breech. I'm using the skewer method that Stan used on the other end of his pistol to get the hammer information. This test uses the same ohmmeter switch circuit that was used to get the valve timing info in a previous post. A piece of wire wrap is at the end of the skewer and it's pushed in until it makes contact with the end of the pellet and closes the circuit. It's then backed out until the circuit just opens. The pellet only has to travel a few thousandths to make contact. This arrangement is very reliable and surprisingly durable. I've used the same stick and wire for several shots and haven't had to replace anything yet. This test method obviously has it's hazards but the metal cylinder and trap keep things under control.

To synopsize the DSO image I'll just point out that the pellet starts moving 375 µs after the valve begins to open. The pressure behind the pellet at skewer contact is 284 psi. The timing for this experiment is based on the valve actually starting to open and not when the hammer strikes the valve stem or bottoms out. That information can be gathered as a separate experiment.

When we get further along with all of the test data there will be a set of Lego blocks that can be used to put the whole time sequence together to make one clear picture. I think these smaller parts are much easier to digest.

[WhatUPSbox?]

I like your probe approach. I was looking at a very thin probe (only weighed 4 grains) that the pellet could push to initiate the transition time with a laser sensor at the muzzle. Your approach can time not only start of pellet motion but potentially the acceleration down the barrel by moving the contact point outwards in sequential shots.

Interesting, the previous (213) pressure curve achieved peak pressure in about 500 micro-sec, in this curve it is more than twice that

[George Schmermund]

Thanks for catching the discrepancy, Stan. I'll redo the tests and see what's going on with the measurement numbers. It must be the new DSO. I'm sure that it can't be the beer.

[George Schmermund]

The only thing that comes to mind as a possible reason for the difference in the 2 curves is that the stick and wire are causing the effect. If you look at the probable time of exit on the #213 image it would be about twice as long.

I did a few shots this morning using a 4 event average from the acquire mode. The averaging does some nice smoothing on the P/T curve as shown in the image. I'm confident that the system is working properly.

The incremented timing for acceleration measurements is a good idea except for the air plug in front of the pellet. I doubt that it technique would give real world timing info unless the stick and wire influence could be reduced to a very small number. I was thinking about using a straw in place of the stick in order to reduce the effect of the barrel air column as the distance between the trigger wire and the pellet increases.

Your 4 grain probe is interesting. It would have to be very stiff with a small cross section. I suppose that if you could place one end into a hollow point pellet to keep the probe centered and then shoot it straight up it might work. Can you give us more details about your plan?

[WhatUPSbox?]

Interesting, I was guessing that the pressure system didn't know about the probe&wire until the pellet moved, and yet the 221 curve in 375 micro-sec gets to 2.84v at the time the pellet moves and the 224 curve gets to about the same voltage in about 250 micro-sec and the pressure rise is about the same on the 213 curve.... interesting.

For the incremental method, a thin gauge (.020?) solid wire may be sufficient and would block a minimal portion of the cross section. The straw approach should work as well. The one I tried was just a little too big for a .22 barrel....I'll order a few more manly umbrella drinks to look for a smaller straw.

For the lightweight probe, I planned to use some high tech garden stuff. We have some Mexican grass tree https://www.monrovia.com/plant-catalog/plants/815/mexican-grass-tree/ specimens. When dry the tip sections (~.070") weigh about 4-4.5 grains for a 7" section for the 2240 barrel. I don't need them to survive, I was just going to use them to trigger the start of motion by unblocking the laser trap at the muzzle. After that it can come apart. The pellet itself would have its normal timing interval going through the trap. I figured a 10 gn wadcutter  + 4.5 gn probe = ~ 14.3 gn CPHPs. All that said, I like your blocking probe better.

[George Schmermund]

I needed a diversion today that would allow me a play session with the new DSO. The decision was to drift back to some of the barrel harmonics experiments. Since the measurements were already done to find the natural frequencies of the mounted 8" barrel it was easiest to start there. 2 sine wave generators were set up to reproduce the 2 frequencies that the accelerometers recorded. The information about those experiments is somewhere in this thread.

