.22 LDC bore clearance study print files

[WobblyHand]

People have been working on active noise control, since the 1936 patent came out.  Algorithms that remain stable for impulsive noise are being worked on.  Apparently, it is not an easy problem to solve.  Everyone, from car makers, to headphone makers would like a solution that worked better.
https://www.cambridge.org/core/journals/apsipa-transactions-on-signal-and-information-processing/article/recent-advances-on-active-noise-control-open-issues-and-innovative-applications/9E27156562A24026ECD6F6A49A54F53A
https://pubmed.ncbi.nlm.nih.gov/24363493/
Not going to lie, I understand the basics of this, but this is far from my expertise.

[subscriber]

I think that the closest we have to destructive interference in airgun moderators is Tesla Valves; turning the air flow against itself.  With an engine muffler you don't want to restrict the gas flow; you just want to reduce the pressure vibration to a state of steady flow.  The latter would also work for an airgun muffler, but mainly it is the pellet's "flow" than must not be restricted.

 I think that active sound cancellation, including electronically needs to listen to the sound in order to respond to it.  A shot's sound has rapid onset and changes continually after that, before dying out.  Probably not enough time for a system to track the sound and cancel it.  Unless it can learn from the previous shot sounds.

An engine muffler is dealing with gas pulses from multiple firing events.  There is a frequency that depends on the engine speed and number of cylinders (and ducting branches from them).  I think it is easier to arrange destructive interference of sound waves where there is repetition in the sound driver.  A gunshot is a multi-spectrum square wave with some reverberation.  More like a car crash than the engine sound.  Although a vehicle crash is in some ways like banging on a drum; hard enough to damage it.  Perhaps a machinegun sound could be managed more like an engine muffler.  If you are familiar with the sound of a Gatling at 4000 rounds per minute, that has a much more musical sound to it.

Rather than state what equipment would be ideal for this project and then giving up because we don't have access to it; all I insist on is not saturating the microphone; and that the meter placement is consistent.  If people cannot tell the difference between shots using different moderators by ear, then frankly, the accuracy of the sound meter matter much less.  The reason for measurements is to predict what people hear, and are likely going to find acceptable/unacceptable.

One reason to care about absolute loudness is to determine the threshold where people notice a difference between sounds.  Important, but a different study.  I am more interested in relative measurements produced by multiple unbiased people, using available equipment.  I am very interested in their subjective impressions, as ultimately, it is how loud your wife thinks your airguns are that matters.  Perhaps trumped by how loud the neighbor's wife thinks they are.

The point of this study is to examine the school of thought that, for maximum sound attenuation, muffler baffle bores should be as tight as possible.  Based on what I have seen, baffle bores can be larger, with a smaller loss in  performance than one might expect.  The fact that mapping this function may show kinks in the graph due to the effect of "pre-curser gas" ahead of the projectile adds another dimension, but points the same way:  That the ideal baffle bore diameter for caliber and power may not be "as tight as you can make it" (short of clipping baffles)


From that noise cancellation paper, snoring represents some of the same challenges as a gunshot.  It changes too rapidly to track; and has multiple components to the sound: 

"Another challenge is that a snore is a fast time-varying and intermittent noise. This requires fast convergence algorithms such as those introduced in Section II(L). Owing to the broadband nature of snore, feedforward ANC with a reference sensor is required. This may cause causality problems if the snorer and bed partner are too close to shorten the acoustic delay. In addition, a snorer does not always snore so a robust snore detector to turn the ANC algorithm on and off is needed.

Furthermore, ANC systems do not completely reduce primary noise owing to many physical limitations. In the snore ANC application, which is often used in quiet bedrooms, still annoying residual noise may need special treatment."

[subscriber]

There is the urge to reduce any problem or solution to a single number.  In this case, a peak dB reading for a gunshot - or more precisely, and airgun shot.  Muffled, VS bare barrel; or compared to other mufflers.

When it comes to airgun mufflers, there is a second factor that is arguably more important than the peak amplitude:  Once the peak amplitude falls into perhaps the 85 to 90 dB range; what matters most is if the residual sound still has a "crack" to it, typical of a firearm gunshot sound. 

