An Interesting Dulcimer Experiment

Just share stories or offer advice

Hardwood back VS Soft Wood Back

Postby fortytwo » Wed Jan 07, 2009 11:03 pm

This thread seems to confirm what I began to suspect about 3 years ago when a misunderstanding with a dulcimer I ordered resulted in a poplar back and cherry top & sides - the reverse of what I asked for. It's a mini with 17 1/4 inch VSL and in spite of having the thickest (1/8 ") top and back -of any dulcimer I own it has the longest sustain and is nearly the loudest.

To test that theory I purchased a kit, advertised as walnut with poplar soundboard with the idea of again reversing the build. Haven't gotten "roun-to-it" yet, but maybe in the next couple of weeks. Will post the results when I do.

Wonder what a Tennessee Music Box style with a poplar back and second hardwood in double back style would do?
fortytwo
Senior Mbr (101-500 posts)
 
Posts: 203
Joined: Fri Feb 18, 2005 8:05 pm
Location: Cape Coral, FL 6 mo; wher

Postby Ken Bloom » Thu Jan 08, 2009 8:58 am

I have built several TMBs completely of poplar and they were cannons. Great tone and PLENTY of volume.

Ken Bloom
http://www.boweddulcimer.com
Ken Bloom
Super Mbr (501-2000 posts)
 
Posts: 1819
Joined: Thu Dec 08, 2005 4:55 pm
Location: Pilot Mountain, North Car

Postby rtroughear » Thu Jan 08, 2009 10:41 am

G'day all

I shouldn't even look at this forum before I go to bed - people keep raising matters.

Paul: Clamping the ends of a mountain dulcimer in order to make some measurements is fundamental to what I want to propose as a sound production model for the instrument, which I've only been hinting at so far. It's almost a make or break requirement. I've had a go at clamping, and found it extremely difficult to immobilise the two ends of a dulcimer. In fact, almost impossible. Here's a picture of an attempt to clamp a dulcimer to a large beam (which I could hardly lift; I was away from home and it was the biggest thing I could find). It was clamped nearly to the point of crushing the scroll and end block.

http://i202.photobucket.com/albums/aa91 ... tial04.jpg

To my surprise, the dulcimer sounded louder, and better, clamped to the log - substantially louder and better. More than that, the log itself was vibrating strongly!

So now I'm going to have to dyna-bolt some hardwood blocks to a concrete log, and glue my dulcimer to the blocks, so there's no chance of slipping, or of the concrete bending.

........but


Everyone seems to be trying to explain mountain dulcimers by invoking complex and esoteric mechanisms, when the core basics have not been explained. There's a lot of running, and not very much crawling, and I really don't think we are past the crawling stage yet (in understanding, not in manufacture). So I'll make a basic proposal, well ahead of definitive proof; and in due course, experiment will prove this to be fundamentally true or not.

The production of sound in a mountain dulcimer can be largely explained by treating the instrument as a vibrating bar rather than as a group of vibrating plates.

[Edit: See post dated 11 March 2010 for an update on this hypothesis]

When the instrument is treated as a complex bar, a heck of a lot of things fall into place, at least things that I could get no answers for.

For example, former considerations about how the top and back vibrate resolve into: how do they contribute to the overall stiffness of the box/bar structure. Their individual contributions become less relevant. So, whether there is a back, or a top, or both, it doesn't matter - what matters is the stiffness of the structure, and how that translates into bar-like vibrations. Any large surface will vibrate as part of the "bar" and produce sound (think marimba bar, only hollow and more complex in shape). The reason the topless dulcimer did not lose loudness was because the bottom was still vibrating as much as the top formerly was, and the inside of the back became the new "top".

This box/bar hypothesis deserves a new thread to be presented fully, and I don't want to make a full proposal before I've done some more supporting experiments. But so far, nearly all of what I know about the instrument can be better explained if it is treated as a bar, rather than thinking about which bits are individually vibrating.

