Official Luthiers Forum!

Hi Everyone,
I need a little advice regarding tap tones. I’m building my first archtop guitar with a carved top and back. I’m using Benedetto’s book on building archtops for help. Currently, I’m working on the top and have got to the stage where I am ¼” at the bridge, tapering to 1/8” at the recurve. I’ve cut the “f” holes but I haven’t yet started the bracing. Tapping the top at the nodes, I’m hearing a deep ringing tone around 94Hz.
But I am wondering if this is too deep and I may have gone too far. I just made an Audacity recording and did an analysis of the taps, but I’m having a hard time figuring out what I’m looking at. I’m getting peaks at 87Hz, 139Hz, 180Hz, and 225Hz, as well as others in higher frequencies. The 94Hz that I’m hearing does not appear as a peak. My question is am I still on the right track? Are there target frequencies I should be aiming at, at this sta

ge or should I wait until I get the bracing on? I've attached the Audacity file if you care to look at the analysis.
BTW, the top is a very tight grained bookmatched Engleman spruce that I bought back in 1994

I'm having a problem uploading the file. Here is a screenshot

I'm sorry I cannot help you there but just to say that I'd love to see pics of your work as you progress. An archie is on my list of things to do but I've yet had any time to get to it. I bought Benedetto's book last year for this reason. Is there no discussion in the book on voicing the top?

I'm at about the same stage as you. You have your center thicker than mine, my edges are thicker. The main difference is that your f holes are cut. I like to carve the f holes in some, but at least get the central part close to where it will finish, and then blend in the recurve, and the f holes. But that's with violins; this is my first arch top as well.

I've never been able to figure out audacity. I tried on a PC years ago before I got a Mac, and it was a pain. I can't get it to work at all on my Mac. I had a Japanese flute tuning FFT program that was FANTASTIC, but it won't work on Mac either, It's so old it might not work on anything.

My belly is curly redwood. .35 sg. but it is still at 310 grams now. That seems heavy. It is about 5 mm thick in the middle, going down to maybe 3.5. The area around the f holes is still 5 plus, and the lower bout is still about 4.5 -5 mm, so they both need to go down 1- 5.5 mm or so. I roughed it out a while ago, and started thinning it out yesterday

I tried something last night. I checked the flex at the bridge using a pencil. Holding it at both ends I pushed down. I couldn't see much movement at the bridge, but the area above especially going toward the cutaway moved. Not a whole lot. I'd need to set up an indicator. But the pencil bent a lot, so it could work as a way to test stiffness. If it moves without bending as much it might be good.

My tap tones are: 160 if you balance it on your finger and tap the outside, and 356 when you hold the outside, and tap around the middle. You can hear other things, a dual tone of 240/356 that sounds nice, but maybe my ears do that? Those are the main ones. It will be interesting to see what kind of drop I get when I cut the f holes in. There is a VERY STIFF band 3 grains wide on both sides going right at the inside edge of the f holes. It just misses the bass side, but goes into the treble side f hole.

My back is glued on now, but it was something like 350 Hz and is 350 grams. 2.5 mm at the recurve, and 6 mm in the middle. I just tapped the middle and it still sounds like 356. It isn't ringing, but that is the pitch of the tap. That seems very high to me.

Sorry I can't help about numbers, and the Benedetto book is rather vague on tuning. Even the thickness are not given as they would be on a Strad poster for violin making. You do know why, don't you?

_________________
Why be normal?

I do a 'tech' version of 'tap tuning' using Chladni patterns. You basically get the same resonant modes, but with more information to guide your process.

What you hear when you tap on the top is every resonant mode that can be driven from the tapping point that is not too active at the holding point. In other words, the pitches you get depend a lot on where you tap and hold. There is a whole family of resonant modes that you commonly see on a set of arch top plates (and some that are less common). The lowest pitch one I've got on a couple my 'better' 17" steel string arch tops comes in as low as 55 Hz, and I can get modes all the way up from there. The ones I commonly pay the most attention to; the so-called 'ring' modes, tend to come in at around 180 Hz. This is a prominent pitch that you'll hear if you hold the plate on the edge near the waist and tap in the center: you might want to adjust your holding point a bit to get the strongest and clearest response. In general I seem to get good results on arch tops when I match the pitch of that mode in the top and back plates before assembly. I'll note that, IMO, the shape of the modes is at least as important as the pitch. The shapes and the pitches are both a function of the distribution of mass and stiffness in the plates. IMO you get the 'best' sound when you get that distribution 'right', and the mode shapes tel you when you got that. The pitches give you some indication of the overall relationship of mass and stiffness, which does matter, but seems, to me, to be less important than the distribution, so long as you don't get things too thin.