Based on feedback from people who have tested my custom airgun muffler designs (not the test mufflers posted in the OP) it is the reduction in the crack sound that is most appreciated.

The crude time/amplitude sketch below has two "charts", with the same peak amplitude.  "A" has a very abrupt edge, with an echo; and would be more objectionable than "B", despite "B" carrying on longer.  So, measuring peak dB is very useful when one is more concerned about hearing damage to the person shooting the gun, than in the context of "backyard friendly" airgun use.    Except for very loud airguns,  I am suggesting that a muffler that removes the "crack" may be preferred to one that peaks at a few dB less, but that still has a definite "gunshot" signature to it.

Why then do I ask for dB measurements for this project?  To satisfy those that insist on objective measurement; and hold that a 2 dB difference is significant.  Based on the available readings, it seems that this three muffler test may have uncovered something about baffle bore diameter and projectile clearance that was missed before. Something that a two baffle bore diameter test cannot show.  If anything, the results suggest the trial be expanded further than +1 and + 2 mm on baffle bore diameters, with with 0.5 mm increments.  Perhaps that can wait until someone who has access to better test equipment and facilities volunteers to run such a trial.

[WobblyHand]

Impulsive active noise cancellation is state of the art, it's hard.  But likely not impossible.  It will take quality sensors, transducers, advanced algorithms and fast processing.  That being said, getting 60 dB reduction is a pipe dream.  If you do a lot of hard work probably the best one could do is maybe 35-40dB.  First starting out, might get 20dB.  I've done some similar work in signal cancellation and got close to 38dB reduction, but in the end, one is limited by the information in the dataset, the time duration and your estimation errors.

Simply taking measurements is a whole lot easier.  Yes, I agree that the microphone cannot be saturated, actually there should be at least 10-20 dB of head room before saturation to be have reasonably low distortion. 

The issue really comes down to experimental technique.  If one is a diligent experimentalist, one can get good results.  But multiple people using different techniques or distances tend to make these studies just qualitative rather than quantitative.  It's not that qualitative data is bad, but it simply isn't quantitative.  So if you really want to know a quantitative answer, the setups have to be exactly documented - like writing a lab report sort of thing.  So someone else could actually exactly duplicate the results, with the same moderator (and ideally, the same airgun).  That includes the shooting range characteristics, dimensions and pictures, since I'm sure they are all wildly different.  If they are "set up wrong", whatever that means, they could impact the results.  Also, each AG has a different pitch and tenor, and they probably are altered by our tuning efforts.  This is an N variable problem, and it would seem that one would like to minimize the amount of variables in the beginning.  Then we could explore the differences between AG's.  I'm very open to suggestions on this.  Just blurting out some thoughts.  But if all we want to do is A/B tests that's ok. 

But the tests are most likely are affected by their acoustic environments.  No we don't have to do this in an acoustic dead room, but we do need to be sure the methodology of data gathering is such that we don't have people shooting along a wall or fence, with an acoustic reflector, or in enclosed unattenuated spaces.  That's why photos, drawings with dimensions and things like that are important.  Ideally, the ranges would be in open fields, with no acoustic reflectors.  Not everyone has access to those areas.  All the ones local to me have turned into developments. 

I appreciate what you are trying to do.  And everyone wants to turn a complex problem into a simple number.  I don't doubt you have come up with some nice hardware, it's been apparent from your various postings.  At the moment I don't have a decent way to capture data the way I would like, but I am looking into it.  I could capture data at 96KHz with a card that I have, but my microphone is limited to maybe 40 KHz.  I don't know, I have to review the data.  How the heck I synchronize the data collection, I have to think about that.  I don't want to have any eclipsing losses - that is sampling after the shot has started.  I need to guarantee that I collect the whole shot.  Most of the systems I see, don't prevent eclipsing loss.  That becomes a variable which can vary up to 10dB, which needs to be eliminated.  If you start sampling in the middle of a symmetric pulse, you lost 3dB, as an example.

