Mass loaded top

[charangohabsburg]

Mass loading the top of your plucked instrument can improve its sound.

After I had finished this sopranino ukulele (scale length 245 mm) I found that, like so many other short scaled instruments its sustain was not what it could and should be. Its loudness is just fine, but sustain is a bit on the poor side although it does not sound too shrill. This lead me to the idea of trying to add some weight to the top. As the low resonances in such tiny instruments are almost always a problem (if they are there at all) the idea of adding weight to the top seemed to be more than reasonable to me. So, about two months ago this is what I came up with as a temporary solution:



I placed two rare earth magnets (one inside and another outside) at different locations on the top, experimenting also with smaller and bigger ones, loading the top symmetrically, and so on. I am quite pleased with the most satisfying result you can see in the picture above: two magnets at the treble end of the bridge, the added weight equals 0.85 grams. Interestingly, using four smaller magnets of about the same total weight and loading each bridge wing tip turned you to be quite a bit a disappointment, to say the least. Also, loading only the bass side wing tip does not do the same, which I reckon has to do with the asymmetry in the top doming (the reason of the not symmetrically domed top is another story).

Finally, this weekend I started to do some frequency response analysis before I'll go on to glue in a replacement for the magnets. This will be the weight I'll glue beneath the treble wing tip:



And these are the graphs of the frequency response, without and with additional weight:



Meanwhile you may make (if you like) your conjectures about what exactly is happening when adding the 0.85 grams at the given location I will add some additional streaks to explain what I believe what's going on. Maybe I'll also do some tap-recording with different weights and/or at different locations on the top. Of course, it would be nice to have sound samples of each configuration, but I'm afraid that I am such a lousy player that the differences in recorded sound would be much more due to my never the same playing "technique" rather than to the different weights!
Well, I'll try it anyway.

[simso]

Im not trying to negate your findings, but one of the problems with quantifying a result is you really need to remove the human variable,

At a coursoury glance to me it appears that you have struck the string a fraction harder and the resultant decay is indicative of that fact, now I could be 100 percent wrong and accept that, but to make any result quantifiable you need to make a device that will pluck the string for you, you also need to mind the fact of temperature humidity and string tension.

If you could do this I personally would be interested in the result.

[DarwinStrings]

Interesting stuff learning at this level Markus. So what is happening there? It seems the peak at about 660 has split into two lower peaks. Hmm so is that a top and a back or is it a top and a cross? Seems the only other thing that has happened is that peak out at 1829 has dropped to 1760 whatever it may be. You got any tea leaves?

Jim

[J.F. Custom]

...you have struck the string a fraction harder and the resultant decay is indicative of that fact...

Not so sure on that Steve.

To the best of my knowledge/understanding, the difference in amplitude is quite possibly as you suggest. Assuming however, that the other results are averaged in each mode - as in, multiple readings with and without the weights added, there are both notable similarities and differences.

Mostly, the frequency response is the same - which suggests to me the method/control of tap was very similar in both cases. The difference in a couple of major peaks however is marked and that should not occur I think, simply from "striking the string" (mods - we don't appear to have a 'strike through' option in the BBcode) no - "tapping the saddle" harder. These are the frequencies at which the top is vibrating and should not change that much. Most notably the 550-800Hz range has changed. This I think you will find is the result of the additional weight.

Of course, it would not be the first time I've been wrong though So my opinion only.

Have you done many frequency response charts yourself to get a 'feel' for how it works? Will help to explain why we don't really need to use a mechanical striking method to obtain usable results. Though care in other aspects is required to gain repeatable results.

Jeremy.

[kiwigeo]

Im not trying to negate your findings, but one of the problems with quantifying a result is you really need to remove the human variable,

At a coursoury glance to me it appears that you have struck the string a fraction harder and the resultant decay is indicative of that fact, now I could be 100 percent wrong and accept that, but to make any result quantifiable you need to make a device that will pluck the string for you, you also need to mind the fact of temperature humidity and string tension.

If you could do this I personally would be interested in the result.

I think he's got to be tapping the bridge not plucking the strings...the spectrum doesn't look like that of a plucked string. I thought I'd have issues with wide variations in amplitude when tapping my tops for spectrum analysis work but amazingly Ive found there isn't a huge variation in amplitude or peak frequency for successive runs of tapping. One can reduce any variation by averaging results from a number of runs......I load the runs as a text files from VA into a spreadsheet and manipulate the data on same.

[simso]

Hey Jeremy,

Yep played with them a few times

My point was not really that there is a difference in the display, but merely that it may or may not reflect the achieved result of increased sustain.

I do not doubt that increased sustain has been achieved, if its noticeable then it has been achieved, its the quantifying the end result, to really quantify that it reflects the end result we need to be able measure all aspects of the test, and remnove variables which cannot be exactly duplicated

[J.F. Custom]

...My point was not really that there is a difference in the display, but merely that it may or may not reflect the achieved result of increased sustain.

