Deflection testing

[Dennis Duross]

Anybody deflection test their finished tops under anything similar to string-tension downward force?

My top deflects .04 under 40 lbs of weight, with .625" high x .3125" wide x-braces (60 degree intersection of the braces).

[Dennis Duross]

Typo. My top deflects .06" under 40 lbs, not .04".

[Barry Daniels]

I do deflection testing of flattop plates during the thicknessing process but I have not done archtops. I shoot for a deflection of .25" or more under a weight of 6.6 lbs over a span of 18". This makes for a pretty floppy plate before bracing.

[Dave Stewart]

I've been hanging back a bit hoping for other input. I deflection test my archtop plates & feel it's a valuable COMPARATIVE tool to use with other data (bonks, FFT's, Youngs modulus, weight/density etc) and against previous builds.
I test at 12 1/2 and 25 lbs on a purpose built (ie self-adjusting) bridge, and measure deflection at treble & bass "post" locations. I'm still learning how to equate the deflection results to the final product. FWIW, my last (oval hole) was in the 040"/045" range under 25# (so you're probably in the same neighbourhood), in part because I wanted more flex and I knew I was tageting a lower breakover with a resulting lower downforce of only 28# or so. I'm happy with the result.

[Dennis Duross]

Thanks, Barry and Dave for your replies.

I'm just working on my 2nd bent-top archtop. The first one had a top that was around .185 thick, and deflected very little under 40 lbs of weight. It was parallel braced, .5" high by .25" thick braces.

This time the top is around .15" thick.

Dave, your oval-hole archtop---she's a beauty.

[Michael Lewis]

I'm curious to know what specifically you are looking for with the deflection test. Should the top just start to deflect under normal string load? Or should it deflect more? Or less? Why? I am not being facetious, I haven't used deflection testing on my instruments but had thought about it many years ago in the 80s.

I can see testing a known superior sounding instrument and trying to replicate that result, but there are so many other elements in the "equation" it would be rather difficult to make an equal instrument from the information you get from a deflection test. This process is sounding like scientific work and will need some accurate and comprehensive record keeping if it is to be useful.

[Dave Stewart]

Michael, you're right....for me it's just another tool, the assumption being "it's better to have measurable information than not have it". Flexibility is definitely part of the equation (although how & where you thinned the plate to get that flexibility are as well). I think all builders do "deflection testing" as part of the ongoing assessment while building, intuitively (eg thumb pressure) or quantitatively measuring. I keep a lot of "details" of each build on a summary spreadsheet that I'm always referencing to compare the physical facts between instruments to the end result. Flattop makers can "thin the plate 'till you hear the floppiness..." & such things to target flexibility .....tougher for archtops.

[Mario Proulx]

Don MacRostie uses deflection testing as part of his mandolin builds, and wrote a good article(I think for the GAL?) some years ago.

After reading it, I made up a similar jig and tested both my existing mandolins and some plates I had just finished carving, and all fell within Don's specs(that he's arrived-at by testing vintage mandolins of note). Fluke? Maybe....

I see it as a method to check whether you've gone too far with a particular plate, so you can then make the decision to either scrap it, or brace it a bit stiffer. There's so much more happening that won't affect deflection BUT WILL affect tone and response, that my belief is that it's not really a method to use to replicate the sound/tone/response, though it appears to work for Don!

[Craig Bumgarner]

I study, build and play Selmer style guitars which have floating bridges and tailpieces like an archtop, but bent tops, not carved. The simple ladder bracing is quite different than most archtops and the light strings and low break angle at the bridge I use make for much lighter bridge loads, so I don't think my numbers would apply to your project or anyone else's for that matter, but the principals are the same.

I've been using deflection testing to get the overall top response in the ball park with known good instruments of similar build. For better or worse, I apply weight only in the bridge area and read deflection midway either side of the bridge. When I build a guitar after a particularly good example (most recently, I have been trying to capture the 1950/60s sound of Jacques Castelluccia's guitars), I take all the standard measurements, bracing details AND deflection tests (I also test the backs). I previously used 10 and 20 pound weights, but after scarring an owner of a 60 year old guitar when we got a quite visible .065" of deflection under 20 lbs, I changed my approach. I now zero my test jig out with the fully strung, tuned up guitar. I unstring it and measure the rise of the top. I then use enough weight to put the top back down to zero. So effectively, I have the amount of deflection under static string load and don't deflect the top any more than it does under normal string tension. On my own guitars, I go a little beyond that just to see what happens, but building to a known string load deflection at the bridge ALONG WITH a close physical duplication of the good example has produced strikingly similar sonic results for me.

I test braced tops clamped to the sides before gluing, again after gluing, and again after the back is installed (I glue top first so I can easily shave braces, then back). It is worth noting I don't get much variation in these three steps. When I first started deflection testing my own work, the deflections on my tops were WAY stiffer than the deflections on the good example guitar. This was somewhat expected as I had a suspicion I was overbuilding, but it was interesting how much stiffer, my deflections were less than HALF of the example guitar. It took considerable brace shaving to loosen the top up, more than I would likely have had the courage to do without the target the deflection tests gave me.