What I want to show in this post is the combined output of the sensors. This is real world stuff and not spread sheet voodoo. When an impulse of sufficient force, whether by firing the gun or striking it with an instrumented hammer, is injected into a cantilevered barrel it will resonated at specific frequencies. As demonstrated in the previous experiments the CP2 breech mounting allows 2 distinct modes of vibration. I doubt very much that this is a condition unique to this gun.

Anyhow, these 2 harmonics can be accurately simulated by the 2 signal generators as mentioned above. When these signals are combined simultaneously and plotted on a Cartesian coordinate system a Lissajous pattern can be drawn to show where the barrel is pointing  at any given moment in time. To interpret the graph assume that you're facing the end of the barrel of an unloaded gun that has just been struck be an instrumented hammer.

The images show a lot of good information about time and the positions of the muzzle activity. Across the top of the o'scope's screen is a time vs amplitude trace of the 2 sine waves as would be normally seen using any conventional o'scope. It's obvious that the 2 signals are out of phase because they're at 2 different frequencies. The lower square display is what the 2 signals look like when they're combined to show an amplitude vs phase angle plot. I've chosen to show only a 10 ms window of these signals for clarity because the plot gets much more crowded and confusing with longer time intervals. The cursor markers in the first image can be seen at each end of the 10 ms window of the Lissajous pattern.

The 2nd image shows the markers at the start and exit time of the averaged pellet shots from the previous post #224, i.e. 2.86 ms.

Please keep in mind that the magnitude of these Lissajous patterns are exaggerated in these examples, but the patterns are very real. Also keep in mind that the 2.86 ms window can be moved around to any position on the pattern.

 

[WhatUPSbox?]

George,
If I understood correctly, what you plotted here is an example using a function generator source. In the real case, those will be replaced with your accel outputs. How does one get from the accel plot to where boresight is pointing?

[George Schmermund]

I guess that the previous post was a bit cryptic in it's presentation. I'll try to demystify it. First we can go back to post #166 and see that the two accelerometers have well defined high Q peaks at 104 Hz and 164 Hz. These are the real signals from the barrel at its resonances which are simultaneous, but 90º apart. Since we now know these frequencies from the gun, it's an easy matter to use a pair of signal generator to do a bench test and time window the measurements with them. Stan, this method is a juxtaposition of your comment "In the real case". The signal generators are now replacing the accelerometers for doing these measurements.

The starting point for both sensors is before the shot is fired which would mean that the end of the barrel is at rest and there is no signal from either sensor. In this bench test we're running both of the generator signals continuously, but they can represent the barrel at rest when both signals cross the zero voltage point at the same time. This is a fleeting moment, but neither signal has any magnitude there. The only time that the 2 signals (or the barrel) come to the zero or rest point again is about 32.6 ms later. This is long after the pellet has left the muzzle. The 1st image shows what the 2 sine wave look like in a conventional o'scope mode. The 2 cursors show the time of the zero crossing points.

The 2nd image shows the XY mode of the 2 waves as they cross the zero point where the barrel is at rest. The time vs amplitude trace at the top of the screen shows the crossover point of the 2 waves at zero magnitude. The XY plot shows both the #1 and #2 cursors on top of each other at the zero magnitude point for each wave. The scales on each side of the XY plot represent the plus and minus values of each of the waves. the yellow Lissajous pattern is the continuous time line for both waves.

The 3rd image shows the same traces at the top of the screen, but with one of cursors moved from the barrel at rest point to the 2.86 ms point. That's when the pellet leaves the muzzle and also where the barrel is pointed at that moment.

I'll reiterate that the magnitudes of barrel displacement are exaggerated, but the changes of position in time are real. I also find it interesting that the barrel will never cross the "at rest" point again during the shot cycle.

I hope this helps in the interpretation of the images.
     

[George Schmermund]

The instrumented hammer is becoming a preoccupation. One of the requirements for proper use of the device is to be able to calibrate it. I've been going round and round with all of the possibilities. Most of the standard methods that are easy seemed to fall short in one area or another. While reflecting upon the situation during a beer induced meditation I was able to recall that there was a resiliometer up in the attic. Hastily untangling myself from the Yogic Sleep pose I managed to limp up the ladder into my hardware cornucopia and retrieve the apparatus. It was just how I remembered it, though not exactly were I remembered it having been.

