Hacking the CP-2

[AmBraCol]

Can't speak for the rest of the lurkers, but my ability to contribute to your posts in a constructive way tends to be limited - but am definitely along for the ride, looking to learn from the experiments others do.

They say there are three kinds of learning.

1.  Learn from reading.
2.  Learn from watching others.
3.  Take a leak on the electric fence yourself.

I tend to the first two methods... 

[WhatUPSbox?]

George,
I'm enjoying your multi-stop journey delving into airgun characteristics and the bottomless toolbox of instrumentation that you've brought to the task. So far, I think the glass lathe is my favorite.

[George Schmermund]

The new o'scope arrived today and it's a beauty! They've changed the front panel layout and color. I like it even more. The new firmware version is stated to be V1.19. I was mainly interested in getting 4 channels to make simultaneous measurements and a good XY mode for Lissajous patterns. Using just test signals to take the instrument on a shakedown cruise proved to be quite rewarding. I'll post some photos as some as I'm sure of what I'm seeing.

These next measurements on real airgun stuff are going to be very interesting and easily worth the investment in new hardware. 

[Rob M]

awesome, pics when ya can

[George Schmermund]

Having 2 accelerometers on the end of a barrel has given me some head scratching moments recently. Being able to measure the barrel vibrating in both the horizontal and vertical directions simultaneously has produced a lot of unexpected information. Since the accelerometers are very sensitive in only one axis and have negligible response to anything transverse to that axis, I was surprised to see a large signal coming from both transducers when the gun was fired. Using the tapping hammer I was able to excite either the vertical or horizontal modes individually. Tapping the barrel at a 45º angle between the sensors again produced a pair of ~ equal amplitude signals. The surprise was the consistency of the signals. One of them was at 104 Hz and the other was at 167 Hz on the same piece of 8" barrel mounted in the same hole in the breech!

I double checked the mounting screws and they were tight. Using the third transducer, which was mounted on the magnet, I could move it around and confirm that the only node was at the breech end of the barrel where it was clamped. Having become nonplussed, I fell back to the safe haven of a few beers.

After regaining my composure, I ventured back onto the battle field. It turns out that the end of the breech where the barrel mounts has 2 modes of flexibility. The vertical mode makes the breech act as a flexure point where the pellets are loaded. The horizontal mode has more stiffness because the whole breech is secured better and is also tightly straddling the main tube. This accounts for the two well defined frequencies.

The photo shows where the problem lies.

[WhatUPSbox?]

George,
For the ~8 inch barrel, the first mode, cantilevered should be about 230 hz
http://www.amesweb.info/Vibration/Cantilever-Beam-Natural-Frequency-Calculator.aspx
http://www.amesweb.info/SectionalPropertiesTabs/SectionalPropertiesHollowCircle.aspx

I know, I know....spreadsheets....but sometimes they are useful

So a couple of additional areas of compliance might be: 1) the two orings at the barrel/breech interface may not be compressed by the two set screws so the barrel is not cantilevered and 2) I don't remember how the block is attached to the tube. If it is effectively cantilevered itself through the section under the loading shelf, from an attach point closer to the bolt, then that will drop the frequency as well. Both of these would be different in the two axes.

To check out your instrumentation, a simply supported configuration may be cleaner, though you would need to move your accel, and the frequency will be above 600 hz
http://www.amesweb.info/Vibration/Simply-Supported-Beam-Natural-Frequency-Calculator.aspx

Or you could clamp the barrel directly

....Still waiting for that first screenshot with 4 traces on it

[George Schmermund]

I went back and tested the barrel again. The measurements show high Q peaks at 104 Hz and 164 Hz. The measurement bandwidth is 4 Hz to 1 KHZ. The vertical display is 60 dB. An interesting dip on the right side of the 104Hz peak is a partial phase cancellation from the 164 Hz peak. The very low Q humps around 800 Hz are at least 20 dB down and of no consequence in these tests. The CRT screen is not as easy as an LED screen to get a good photo from without some effort with camera settings, but this should do for a quick test.

The 2515 analyzer comes up with pretty much the same numbers as the o'scope. Someone should now check their spread sheets for sanity. Thanks for the nudge, Stan.

[George Schmermund]

I can go back to the Complex Modulus Apparatus (CMA) to test just the barrel as was done on the Vigilante 10" barrel in the other thread. That will give complete isolation for the natural frequency test. That exercise was enjoyable in itself.

The thing to keep in context here is that I'm trying to do real world measurements on the gun. The tests presented so far are those of the barrel as it is used when firing the gun. The rest would be academic, but still interesting.

[WhatUPSbox?]

Someone should now check their spread sheets for sanity.

