Tap tuning

[Jim Baker]

Hi folks. I'm building an octave mandolin based on a scaled up mandolin design from Graham MacDonald. I have the top and back carved and would like to refine the top before assembly. Has anyone experimented with tap tuning using an electronic tuner? When I tap the top my tuner reads a little sharp of "B". Could this possibly be valid? I'd appreciate any comments. Thanks

[Martian]

I will be intrerested to see your responses, I did buy a peterson tuner a few years back, but I really do not know if it is valid or not, I would tap and then if I had bars that read B+30 cents, I would just bring them down to B -0 , but I really can't say if it was not a waste of time. as the target note seems real arguable. I started de-flection tuning instead , but among a couple of my builder friends, it was also arguable. maybe it is meant to be argued. lol good luck. Should also be noted , I am an extreme newbie, as only have built 3 mandos, and getting ready to build 4th

[j. condino]

arguing about snake oil....

[peter.coombe]

Tuning to specific notes is a waste of time. Try and correlate specific notes to the final sound and you will get a random pattern (I have tried). So snake oil is correct. However, the relationship between the top and back is important, but what relationship you need to aim for in an octave mandolin is anyone's guess. Just build it. No harm in taking measurements while you build it, but you will need to build a lot of instruments and take a lot of measurements before you can make any sense of the measurements.

[Graham McDonald]

If you have it carved to around 1/4" thick in the centre and 40% thinner in the recurve, using 1/4" wide X braces 3/8-1/2" high at the crossing tapering to 1/8" at the ends you should be in general area of what is required, but there are lots of variables in arching, graduations and brace carving which will confuse things. I am listening for clarity rather than a particular pitch.

I put a new carved soundboard on one of my bouzoukis last year and it tapped at G#3, but not very clearly. The new soundboard sounded fine when strung up, but it was a slightly larger and deeper body that I made in 1990 than I would usually use. The tapping was done holding the soundboard at around the 10 o'clock position neck end down and tapping around the 4 o'clock spot. It can be tricky finding that node (or is it an anti-node) to hold and to tap. A different, higher note holding it where the tailblock is glued.

You can always string it up without any finish and do some thinning probably towards the edges under string tension.

Cheers

[Tavy]

What others have said about specific notes... you can learn stuff IMO by recording the taps into audacity and plotting the spectrum out - but that's more a case of "do these resonance work together well", and "are there modes that are currently inhibited" and so on.

[SBJ]

What are the rest of the dimensions of your OM (scale, depth, width, bracing, etc.)? I'm planning to build an OM after a couple more mandolins.

[Ivan Kelsall]

I have to agree with Peter Coombe - Tap tuning or Flex.testing are 2 methods of 'trying' to establish some form of control over the ' final tone ' of the finished instrument. However,trying to corelate any measurements to the final outcome,i feel would be pretty difficult,& IMHO,would only apply to a 'specific' instrument - the next one along would be a whole new ball game, as the woods etc. add their own 'variations' to the equation.

Jim Baker mentions that his top produces a 'note' a tad sharp of 'B' (which B ?). But is this the 'optimum' for his build ?. The only way to find out is to build it & see (hear). This is where lots of experience comes into play. Being able to judge the woods & having a good idea what you're aiming for to begin with. I have to admit,that if i were to build any style of mandolin,no matter how crude - i'd have a million questions,as i haven't clue re.how to 'test' a top / back.

It's very much different in engineering,using materials which are for the most part almost identical in their physical / structural properties,where you can predict with ''some degree of certainty'',the properties of a finished item made from them. However,even those materials have to be tested to establish if they are within their specification before they can be used - just in case !!,
Ivan

[pelone]

Is "tap tuning" akin to a voodoo, shamanistic approach to trying to reach the best outcome. When one might consider the whole being a sum of the parts---perhaps trying to take each individual component and tap them to the same note would then be the next logical step. However, taping binding, bracing, tops, bottoms, sides, neck, fret board, bridge, nut and tuners to a B, for example, is not likely to contribute to a productive exertion of effort. I would suspect, based upon the reporting of Dr. Cohen, that a top that is tapped to B is not likely to be a B when it is joined to the sides of the instrument.