That is the BASIC proposition. Overlaid on that will be all the subtleties and mysteries that all wooden musical instruments exhibit - air vibrations, air/wood interactions, localised wood resonances, etc, all too complex for the mind to grasp. Mostly, we seem to have been addressing the subtleties, which is fair enough when talking about the distinctions between two dulcimers, and what needs to be done to produce a charateristics type of sound. But that has never explained to me how the instrument basically works. This is an attempt to provide that underpinning explanation.

I don't fool myself that this simple proposition will explain everything, but it has already got me thinking along different lines when I make my next instruments, so it already has practical implications for me.

So there's something to think about while I'm sleeping...

Richard T
Last edited by rtroughear on Thu Mar 11, 2010 7:50 am, edited 3 times in total.
rtroughear
Super Mbr (501-2000 posts)
 
Posts: 603
Joined: Tue Jun 05, 2007 8:18 am
Location: Brogo near Bega, NSW, Australia

Postby harpmaker » Thu Jan 08, 2009 10:56 am

In regards to clamping it solidly, try this:

Take a pair of posts about table top high. Set them the appropriate distance apart with the botom ends in buckets of concrete or stone & clamp the dulcimer onto the tops of the posts. Steel posts would work good. You can lock the pots together solidly but with minimal sound generating surface with a simple X brace that has a center bolt, like used on metal shelving.

This would firmly isolate the instrument in the air and also eliminate the possum board effect your workbench is adding to the results.
harpmaker
Dulcified! (>2000 posts)
 
Posts: 4687
Joined: Mon Feb 10, 2003 12:33 am
Location: Pershing, IN

Postby rtroughear » Thu Jan 08, 2009 11:32 am

Dave

That's not a bad idea for solidly mounting an instrument in space, but I need to make sure the whole instrument can't flex. Even steel star pickets bolted between the tops of the two posts might allow some flexing. But it's worth persuing.

A bar needs its end to be free for it to vibrate; so rigidly fixing the ends should basically kill most of the normal sound of a mountain dulcimer if it is acting like a bar. It might still vibrate like a "string" in that configuration - I'll have to find out about constrained end bar vibrations compared to free end, but it is unlikely to be a normal dulcimer sound in frequency spectrum and amplitude. If sufficiently immobilised, putting a pickup on the ends should show much less vibration than when free.

I'm not so worried about testing on the bench. The instrument is isolated from directly passing energy to the bench, and any reflections from the back, off the bench, are largely away from the mic which is directly overhead. I'm looking for gross effects anyway.

Richard T
rtroughear
Super Mbr (501-2000 posts)
 
Posts: 603
Joined: Tue Jun 05, 2007 8:18 am
Location: Brogo near Bega, NSW, Australia

Postby PaulC » Thu Jan 08, 2009 9:16 pm

rtroughear wrote:Dave

That's not a bad idea for solidly mounting an instrument in space, but I need to make sure the whole instrument can't flex. Even steel star pickets bolted between the tops of the two posts might allow some flexing. But it's worth persuing.


If you have a wood lathe, you can suspend it between centres....


A bar needs its end to be free for it to vibrate;...


Not true, nor does a rod with one fixed end represent a dulcimer. A bar can, and will vibrate between fixed points. This is what you have in a dulcimer. The fretboard will vibrate in some net sympathy with the string action. A vibrating string fixed at the ends produces a standing wave at some given frequency, which moves air, AND it produces a image frequency between its fixed nodes (bridge <--> nut), so that the entire instrument is resonant and constitutes a standing wave pattern tuned to some note. In other words, the vibration pattern is a complete circle from nut to bridge and back to nut again, all resonant with the tuned frequency of the instrument, though the frequency of any individual segment may be different due to different densities/propagation. If the boundary propagation between materials was perfect, ie elastic, then the instrument would sustain for ever, but since wave propagation involves losses, the sound fades (and the standing wave pattern changes frequencies too, if you could see it. IF you have a strobable light, you may be able to see the standing wave pattern in the strings. You can see in some lighting conditions, and you can see the wave pattern change with frequency as it decays.)