Yes, that is a great point. The web is full of calculators and databases of all sorts. In this example I found those after just a quick google search. In reality, I have no idea who owns that site and how thorough they are in proofing their tools. This is the general risk of using other peoples spreadsheets, calculators, material data etc. In this case I checked their cantilever results against other handbook derived results and they were consistent. I also like that they show the equations used and reference the handbook source. For this application the frequency results are not, in general a safety issue and the material properties for stiffness are pretty consistent for steel. On the same site they have other calculators for stresses in pressure vessels as an example, as well as strength properties for materials. These need to be approached with caution. Equation errors, misunderstanding idealizations of the handbook approach, scope of what the factors of safety cover, and uncertainty in process driven material properties can all lead to an unsafe design decision.

The thing to keep in context here is that I'm trying to do real world measurements on the gun. The tests presented so far are those of the barrel as it is used when firing the gun. The rest would be academic, but still interesting.

Agreed, I always try to have a handbook or back of the envelope evaluation for any configuration I'm looking at. It tends to provide a good bounding value for either instrumentation or detail analysis results. In this particular case the calculator provides 1-D results for a uniform tube, fixed to an infinitely stiff mount. In predicting your test configuration, only the first mode prediction is useful, you will have modes in other directions that will show up after the first mode. You also have minor lumped mass at the tip and the mount is not infinitely rigid. All of these will drive the frequency down. So given all that, if your results showed a first mode above the ~230hz, I would be scratching my head. The fact that they are significantly lower, starts the search for the extra compliance. This can be useful for the actual barrel study. For example if the compliance is in the barrel setscrew/oring mount, then it may respond to torque values, additional shimming, or material choice for the orings. If the compliance is due to cantilever/softness in the mounting block then additional shimming or clamping may influence the results.

None of this probably matters for the accuracy with the 8" barrel in pistol form but may be more significant for the 18" barrel.

Finally, I find comparing analytical predictions with tests result to be lots of fun (at least if the test results didn't include one's safety)

Carry on...great stuff

[George Schmermund]

The rest of the parts for the pressure calibrator arrived today. Everything went together without a hitch or leak and I'm very pleased with the outcome so far. The only downside is the painful reliance on... you guessed it... online calculators and spreadsheets! To arrive at the actual pressures above CO2 in the liquid/gas phases is like throwing darts at a board. Well, I'm OK with that at this point because the differences between various sources isn't enough to be a problem with the measurements right now. The real solution is a dead weight tester. The one that I have now doesn't go high enough to for these pressures so I'll just keep an eye on eBay.

These small Endevco transducers are really nice and the older Endevco 4428A  conditioner/readouts that I've been collecting are very good and fairly plentiful on eBay. Their analog output makes them compatible with the way I do most of my measurements. There are some perks to being from an older technology era. 

[Rob M]

looking great

[WhatUPSbox?]

 I really like that co2 adapter system...can't think right now what I would use it for but it is on the list

[WhatUPSbox?]

Continued to poke at how far one can take hobby level instrumentation. I dug out a 2 axis accel dev board I've had forever. It uses a now obsolete ADXL202 mems type sensor. Similar boards currently use the ADXL335. They are cheap, light, don't require conditioning electronics and provide analog voltage output. I first set up a 2240 barrel in a suspended free-free (first picture) to see if the output was useful and played nice with the spreadsheet prediction. It did pretty well, the predict was in the range of 1.53 to 1.55 Khz and the results came in at 1.47 Khz. Also note the George recommended tap hammer was acquired and used...worked great. The board mounting to the barrel worked well but may need to be changed for any shot response tests.

I also looked at the cantilever configuration, primarily to see what drives the results away from prediction. The second picture shows the same barrel clamped into a v-block, in what appeared to be a rigid mount. With 6" cantilevered, the predict for a bare barrel is around 380 hz, with the 4gm sensor at the end, it should drop to around 350 hz. The tests showed 313 hz in the horizontal and 265 hz in the vertical. I'm guessing some of that compliance is in the line contacts in the v-block, and some is in the bench top. I was surprised at the end mass effect, as well as the what I thought was a stiff mount. But mostly it was to shake-out the board so to speak.

I also tried the miniature microphones for these tests. They did well in the free-free test with similar results. The cantilever test was much more difficult with ambiguity on the directionality of the response.

I think the microphones can work well at timing sharp events, their low cost and tiny size make them worthwhile.
For frequency, the hobby accelerometer boards can provide useful data.
For anything requiring higher fidelity or amplitude accuracy, the instrument grade accels that George is using are the way to go.

Fun stuff

[George Schmermund]

Stan - Nice experiment. It's pleasing to see someone else getting dual simultaneous modes out of these simple tests. It must all lead somewhere.

I did a simple test this morning to evaluate the difference between the the way we might be using our hammers. I converted mine into an "instrumented" hammer for this test. The demonstration is to show how the hammer's impact tip material will effect the force/time impulse. I used a rubber cap like the one's use to protect things like threads on fittings and such. This one has a nub on the end that makes it ideal for getting good taps.

