Hacking the CP-2

[George Schmermund]

Back in April (post #180) I was playing with piezoelectric sensors and came up with a donut type that could be placed at the muzzle. I wanted to use it as a simple DSO trigger signal generator for when the pellet left the muzzle. Due to the number of distractions that I constantly surround myself with, the detector got sidetracked.

This morning I was shuffling things around on one of the test benches and the device rematerialized. Now that I'm consumed with pressure measurements, the original test as a pellet detector had now transmogrified into a slick and simple way of detecting and measuring pressure pulses at the muzzle.

It was a simple task to put it in front of the barrel and take a few shots through it. I was surprised at how well it worked with no changes to the original geometry. The duct seal putty is there to damp out the resonances of the disc, washer, and vise. The detector is in a vise for these first test runs, but it may get an smaller adjustable stand of its own. A pair of these would make a good muzzle chronometer. The need to move a typical chronometer away from muzzle because potential gas disturbances, particularity with CO2 guns, is negated because the detector ignores the pellet and just measures the impulse disturbances.

This sensor is fast and sensitive and I can see many hours of further distraction using it. Of great pleasure right now is to see how well presented the initial barrel air plug blast is captured. This plug is the same one that is shown in Stan's video (#181) as the very early shock wave. The shock wave has some energy number associated with and may be useful for the spreadsheeters out there looking for hidden drains on efficiency.

The last pulse coming out of Stan's video must be hammer or valve bounce and is also detectable on the P/T curve and the piezo signal in one of the DSO images. I'm still very impressed with Stan's excellent video post!

Rather than describe each DSO image and photo I'll just give a general idea. The photos should be self explanatory, though the DSO images are probably only a gourmet mystery meal for Stan. Others are also welcome to come in and help yourselves. The images are a sort of medley of several shots. Some are with and without pellets as the powerplant pressure drops. Some are with a new powerlet installed, again with and without a pellet.

I'll clean up the experiments with this sensor and present more coherent info sooner or later. Right now I'm going out back to shoot off some contraband firecrackers and shout out some July 4th hurrahs!

Lloyd - Welcome aboard!

[WhatUPSbox?]

George, that is providing some nice data, I'm glad you rediscovered it. Were you ..errr.. running out of gas on the first 4 DSO's? The transit times are stretched and the pressure curve is depressed.

The last two traces are intriguing. Do you think you could fit your instrumented pellet catcher behind the sensor to get a timing mark for the pellet (or maybe a broken wire across the muzzle)?

Note my only contribution to the 181 video was a bit of google time.

[George Schmermund]

Stan - The last 2 images are with the sensor moved right up to the muzzle. That's why the air plug impulse is so attenuated.  The timing should be very close to what the force transducer has measured.

The photos with the sensor at the muzzle weren't very informative about its build and arrangment so I didn't include them.

[WhatUPSbox?]

I was wondering if the pellet was leading the activity around 2.7 msec or if there was a significant  chunk of high speed air leading it. Kind of gets to the airmass discussion.

[George Schmermund]

If you go back and watch the video at .25 speed I think that most all of the information may be there or at least enough to remove a lot of the guess work.

It still leaves a lot to be deciphered, but that's why were doing this, I hope.

[WhatUPSbox?]

yes, even at that speed I am always surprised how quiet it is between the smoke ring and just before the pellet. 

[George Schmermund]

I did a few shots using the piezo sensor and a wire at the muzzle. The sensor was given an extra dose of putty to reduce the sensitivity to resonating. The wire was a truncated version of the wire probes with only a 90º bend at the business end that crossed the muzzle. This makes the wire easy to form and needs to be electrified at only one end. This also allows the disc and wire to be placed in close proximity to each other.

There was not a lot of time spent on the positioning of the 2 detectors, but this looks good for now. The DSO was triggered on ch 1 (pressure), ch 2 is the wire, and ch 3 is the disc.

[WhatUPSbox?]

That is nice data. Interesting contrast to the no-pellet trace (DS0008). Looks like not much happens at the muzzle before the pellet gets there.

Doing a little wild speculation and hypothesizing.....
In DS0008 the piezo rings at about 0.6 ms. For an 8 inch barrel, that's right around 1100 ft/s ... Initial shock from the valve opening?