The bridge and nut are nodes - points of complete cancellation of waveform. The number of peaks and nodes between the bridge and nut are related to the frequency of the entire instrument. The string is just the adjustable part, and the part that contributes the greatest amplitude individually. However, its the sum total of the instrument's amplitude that contributes its volume.
PaulC
Senior Mbr (101-500 posts)
 
Posts: 189
Joined: Thu Nov 06, 2008 9:43 pm
Location: Southern Ontario

Postby harpmaker » Thu Jan 08, 2009 9:47 pm

rtroughear wrote:Dave

That's not a bad idea for solidly mounting an instrument in space, but I need to make sure the whole instrument can't flex. Even steel star pickets bolted between the tops of the two posts might allow some flexing. But it's worth persuing.

A bar needs its end to be free for it to vibrate; so rigidly fixing the ends should basically kill most of the normal sound of a mountain dulcimer if it is acting like a bar. It might still vibrate like a "string" in that configuration - I'll have to find out about constrained end bar vibrations compared to free end, but it is unlikely to be a normal dulcimer sound in frequency spectrum and amplitude. If sufficiently immobilised, putting a pickup on the ends should show much less vibration than when free.

I'm not so worried about testing on the bench. The instrument is isolated from directly passing energy to the bench, and any reflections from the back, off the bench, are largely away from the mic which is directly overhead. I'm looking for gross effects anyway.

Richard T


I guess I am getting lost as to what you are trying to achieve with your experiments. It looked as if you were trying to lock the ends down as firmly as possibly, which to me at least, would obviously affect the sound.

Don't underestimate the possum board effect. I have seen player take a dulcimer with a simple sound board transducer attached and set it on a possum board with a noticeable increase in the sound out put at the amp. Our WAG explanation at the time was that possibly somehow the possum board was reinforcing the vibrations of the total instrument.
harpmaker
Dulcified! (>2000 posts)
 
Posts: 4687
Joined: Mon Feb 10, 2003 12:33 am
Location: Pershing, IN

Postby rtroughear » Sat Jan 10, 2009 10:39 am

Paul:

When I say a bar needs to be free at both ends to vibrate, I’m referring to a free bar, which is the model I’m proposing. Clamping the dulcimer was an attempt to deny it the opportunity to act as a free bar.

If I can summarise what I read you as saying:

“A dulcimer acts as a bar fixed at both ends. The strings, being fixed at the bridge and nut of the instrument, cause the fretboard , and in fact the entire instrument, to vibrate at the same frequencies as the string is vibrating. This takes the form of standing wave(s) in the body of the instrument, which will last until the energy losses in material junctions cause it to die out. The bridge and nut are fixed in place relative to the strings and body, and the number of vibrational peaks between them (in the body of the instrument) is set by the resonant properties of the instrument. These standing wave vibrational peaks in the wood of the body are what produce the sound we hear, with some smaller contribution from the string itself. There may be some non-linearity in frequency as the tone decays.”

Well that’s the first attempt I’ve seen here, or anywhere, that tries to explain how a mountain dulcimer works. It’s actually not much different from what I’m proposing except you claim a fixed bar model and I propose a free bar. But I don’t buy it - although I would if you produced evidence to show it works that way.

I don’t buy it for the main reason that the ends of a dulcimer are not fixed at all. They are as mobile as any part of the instrument, and that adds support for a free bar model. You can tell this by just holding an end whilst strumming. Have a look at www.kettering.edu/~drussell/guitars/hummingbird.html for an interesting animation of a guitar experiment showing the movement of the headstock and neck acting as a bar.

Image

Also, though not relevant to one model or another, the bridge and to a lesser degree the nut, and therefore the string ends, are not really fixed in relation to the body of the instrument. They can move around quite a bit. Not as much in a dulcimer as in a guitar, but still mobile relative to the body (we are talking microns here).
See http://www.speech.kth.se/music/5_lectur ... tuned.html for a brief discussion of this.