[WobblyHand]

Why then do I ask for dB measurements for this project?  To satisfy those that insist on objective measurement; and hold that a 2 dB difference is significant.  Based on the available readings, it seems that this three muffler test may have uncovered something about baffle bore diameter and projectile clearance that was missed before. Something that a two baffle bore diameter test cannot show.  If anything, the results suggest the trial be expanded further than +1 and + 2 mm on baffle bore diameters, with with 0.5 mm increments.  Perhaps that can wait until someone who has access to better test equipment and facilities volunteers to run such a trial.

There very well could be a sweet spot for the clearance needed.  To find it empirically, will take some very methodical and repeatable testing.  Not all breakthroughs come from pondering things, sometimes you just plain got to try it out and see what works better.  Wasn't that what Edison said, success is 2% inspiration and 98% perspiration? 

[WobblyHand]

AGN has a tough crowd.  They can be pretty harsh.  Sorry you are dealing with that.  Things are slower here, but at least we don't pick many fights here.  The moderators tolerate a lot less "stuff" here and aren't afraid of the ban hammer.

[subscriber]

There very well could be a sweet spot for the clearance needed.  To find it empirically, will take some very methodical and repeatable testing.  Not all breakthroughs come from pondering things, sometimes you just plain got to try it out and see what works better. 

I agree 100%.

I predict that the ideal bore size depends on power level and how it is achieved:  A short burst of air at higher air pressure, VS long valve dwell at lower air pressure.

If you set up your .22 PCP to shoot your favorite pellets at 30 FPE, then fact that a different pellet at a different tune and velocity might favor another muffler bore diameter may be moot.

I would argue that any dB differences less than 3 dB are academic, but some people on AGN forum are adamant that, "every dB matters".   Reminds me of a Monty Python song

To complicate matters academically, my muffler designs usually do not have constant bore diameters from end to end, unlike the ones designed for this study.   I usually choose a baffle bore diameter for the first baffle, then step the bores open to form a 0.5 degree included angle cone.  So, the longer the muffler, the wider the front baffle  bore gets.  This is purely to guard against potential angular misalignment, and prevent clipping.  If I were to taper the three muffler set baffle bores, that would add another wrinkle to the test.  It also might confuse the general audience, so I went with short mufflers and constant baffle bore sizes per muffler.

[WobblyHand]

I took your advice and tapered the bore in my moderators.  Think it's only a 1/4 degree though.  Since I can machine stuff, I can run closer tolerances if I want to.

As a radar engineer I fought for every dB in receiver sensitivity.  It made a huge operational difference.  Like the difference between crashing and dying or not.  So I understand where some folks are coming from.  I'm not sure that a dB matters for this application.

[WobblyHand]

I predict that the ideal bore size depends on power level and how it is achieved:  A short burst of air at higher air pressure, VS long valve dwell at lower air pressure.

This effectively is describing the characteristics of the air burst, the impulse shape, which has an extremely high influence on whether it sounds objectionable or not.  A true impulse has energy spread everywhere.  Fast attack signals have a lot of spectral splatter as well. 

By having a shaped pulse, ie it's envelope, we can alter its sonic acceptability.  Think about this as a filtering problem (and I am not sure this is a good analogy,) if we could 1) shape the pulse (air blast), we could contain the spectral information to a region, and 2) we could design a filter network to suppress that frequency range.  Writing this down has given me an idea to look into.  Don't know if it will lead anywhere...

What I found was quite interesting, basically a chapter in a book where acoustic stuff is related to electromagnetics.  Now, I know nothing (not much any ways) about acoustics, but I used to know e-mag cold.  But I'm rusty and out of practice.  Need to read some more and see if my idea pans out.  Might not be earth shaking, but could help me understand a methodology to make things better.

[WhatUPSbox?]