In DS0019 the little blip on the piezo curve is 1 ms out. Back in DS0003 (post 245) it took ~ .4 ms for the pellet to move so that gets back to .6 ms. So out here on the limb it looks like when the pellet breaks stiction it generates a mini sonic shock that the piezo picks up....

......unless of course it isn't

[George Schmermund]

I was curious about what the actual rise time capabilities of the pressure sensor might be in relation to the valve starting to open. The DSO was zoomed in to get a sub-microsecond view of the first start of a pressure rise. There's more that can be done to refine this measurement approach, but I find the info useful for now. This is just a quick snapshot.

The time scale is obvious from the grid. The pressure scale is the usual 100 psi/V.

[George Schmermund]

As long as the zooming exercise is ongoing I thought I'd look at the P/T curve more closely along the peak. It turns out that what I had previously suggested was probably turbulence while the valve was open is actually periodic.

The DSO images are of 2 different shots without a pellet. The first one is without any signal filtering and the second one is with a low pass 40 KHz digital filter from the DSO. I'm starting to get a feel for the capabilities of the filtering app that the DSO provides. With the filter it's clear that the ~ 20.8 KHz signal is being modulated by a lower frequency signal.

The 20.8 KHz signal is also on the peaks of P/T curves when a pellet is shot. My idea on this outcome is that there is a whistle being formed by some fixed part geometry somewhere in the power plant. Maybe it's the transfer port. Maybe it's the valve assembly. The signal's amplitude actually gets larger as the power plant pressure decreases, but the frequency stays the same. I'm liking this scope more every time I use it.

[WhatUPSbox?]

George, it is great to see you are back at it. All I've done during our recent heat wave is make a list of the cases I want to run. Note to those of you in the south or southeast, replace the word heat wave with pleasant summer morning.

I love the data you get from that pressure sensor

[George Schmermund]

I've replaced the image in post #328 with the proper image that's shown here. I consider this test to be a milestone measurement regarding when the action really starts.

[George Schmermund]

Stan has mentioned the use of a quadrature cell and laser to do some barrel harmonics testing. I'm of the opinion the the information is obtainable with the use of a pair of very low mass accelerometers. Being a test and measurement type of guy, I'm always interested in trying new measurement techniques. There are few things in life, at my age, that are as satisfying as a pint of locally brewed IPA, and the correlation of 2 different testing methods that produce complementary experimental results.

To that end I decided to invest in a Keuffel &Esser 71-2627 quadrature sensor with target. It was another eBay sleeper ($20). The sensor is simplicity itself and works essentially the same as a group of solar cells. It's activation requires only a collimated light source and a mirror. The output signals can be routed to the previously discussed NIM instruments that were going to be used with the LED barrel timing experiment. The processed signal can then be routed to the DSO and be displayed in the XY mode as a Lissajous pattern.

The measurement can be made with a conventional light source and some optics. As a gesture to Stan's enthusiasm for the quad detector technique I'll fire up a vintage He-Ne laser for the test.

[WhatUPSbox?]

$20 ??!! Wow, that is a great find.

I think the optical data complements the accel results, providing slope data that the accels don't measure.

The old school HeNe will be great to see.

[George Schmermund]

Stan - Could you expand on your slope comment?

[WhatUPSbox?]

George,
I was envisioning a small mirror rigidly attached to the muzzle and at right angle to the boresight. With the laser source and detector on or near the breech (your clamping looks pretty rigid), the laser dot on the detector should only (ideally) respond to changes in tilt of the mirror and be insensitive to lateral motion. If you can synchronize the detector output to the pellet exiting the muzzle, the change in the tilt of the mirror should represent the change (compared to pre-shot) in the slope of boresight at the muzzle at the time of pellet exit. This pointing angle change would be one of the contributors to errors at the target (the other possibly being the muzzle lateral velocity from the accel data).

Some evaluation of the sensor for an optimum spot size will be needed. You could also add a second rigid mirror at the breech to set up multi bounce to increase the motion on the detector. Of course with longer path lengths you are more susceptible to air convection from temperature gradients caused by the presence of say a large cold beer.

[George Schmermund]

The distraction of too many projects has taken it's toll on airgun testing of late. One of the things that has brought me back to contemplating some new testing has been the results of earlier barrel harmonics measurements. The results contained in 2 DSO images that were originally shown in post #288 back in June have haunted me ever since.

