Mystery of misbehaving F3 note.

[Łukasz Martynowicz]

In the pitch stability thread I’ve mentioned my misbehaving F3 note. I’ve done some more tests and I’m fairly certain that I’ve managed to norrow down the likely cause/causes. Attached is a screenshot of spectum analysis from Audacity: on the left are two repeats of LH/RH comparison (I have two F3s) with endplates mounted (a complete instrument) and on the right is LH/RH comparison without the endplates (only bellows and single reedpan/action board on the bench. It is the LH note that is problematic. Volume difference in the bench part is only partially caused by my failure to recreate exactly same pressure (this is large box, 8 2/3”, and is as such is VERY responsive), this note does sounds percievably louder during normal play, but the 2khz problem persists after decreasing the volume by decreasing padhole diameter.

 

First a comprehensive list of all manipulations that do not affect this problem:

reed and valve: this problem does not follow the reed when I switch reeds between sides

chamber volume and coupling: chambers on both sides are the same and introducing a controlled airleak does not change anything

chamber placement on the reedpan/relative to the fretwork: since this is hybrid Hayden I have 4 chambers of this size all over the reedpan and moving the reed around does not alter this particular tone quality.

padhole size only affect overall volume

baffles, of any kind, outer, inner, partial, full, even nearly airtight, do not alter this particular tone quality. This also includes antlers/hand placement.

 

Now, the only manipulation that switches the side of the problem is switching endplates. But as I wrote above, it has nothing to do with reed placement relative to fretwork/hand. 

 

So, if I did not ommit something else, there is only one (and a half ) possible culprit. Both endplates share the same pattern, except for number of button holes and thus a volume of solid wood left in the LH endplate (this also means larger continuous (I’m not sure about the word here, entire endplate is obviously continuous, what I mean is „without any hole in it”) area in the center of the endplate). Other than that („and a half” ) there might be a slight difference in the thickness of shellac coating on the LH as I had to make some revisions to it. When tapping on the endplates they do make different sound, with LH being slightly lower.

 

Now three questions to more experienced builders:

have I forgot to test something else

have you encountered a similar problem

do you think that removing wood on the inside of the endplate will suffice or I have to bore another hole through the surpluss continuous area.

 

This problem is the most audible with F3s, but I now think it also influences other notes slightly, up to G#3, just not to the point of dominating the chord/octave/side ballance so it have slipped my ear before yesterday’s investigation. It might also be a part of the reason why my F2-A2 range has a clear trumpet quality to it.

 

One last curiosity, I have encountered the same problem of F3 standing out on entirely different box, a 7”, 3d printed hexagon with normal alluminum accordion reeds and not brass DIX reeds as in the box in question.

Edited September 8, 2021 by Łukasz Martynowicz

[d.elliott]

To my utter shame, your highly decorative printouts mean nothing to me, I have never used Audacity. but a sound file might help me, meanwhile I shal have a look at your pitch & stability thread

[Łukasz Martynowicz]

3 hours ago, d.elliott said:

To my utter shame, your highly decorative printouts mean nothing to me, I have never used Audacity. but a sound file might help me, meanwhile I shal have a look at your pitch & stability thread

Here is an audio source for the left part of the screenshot. 1st and 3rd is the bad reed, 2nd and 4th is the good one.
 

 

ZOOM0001_MS120.wav

[alex_holden]

I don't have a solution, but I have noticed on traditional-style instruments that you get similar tonal differences between chambers that have the pad close to an outer wall and ones that are closer to the centre of the pan (what I call "inner chambers"). I've not yet figured out why it happens or how to prevent it, other than making the instrument bigger and moving the chambers to the outside of the pan. The effect seems stronger when the affected reeds are lower in pitch.

[Łukasz Martynowicz]

4 hours ago, alex_holden said:

I don't have a solution, but I have noticed on traditional-style instruments that you get similar tonal differences between chambers that have the pad close to an outer wall and ones that are closer to the centre of the pan (what I call "inner chambers"). I've not yet figured out why it happens or how to prevent it, other than making the instrument bigger and moving the chambers to the outside of the pan. The effect seems stronger when the affected reeds are lower in pitch.

At first I thought so too, as I remembered discussions from years ago about those tonal differences you write about. But this particular quality does not change when I move this note around four different chambers on the LH reedpan. 
 

Now regarding what you describe I always wondered if it is because in traditional concertinas the back wall of the chamber is decoupled from the pan, while with inner chambers all walls are connected (either glued or carved from a single block altogether).