Anyhow, the plan is to use it in reverse of it's intended application and measure the impulse force instead of the rebound height. This procedure gives me a calibrated weight impact on the hammer and a DSO capture of the event. I tried a few drops and it really works well for this repurposed application.

The photo shows the setup.

 

[WhatUPSbox?]

George, Thank you for the extra clarification on the harmonics plot. I guess my question was more on the relationship of the measurements to changes in boresight. My understanding is that boresight error at the time of pellet exit is the slope of the tip of the barrel relative to the breech and possibly any lateral velocity the muzzle may have at that time (I need to pull out the dreaded spreadsheet to see if the second one is negligible). Your configuration provides acceleration data. To get to lateral velocity and/or slope, some time history is needed. It is not clear to me what the time history should be given the hammer and pressure rise sequence. To get the slope some mode shape consideration comes into play.

All very interesting. I was wondering if an optical test of boresight could be done, perhaps using something like a quad-cell (expensive).

[WhatUPSbox?]

I did a little more exploration with the 2240 setup. In the spirit of you can't have too many wires, I added a few (image 1). The setup has the laser gate on the hammer probe set to trigger on valve impact (rather than on hammer release) this is channel 4 (dark blue) in the traces. The microphone is still attached to the tube near the valve (channel 1 - yellow). I added a laser gate near (.25") the muzzle (channel 4 - purple). I also included the broken wire at the muzzle (channel 2 - light blue).

Image 2 shows the overall trace. The hammer probe and the mic are in pretty good agreement in detecting hammer/valve impact. With the valve pressurized, I did not see any of the bounce on the hammer probe.

Image 3 shows some of the timing details. The valve duration was typically just under 2ms, pellet exit at about 2.7ms. I did have some additional trips on the laser gate at the muzzle. The powerlet was new and there was probably more CO₂ spray than before. The pellet preceded the additional trips as you would expect with CO₂ spray. I'll have to go figure out what the large signal is in the mic at the tail end of the sequence, it is at about 350hz.

Next is to go try some of the options for detecting pellet start to complete the sequence.

[George Schmermund]

Very impressive measuring arrangement, Stan. I hope your wife lets you put it up on the mantel when your done testing. Your DSO images are becoming  intrepretable with your new sensor array. The screens are much easier to decipher now that I'm looking at pretty much the same displayed info on my scope, it's just in different places on my screen.

The broken wire detector signal is certainly clean. How have you removed the hammer travel to get the valve info?

The microphone appears to hear 2 different impacts. Are they the stem and then valve's face? The larger slow signal may be the mass/spring system of the mike and rubber band.

A pressure transducer behind the pellet at the breech would a good addition a this point seeing how there's room left to festoon some more wires. Then adding a pellet release detector would seem the be all that's need to complete the kit.

For clarity, how are you using the terms slope, lateral, and boresight? I'm pretty sure I know, but I don't want to guess and confuse the discussion.

[WhatUPSbox?]

The hammer probe has two options: trigger on release or on valve contact. For valve contact, I adjust the position of the 2240 with the cross vise so the left edge of the flag on the probe is just tripping the laser gate when in the post fired position, i.e. the hammer resting on the valve stem (the hammer spring has preload). If you look closely you will see the red laser dot just on the left edge of the Al flag. When the hammer is cocked, the flag moves to the left and the beam is obscured. When the hammer is released the beam stays obscured until it travels past the valve stem contact point. The signal then stays high until the valve closes and pushes the hammer away from the contact point. There is a limit to the accuracy given the probe & flag hardware used, but in a half dozen shots, the probe signal and the mic signal were within 0.1 ms of each other (std dev 0.055). With a little bit more care, that could probably be tightened up.

I don't think the large slower signal is the mic mount, it has sticky tape under it and is fairly firm. If it came loose it would be very slow. It may be how the 2240 is held in the vise. 350 hz is still pretty firm. I didn't want to clamp the gun by the tube like you use because I wanted to give the mic a chance at hearing the valve impact. Since it is after the pellet exit point, I set it aside for now.

The pressure transducer is a longer term project. I did find a low cost 1000 psi unit that might work but it hasn't started its swim across the pacific yet. The home grown (literally) pellet probe is in work.