I covered this to the extent I could at the time in a three-part series of articles in 'American Lutherie', # 28-30, back in '91-92. They should still be available in the appropriate 'Big Red Book' from the GAL. I've learned a bit since (or so I hope!), and would be happy to help out with some updates.

A friend of mine has an Andriod app out in beta testing now that he calls 'Luthier Lab'. It has a bunch of useful tools, including a signal generator and spectrum analyzer. You can set up 'projects' using various outlines (which you can customize), and it will calculate arch patterns using the 'curtate cycloid' model. You can use the camera on your tablet to get photos of mode shapes and keep them in a file along with all the other info on the project. There are other goodies in the package, and he's working on refinements even as we speak. You might try to find it on line.

I built a Chladni resonator, but I haven't had much success getting any patterns with it. Mostly what happens is that, all the tea that I am using, ends up in the middle, because of the scoop of the archtop. I do get some frequencies that really excite the tea, whereas at other frequencies the vibration is only moderate. The problem is none of the excited frequencies seem to match with the Audacity peaks. I've included a pic of the resonator,. Perhaps you have some suggestions on how to modify it to get better results.

The trick with getting the patterns is to drive the plate in an active place ('antinode') and support it at non-moving 'nodes'. You seem to have your speaker in the center, and it's not at all clear how you're supporting the plates. It's really just the same thing as tapping: again, you get different sounds depending on where you hold (which node you support) and tap (which antinodes you drive). The signal generator just allows you to drive one mode at a time by isolating the frequency.

I use a hand held speaker, generally a 4" or 5" unit. It really helps if you can use a small speaker that won't overlap the node lines by much. The node lines are like the pivot on a see-saw; the motion changes phase as you cross the line, and driving over the node is like pushing down on both ends of the see-saw at the same time.

The trade off there is with power handling. It takes 15W or so of power to drive an arch top top or back strongly enough to get good modes, and it's hard to find small speakers that can take that. Keep in mind that the rating on a speaker is usually a 'maximum intermittent' capacity: if it says '20W' that means it can take that sort of power for a few seconds. Continuous ratings are usually about half that. Also, pure sine waves, such as your signal generator should produce, have higher peak-to-peak voltage at a given power than what the speakers are rated at. I run 30W max through my old signal generator, and burn out speakers rated at 40W after a few years.

I use 4 small pads of foam glued onto wood blocks to support the plates. The wood pads just help keep the foam ones from falling over. In use, you place the pads around the edges at the places where you expect to see node lines run off the edge, and drive with the speaker in between pads. Start with the signal generator set at a lower pitch than you expect to need, and scan upward slowly, watching the plate closely for any sign of a wiggle. When you start to get a pattern move the foam blocks around to the places where the motion is least, and adjust the frequency to get the most activity. Once you've gotten the clearest possible pattern you can get a pretty exact reading on the pitch by turning the power down to where the glitter or whatever just barely moves.

Learning to find the modes is probably the hardest thing about this. One thing that really helped me was getting a piece of expanded styrene bead board and finding the modes on a square piece. Unlike the usual insulation board the expanded bead board has the same stiffness in all directions, so you get 'closed' modes on a square piece. It's cheap, easy to drive, with high losses which keeps the band width of the modes large: they show up over a wide frequency range which makes them easy to find (don't even try glass...).

I use aluminum glitter. Mark Blanchard uses dust from under his table saw. My band saw makes dust that is too fine. I don't know anybody who uses tea leaves or portulaca seeds any more.

Daniel, I realize I was no where near you, mine looked BAD. Wasn't nearly close to finish. Sorry to equate mine with yours, yours looks good. And still closer to finish than mine is now. I still have no FFT numbers. No glitter. It does weigh 248 g, but the edges are quite thick. No diagonal braces yet. I did check the tap tones with an online tone generator. It seems to be around 200. If I balance it in the middle, and tap the outside, it is about 100 or so. Those are the main notes, you can find others. One is about a fifth up; and a rumbler, maybe a third or fourth lower.