[peter.coombe]

Not quite voodoo, but close. How most people understand tap tuning if they are reading Siminoff's chapter on tap tuning, then yep voodoo. A top that is tapped to B is not going to be a B when the sides are glued. It will change again when the back is glued on, also when the neck is glued, the tailpiece installed, the tuners installed, the bridge installed etc tc. Absolutely everything contributes to the modes of vibration of the completed mandolin (has been pointed out by Cohen and Rossing in their paper), and you can't just isolate one component and tune that to a frequency and not expect other modes to be affected, because they are. So the stiffness and mass of the top that can be measured by tapping when a free plate is only one thing that contributes. All the other things will be different in the next mandolin because wood is so variable. So tuning the top (or anything else) to a particular frequency gets lost in all the noise introduced by all the other parts. Not only that, but a top (or back) does not just have one modal frequency you can tap to, there are many and it is incredibly difficult to get two tops with all the same modal frequencies because wood is so variable. The only way that tuning to a particular frequency could be useful is if everything else is absolutely the same, and that never happens from one instrument to another, far from it. However, relationships between frequencies are useful because they are not affected by the noise so much, but even that has limitations, and you don't know what relationships to aim for until you have made and measured a lot of mandolins.

[pelone]

What he said.

[Ivan Kelsall]

I think that Peter Coombe in post #10 has explained what we all most likely know & understand,that you can't isolate individual components of a mandolin & try to 'tune' them,when the sound of the finished instrument is dependent on the way that the all the components interreact,& they certainly won't maintain their 'tuning' after assembly.

I know that in other threads re.'tap tuning',that John Hamlett has said the he uses the 'flex test' method to try to gauge when his tops / backs(?) are at their optimum thicknesses. That makes far more sense to me as ultimately,it's the 'physical' attibutes of the carved tops / backs that determine the outcome,
Ivan

[sunburst]

...I know that in other threads re.'tap tuning',that John Hamlett has said the he uses the 'flex test' method to try to gauge when his tops / backs(?) are at their optimum thicknesses...

No sophisticated measurements of stiffness, just judgement and experience. I weigh tops and backs and try to get a good relationship of weight and stiffness to gauge how to carve the top and back for good coupling. Thickness therefor, as long as it is within a range that I consider normal, takes care of itself and varies from one piece of wood to another. I also vary arch height according to density and stiffness of wood.

I abandoned all "tap tuning" attempts many years ago after decreasing amounts of information recording. Basically, I tapped, adjusted, recorded the results, meticulously at first and gradually less until I quit. I never found any correlation to anything in the final sound of the mandolin, so I gave up trying. I concluded that tap tuning is of no use to me.

[Jim Baker]

What are the rest of the dimensions of your OM (scale, depth, width, bracing, etc.)? I'm planning to build an OM after a couple more mandolins.

Scale length 22.5", body width 14", body length 16", body depth 2.5". Top is carved sitka, back is carved maple. I'll follow up when I string it up. X bracing top and back.

[Jim Baker]

If you have it carved to around 1/4" thick in the centre and 40% thinner in the recurve, using 1/4" wide X braces 3/8-1/2" high at the crossing tapering to 1/8" at the ends you should be in general area of what is required, but there are lots of variables in arching, graduations and brace carving which will confuse things. I am listening for clarity rather than a particular pitch.

I put a new carved soundboard on one of my bouzoukis last year and it tapped at G#3, but not very clearly. The new soundboard sounded fine when strung up, but it was a slightly larger and deeper body that I made in 1990 than I would usually use. The tapping was done holding the soundboard at around the 10 o'clock position neck end down and tapping around the 4 o'clock spot. It can be tricky finding that node (or is it an anti-node) to hold and to tap. A different, higher note holding it where the tailblock is glued.

You can always string it up without any finish and do some thinning probably towards the edges under string tension.

Cheers

Thanks Graham. I'm following your mandolin project book and scaling based on other octave mandolin dimensions. I guess I'll find out in time if it will work. My mandola turned out okay and I followed the same principles.