Ah. Well, this is very different to how I read your post and I could have saved my fingers a lot of typing!

What effect those frequency changes are having in terms of tone, sustain etc is beyond ability to comment on. But the additional weight is having some effect on the resonant frequencies and where they occur - in turn, effecting the way the instrument functions or performs as a whole. This in the same way Trevor adds mass to the sides. Trevor has done enough testing to narrow down targeted frequencies for the results he wants to achieve. As it appears this is Markus' first experiment, it is unlikely he has a set target, but simply used trial and error listening to the effect along the way until he noted it seemed an 'improved' performance in the way he wished.

An experiment still with merit, if not 'scientifically' tight.

I think he's got to be tapping the bridge not plucking the strings

Aye, as I suggested here -

...from "striking the string" (mods - we don't appear to have a 'strike through' option in the BBcode) no - "tapping the saddle"...

Cheers,

Jeremy.

[charangohabsburg]

Steve, I am only looking at the location of the peaks along the frequency axis. I have shifted the two curves "vertically" (along the dB-axis) for better visibility (no overlapping).

Jim, I don't think that the green peak at 660 Hz (measured 658.8 Hz) split up into the two blue peaks at 604 and 772 Hz. I rather believe that the 660 peak is the top resonance T(1,1)2 which with the addition of the weight drops to the blue 604 peak, and that the blue 772 peak which I assume is the back resonance T(1,1)3, maybe already is there in the green graph at about 763 Hz, or maybe was totally absent because the back could not be driven at all by the top.

The picture below is the same as the one in my former posting, only that I have added several faint vertical lines indicating some minor peaks in the graphs that remain just the same (the two greenish faint vertical lines indicate where minor peaks in the green graph convert into shoulders in the blue graph).



The next higher peak above "green 660" I can observe it moves down with adding mass to the top is the one at "green 1830 Hz" you had mentioned, which goes to "blue 1755". I suppose this must be the next higher top mode, most probably the cross dipole T(2,1). I have some poppy seeds and a noise machine but have not run it yet on this sopranino uke.

Just in case someone is interested in the exact locations of the peaks and/or in what happens above 2000 Hz I am adding here the two zipped Visual Analyzer data files.

[kiwigeo]

I ran a bunch of tap tests on my current build (falcate classical) with different sized weights (12, 27 and 32 gram) stuck to the bridge and I noticed a definite drop in amplitude and decrease in frequency of Helmholtz and main top peaks.

I'm at work at the moment but will try and post up the relevant spreadsheet plot.

[charangohabsburg]

As it appears this is Markus' first experiment, it is unlikely he has a set target, but simply used trial and error listening to the effect along the way until he noted it seemed an 'improved' performance in the way he wished.

I have done some tap-graph recordings with VA before, but only on guitars. This is the first small instrument I am looking at its resonance frequencies, and it is the only uke of that size I own.

So, I have noticed the improved sustain when playing with the added weight, and was wondering how the resonant frequencies had changed by adding the weight. I am also noting a "mellower sound" with the added weight, but just by listening I am not able to tell if this new mellowness is just due to the increased sustain or if other factors play a role there.

I am only observing, and I see that not only the top frequency dropped (which of course can be predicted without measuring anything, and that, as a surprise an additional peak got visible, which I assume is the back resonance T(1,1)3. Of course, I can not say if this "new peak" is responsible for the increased sustain I can notice with the added weight, and I also can not say if this "new peak" (blue 772) is the sole responsible for the mellower sound.

[charangohabsburg]

I ran a bunch of tap tests on my current build (falcate classical) with different sized weights (12, 27 and 32 gram) stuck to the bridge and I noticed a definite drop in amplitude and decrease in frequency of Helmholtz and main top peaks.

I'm at work at the moment but will try and post up the relevant spreadsheet plot.

That sounds promising Martin.

I have not compared the amplitude of the peaks, but I have observed that weights over 2 gr. deadened the sound quite a bit, whereas the 0.85 gr. I will glue in give me the possibility to pluck the uke much harder than with no weight added.
I also have to say that this dwarf is built on the light side. Its total weight is about 160 gr., the bridge (including saddle) weighs 1.7 gr., top thickness is about 0.8mm, the sides about 1mm, and so on...

[kiwigeo]

Ok here's a plot of results from a falcate classical guitar (bound with no bridge) and with different weights blue tacked to bridge area. Also presented is spectrum for raw unbound guitar. For each sized weight (20 cent pieces) I took average of 5 runs done with VA (ten taps captured per run). Looking at my results there isn't a huge drop in amplitude but you can see a clear drop in peak frequency with larger weight. Note that for these tests I had the guitar sitting on four foam blocks rather than holding it in playing position......this explains the weird looking 12grm plot below 150hz (not present when holding the guitar)

img-130630215514.pdf

(212.21 KiB) Downloaded 548 times

[matthew]

But Markus, what is your ear telling you about that weight? You didn't really seem to say, and I would have thought it was an important observation to make ...