The red trace is the impulse waveform with just the bare steel hammer tip and the aqua one is a tap with the hammer's tip covered with the rubber cap. As the photo shows, the impulses have very different time constants. The rubber cap extends the impulse time which allows much better low frequency excitation. By changing the tip material you can make simple mechanical filters for this type of testing. You might want to give this a try.

[WhatUPSbox?]

Interesting, I'll have to try out some different tips. I just tried to get a clean tap without a double tap.

[WhatUPSbox?]

OK, another little packet made it over the pacific ocean. The thought was to replace the broken wire timing approach with some form of a light gate. I found a laser receiver module (first image) that met my criteria of cheap and easy to use. I added a laser source ( a dollar store type pointer in module form) to set up a light gate across the muzzle (second image). The response time of the receiver is pretty good (third image), less than 50 nano-sec.

I tried a few locations but ended up about .25" in front of the muzzle. The fourth image shows the response of the Destroyer pellet fired from the 2240. One interesting note is the Destroyer pellet is .228" inches long, the little point in the center makes it .257". So if you take the 46.4 micro-sec transit time it works out to 410 and 462 ft/sec respectively. I'm not suggesting this is chrony quality data but it is nice to be in the ballpark.

Right now it looks like a good low cost addition to the toolkit. I'll have to run it with the broken wire to confirm it is doing what I think it is.

A couple of side notes. I had it just grazing the muzzle and had some additional transients on the trace. Turns out with the 2240 held in a lightweight panavise clone, I was catching transients of the barrel recoil. Interesting, and may be useful later to time recoil to pellet exit. Also CO₂ generated clouds showed up once or twice (I have the power adjuster turned in a bit). This could be more interesting info if timed to pellet exit.

Next is to fold this in with the rest of the timing instrumentation and then read the tea leaves

More fun stuff. 

[George Schmermund]

OK, Stan, now that you've thrown down the cheapskate gauntlet I have no choice but to follow suite. These piezoelectric transducers came in the mail today all the way from China, too. That's how far I'm willing to go to meet your challenge! These are the 10 PAK of 35mm disks for $1.78, free shipping. There is a 20 PAK of 12 mm disks on the way that cost $1.49 , free shipping.

To keep the playing field level I used the flying wire and Scotch tape attachment method. Also included was the odd piece of cat hair into the mix as a flourish. Kitchen scissors were used in an indiscriminate way to make a smaller sensor. These disks can sure "take a licking and keep on ticking" (plug for Timex).

The signal from the sensor was connected to the o'scope with a 10X probe with no conditioning. That shows a pretty good output for a small tortured device. The sensor and tape weigh in at less than 1 gram. This could easily be cut in half again, obviously. The o'scope shows the same fundamental frequency (in one axis) as the other accelerometers.

I should have chosen a higher contrast background for the before and after disk photo, but I think that it still works. I'll blame shift the oversight to the beer.

PS - I just saw your new post as I was posting this one. Nice work. I am surprised at how clean the break points are in your o'scope images.

[WhatUPSbox?]

I don't want all this talk of low cost to imply that I skimped on the optical chopper used to evaluate the detector response......

[Cindi]

George and Stan: Both of you "enablers" are making my piggy bank nervous! Every time you have interesting data, you are showing off your wonderful multi channel DSOs. A DSO was not in my budget this year.
I did appreciate the helpful bargain pointer, however I'm not going to buy a tiny & cheap DSO either, my eyes are not happy with tiny displays, and I get frustrated with fiddly little controls.

Back to sensors:
I did find an Endevco 8511a-5k at a low cost and purchased it. Still looking for a 8510B-2000 at an affordable cost (list price is around $1,100).

At a cost between the broken wire and the Laser Pointer+Laser Detector combination are Slot Type Optical Photointerrupters with a large gap. The sensor slot is much smaller than a .177 pellet. A pair of these could be the basis for an end of barrel chronograph.

Example 1:
Digikey P/N 160-1936-ND  $1.11 15MM gap (Horizontal) and 1OMM clearance (Vertical),  Photo transistor output

Digikey P/N 160-1936-ND  $1.11 15MM gap (Horizontal) and 1OMM clearance (Vertical),  Photo transistor output

Example 2:
Digikey P/N 365-1768-ND,  9.5MM gap (Horizontal) and 1O.8MM clearance (Vertical)
  TTL Output (buffered totom-pole) [p/n OPB910W55Z $5.35]
  Open Collector output version available [p/n OPB901W55Z $5.33]

[WhatUPSbox?]

Cindi,
Yes, I'm looking at the photo interrupters as well. Started with the laser detector just because it is more general purpose....and it uses a laser!

If you can wait for the transpacific they are less than a cup of coffee (I paid $3.28 for two and free shipping). The lasers that run straight off of 5 V are $1.30.

I have not found a low cost, fast pressure transducer in this pressure range.
 
As far as the DSO is concerned, I blame George. My old analog scope will be looking for a new home. I looked at the PC based scopes but decided I wanted a standalone.