Bruce, Since the sound measurement discussion has come up in a number of threads, I researched it a bit and did some of my own measurements. Here are some observations, though with your data processing background you may be entering at a higher level.
Initially, I worried about the response time of the microphone, sampling rate of the soundcard, etc. I set up an OK microphone (imm-6 Dayton), a 96 kHz soundcard, and captured a bunch of traces using Audacity SW. Based on the short duration of the transient, I was thinking that the long sampling windows of the sound level meters (Fast=125ms, Impulse= 30ms I think) must not be applicable.
I then started  looking at the other side of the problem, the human hearing response (psycho-acoustics). First, there is very little data on the hearing response to impulse sounds. There are studies on hearing thresholds of tonal sounds, there are studies on ear damage from impulse sounds, but very little on the ability to hear an impulse sound. What I did find is that the ability to hear short duration sounds drops off below 200 ms and especially below 50-100ms. Down below 50ms, the ear acts like an integrator and it is the energy, not the peak that is important.
So now I'm back to thinking that the 125ms Fast setting on the sound meter may not be too bad and that I probably don't need to worry as much about capturing the sub-ms details of the waveform.
I would be interested to hear your conclusions.

[subscriber]

I see the difference in pressure x duration as taking place in a sealed system.  Sure, there is some noise generated before the projectile leaves the muzzle, but mostly it is the volume of air expelled and the muzzle pressure that drives muffler design.

To achieve a given FPE requires a specific average pressure.   Using very high input pressure and a short valve dwell drives  the bulk of projectile acceleration near the breech end results.  This results in a lower volume of air exiting the muzzle at lower pressure.  Using a low airtank (or regulator) pressure and a long valve dwell to achieve the required average pressure, ends up with a greater air volume leaving, at a higher muzzle pressure. 

The difference in sound is obvious when you shoot an unregulated PCP from a tank pressure so high that the velocity is lightly depressed for that hammer strike.  Then continue shooting until the velocity has peaked and dropped below the starting velocity.  The latter shots are loud, if with a lower tone - less crisp than shooting very short valve duration.

I wonder if the valve duration sound is emanating from the airtank at the input end, because it can't be coming from the muzzle, until the projectile exits.  All I can say is the reason why a short burst of higher pressure air is more efficient is the same reason its muzzle blast is less objectionable:  The air expansion ratio is higher, if the valve closes earlier.  If the valve closes after the projectile has passed half the barrel length, efficiency suffers.  Closing the valve as the projectile exits the muzzle has terrible efficiency and makes any kind of muffler work very hard.

[subscriber]

Bruce,

If you were simply stating that the projectile uncorking from the muzzle produced very high rise rates and that the way the ambient air is slapped by that is what produces the objectionable noise, I agree.

It is like a belly flop making a slapping sound, compared to a streamline dive.  If only we can taper the air pressure leaving the barrel muzzle to make it less of a square wave our ears could not mind so much.  Other than tuning the PCP for lower muzzle pressure, the only other way is to use a muffler - assuming that a four foot barrel if off the table

[subscriber]

Stan,

The sound duration effect on human perception has a wrinkle:

A moderator takes the edge of muzzle blast, and lowers the peak by stretching the event duration.  So, are we making a very short loud sound seem louder by stretching the duration?

One clue about how sharp and loud the original sound was is the presence or absence of noticeable echo, and the duration of the echo.  If I hear a gunshot in the distance, it is not how many dB it is at my location that draws my attention, but the duration of the echo.  Ditto for videos of guns being fired.  Gunshot sound capture has the level suppressed automatically, and it does not hurt our ears to watch the video, the same way watching an arc welding video does not hurt our eyes.  But there are other clues about how loud or bright the sound or light are.

Obviously echo is very much dependent on what surfaces are in the vicinity, but for a given location, the more obvious the echo, the louder the sound is assumed to be at its source.  The video below is of a trebuchet launching a supersonic projectile.  The crack in the open field is obvious and a "rolling echo" is reflected back for a long time, even in the absence of hard surfaces such as walls.  I don't need a dB meter to tell me that passers by will think someone is shooting a firearm not far away from them.  Generally, this sound is exactly the opposite of what we want from our airguns: 

Copy and paste link to start at the point of interest, 14 minutes 44 second in:
www.youtube.com/watch?v=gdXOS-B0Bus&t=884s

[WobblyHand]

If only we can taper the air pressure leaving the barrel muzzle to make it less of a square wave our ears could not mind so much. 