Considerations of sound efficiency and frequency non-linearities are probably not very relevant to a basic sound production model. I’d see those as operational factors in designing actual instruments.






Dave: I’m setting out to prove, or disprove, the proposition that a mountain dulcimer vibrates in the way a free bar would (although in a more complex and less predictable way), like the pictures on the right hand side of the diagram below. (Not the fixed bars). So far I’ve found some evidence to support the idea, but nothing that disproves it.

Image


One test to investigate the box/bar proposition would be to absolutely fix the ends of a real dulcimer so there is no possibility of them moving. That’s what I’ve tried unsuccessfully to do so far. I could bolt the instrument to the garage floor, but then I couldn’t get it on the bench for measurement under standard conditions. I’ll probably use a concrete house piling next.

For normal acoustic measures of the sound field around a dulcimer, I don’t want to constrain the ends, and will suspend it vertically from the ceiling using elastic straps to stop it moving around, and a wire frame surrounding it to mount microphones. I'll also have to devise a reliable string striker that can reproduceably strike three strings. All of this has to wait until after April when I have more time.

Richard T
rtroughear
Super Mbr (501-2000 posts)
 
Posts: 603
Joined: Tue Jun 05, 2007 8:18 am
Location: Brogo near Bega, NSW, Australia

Postby PaulC » Sat Jan 10, 2009 11:27 am

Well that’s the first attempt I’ve seen here, or anywhere, that tries to explain how a mountain dulcimer works. It’s actually not much different from what I’m proposing except you claim a fixed bar model and I propose a free bar. But I don’t buy it - although I would if you produced evidence to show it works that way.


Our argument is largely semantic, due to the real world constraints of the instrument (ie, whether it is a free bar or fixed bar). A dulcimer is composed of asymmetrical components, especially in regard to the energy imparted to it to produce a tone, and the energy bled from it to dissipate that imparted energy.

if you were able to suspend the instrument unfettered in free space and strike the string(s) forcefully, you would see the entire instrument vibrate at a frequency that is the net sum of the various resonant components, plus the wave mechanics associated with the various reflections off the multitude of non-transparent acoustic surfaces that exist. Its all of this that create the aural experience and decay in sound production over time.

Actually, I believe pursuing the vibrational properties of the dulcimer alone, is something of a red herring. The aural properties, which ultimately is what we concern ourselves with, in my view, are a greater product of the reflective properties of the components of the dulcimer and only secondarily, the vibrational properties. I think that is the significant conclusion from your acoustically topless dulcimer.

I say this because the observable difference (or lack of it) has to do with vibrational amplitude, not frequency per se. Frequency is important in terms of the "match-up" with materials, but the observations are concerned primarily with amplitude, not frequency.

The frequency character of the sound is tempered by the varied reflective (and vibrational) properties of structure and materials, to be sure, but since the structure and materials are held to a fairly limited range, that variability is less significant in the overall sound production, i.e. nominally equal.

A free-ended bar is a theoretical construct, btw. It doesn't actually exist. The end of the bar, because of the boundary between the material and free space, actually reflects some energy, imposing additional wave mechanics on the established standing wave. Where the wavelengths are small relative to the boundary, the effect is well known and demonstrable (cf RF energy in antennas, as but one example). I'm not in a position to assess whether the reflection is negligible at audio frequencies, but it does exist.

Ultimately, in your quest to explain a dulcimer I believe what you will come down to is an investigation of the amplitudinal interaction of vibrational, reflected and transmissible frequencies, with the latter two components having greater contribution than you may realize.
PaulC
Senior Mbr (101-500 posts)
 
Posts: 189
Joined: Thu Nov 06, 2008 9:43 pm
Location: Southern Ontario

Postby rtroughear » Mon Jan 12, 2009 10:13 am

G'day all

Paul: This is all just unbridled speculation on our part – the sort of thing luthiers love to engage in; and the sort of thing I wanted to avoid because it doesn’t provide practical information that helps builders decide which piece of wood to use, and how to shape and place it.