I'm loosely using boresight for the direction defined by the last inch or so of the barrel when at rest. It has some fixed relationship back to whatever the sighting system is established during the zeroing process. I use boresight error for the change in that relationship at the instant the pellet is passing through that section on the way out the muzzle. This is an angular error so I used the term slope to capture the change from the at rest condition, of course it has two directions involved. I used lateral for the two directions at right angles to the barrel axis. These are the axes which your accels are aligned with. I'm starting with the assumption that angular errors are more important than translation, but that can be verified.

I noticed that your screen shots still have the save menu overlay. On my DSO there is a print or hardcopy button on the front panel. If there is no printer attached but there is a USB thumbdrive in place it dumps a screenshot to the drive without bringing up any menus (you can preselect the image format) maybe yours has something similar.

[George Schmermund]

For clarity I'd like to show this graph that I've copied from the internet. It's a generic plot of the relationship of acceleration, velocity, and displacement. If we were to drop a vertical line through the graph at any point along the time line the intersections would give the instantaneous values of A, V, and D and how they are related to each other. Therefore, taking the integral of acceleration once will give the velocity and taking it twice will give the displacement at any point on the acceleration timeline.

The trace of the two accelerometer signals plotted simultaneously on the same XT timeline shown at the top of the DSO image is also displayed in the XY plot as a Lissajous pattern. Where the cursors are shown on the XY pattern are in fact the exact position of the barrel at the assumed exit time of the pellet if the double integral of A is taken. This displacement  position is in reference to the origin or rest position of the barrel as displayed by the upper trace.

The position of the pellet on exit has to be determined by the exit time in relation to the the start of the barrel movement. The clock really starts when the barrel begins to oscillate. The pellet exit time has to be referenced to the start of that clock. I'm a few pints into this post so please point out any errors or oversights that may have gotten incorporated into it.

[George Schmermund]

Stan, you're right! There is a hardcopy button on the front panel. I had to get a flashlight to read the label, but It works just like you said. I like this DSO even more now that I can copy screenshots without the menus in the way. Thanks for the tip!

[WhatUPSbox?]

George, I noticed my DSO (with royalties to Sir Isaac I'm sure) will do an integral of the waveform, though I'm not sure it will do it twice, probably need to pull the data out for that. Yours may do that as well.

On the other hand I was wondering if you could also use a direct displacement measurement of the muzzle when you do the pellet shot runs. The VRT you were using in Jan is a displacement measurement isn't it? I may have a cheap hall sensor that I'll tinker with. I was even thinking about your phono cartridge suggestion earlier.

Glad you found the print button. Some of these labels are definitely meant for younger eyes than mine

[George Schmermund]

Doing a direct displacement measurement would be interesting and possibly useful, but you would have to contrive a way of measuring the instantaneous time at any point along a dynamic displacement curve. It will be hard to beat an accelerometer, or better yet a pair of them, for making this measurement. My DSO has the advanced math capabilities, but the vintage B&K 2515 may be a better instrument for these measurements because it's designed to work with accelerometers.

After playing with the piezoelectric discs the crystal phono pickups look far less attractive. Being able to cut and shape the discs provides almost unlimited possibilities for making simple vibration measurements. The prices on eBay are ridiculously cheap.

[WhatUPSbox?]

The thought was the VRT could provide a real time displacement measurement and then a broken wire or any of the other pellet exit indicators could provide the time marker. With a little bit of history on each side of pellet exit, the velocity can also be obtained. It would be an interesting comparison to trying to integrate a complex and perhaps noisy accel trace of a pellet shot.

The phono cartridge thought was going after direct displacement and velocity measurement, not a force measurement like the piezos.

[George Schmermund]

The VRT is considered to be more of a detection device than a measurement device. it's used where detection of vibratory motion needs to be done without contact or adding mass to the DUT. It's not used where absolute amplitude is needed. Also, it's a velocity detector and would still need the signal to be integrated once to get displacement. That gives it 2 strikes right out of the gate.

Virtually all common accelerometers in use today use either quartz or PZT piezoelectric or piezoresistive crystals. The accelerometers use a seismic mass attached to the crystal and the force detectors use a preload on the crystal. The crystal phono pickups use neither and are generally torsional sensors. The discs allow experimentation with all 3 methods of generating a signal. They're too cheap not to play with and you can determine your own favorite method.

I'm glad that you're giving thought to all of this. Keep experimenting and posting your results!