[Ivan Kelsall]

John Hamlett's post sums it up perfectly for me - ''There's no substitute for experience !'',
Ivan

[Chris Pritchett]

I've now built 11 instruments, 7 mandolins and 2 OMs so still consider myself a "newbie" still. Like others, I have tried the 'Tap-tuning' route albeit a via a more 'empirical' method. I'm reserving judgement, so far, as to whether aiming for 'target notes' for each tone bar, and the top assembly generally is either practical or critical. My method of carving (contouring mechanically before hand finishing) means I am getting consistent profiles and weights for the plates. However, I have had quite a variation in tunings from almost identical plates. The main variable has been the tonewood used, building in limited numbers has meant that my timber sources have not been consistent. A reasonable conclusion is that differences in density and stiffness are the probable cause of this variation.
So, should I abandon consistency of weights and profiles and strive for some 'theoretical' target tunings? At the moment I think not.
it seems much more important to find the potential in each piece of wood and in the relationship of each part to the whole. By which I mean striving to get each part of the top and back to resonate and to continue to revise the final profiling with the instrument strung up in the white. I seem to be achieving consistency in the general quality of my instruments albeit with some tonal variations which I'm sure is down to the individual character of the wood.

[graham darden]

"I also vary arch height according to density and stiffness of wood."

John,
How is the arch height related to wood stiffness and density. Does stiff and dense call for more or less arch height than flexible and less dense. And vice versa. Do your #s stay in the same range for the different arch heights?
Graham

[sunburst]

If we have two pieces of wood of the same density and stiffness and we carve each to the same thickness, one flat and one arched, the arched one will be stiffer as a unit. Within reason, we can stiffen or "loosen" a piece of wood by varying the amount of arch. We obviously don't stiffen the wood itself by increasing the arch, but we can slightly stiffen the unit that is the top or back.
I vary arch height only by about 2mm. Stiff top wood usually gets about a 14mm center while less stiff top wood gets around 16mm in the center.

[peter.coombe]

There is a lot of educated guessing and trial and error involved in music instrument building. Some people call the educated guessing intuition. It is empirical in that you don't need to understand how the changes you make actually work, just that they work or don't work. This does work in that with experience you can improve the sound of the instruments you make, but it is very slow and you need to make a lot of instruments, so as already said, there is nothing like experience. That is certainly true, but there is another approach that some guitar makers have taken. There has been a lot more acoustic research done on guitars than mandolins, and with the publication of Australian luthiers Trevor Gore and Gerard Gilet's books things have certainly changed in the guitar world. I would describe the books as a practical implementation of the guitar research using sound engineering principles. There is a fair bit of maths involved in he books, and that puts off some people, but their methods do work. Trevor has a PhD in engineering so knows his stuff. Briefly summarising, it involves measuring the modes of vibration of an assembled instrument and they have devised methods to adjust the frequencies of the modes of vibration which doo affect the sound. They have also come up with a method of measuring a top blank and calculating the optimum thickness that will result in a 2deg rotation of the bridge which is about optimal for good sound. The basic idea is to adjust the frequencies so that they sit in between notes on the scale, the relative frequency between the main top mode and the main back mode is also important for good sound. There was some degree of trial and error in devising what was optimal, but once you have the measurements it can be reliably repeated. Unlike free plates, in an assembled instrument the absolute frequencies are important. If one of the main modes of vibration sits bang on a note, then you get a wolf note and peaks and troughs in the sound. It can also causes intonation problems. If you want to get into all this measurement stuff, read the books. Australian luthier Graham Caldersmith has also written a paper that summarises much of the guitar research.

Why guitars, aren't we mandolin makers? Well Cohen and Rossing have shown that mandolins vibrate similar to guitars, but the frequencies are higher. Therefore the guitar stuff is relevant, or at the very least a reasonable assumption is that some or most of the guitar research can be applied to mandolins. So why not apply the Gore/Gilet methods to mandolins? That is exactly what I have been doing recently with my pancake mandolins, and the results have been encouraging to say the least. Big improvement in sound, and the improvements are consistent. I have now made 2 mandolins and a mandola and they all show the same improvements. The sound of the flattop mandola is astonishingly good for a flattop instrument. Blindfold you would not know it was a flattop. There is a problem in implementing their methods in mandolins, however, in that there is no access to the interior so you can make the adjustments. So the frequencies are a bit hit and miss. This is where experience comes in, and I have been able to nail the frequencies so they are in between notes. The biggest change has been the relationship between the top and back, and I have been able to nail that as well by measuring the free plates. There does seem to be a correlation between the relative frequencies I measure with free plates and the relative frequencies in the finished instrument. So I can predicably get a consistent relationship between the top and the back in an assembled mandolin. Read the guitar stuff and also of course the Cohen and Rossing research papers, and then you will get a much better understanding on how these complicated little things actually work. This stuff is not snake oil, it is based on science.