Guitar, violin, mandolin, ukulele and other instrument makers have access to a basic understanding of how those instruments work, as a sort of ground zero starting point. Mountain dulcimer makers don’t – that’s what I hope might come out of experiment. It won’t take any of the future fun out of speculating about the mysteries of the instrument because, as with guitars and violins, once we get past the few defining lower resonances we’ll be in uncharted and uncontrollable territory. Perfect for discussion.

So I’m trying to apply Occam’s Razor (or the KISS principle, if you like). Approaching the matter from the standpoint of wave dynamics is too inaccessible and esoteric for most makers, and I’m not convinced it’s a workable approach anyway. I’ll stick with the familiar and measurable variables of mass, stiffness, frequency etc. in my experiments and explanations. Secondary effects such as the contribution of internal sound reflections and the materials' acoustic transmissibility I’ll have to leave to future generations. Those effects are unlikely to be part of a basic model (I speculate).

I would'nt say the difference between a free and fixed bar model is only semantic - real pieces of wood vibrate quite differently depending on whether the ends are allowed to move, or not, and that affects the sound in both amplitude and frequency.

The measurements I’ve made so far have all been acoustic; vibration measures will come in due course. I presented only the sound level (amplitude) results because I didn’t think this forum was a sensible place to present dozens of graphs of frequency spectra. The required explanatory accompaniment alone would bore everyone to death. I don’t yet know how I’ll make available the details of what I’ll do.


But Folks: None of this is proved one way or the other - the mountain dulcimer as a bar, fixed or free, or as a collection of independent vibrating surfaces, or any combination in between. One day some light might fall on the matter, but don't stop buildng in the meantime, or put off starting. I don't think you can make a bad mountain dulcimer.

Now I should get back to the twenty two half-finished ukuleles in the shed.

Richard T
rtroughear
Super Mbr (501-2000 posts)
 
Posts: 603
Joined: Tue Jun 05, 2007 8:18 am
Location: Brogo near Bega, NSW, Australia

Postby Andreas » Sun Feb 01, 2009 10:05 am

Hi Richard,

Is the following idea perhaps useful: glue a reflective top (foil) to the surface of interest. Shine a laser at various spots, on the foil, producing a number of reflected spots, and film these reflections. The amplitude of motion of these spots indicates whether the corresponding spot on the dulcimer is a node, antinode, or something in between. A long exposure time would effectively integrate the motion, a short (casio now makes a very fast digital camera, I believe) exposure would reveal the details.

Another simplification could be made by making the dulcimer a rectangular box, or even a solid beam. I believe vibration models exist for such geometries.

The contactless nature of the measurement may be useful.

Andreas
User avatar
Andreas
Super Mbr (501-2000 posts)
 
Posts: 1026
Joined: Thu Oct 02, 2008 7:59 am
Location: Sweden/California/Germany

Postby rtroughear » Thu Feb 05, 2009 3:21 am

Andreas:
It would be good to discover and catalog characteristic vibration patterns of the mountain dulcimer, and maybe the method you suggest could work. But first consider this (which is relevant to any method that measures the vibration at a point on the dulcimer surface):

1. The vibration pattern will be complex over the whole of the instrument body, so readings at a matrix of measuring points would need to be taken on the top, back, sides and end blocks. eg.

Image

2. The vibration at each point will be frequency dependent; so multiple sets of readings would be needed over a range of excitation frequencies. In practice readings are usually taken at identifiable resonant frequencies at which the largest vibrations occur, and which shape the sound timbre. There might be between five and ten of these below about 2000Hz.

3. The vibration patterns measured on a dulcimer may be unique to that instrument ie. not representative of dulcimers in general. So a number of instruments, of the same basic design, really should be measured to confirm that the vibration patterns are characteristic of that design, and not just one instrument. For every different design there are likely to be a different set of vibration patterns (or modes).

We’re starting to get to a serious investment in measuring time here, and unfortunately I think the likelihood of a generalised outcome is low.