[alex_holden]

6 minutes ago, Łukasz Martynowicz said:

At first I thought so too, as I remembered discussions from years ago about those tonal differences you write about. But this particular quality does not change when I move this note around four different chambers on the LH reedpan. 

 

Do you mean, if you put an F3 reed in a particular chamber it sounds poor, but if you swap an E3 or F#3 reed into the same chamber without changing anything else it sounds fine?

 

9 minutes ago, Łukasz Martynowicz said:

Now regarding what you describe I always wondered if it is because in traditional concertinas the back wall of the chamber is decoupled from the pan, while with inner chambers all walls are connected (either glued or carved from a single block altogether).

 

I don't know. I have tried adding chamois gasket material to the back wall of an inner chamber (on the theory that maybe the bellows frame lining is absorbing some harsh harmonics), and I couldn't hear any difference. Also @Dana Johnson's reed pans seem to be routed from a single block and have solid back walls, even on the outer chambers.

http://www.kensingtonconcertinas.com/photo-album/5-inch-reed-pans.html

 

My current best guess is that it has something to do with how the sound resonates inside the action box. I tried fitting a small dummy wall inside the action box next to an inner chamber pad, with again no noticeable difference.

[Łukasz Martynowicz]

1 hour ago, alex_holden said:

 

Do you mean, if you put an F3 reed in a particular chamber it sounds poor, but if you swap an E3 or F#3 reed into the same chamber without changing anything else it sounds fine?

 

 

I don't know. I have tried adding chamois gasket material to the back wall of an inner chamber (on the theory that maybe the bellows frame lining is absorbing some harsh harmonics), and I couldn't hear any difference. Also @Dana Johnson's reed pans seem to be routed from a single block and have solid back walls, even on the outer chambers.

http://www.kensingtonconcertinas.com/photo-album/5-inch-reed-pans.html

 

My current best guess is that it has something to do with how the sound resonates inside the action box. I tried fitting a small dummy wall inside the action box next to an inner chamber pad, with again no noticeable difference.

Only F3 reed and maybe G3 reedto lesser extent behave like this and perhaps F2 and G2 are affected but those do not have direct counterparts on the RH side so I can’t easily verify this. If I move this F3 reed to other four chambers of same dimensions the problem stays with the note, not the chamber, despite not being related to the reed itself. As I wrote, this problem stays with the instrument side, nothing else seem to affect this particular rouge frequency spike. Of course moving the reed around changes the tone as you would expect, but more on the bright-vs-cassotto dimension. 
 

Dana’s reedpan design indeed falsifies my intuition. 
 

Same as you, I have also tried fitting a dummy wall behind this pad, and same as you got no results.

 

 

[dabbler]

I'm not a master builder but here's a thought.  Maybe the resonance of your fretwork is tuned such that it absorbs energy from the F3 frequency.  That is, the fretwork is acting like a tongue drum tuned in a way that interferes with the F3.  You could probably test this by sticking hard modeling (plastalina) clay or mounting putty to the fretwork to alter or eliminate the resonances.

Edited September 10, 2021 by dabbler

[Geoffrey Crabb]

Being a simple concertina maker/repairer (now retired) and a technology Luddite, I am somewhat confused by the following part of your statement which seems contradictory:

 

23 hours ago, Łukasz Martynowicz said:

 If I move this F3 reed to other four chambers of same dimensions the problem stays with the note, not the chamber,  despite not being related to the reed itself. 

If,

    1. 'the problem stays with the note, not the chamber'. This suggests that there is some anomaly with that particular reed. 

 

   2. 'despite not being related to the reed itself.'  have you actually proved this by careful visual examination and comparison with the other ,non affected, F3 reed?

 

If all the gaps etc are the same, perhaps the profiling is different. 

 

Just a thought

 

Geoff

 

 

Edited September 10, 2021 by Geoffrey Crabb

[Łukasz Martynowicz]

1 hour ago, Geoffrey Crabb said:

Being a simple concertina maker/repairer (now retired) and a technology Luddite, I am somewhat confused by the following part of your statement which seems contradictory:

 

If,

    1. 'the problem stays with the note, not the chamber'. This suggests that there is some anomaly with that particular reed. 

 

   2. 'despite not being related to the reed itself.'  have you actually proved this by careful visual examination and comparison with the other ,non affected, F3 reed?

 

If all the gaps etc are the same, perhaps the profiling is different. 