It seems like your numbers are in the ballpark.

I forgot to measure what the tap tone was before I cut the f holes to see how much it dropped. Maybe from around 300 to a little over 200? It does give in to pressure on the bridge area a little now. Still bends the pencil white a bit.

Just so I don't look like a hack. I have a habit of showing mistakes, and bad photos!

_________________
Why be normal?

Hey Alan, thanks for explaining how to get good Chladni patterns, unfortunately I moved on to gluing in the braces before I read your explanation. So I guess I'll use it on the next top I make.
Now that I have the braces on, my next step is to tune them. I've watched a couple of videos on this but I am more confused as ever. In some, the luthier emphasizes striving for one even ringing tone throughout the sound board, in others the luthier emphasizes trying to obtain a variety of ringing tones throughout the sound board. Other luthiers focus on tuning the tone bars to a particular frequency. Others try to get the top to deflect by a certain amount and focus less on the tone. The Benedetto book is really not very helpful on this subject. When I do a frequency analysis in Audacity, I get different peaks depending on whether I am tapping the tone bars or the top surface. So I'm really not sure what I am supposed to be aiming at or what method I should use.
Do you use any of these methods or combination of methods and if so, why. Thank sort your help.
Ken, it looks like you're getting there. It's good to be working along with someone on the same project.

I recommend not overthinking this and going down this rabbit hole. On my first archtop I built it to the dimensions given in Benedetto's book and did the tap tuning thing once the box was complete. It turned out great. On your second one you can maybe dig a little deeper. Beware of paralysis by analysis.

This is my first guitar, I made a half scale for the grandsons, but it was just a quick, bang, zoom, done thing. I've made about a dozen violins, and have about 4 started. Violinmakers are much like guitarist seem to be. They have a method, and that's what they do.

Many are into tap tones, and Chladni patterns. They measure EVERYTHING. They know every mode, and some even seem to know how to move them around. I use tap tones on free plates just to get an idea on the stiffness of what I have. Archings and thicknesses are only part of the stiffness profile, the other part is the wood; the way it is cut, and the way it grew. I use tap tones and weight to get a RELATIVE stiffness that should compare things that are basically alike, like backs and bellies of guitars, or violins.

I don't have any info on guitar measurements. My belly now, without tone bars is 200 Hz squared times 248 grams divided by one million. 9.9 I don't know if it is stiff or weak, but it is my first number. I do my main tuning with the plate held on the edge, just like it will be on the instrument. In fact, most of it I do when the box is all glued up. I do like to get the center area, especially on the belly with the bass bar, or on a guitar the tone bars, somewhat even in tune. I like to glue or clamp it to a board with a hole cut out for access to the inside. The guitar making practice of gluing the belly on first would seem to work great for this, but the high sides might make carving the braces difficult.

I would think that a belly would have a bunch of tones all over the place. How you get them, and how you find them, I don't know! I do area tuning something like what one violin maker does. It seems to have gotten its start in Harpsichord making, as the only two people I've seen that do it were harpsichord makers first. It is basically making the belly, and back, homogenous in tone over a smaller are than the entire plate.

The method seems to be off his site now, but the hints are there. It is almost spiritual in nature, so it might seem pretty weird. I don't carve away at individual growth lines. I only use it in a general sense. I guess you would get it, or you won't.

http://keithhillharpsichords.com/on-musical-instrument-making#hints

Contrary to their method I don't try for certain tones, and certain musical steps. I just go for even in an area. I think that the arching gives the different tones. It does seem to free up the tones. It does find thick spots without using a thickness caliper.

I think deflection might give an idea as to what kind of tension you would need to drive it. But you'd have to know what kind of movement, and what kind of weight gives it, to come up with anything usable. I saw somewhere that someone used a set amount of deflection with a 37 pound brick! Unless you've kept track of this, and keep careful records, it is just kind of good to see that the belly at least flexes across the entire area.

You're using a cross brace? I'm going with parallel bars. I have a formula for getting the shape of them on violins. I'm going to try the same thing on the guitar and see how it works. A variation on a violin maker from Chicago. Your max height is added to the thickness of the belly. Let's say 16 mm + 5 mm. So 21 is the middle. 1/4 of the bar away we divide by 1.11 to get 18.9 -5 (if it's still 5 there) gives 13.9. The next step and 1/2 is 1.11 squared, so we end at 10.38. Three quarters over is 1.11 cubed for 6.22, and the end is 1.11 to the fourth, so it is 2.38. Linear regression? I don't know, but It seems to work.