[charangohabsburg]

Hello Matthew,

Well, I thought I really had said it. With the additional weight it's got more sustain and a mellower sound.
And yes, of course the important thing is the sound of the instrument. I didn't finish this ukulele to look at its spectrograms and Chladni patterns.

[charangohabsburg]

With the additional weight it's got more sustain and a mellower sound.

Me, trying to play something:

- first minute with the additional weight,
- during the second minute the same theme without weight


youtu.be/

This or next week I'll finally glue a little weight (also 0.85 grams) to the inside of the top which will be the definitive end of experiments with this ukulele.

[Trevor Gore]

Not sure how I missed this one first time round...

Anyway, what I see with the added mass:

1) The mass has been added on what I would expect to be an antinode of the cross dipole
2) The cross dipole has dropped in frequency, emerging from a combined peak with some other mode, possibly a long dipole, hence revealing the two peaks
3) These graphs don't say very much about sustain; a waterfall plot is required for that
4) The extra sustain will be due to the flywheel effect of the extra mass

Have to say, I'd struggle to pick it from the recording in a blind test, though when I A-B'd some sections there seemed to be a clear difference. But was it the instrument or the player??

Of course, the difference might be more obvious to the player than an audience.

[charangohabsburg]

3) These graphs don't say very much about sustain;

I never have alluded to such a thing. The difference in sustain here is just what one can hear more than clearly, even through Youtube and cheap computer speakers.

Have to say, I'd struggle to pick it from the recording in a blind test, though when I A-B'd some sections there seemed to be a clear difference. But was it the instrument or the player??

Of course, the difference might be more obvious to the player than an audience.

I have recorded the same stuff over and over again in order to pick the pair of recordings where I did make the least mistakes. Listening through the whole stack of recordings made it obvious that the difference in sound is very consistent: the recordings without mass loaded top always sucked more than the others. Not one exception. However, I really would hate to let you suffer by serving the whole load of recording attempts.

One problem with comparing sound is the time between listening to two samples. It is not easy to memorize quality of sound over more than a few seconds. So here we go with a much shorter sample pair, not showing or saying yet which is which:


youtu.be/

[DarwinStrings]

Hmmm, it does sound like that bass note sustains a little better in the first sample.

Jim

[Trevor Gore]

3) These graphs don't say very much about sustain;

I never have alluded to such a thing.

Yes, I know, Markus. But others did:

My point was not really that there is a difference in the display, but merely that it may or may not reflect the achieved result of increased sustain.

My system won't properly download your 7 second clip, (keeps hanging for some reason) so I've only heard the first two seconds. But I remember that section from the first video and there is a huge difference there. So big that I'd put it down to the fingering!

So here's a real test of my long term audio memory. If it's not the fingering, I'm thinking that the first two seconds is without the added mass.

[charangohabsburg]

3) These graphs don't say very much about sustain;

I never have alluded to such a thing.

Yes, I know, Markus. But others did:

My point was not really that there is a difference in the display, but merely that it may or may not reflect the achieved result of increased sustain.

OK, I misinterpreted your statement, Trevor. Obviously, my long-term text memory failed.

But I remember that section from the first video and there is a huge difference there. So big that I'd put it down to the fingering!

I may be a bad ukulele player, but not that bad!

Actually, it would be pretty hard to make it sound like without added mass when the mass is added. Something like fingering right over the frets instead of in the spaces would be needed, but without muting the uke completely. I also tried to do that, but I'm not good enough to get even close to that, it is much more difficult than let's say playing harmonics over the 5th, 4th and 3rd fret on the guitar.

The short section in the second comparison is where the difference gets most obvious: everything played more up the fretboard than fret four sounds bad without added mass.

So here's a real test of my long term audio memory. If it's not the fingering, I'm thinking that the first two seconds is without the added mass.

Yes, the first half of the 8 second clip is without added mass.

[Trevor Gore]

I may be a bad ukulele player, but not that bad!

With my dumpy, gnarled fingers everything I attempt to play on a uke sounds damped and muted! Apologies for the psychological projection!

I finally figured why my system wouldn't download your short clip properly. It was receiving a Windows "Automatic Update".

#$%^ !

[charangohabsburg]

Hmmm, it does sound like that bass note sustains a little better in the first sample.

If you refer to the short sample this is because in the second half the higher notes sustain as well. The lower notes (especially the not fretted strings) are not a problem, but the higher ones sound really bad.

I finally figured why my system wouldn't download your short clip properly. It was receiving a Windows "Automatic Update".

#$%^ !

If you don't have a problem Windows will make one (or two).

amour01.jpg (34.5 KiB) Viewed 23298 times

[Trevor Gore]

There should be a caption competition for that picture.

I'll start clean and simple: "Ebony on a rainy day"

Bob! Get ready to pull the plug....

[charangohabsburg]

There should be a caption competition for that picture.

I'll start clean and simple: "Ebony on a rainy day"

"No Action"

[kiwigeo]

The day Microsoft started making guitars.....