In post #290 I commented:

"What interests me most now is the action in the first 5 ms after triggering the DSO. I didn't use the exit detector on any of these tests, but there is no chance that the pellet hasn't left the barrel by this time. Most intriguing is that the the first 2 cycles of both the hammer tap and the pellet shots are at a higher frequency than the natural decay frequency of the barrel. This would mean that those first 2 cycles would be what really determines the position of the muzzle when the pellet exits and not the calculated or measured first resonance mode usually represented as a continuous or exponentially decaying sine wave."

In both the hammer tap and the powerplant excitation a higher frequency oscillation can be seen to occur within the first ~ 3ms. The pellet can still be in the barrel during this time window and therefore never actually sees the real barrel harmonic motion. Since I didn't have a solid explanation for this initial barrel activity, other than seeing it was there, the data was generally ignored and things moved on.

Inadvertently, I've recently run across a wealth of published information about "forced oscillations and resonance". Of particular interest is a set of lecture notes that can be searched at:

Notes on Diffy Qs
Differential Equations for Engineers
byJiˇríLebl

In section 2.6, figure 2.7, I found the answer to my measurement dilemma. It's now clear (to me, at least) that the pellet experiences mostly the forced oscillations of the transient or impulse and not the barrel's natural frequency!

Spreadsheet rebuttals are always welcomed.

[WhatUPSbox?]

George, I understand the interference of projects. Thank you for rescuing my brain from the - setting pavers and painting wood - cage it has been trapped in lately.

A while back I ran some numbers (spreadsheet...sorry) for just the CP-1 barrel fixed at one end. There was a mode at ~1,400hz.  With the extra compliance in the breech, it would come down some and should be what you are seeing in the unfiltered tap test data.

Characterization of the forcing function over the ~3 msec the pellet is in the barrel is an interesting question. An ideal impulse is a single event and rings all the barrel harmonics. In the firing cycle there are several events in that 3 msec period: hammer strike, pellet release, pellet accelerating down the barrel, etc. How these interact with the barrel modes is not an easy assessment. The accel is sensitive and picks up a lot of the noise sources in the cycle. The really high frequency stuff won't generate a lot of displacement and probably won't create much lateral velocity. If your test rig is operational, getting a trace with the pressure curve and the accel data, and maybe your pellet exit trip wire would help shed light on whether any of the cycle events line up with the accel response. Eventually may have to introduce the DSO's low-pass filter to get rid of the very high frequency noise.

Unfortunately, I can't try this with the 2240.

[George Schmermund]

Now that I know what I'm looking for the measurements can be redone using the DSO filtering capabilities.

[George Schmermund]

The barrel harmonics project is getting increasingly interesting now that I've gotten better at using this new DSO properly. I've shifted my enthusiasm for analog signal filtering to becoming much more enamored with this scope's digital filtering capabilities. Being able to switch the DSO's filter in and out even after the signal has already been captured is really valuable for these vibration measurements.

I'm starting to get a better idea about the nature of the how the initial force is being applied in these tests. The definitions of these forces are becoming clearer in my mind and how to most properly use them. The reading that I've been doing lately has narrowed down my impression of what a transient is compared to an impulse. My current understanding is that a transient event is something (a force) that disturbs a system in a way to momentarily alter it's steady state condition. On the other hand an impulse is a single event that would look like a step function where the energy is deposited as a single nonrecurring event. A steady state condition in these barrel tests could be one where the barrel is at rest or one where it's running at one of its natural frequencies which would be the case of simple harmonic motion.

Stan has suggested that a transient event could be initiated by the sum of several forces that occur during the brief firing cycle. He also pointed out that an impulse could be initiated by a single hammer tap on the barrel. Both situations are capable of producing forced oscillations of the barrel that are unrelated to the barrel's own natural frequencies. These forced oscillations only last a few milliseconds before the system restores itself a to steady state condition, but that time interval can be close to the time when the pellet is still in the barrel.

Today's testing has captured some very interesting information that may reshape certain notions about what is really going on with barrel harmonics. I'll leave things here for now because I need to sort out all of the DSO images and also want to here any feedback about my understanding of proper nomenclature regarding what and how these things can be  measured and explained.