Hacking the CP-2

[George Schmermund]

OK, Stan, you're not going to out cheap me on this one. Washer - found it in a box. Sensor - 18 cents. Post-it notes - poached from my wife's desk.

This device is going to be a muzzle exit detector. It's ostensibly immune to acoustical excitation and the trigger signal for the DSO can be set to ignore the pressure pulse. The pre-trigger signal will still catch the pressure pulse for inspection (maybe).

The sticky part of the pads can be trimmed off of the note part and be used as the expendable part of the transducer. I tried a Scotch tape pull test on the sensor's coating and nothing came off with the tape. The sticky part of the note pad is much less aggressive and should allow many tests if the sensor itself isn't hit by a bullet. A pair of these would make a cheapie chronoscope.

I'll test this out this afternoon and report back.

[WhatUPSbox?]

Well if you can work without a laser.........

I like your design. It will be interesting how the post-it responds. When I was setting up the broken wire I tried Al foil on edge. That blew off without a pellet. Maybe if you ran a case with a piece of thin wire taped across (or a cat whisker).

Tuning the position of the washer relative to the muzzle may create a useful pressure wave sensor that can be combined with a second pellet detector, and some accels to give a fuller picture of the timing at the muzzle.

A while back you brought up Schlieren photography of the muzzle. I found this video. There is an interesting puff that takes place long before the pellet gets there.

[George Schmermund]

Stan - That's the best pellet gun video I've ever seen. That initial shock pulse in the video is the one that I was getting when the Vigilante's muzzle blast was done with the microphones. That was back when we were discussing N waves. The string of pressure pulses that were recorded back then without a pellet being fired are the same thing, but with the air plug out in front of the CO2 plug. The strings that were shown back then, including the heavy duster gas injection was also captured in the string of pulses, though nothing like this.

The microphone's o'scope images are like a Hilbert transform slice through this video. It would be very interesting to know the details of the video's timing sequence. The tail-end image may be hammer bounce.  I'm actually astounded by the density of information you've just presented.

After watching your video ten times I feel totally defeated and deflated in all of this test and measurement stuff. Beer won't do it tonight. Where's my gin bottle?

[WhatUPSbox?]

Yeah, I could make that my screensaver.
If you watch it on youtube you can bump up the resolution/size a little bit and play it at .25 speed. A little bit more detail pops out. They don't provide much background other than it was shot at a frame rate of 15K. With a guess on pellet size, might be able to derive a speed.

I would not give up hope. I think your piezo-donuts might work out well as pressure wave sensors. If you add the chamber pressure, an explicit pellet detection, and a microphone. Substitute in a couple of accel readings after a few shots and you got a pretty thorough data set.

[WhatUPSbox?]

OK, found one more. 2700 frames per sec...not sure why it says no pellet. looks like one at the 1 sec mark

[WhatUPSbox?]

Cindi,

I found I had a generic photo interrupter. It is too small to send a pellet through but I triggered it with a dremel driven chopper. The generic one had a relatively slow response time (~150 micro-sec). I think the ones you are looking at from Digikey...with a spec sheet...are the way to go. It will be interesting to see if they respond to muzzle blast without a pellet.

I'm with you on the tiny screen & fiddly controls...to take that to an absurd level there is a 4 channel version of that tiny DSO

[George Schmermund]

 Last evening was spent consuming therapeutic doses of tincture of juniper berries and other botanical restoratives. Today I was able to climb back to the helm and continue this voyage into the unknown. I thought it best to start in slowly and get my sea legs again by getting back to hammer/valve timing.

The previous effort on this task was to use an accelerometer attached to the breech above where the action was happening. This choice was expedient, but the transducer picked up a lot more activity then was needed. In an attempt to get a cleaner signal and better timing info, the choice of sensor and placement was moved.