In the guitar and violin world some progress has been made in the science/art of “free plate tuning”, which is essentially a hard-measurement extension of tap tuning of top and back plates for those luthiers who don’t have the 20 years left to learn that method, or who like to see pretty pictures of plates vibrating. An isolated top or bottom is excited into vibration by a nearby loudspeaker; sawdust or something similar is sprinkled onto the plate; and the loudspeaker frequency is varied until the sawdust jumps into patterns of nodes and antinodes. Several vibration resonant modes might be examined for frequency and the shapes of the patterns. The plate thickness or bracing can then be varied and the measurements repeated until the desired patterns and frequencies are achieved. THEN the plates are assembled into the instrument with the hope that the changes in vibration modes, caused by gluing the free top and back plates to the sides, will result in a more favourable instrument. It seems as much an art as a science to me. However, for violins and several types of guitar, there are now known general patterns that at least serve as a starting point for free plate tuning, and end results can be good.

There’s nothing like that for mountain dulcimers, so I’ve measured the free plate resonant modes of the tops and bottoms of the last twenty or so instruments I’ve made in the hope that some common patterns might emerge. No such luck. What I’ve found is:

1. The free plate vibration modes of an unbraced top and back, without fretboard mounted, have some similarity to violin back plate vibration modes and are fairly repeatable across different plates and for both tops and backs ie. provided the plate is the same shape, the vibration patterns are basically the same, but the mode frequencies depend on the plate density and stiffness. Here's a top and a back from two different dulcimers - no bracing or fretboard, just the thin plates:
D36_TopModePlainPlate__06.jpg

D37_BackModePlainPlate_06.jpg


2. When braces and fretboard are added the vibration modes change completely in patterns and frequency. The patterns are heavily influenced by the position of the braces (keep in mind this is in the free plate, not the assembled instrument).

3. The vibration mode patterns and frequencies are different for each bracing pattern. Shaping the top braces may lower the resonant modes by only about ½ semitone, because the top bracing is completely overshadowed by the stiffness of the fretboard. Adding back braces may eliminate some of the lower vibration modes and may raise the resonant frequencies by roughly one octave over the unbraced back, and shaping the back braces has more of an effect than on the top, but still small.

Overall, I haven’t discovered any generalised vibration patterns for the free top and back plates, or data to link the free plate vibration modes with the completed instrument resonances, or any correlation between different bracing patterns or vibration patterns and subjective auditory results.

I think this means that in the mountain dulcimer world the variety of instrument shapes, building methods, and design details probably preclude the assembly of a set of standardised top and back vibration patterns.

I don’t do free plate measurements any more because it hasn’t helped me produce what I think are better sounding dulcimers, and I nearly go deaf every time I do it. It might be a helpful method if I froze the design and spent the next decade making only that style, and systematically studied each one as I made it, and then applied some standardised listening test or hard measurements correlated with known listening preferences. But then it might not.

The same vibration method can be applied to the completed instrument – at least to the top and back because they're generally flat. I’ve only done a little of that because changes can’t really be made after the instrument is finished, and also I don’t know a method of correlating the vibration patterns with the sound quality. In its place I measure the acoustic resonant properties of the box and the enclosed air by tapping with a small rubber hammer (a pencil eraser on a stick), and sweeping sound inside with a small loud speaker and analysing the frequency spectra of the resultant sounds. These end up more generalised over all designs and construction methods, and between makers – but still no identifiable correlation between resonant peaks in the sound spectra and quality of sound of the instrument. But I continue to do these measurements on each instrument because it’s easy to do (just a microphone, a lap top and a knuckle is all that’s needed). One day some general insight might dawn on me if I do enough dulcimers.

That’s about it for now.

Richard T
Last edited by rtroughear on Fri Mar 25, 2011 6:12 am, edited 1 time in total.
rtroughear
Super Mbr (501-2000 posts)
 
Posts: 603
Joined: Tue Jun 05, 2007 8:18 am
Location: Brogo near Bega, NSW, Australia

PreviousNext

Return to Making Dulcimers

Who is online

Users browsing this forum: No registered users and 4 guests