Tuning a flattop will be a whole other can of worms. I haven't done that yet. Maybe this will help some. Maybe it just confuses. I'm not normal. I admit to that. As Barry just said, glue it up, and get something done. Tuning from the outside works great. You can even string it up and see how the strings that you want to use sound on it. I don't think that extensive tuning of free plates will accomplish a whole lot.

Others may disagree.

_________________
Why be normal?

Ken, your violin building experience will definitely come in handy for construction aspects of archtop guitars, but the acoustic science may not be as transferrable. My research indicates that the two families of instruments work very differently. But who knows, I'm interested in seeing what you come up with.

I start with getting 'good' mode shapes on the unbraced top before cutting the holes. When the holes are cut the mode shapes and pitches pretty well go out the window Then I glue the braces on, and tune those. Usually by the time I'm done the mode shapes and pitches are back to what they were before, more or less. Just because you've got the braces on doesn't mean it's too late, although, since you don't know what the modes looked like before you're not in as good a place as you might like to be.

I've used a number of different arching and graduation schemes on guitars over the years. Overall it seems to me that getting the 'right' balance between things like top thickness and arch height, and between braces and top graduations, is the recipe for a good instrument. "Good" for a twenties style acoustic archtop, a big-band box from the late thirties, and a 50s-60s mainstream instrument may be different things, and call for different balances.

What confuses me using the tap tone method is that the frequencies and the amount of sustain vary considerably depending on where I hold the top plate and where I knock. So is there a specific point to hold and a specific place to knock or is it just where there is greatest resonance? I get greatest resonance when I pinch the upper bout at the widest point, around the point the tone bar reaches the edge and when I knock at the bottom where the tailpiece would make contact. Here I get a main frequency of 104Hz. If I knock on the tone bars, I get less sustain and a higher pitch.
If I hold it at the cross of the tone bars and knock from above, I get some nice ringing tones all around the edges but at the bridge the is sounds flat. I guess that is what you would expect. Here is the Audacity spectrum as it looks now.

The 'correct' place to hold and tap depends on which mode you're trying to get. Basically, to get the strongest response you need to hold at an non-moving 'node' line an tap at the 'antinode'; the active place between node lines. Since it's hard to know exactly where the nodes and antinodes are unless you look at the Chladni patterns you have to start by making your best guess about where you think they'll be, and adjust things from there.

Keep in mind that tapping, in theory, puts in energy about equally at all frequencies up to a certain point. You'll be activating all the resonances that are active at the tap point and more or less inactive where you're holding it. That can be a pretty big spread.

The holding and tapping points you describe would tend to favor the #2 mode, sometimes called the 'X' mode, and the pitch your'e getting could be about right for that, depending on how you're doing things. If you were tuning a violin you'd want to match that pitch in the top and back, but on an archtop guitar it's not so important to match that one.

If you hold at the waist, and tap in the center of the plate, you should get a reasonably clear tone. I often get that one at about 180 Hz. Its a 'ring' shaped mode, often referred to as the 'O' mode, and is actually the counterpart of the 'X'. On a flat plate you'd see those two at very nearly the same pitch (the 'O' would be higher by a little bit), but the arch moves the 'O' mode frequency up. On violins the two are often about an octave apart, but not so on arched plate guitars. I talk about this in part 2 of my articles in 'American Lutherie'. At any rate, on arched guitars this seems to be the most important mode, and I find that tuning the top and back to give the same pitch for that mode, with a ring that is as well 'closed' as you can get it, tends to produce a good instrument. Again, IMO the mode shapes are more important than the exact pitches, but it does seem to help to match the pitches of at least some modes between the top and back.

Again, IMO, the arch geometry probably has more to do with how well the top resists sinking than the actual thickness. It's one reason I don't thin the recurve after assembly. I've observed that wide recurves that dip too low tend to sink fast. I want to be sure to maintain a positive slope up from the low point of the arch (which I try to locate just inside the edge of the liners). 'Curtate cycloid' cross arches have been very stable for me. In a sense, arch top guitars do with arch shape and graduations what flat tops do with bracing. It's a system, and you have to learn to work with everything together.