After the success with making the "donut" sensor for muzzle measurements I decided to change the scaling and make one that would fit around the valve stem and get struck directly by the hammer. I'll need to arrange for some type of shock absorber to spare the disk from impact destruction. I'll mill a small grove into the face of the valve for an electrode lead and bring the wire out through a hole in the side if the main tube. This approach should also give good information about hammer bounce.

The photo shows where things are at this point.

[George Schmermund]

Stan - In your free-free test you used rubber bands as a compliant overhead suspension to support the barrel. I've seen several YouTube videos about doing modal testing in a free-free arrangement where they use a soft egg crate foam supporting the DUT. Since you already have the earlier data, what do you think about rerunning the test with the egg crate foam method as a comparison?

[WhatUPSbox?]

yes, I have some soft foam, I'll give it a try.

[WhatUPSbox?]

George, Here is the 2240 barrel on foam. It worked well, though you can only tap it in one direction, not an issue here with enough signal in both channels.
 The predicts were 1.54 khz for the first mode and 4.24 khz for the second. The measured results were 1.57 khz and you can see the second mode out around 4.2 Khz.

Note: I really need to learn how to use the built in fft math function (as well as the others). I am using the trace fft which is near real time but has limited number of points. You can also operate on a stored waveform with much higher number of points. So I'm not sure what the uncertainty is in the numbers above. Of course there is a small uncertainty in the modulus number used for the predict as well. Overall I think it is a decent match and the free-free numbers don't apply directly to a real world configuration, but provide a sanity check.

[George Schmermund]

Stan - Thanks for doing the followup test using the foam. It confirms that the method is a viable shortcut for doing some of these experiments.

I've been playing with my new DSO and grinding through the almost 200 pages of manual. Entry level instrumentation is a lot different than it used to be. I'm slowly getting it to do my bidding, but getting data into a postable form for this thread has been a challenge. I think I've got it working, but I can't seem to get a screen shot without having the menus captured too.

Being a computer dummy doesn't help, but I like this scope too much to be overly disappointed. I'll just keep after it and see where it ends up. I'm getting ready to do some multi-channel tests in the shallow end of the pool with it. I should have something that isn't just a camera screenshot sometime this week if I stay busy.   

[WhatUPSbox?]

Yeah, how to turn off a display item was on page 147 in mine. The soft menus don't turn off. There is a convenient print button that grabs a screenshot to USB but any cropping is done in the PC.

[WhatUPSbox?]

OK, down another level in this rabbit hole.

I was thinking of ways to monitor what the hammer is doing. The hope is to tie it in with some of the other measurements and start to fill in the timeline. George is looking into the piezo-bumper on the valve face approach, so I thought I would see if the skewer probe concept can be stretched any further. I changed the probe from being a continuity type signal to breaking an optical beam. I added a flag to the exposed end of the probe. The width of the middle position (Step 2) approximates the static stroke of the hammer. The other two are a little bit on each side of that length. The laser/detector combo is used as the detector. The first image shows the parts involved overlayed on the 2240 (thank you to whoever made that great section view). When the hammer is cocked, the flag moves to the left and the laser dot is just at the right edge of the flag. The probe tip inside is trapped between the hammer and the spring so it follows the hammer. When fired the flag is pulled to the right and the laser is blocked until it reaches the left side of the flag. In the Step-2 position this roughly corresponds to contact with the valve stem. Any time spent with the laser exposed (signal at 5V) the valve is compressed (approximately).

Image two shows the test configuration. The 2240 is held in a cross vise that allows me to start each run with the laser beam just on the verge of being broken. The laser/detector are on a scissor-jack platform that allow me to move it up and down to use the different steps on the flag. The laser beam and detector have a finite width that effectively shortens the flag width. I estimate it at 0.05" but I still need to measure it. (it amazes me that these <$1 laser modules can be focused). These initial shots were dry fired without CO2. I want to add the microphones and see if the hammer motion is consistent with anything in the mic traces. Without CO2 pressure I'm guessing there is a lot more bouncing than in real shots. For anyone looking at the results, only the initial hammer flight would relate to a CO2 powered shot. Again, this is only to try out the test method the actual dry-fire data is not relevant real world performance.