I would also mention the guitars I study and build are pretty simple in design. For instance, nearly all have simple ladder brace systems and no graduation in the tops. They also have floating bridges with a tailpiece which to me lends itself to this kind of testing. Things could get a LOT more complicated with fixed bridges, graduated tops and sophisticated bracing systems.

CB

[Steve Senseney]

Interesting comments Craig!!

[Mario Proulx]

Things could get a LOT more complicated with fixed bridges, graduated tops and sophisticated bracing systems.

They do! Yours' is a great example of using a method to help replicate a known, on a relatively simple instrument. I say relatively, because there is no such thing as a simple instrument, especially given that we build them with wood, a medium that itself can and will vary a LOT!

But with carved, arched plates, oy! The variables are impossible to fathom. We can change the height of the arch, and get a totally different tone, with the same deflection. We can change the graduation, and again get a completely different tone, with the same deflection. Add braces, tone bars, etc... Even the sound hole(s) location, shape, and size(s) will change the deflection.

What deflection testing is excellent for, is how you used it to find that you were overbuilding, and it showed you how much lighter you could go. That right there, makes it all worthwhile!

[Dave Stewart]

When I first started deflection testing my own work, the deflections on my tops were WAY stiffer than the deflections on the good example guitar. This was somewhat expected as I had a suspicion I was overbuilding, but it was interesting how much stiffer, my deflections were less than HALF of the example guitar. It took considerable brace shaving to loosen the top up, more than I would likely have had the courage to do without the target the deflection tests gave me. CB

Yeah, I wish there were more examples out there for archtops. Don & some others have done it for mandos, but for archtops there's virtually no deflection info to even ballpark it. Makes sneaking up on it slow going. (Maybe I should thin one to the deflect/collapse point, then back up .....or not)

[Mario Proulx]

(Maybe I should thin one to the deflect/collapse point, then back up .....or not)

I did just that with a mandolin way back when... It had a removable back, and I just decided it would give me all it had to give... <g> It was an eye opener, first as to what happens to tone and power, then by --how-- it collapsed.

Very educational. I later re-topped it and it's lead a good life since.

[Craig Bumgarner]

IMHO, the only reliable way to get this kind of data is to test guitars very similar to what you want to build, using your own test rig. Or at least initially. Course getting your hands on those good examples can be tricky. My current plan is to go to the Django in June festival in Northampton, MA in a couple weeks and ply the owners of cool guitars with drink and other vices while I have my way with their guitars. I hope to add a dozen or two to my knowledge base that way.

[Michael Lewis]

I generally work to known arching and graduation numbers, and generally get reliable results. As always, some turn out better than others. From doing repair work for a number of years I have learned that some really shoddy construction can hold together under the stresses of being an instrument. Also by seeing some instruments and designs by other luthiers like Steve Klein, ErvIn Somogyi, and Ken Parker which amaze me how little structure you can get away with if you know what you are doing. Add to that the experience of making some instruments too thin, and I get a reasonable perspective of what to do to make good sounding and playing instruments that will stay together for a good while. It also helps to have access to great instruments that you can measure and play. Thickness gauges are really handy for the latter.

[Dennis Duross]

My concern really just has to do with what amount of plate deflection under string load should be in the safe range? Structurally speaking.

[Michael Lewis]

From a structural perspective the less deflection the better. The trade off is structure for sound, but the art is to find the minimum structural sacrifice for the maximum tonal response. Your ingenuity and intuition should eventually gain you a design /recipe that works well for you.

[Craig Bumgarner]

......Thickness gauges are really handy for the latter.

The Hacklinger thickness gauge is crazy expensive, but there is no substitute for measuring plate thickness on assembled instruments, at least that I know of and I did some looking before buying mine. Having taken the plunge though a couple years ago, I now believe it is an essential tool for anyone looking to evaluate the construction existing instruments. Without it, you are just shooting in the dark. Many builders say they build to stiffness not thickness and while that certainly makes sense, it is hard to evaluate stiffness in a completed instrument. Judging stiffness AND its implications is not intuitive so we have to calibrate our senses on this. Thickness, deflection, tapping for resonance, driving tops with frequency generators, thumb pushing assembled instruments, etc. on lots of instruments is the only way I know.

[Dave Stewart]

......Thickness gauges are really handy for the latter.

The Hacklinger thickness gauge is crazy expensive, but there is no substitute for measuring plate thickness on assembled instruments, at least that I know of .........

Although I have no affiliation, recently, a new instrument has been in development that has some attractive features (eg automatically measuring & recording the measurment at the appropriate location on an image of the instrument.) If you're in the market, this should probably be a considerstion.
His website is http://www.magicprobe.net/ (You can follow some of it's development here http://www.mandolincafe.com/forum/showt ... e-Improved)
Or, you could just make your own Hacklinger.

[Dave Stewart]

Speaking of Hacklingers, CLOSEOUT SALE AT STEWMAC