OK, so that makes a lot of sense now. Thanks for explaining that.
But now I realized that I have another problem. I may have to remove the braces anyway. I noticed that my top plate has warped. The top and bottom bouts have turned up about 1/8”. I can’t tell if it warped before I glued in the tone bars in or if it happened after. If I clamp the top plate down to the sides, there aren’t any gaps. but there is some tension put into the top which will probably raise the pitch . I may need to plane the braces off. It might be a good exercise to do as I can keep tapping and listening to the pitch drop and figure out when I have gone too far. Any recommendations?

I little warpage is not a big deal. If you can clamp the plate flat to the rims without undue pressure then glue it on and proceed. Remember what I said about paralysis by analysis.

I don't know If I'd remove the braces for that. Many of the violins I've carved developed potatochipitis. They bend up at the sides. I have been roughing the inside first; close to size, then rough the outside, and let it set a while. That seems to help it. Most of the stock is off, and I can then flatten it, and it doesn't get too bad. As long as you can glue it flat with moderate pressure, it should be fine.

An 1/8" of rise on the sides shouldn't add that much tension.

If you DO want to take them out, (I wouldn't) I have a suggestion:

Many makers clamp the belly to a flat plate for fitting, and gluing on the bass bar. I noticed more on the guitar than on violins and violas, that when gluing the diagonal braces yesterday the belly will flex while fitting and gluing when it isn't held. I got it done, but I think the next time I'll make up a holding fixture. I can use it for tuning besides. I haven't trimmed them down yet to see what it has done to the sound. Untrimmed the sound is COMPLETELY different. We'll see tomorrow maybe.

_________________
Why be normal?

This morning, it was raining and the humidity was really high, so I took the top plate and set it down on a garbage bag so that the humidity could get to the outside easier than it could get to the inside of the top plate. My workshop is not climate controlled. Two hours later and my warp has almost disappeared. Problem solved. I noticed that the tap tones have lowered slightly too. Amazing how much a little moisture can change things. I think I will store the top plate inside the garbage bag until this humidity clears.

Ken, it is quite a change in pitch and ring when you first get those braces on.

I planed the bars down to what I was going for. The bass side is 12.5 mm and the treble is 11.75 mm. Maybe a little wide at 8-9 mm. The "ring" mode seems to be about 180. You can hear other tones that blend with that above and below it. If you let the tone bars rest on your hand, and tap the edges you get notes from around 90-120.

It's heavier than I expected at 260 grams.

I think I'll clean the inside up some, put the black. veneer on the f holes, and then glue it on the guitar. Then I'll see what kind of tone you get when you hum into it.

My basement has a dehumidifier, but it is still fairly cool down there, it isn't heated, and the dehumidifier doesn't do much until it warms up. Maybe in late July this year the way the weather has been! It stays warm into the fall.

_________________
Why be normal?

I'm considering gluing on the back with a paper spacer, so I can remove it later. I think I'm happy with the top and the back - both ring nicely with a lot of different notes but I am curious to hear what it is like together. Putting a thin paper seal in the glue joint is a common practice in wood turning and pattern making, allowing the glue joint to be easily broken. Is this something guitar makers use? I want to be able to tune the "f" holes, then oversize them for the binding. binding the "f" holes after it is permanently glued is a pain.

I'm gluing a thin black wood binding on mine now, and even OFF the body it is a pain. The slightest gap can buzz. I'm using some fish glue a guy who runs a violin shop in Flint gave me. Sticky, but you don't have to work in seconds. Blue painters tape as a clamp. Some kind of thin tape would be great for the terminal holes, or I could just pull out the scissors, and cut thin strips instead of tearing them? Ah, that would make so much more sense.
I thought about tuning the f holes, but I'll just see what comes out. I don't know how much influence size really makes. Shape and placement might make as much of a difference, and the general stiffness of everything, and the volume enclosed in the body.

I don't like taking tops off violins. I've done it, and they come off without paper. Paper might help, but then you have to scrape the paper off later; or maybe not. Guitars do have those bindings. Someone else would have to answer you there.

I know what you mean about wanting to hear it. But there is still bindings, and fingerboards and frets, and tuning machines to do after that. Just to hold it as a guitar would be cool.

_________________
Why be normal?