Image 3 shows the trace with the middle (step 2) position being used. I listed my guesses of what is happening and the time line for each. Note from the earlier tests with the skewer probe, the pellet is gone by 15 ms or so. Again once the trace starts, each time it reads 5V the laser is unblocked on the left side of the flag.

In image 4 I overlayed two separate shots (using step 2) they are surprisingly repeatable.

In image 5 I used the full width of the flag. As expected the first leg is longer since it now includes some of the valve compression. The number of bounces is less since the smaller bounces stay obscured.

In image 6 I used the shortest (step 1) part of the flag. Again, the number of bounces shown is less because the smaller bounces don't compress the hammer spring enough for the beam to be obscured. Note that the first shot didn't achieve the expected shorter first leg (the repeat did though) I think this is Murphy taking his crack at the repeatability claim.

It will be interesting to see if this probe approach helps clarify what happens and when between hammer release and pellet exit. Still trying to figure out how to fully check out this measurement and whether it is useful. Comments are always welcome.......wife still shakes her head when she walks by.

Fun stuff

[George Schmermund]

Stan - Your experiment looks interesting. Except for a small error in step 9 of your guess list, you seem to be on the right track with this method. The ~ 18% trigger offset might be do to the compressed time scale view or else something is not resetting to the same place in the hardware.

It's nice to see a different approach to these measurements. I look forward to your progress.

[WhatUPSbox?]

George, yes I think step 9 is the hammer coming to rest on the valve stem....I have no idea if the valve opens. It is also ambiguity in the exact size of the blocking flag. I expect the small bounces to go away once the system is under pressure

Note the plots labeled as overlays are just graphical overlays meant to qualitatively compare the repeatabilty. They were overlayed in a graphics program. (yes I need to learn how to store and retrieve waveforms). The 18% difference in the last (step 1) results were not the trigger points but the measured duration of the first hammer motions.

[George Schmermund]

Thanks for clearing up the overlay anomaly. My comment on the step 9 point was really about a simple arithmetic error. It now seems petty to even mention it, but that's what peer review is supposed to be about. Numbers are different than things likes spelling errors and grammatical mistakes, unless the latter causes confusion. I'll attempt to defocus my  perusal of postings and pay attention to more egregious things.

[WhatUPSbox?]

Thanks for catching that. I agree number typos should be addressed

[Cindi]

Some of my most enjoyable experiences in Research and Engineering was the interaction with my peers even if it just about attention to minor details, it made us all more mindful of the complexities & subtleties of our work, peer review helped my team achieve the best results.
The two of you are courteous and show each other respect, I have been enjoying this thread (it's the first one I look at in the morning) and hope to have some experiments of my own going in the future.

Best of the Day to both of you
-Cindi

[WhatUPSbox?]

Another incremental step, I added the microphone in contact with the 2240 tube in the valve area (image 1). Contact is maintained with double sided tape. The rubber band just keeps it in compression and avoids having it knocked off. I also made a new flag for the probe that more closely spans the two trip points for the two static positions of the hammer.

Image 2 shows the trace for a no CO₂ dry fire. The middle purple 5v line is the contact with the valve. Looks like both the probe and the mic compare well on the timing of that event. Image 3 just zooms in. the difference between probe trace step and the mic event is about 80 micro-sec. That's probably mostly in the probe trigger position uncertainty. System is too sloppy to improve on that without doing some extra work. It will be interesting to see if the mic results are as distinct once there is CO₂ in the system.

Yes Cindi, I've always enjoyed peer reviews and robust technical discussions. Though they seem to be a bit of a fading art.

[WhatUPSbox?]

George,
Do you think the valve in the pressure calibrator you assembled is a good enough long term seal that it could be a useful source of compressed gas for blowing out dust, etc. when away from a compressor?