 

Just a thought

 

Geoff

 

 

As I wrote above, there are two F3 notes in that box, one on each side. When I swap reeds between sides the problem stays with the side of the instrument, not any of those two reeds (four tongues, this is a hybrid).

 

1 hour ago, dabbler said:

I'm not a master builder but here's a thought.  Maybe the resonance of your fretwork is tuned such that it absorbs energy from the F3 frequency.  That is, the fretwork is acting like a tongue drum tuned in a way that interferes with the F3.  You could probably test this by sticking hard modeling (plastalina) clay or mounting putty to the fretwork to alter or eliminate the resonances.

That is my current theory, and I have made mass tests by adding putty to both sides, to either fix LH or recreate the problem on RH, but to no avail (but perhaps I have simply used too little added mass, such composite system probably acts differently to solid wood and I only added enough to cover the difference in geometry). But I figure this is a geometry issue if the endplate rather than mass of the endplate (the area of the „tongue drum” part). My current plan is to 3d print temporary endplates of the same geometry and work from there with carefull alterations in hope to pinpoint the cause and then apply results to the real endplate. But first I’ll try to butcher the endplate on this other box showing the same symptom, as it is a „single serving” 3d printed travel box made for this year’s vacations, now sadly behind me. I’ll report results of this butchering in a day or two.

 

[Łukasz Martynowicz]

Ok. I have revisited the problem with a proper real time spectrum analysis app and have some findings and a good enough workaround but I still have no clue about why the difference is there.

 

So, the problem does not originate in the endplate/action box, it just so happens that with F3 it emphasises the problem so much that it was unacceptably large and could not go by as "flavour". Now that I know what exactly is happening I could check all notes and as it turns out it does affect not only F3 (it is just the most offensive note), but assorted other notes, mostly on the LH side, with just a couple on the RH side, because lower notes are generally more affected by this. What is happening is that some notes have 2nd partial reduced and this energy goes into several upper partials, creating a gap and a "trumpet like" timbre. Only a single note, A3 is affected equally on both sides of the instrument. My best guess about why those differences are there is that either the amount of leftover reedpan material or exact chamber neighbourhood (layout of surrounding voids) interferes with some frequencies. So Alex, it turns out that this is related to what you wrote earlier, but it is not directly related to distance to outer edge, or anything else that I could pinpoint. At this point I simply accepted this phenomenon and moved to finding a workaround.

 

On 9/9/2021 at 7:42 AM, alex_holden said:

I don't have a solution, but I have noticed on traditional-style instruments that you get similar tonal differences between chambers that have the pad close to an outer wall and ones that are closer to the centre of the pan (what I call "inner chambers"). I've not yet figured out why it happens or how to prevent it, other than making the instrument bigger and moving the chambers to the outside of the pan. The effect seems stronger when the affected reeds are lower in pitch.

 

What I've found worked well enough was to increase the depth of chambers in question by 50% (I've tested this in 10% increments) which promoted lower partials just enough while not muffling too much upper partials. I've fixed three notes at this point so it's a reliable enough solution.

 

So now I have an instrument that requires that chambers on each side of the instrument differ substantially in order for two identical reeds to sound uniformly enough. Odd.

Edited September 12, 2021 by Łukasz Martynowicz

[ttonon]

Hi Lukasz, Audacity enables the calculation of a frequency spectrum showing a rectilinear graph, with sound intensity on the vertical axis and frequency on the horizontal axis.  I believe that such a representation is much more useful than the kind of spectrum you posted here.  If you can't figure out how to get that graph on Audacity, let me know and I can perhaps lead you through it.  

 

More basically, I don't know what your "problem" is.  Apparently there's a problem with the sound of two different F3 reeds when alternatively mounted in the same cavity.  Is this correct?  If so, could you please explain in detail the problem in the sound?  Do you think you see the "problem" in the spectrum you posted?  If so, please explain.  In the "Bellows pressure and musical pitch" thread, if I recall, you thought that the poor sound was because of a pronounced harmonic and I suggested you do a spectrum analysis.   Do you still think the  problem is with a pronounced harmonic?  If so, the kind of spectral graph I describe will be most useful.  

 

Regards,

Tom Tonon

www.bluesbox.biz

Edited September 12, 2021 by ttonon
Signature

[d.elliott]

From time to time I have experienced the this sort of problem on various instruments, I have never truly understood the issue but I have found that the following has helped. running a 400 frit file round the frame vent and across the top surface of the frame to ensure a clean edge around the reed vent, and the removal of any burr, then polishing the flanks, and tip of the reed tongue to ensure no burr or accumulation of dirt or even corrosion. Then re-centralising the tongue in the vent.