[fscotte]

Most important is the arch, and everything about how its shaped.

All else is secondary and determined by how the arch is carved.

[CarlM]

Particularly for John Hamlett but others please chime in, how do you tell stiffness to determine how to make the arch before the board is carved? Do you have a test rig? I see how you can weigh the board and check density. Or do you rough the shape and check stiffness before establishing the final arch?

[Marty Jacobson]

I believe (emphasis on believe) that if you have a top plate with the same mass and same FFT plot that it will function identically, or at least very similarly, to another top with the same mass and FFT plot. To get the FFT plot, you hold the plate so you get a nice ringing sound, and then tap around in various places while recording the response. Or at least that's how I do it. Then in Audacity or some other program, plot the frequencies and their amplitudes using an FFT or other spectrum analysis tool. This tells you two things: where the peaks are, and what their relative amplitudes are.

From this, I was able to make a carbon fiber mandolin top plate which behaved very similarly to a wood plate without too much effort. Matching up wood plates is easier, especially if using woods of similar densities and other properties.

Arching and all that will impact stiffness, and so in return that will change the FFT plot. So unless you changed something dramatically, like height of the arch or overall graduation scheme, you should be able to get another good instrument from information gained in a previous build. The first build (or dozen) is the trickiest, since you're still figuring out what it all means.

Flexing, candling, measuring, and all that is very helpful, and necessary. The FFT plot and mass are just two additional metrics that help you know when you're done.

It's really encouraging to have a plate that you think will sound amazing from your qualitative sense of it, from all the flexing and such. And then to get the numbers which match up with a killer instrument you made previously, you gain a good sense of confidence that it will be as good as you feel like it will be.

[sunburst]

CarlM, there are a few ways that I determine arch height.
-I have a lot of red spruce from a similar source and I've been using it for nearly 30 years. I pretty much know what to expect from that wood.
-When I use softer, lighter spruces, like euro' or engelmann, I use a higher arch.
-for unfamiliar wood, I start with a high arch and as I carve and begin to get a feel for stiffness, I can lower the arch if I think it needs it. For example, I'm building an F right now using a very old one piece top. The wood was given to me by a friend at Galax last year. I don't know where the wood came from originally, I assume it is sitka because of the with of the vertical-grained piece, but it is wood that I know very little about, other than it feels light and lively. I started with a fairly high arch and ended up keeping the arch fairly high because the wood was feeling light. The weight of the completed top confirmed that it is wood of low density and good stiffness. I could have lowered the arch a little more, perhaps, but I think it will be fine. I'm also building an A using red spruce with compression grain. It is probably the hardest, densest, heaviest top I've ever carved. I'm using a low (by my standards) arch and the thinnest grads I've ever used for that one, yet it weighs more than any red spruce "A" top I've carved. It will be interesting to see how it sounds...

[peter.coombe]

FFT is just another method of measuring free plate modes of vibration. It gives you more information than Chladni patterns, which is what I do (and I also measure mass), but it is essentially the same thing as tap tuning or measuring Chladni pattern frequencies. I started measuring free plates long before cheap FFT software was available.

I agree that if you match top FFT and mass in tops they will behave the same, but only if everything else in the instrument is the same. The matched tops will not function identically if the backs are different because there is coupling between the top and the back and the enclosed air, and remember, the instrument vibrates as a whole, you can't separate out individual components once it is assembled. So for it to function identically in the assembled instrument, the mass of the head and tail blocks, the mass and stiffness of the sides, the mass of the tailpiece, the mass of the tuners, mass of the bridge, FFT and mass of the back, mass and stiffness of the neck, air volume, soundhole area etc, all need to be identical. Quality factors also may need to be identical. That never happens. I have measured before and after the tailpiece was installed and the main air mode, main top mode, and main back mode all go up in frequency. That surprised me. Similarly with the tuners. That also surprised me. Installation of the bridge and strings cause the modal frequencies to go down, as expected. The top is only one part of a vibrating system, and the rest of the system influences how the top vibrates. So, you might need to re-examine your belief. It ain't that simple.