Shaping the sound/air pressure is the idea. 

[WobblyHand]

Stan, measuring the sound through a filter will distort the original information, but one can get some kind of repeatable information, a value.  But if we want to understand how to modify the moderator to get the characteristics we'd like, I'd personally go for minimal filtering (obviously we need to prevent inadvertent Nyquist foldover).  I'm starting to think about the moderator as an acoustic filter.  The size, shape, and spacing of the baffles are changing the acoustic filter response.

I need to set up some acoustic sampling rig in the next few days.  My casual aural observation is that the report with and without a projectile is pretty close for a limited set of subsonic shots.  I'd like to instrument that to know if that is true or not.  If it is "mostly true", it would make designing an acoustic filter, aka a moderator a little easier.  If I could make a filter analogue then I could test it with different pulse shapes and amplitudes, an audio workbench, so to speak. 

[WobblyHand]

This chapter gives me some hope that I could somehow design an acoustic filter using my former E-mag knowledge.  https://phys.libretexts.org/Bookshelves/Electricity_and_Magnetism/Electromagnetics_and_Applications_(Staelin)/13:_Acoustics/13.02:_Acoustic_waves_at_interfaces_and_in_guiding_structures_and_resonators
Used to design microwave filters using apertures in waveguides.  Guess what a moderator is?  An acoustic waveguide.  What we would like is a Gaussian like shape, as opposed to impulses or a reverse sawtooth.  |\

Now to see if I still can do this stuff, it's been decades...

[subscriber]

I think an airgun moderator is more than just an acoustic filter.  I think the last baffle chambers are closer to pure acoustic filters, while the fist baffle chambers function primarily to partition off and expand discreet volumes of air to lower their bulk pressure.  Thus turning the large pop at the barrel muzzle into several smaller pops of diminishing intensity.  The air pressure oscillation from all of the pops are superimposed and filtered to some extent by the chambers down stream.  Then the sound is filtered and partially blocked as you move towards the front.

Obviously, there is no sharp cut-off in function when considering a moderator that has only four chambers.

A related study that could be very useful and fun is to make moderators that have the same cylindrical internal volume, then compare performance based on the number of baffles and their spacing.

The thread in the PCP gate "Taming the bark of an Avenger" has a large body of work in which many moderator designs were tried.  It was there that I developed my own benchmark that I dubbed "plume baffles".  There the spacing and angle of the conical baffles shifts continually.   The early baffles are not slanted as much and are closer together.  This is because the design tracks the bulk pressure of the air as it flow behind the projectile radially towards the ID of the casing.

As the air moves forwards it expands and needs more room.  This goes along with a drop in bulk pressure.  Bulk pressure is what drives the air radially to be more effectively engaged with the next baffle "wall"; away from the exit.  It is this function of bulk pressure that makes some moderators more effective with a projectile than dry firing. 

If there are too many baffles, spaced too close together, I think that air flow takes a shortcut between baffle bores, rather than spreading out to be more effectively engaged by the baffle "walls".  After the bulk pressure has dropped and airflow down the baffle bores is at its maximum velocity, bulk air has considerable forward momentum.  This important, because there is less force spreading the air radial and more inertial resistance to spreading radially.  What my increasing baffle slant does, is help the air flow stay attached to the back side of the cones, rather than break away to continue on down the baffle bores as a narrow "jet".   The lower bulk pressure and the faster the air flow at lower pressure, the more the baffles need to look like a rocket nozzle. 

We assume that the same amount of air is released dry firing, but the resistance of the pellet probably affects  valve duration, and the shape of the leading column of air entering the muffler.  Certainly, the presence of a projectile should act as a moving plug to help drive air into the baffle space, towards the ID of the outer tube wall.  So, any study that insist on the scientific method to represent shooting airgun projectiles, should include shooting airgun projectiles - in my opinion.