 

I can recall changing the reed tongue in one instance because the reed had been filed in such a way that the filed surface was far from parallel to the under side and that the filing seemed to be closer the tip than I would have expected. This did work, emphatically!.

 

I suspected that the reed was beating with a figure of eight path, and that there was possibly a slight secondary flex to the reed. I am talking an old and traditional reed here.

[Łukasz Martynowicz]

10 hours ago, ttonon said:

Hi Lukasz, Audacity enables the calculation of a frequency spectrum showing a rectilinear graph, with sound intensity on the vertical axis and frequency on the horizontal axis.  I believe that such a representation is much more useful than the kind of spectrum you posted here.  If you can't figure out how to get that graph on Audacity, let me know and I can perhaps lead you through it.  

 

More basically, I don't know what your "problem" is.  Apparently there's a problem with the sound of two different F3 reeds when alternatively mounted in the same cavity.  Is this correct?  If so, could you please explain in detail the problem in the sound?  Do you think you see the "problem" in the spectrum you posted?  If so, please explain.  In the "Bellows pressure and musical pitch" thread, if I recall, you thought that the poor sound was because of a pronounced harmonic and I suggested you do a spectrum analysis.   Do you still think the  problem is with a pronounced harmonic?  If so, the kind of spectral graph I describe will be most useful.  

 

Regards,

Tom Tonon

www.bluesbox.biz

There’s a .wav file with this problem attached to one of my previous posts. I have now switched to real time app and indeed the freq/volume graph is what I’m using. The problem is with missing/reduced second partial and shifting this energy into 2-4 higher ones (how many depends on the note in question) and as I wrote in previous post, it turned out this affects not only F3 but few other notes as well. F3 was most audible because the endplate emphasised this further in case of this note. 
 

And the problem is not with two reeds alternatively mounted in the same cavity, but a single reed alternatively mounted in two identical cavities. I just have two of those reeds, also identical, so I can swap them around. The problem stays with LH cavity. 

 

1 hour ago, d.elliott said:

From time to time I have experienced the this sort of problem on various instruments, I have never truly understood the issue but I have found that the following has helped. running a 400 frit file round the frame vent and across the top surface of the frame to ensure a clean edge around the reed vent, and the removal of any burr, then polishing the flanks, and tip of the reed tongue to ensure no burr or accumulation of dirt or even corrosion. Then re-centralising the tongue in the vent.

 

I can recall changing the reed tongue in one instance because the reed had been filed in such a way that the filed surface was far from parallel to the under side and that the filing seemed to be closer the tip than I would have expected. This did work, emphatically!.

 

I suspected that the reed was beating with a figure of eight path, and that there was possibly a slight secondary flex to the reed. I am talking an old and traditional reed here.

The strangest thing is, this is not because of anything with the reed itself.

 

Edited September 12, 2021 by Łukasz Martynowicz

[Łukasz Martynowicz]

11 hours ago, ttonon said:

Hi Lukasz, Audacity enables the calculation of a frequency spectrum showing a rectilinear graph, with sound intensity on the vertical axis and frequency on the horizontal axis.  I believe that such a representation is much more useful than the kind of spectrum you posted here.  If you can't figure out how to get that graph on Audacity, let me know and I can perhaps lead you through it.  

 

More basically, I don't know what your "problem" is.  Apparently there's a problem with the sound of two different F3 reeds when alternatively mounted in the same cavity.  Is this correct?  If so, could you please explain in detail the problem in the sound?  Do you think you see the "problem" in the spectrum you posted?  If so, please explain.  In the "Bellows pressure and musical pitch" thread, if I recall, you thought that the poor sound was because of a pronounced harmonic and I suggested you do a spectrum analysis.   Do you still think the  problem is with a pronounced harmonic?  If so, the kind of spectral graph I describe will be most useful.  

 

Regards,

Tom Tonon

www.bluesbox.biz

 

For your convenience, here are audacity screenshots, first is the bad one.

[ttonon]

Hi Lukasz, I just now heard your wav file and I agree, the 1st and 3rd sounds seem restricted, and the 2nd and 4th more open or free.  In my experience, such a restricted sound occurs when the mounting of the reed - most often the geometry of the cavity - begins to acoustically interfere with tongue vibration.  For instance, if the port is too small, or even if the leather valve isn't opening completely enough.  Alternatively, there could be some defect in the construction of the reed, such as a burr, or other way to interfere with tongue vibration.  

 

Your guess that the problem lies in a restricted second partial seems to be borne out by the spectrum.  In the bad note, that partial is about 13 dB below the fundamental, and in the good note, it's only about 5 dB below the fundamental.  I'm not sure there's evidence in the spectrum for your claim that the energy of that second partial is fed into the higher partials, which from my experience would be an odd occurrence, knowing what I know about the conversion of pressure pulses to audible sound. 

 

We can notice that the bad note has also a greatly reduced 8th partial, compared to that of the good note, but I doubt that has anything to do with what we hear, since the pitch of that partial is about 22,000 Hz, far above our normal hearing range.

 

However, your discovery of this reduced second partial may give us a clue.  If you'd go through the trouble of presenting here all the exact 3-D dimensions of the cavity and the size of the port, I'd be glad to take a look at it, possibly finding an identification of the second partial frequency with some way the cavity could resonate.  If it's a simple rectangular cavity, it should be easy to do.  

 

Best regards,

Tom 

 

 

 

 

 

[Łukasz Martynowicz]

1 hour ago, ttonon said:

Hi Lukasz, I just now heard your wav file and I agree, the 1st and 3rd sounds seem restricted, and the 2nd and 4th more open or free.  In my experience, such a restricted sound occurs when the mounting of the reed - most often the geometry of the cavity - begins to acoustically interfere with tongue vibration.  For instance, if the port is too small, or even if the leather valve isn't opening completely enough.  Alternatively, there could be some defect in the construction of the reed, such as a burr, or other way to interfere with tongue vibration.  

 

Your guess that the problem lies in a restricted second partial seems to be borne out by the spectrum.  In the bad note, that partial is about 13 dB below the fundamental, and in the good note, it's only about 5 dB below the fundamental.  I'm not sure there's evidence in the spectrum for your claim that the energy of that second partial is fed into the higher partials, which from my experience would be an odd occurrence, knowing what I know about the conversion of pressure pulses to audible sound. 

 

We can notice that the bad note has also a greatly reduced 8th partial, compared to that of the good note, but I doubt that has anything to do with what we hear, since the pitch of that partial is about 22,000 Hz, far above our normal hearing range.

 

However, your discovery of this reduced second partial may give us a clue.  If you'd go through the trouble of presenting here all the exact 3-D dimensions of the cavity and the size of the port, I'd be glad to take a look at it, possibly finding an identification of the second partial frequency with some way the cavity could resonate.  If it's a simple rectangular cavity, it should be easy to do.  

 

Best regards,

Tom 

 

 

 

 

 

 

That's exactly the mystery though - ALL tested cavities (this is a hybrid, I have more than two cavities of same dimensions), good and bad are 39mm x 14mm x 10mm and ports are 10mm diameter (and I also tested it with 12mm, 9mm, 8mm and 7mm ports). Graphs you see above are identical if I switch the reeds side to side or if I test any other identical chamber on the LH. I have two reeds (four tongues, four valves) and the problem always stays with LH side of the instrument for F3 note, with or without endplates, in both airflow directions and with all possible flips and orientations of those four tongues. Some of those chambers identical chambers are outer (basic two in question are both outer, but one of them is perpendicular to the closest outer edge and one is at 45 degrees) others are more or less inside the reedpan.

 

Regarding "feeding into higher partials" - audible volume of the bad reed is louder than the good one (though dB meter stays at +/- 1dB spread between sides), so I figured that this energy has to go somewhere and on those graphs you have bit different distribution of dB levels of higher partials. You're the expert though, I'm just a more and more confused amateur builder looking for solution even more than the cause.

 

Now the trick with increasing depth of the bad side chamber by 50% does change the spectrum enough for those notes to blend into the accompaniment well enough to pass as a flavour instead of the right out interruption, but this is a not ideal workaround.

 

 

Edited September 12, 2021 by Łukasz Martynowicz

[ttonon]

Lukasz, from my calculations, I see no possible resonance interference with the tongue vibration.  This applies to both Helmholtz and quarter-wave tube resonance.  

 

I thus cannot make a suggestion here, only I'm still not clear on what the basic problem is.  Are you saying that all the F3 reeds you put into the "bad" cavity show the same problem?  Also, that when you put those same reeds into a different cavity but with the same dimensions of the "bad" cavity, there is no problem?  

 

Best regards,

Tom