Shallow cone angles drive longer baffle spacing as a mathematical certainty.   Shallow baffle spacing will work early on where the bulk air pressure is higher.  But there will be less expansion volume between baffles that are greatly slanted per inch of muffler length.  As muffler length is not "free", and most people prefer shorter mufflers, it seems to make sense to me, to start with more upright baffles cones, and then slant them more towards the front.  Because it helps with air management and because it helps with sound filtering:

I see the sound filtering aspect of muffler design as multi-dimensional:   
The way I slant my baffle cones should also reduce pressure waves bouncing between parallel surfaces, compared to say all 60 degree cones (as used in the test mufflers of the OP).   

Then, I see spacing of the cones as part of noise filtering:   If you want to hear people talking on the other side of a door, you not only get close to the door; you put your ear close to the key hole.  Here, key hole is an analog for the baffle bore ahead of the noisy air in any give baffle chamber.  The closer you are to the hole in the baffle, the wider the angle of sound that escapes through that hole.  The further away from the hole, the smaller the hole seems to the radiated sound source.

An analog of this would be looking out of a window through a small gap between curtains.  If you stand far away from the curtain, you see mostly curtain with very little information seen through the gap in the curtain. If you stand close enough to the gap in the curtain, the curtain block almost none of what can be seen on the other side.  I see this as angular filtering from a radiating source, as a function of stand-off distance.  This perhaps explains why Weihrauch moderators are so effective, despite having only three chambers.  The first is a large chamber (capacitor) to drop bulk pressure.  The next is a flow resistor (that works like its electrical analog).  The last is another larger expansion chamber that has a very long stand-off distance, to make it seem as if the baffle bore at the end is very tight.  Certainly with regard to noise trying to escape the chamber.

So, I think that modelling a muffler would be pretty tricky, but useful if you can do it.  A test of a math model, or set of models is if they predict what happens when you test actual moderators.  If the math model does not predict the empirical results, then the math model may be incomplete.  Some important factors may be missing, because no one knew they were important.   Or the concept of how mufflers work is incomplete, or dare I say, wrong.

The fact that there are so  many baffles types that work really well, suggests there are several overlapping principles at play.  Some design approaches favor one principle, and other designs favor one or more other principles.

"More baffle are better" is countered by looking at the three chamber Weihrauch approach. 
You must "not use constant baffle spacing" is countered by the benchmark of quiet, the Marauder. 
Similar arguments can be made for baffle angle.  Some conical and some flat designs that work very well, can be found with either approach.  My plume style baffles seem to fall in the middle somewhere; and seem to work rather well.  Not just at lowering dB, but at reducing the inherent irritating "crack" signature.  Handsome is as handsome does.  Technology is most interesting, if it works.  Else it is interesting only to avoid emulating it.  Yes, one learns more from failure and fixing it, than accidental success...

End of chapter 3 billion 

[WobblyHand]

You've obviously been thinking about this longer than I have.  Not sure that all your theories are correct, but you certainly have more practical experience.

 I'm thinking that under some conditions using filter theory could be beneficial.  I very well could be wrong, but I think tuning the aperture size and spacing would very well change the filter response, aka user perception.  If I were to jump deep into this, I'd start with simple straight baffles, so the math is easier, and design an acoustic filter to see how it turns out.  I'd like to determine it's impulse response.  Angled baffles might be a perturbation relative to straight ones.  A lot of EMag problems can be solved this way, so I'm guessing acoustic ones might be solved similarly.  Obviously, there's limits to the aperture size as the projectile needs to pass unimpeded. 

Still mulling this over.  It's an interesting problem.

[WobblyHand]

Don't know how far I will get, but I installed the CfdOF workbench in FreeCAD.  This gives me CFD analysis.  I have used the FEM workbench to do finite element analysis. 

CfdOF might give me some insight.  Or it could be a totally frustrating experience...  I'll find out in a while.  I will make a series of constant aperture baffles in a tube, and attempt to puff an impulse (really a square pulse, I hope) into it and see what happens.  And no, I don't quite know what to do, so I'll be making a series of dumb mistakes - but maybe I will learn something.

[WhatUPSbox?]

Well, If you first model a bare muzzle, it should look like this.