Measuring and calculating specific mobility

[blackalex1952]

I am trying to understand more about the specific mobility calculations in the Gore/Gilet books.
It seems to me, with my terribly disfunctional mathematical skills, that if the top resonance is dropped, the specific mobility drops with it.
Eg if a top resonating at 203Hz (Craig Bumgarner figures for a modern Selmer Maccaferri style guitar that he measured)that has a spec mob of 18.2 is dropped to 193Hz, I get a spec mob of 17.32. That is 10Hz drop at that frequency range.(very grateful for Craig's generosity in sharing those results-they are here on the forum for anyone who does a search-it's a great thread)
I just tested an OM I am into and I have 190Hz giving me spec mob =17.5.If I can get the top down where I would like it to be around 185Hz, my spec mob will drop to 17.08 , not that much of a change (5Hz drop in that range). I have noticed that since adding side mass, although the wolf notes are under control, the guitar doesn't seem as punchy and loud.
OPINIONS and EXPERIENCE PLEASE!
Also, where are you guys ending up with your spec mob figures relative to your subjective experience of the sound of your guitars? Also how high a specific mobility have people been able to achieve and still build an instrument that won't fall apart? Is there an upper limit where tone goes awry in an unexpected way? Does edge thinning and brace work rather than the addition of side mass result in an increase in mobility? Can the lowered spec mob be compensated for with a combined approach? What's going on? Maybe my maths is letting me down here....

[kiwigeo]

Are you actually measuring specific mobility with a measuring jig? ie measuring top deflection?

[blackalex1952]

Yes...using a deflection jig with a 1.25kg weight. maybe a heavier weight would give better results with a larger deflection? BTW I meant to say 180Hz would be my best target as 185 is bang on F# which I know is wrong. But for the purposes of my calc I wanted to see what effect it would have on spec mob to lower f...Just removed the side masses and strung the guitar up. Sounded nice but the G and G# loose a bit of sustain on the fundamental on the sixth string 3rd and 4th fret and 15th 16th fret. Other notes on other strings are pretty good. When I added al lot of side mass, yet to measure the exact weight, but one certainly feels a much heavier guitar in the hands, I don't get a lot of drop in the Visual Analyser taps, and the wolfing still happens. The fact that it is spread over two frets and yet the actual frequency reads between scale tones is a mystery, unless the peak is wide, ie th Q is wide.Without mass 195.8 say 196, and with the mass 191.7 say 192 which for a lot of weight is a drop of 4 Hz. The added mass puts the frequency closer to exactly 1/2 way between F#and G the exact 1/2 way freq would be 190.5.

[jeffhigh]

I am trying to understand more about the specific mobility calculations in the Gore/Gilet books.
It seems to me, with my terribly disfunctional mathematical skills, that if the top resonance is dropped, the specific mobility drops with it.

Depends on how you lower the top resonance.

If you do it by side mass, you increase the effective vibrating area of the top as well, but the specific mobility will drop.
If you do it by adding mass at the bridge the specific mobility will drop without an increase in the effective vibrating area .
If you do it by brace shaving, the stiffness of the top will reduce which will compensate for the frequency drop in the SM calcs.

I would consider that mass (either side or bridge) should only be used for minor trimming rather than major adjustment like the 20HZ drop you really want.

Are you aware that the frequency you use for the calcs is the uncoupled top, measured with a rigid plug in the soundhole?

[blackalex1952]

Thanks Jeffhigh! All my questions basically answered, along with the understanding I came to re changing the deflection figure. As I said, I'm a bit slow with maths...thanks for your very clear thinking.Great and concise practical answer fro you.
Only question left at this point, what is the effect of edge thinning? I would assume that it increases the amount of deflection and lowers the frequency.

[jeffhigh]

Only question left at this point, what is the effect of edge thinning? I would assume that it increases the amount of deflection and lowers the frequency.

Yes, But with less scope for changing the frequency than scalloping braces or reducing overall brace height near the centre of the soundboard.

[blackalex1952]

I have managed to get my t1(1) frequency to around the mid point between G2 and G#2 (102Hz) and my T1(2) to 192Hz which is close to the mid point between F#3 and G3. I am getting the very clunky note on G# and pretty much a similar clunk on G. But only really a hassle on the 3rd and 4th frets 6th string, which is G2/G#2. My specific mobility calc gets me a figure of 16.8.
I'm wondering if the bandwidth of the 102Hz and192Hz peaks are a little wide? Am I correct in saying that as the specific mobility rises, the bandwidth (Q) at that peak would narrow? Does this imply that brace shaving would be the best way to adjust the resonance as the resonance could be lowered to say 180Hz, increasing the specific mobility, narrowing the bandwidth to be more specific to a non scale tone and would also repel the T1(1) to a lower frequency which could then be adjusted to where it needs to be to avoid both a scale tone and an octave between T1(1) and T1(2)? Is there a way to modify the air mode bandwidth? Am I completely off track here?????
The following table shows the difference between scale tones at different relevant frequencies and the mid point average of each frequency. I surmise that these mid points are the ideal off scale targets. It can clearly be seen that as the frequency drops, the difference between those frequencies drops implying a narrower range for the frequency to be for it to satisfy the off scale tone requirement, and to my current thinking a narrower bandwidth for that frequency is needed for that frequency to not overlap onto the adjacent notes and cause admittance problems. Comments and advice please.....

[jeffhigh]

I have no idea how to narrow the bandwidth or if it is desirable, but have you tried detuning to see if the problem stays with the fret or follows the note.
I have had this issue when a fret was a little loose.

[blackalex1952]

Follows the notes not the frets.G2 + G#2 still a problem.Any comments re my speculations?
BTW it's an OM size guitar, sounds fantastic to me except for the wolf notes, CF rio neck very light bracing tapered not scalloped, sitka spruce top, ebony mahogany neck, 2deg bridge rotation, non live walnut back.I arched the top somewhat, a 20ft radius, 15ft radius back. Loud,punchy tight bass,thick mids, clear highs and lots of sustain except for those two bad notes.Other than that it plays very evenly over the entire fingerboard,It drowned out a Cole Clark and a Maton dreadnought in an acoustic session. Out toned em too...and ate a Martin I have here for breakfast. The two wolf notes are sort of acceptable, but I don't like them around the G on the sixth string given the end user's style of music and abilities. I'm probably being a bit picky but want to learn and evolve....

[Trevor Gore]

G2 is at 98Hz and G3 is at 196Hz, with your T(1,1)1 at 102Hz and T(1,1)2 at 192Hz. Play G2 on the 6th string and the first two partials are both in close proximity to resonances, which is likely why you hear some wolfiness around there. Classical guitars are particular prone to this as 100/190Hz is where they naturally fall, with the problem exacerbated by the thick plastic 3rd string. Wolves are never very far away on high mobility guitars. Choosing 100/180Hz helps avoid hitting two partials of the G quite so hard. However, I think it's going to be hard for you to get down there (180Hz), made more difficult due to the low radius top curvature (20' vs. ~33', which is what I use). Wolf notes are often much more evident to the player than to the audience, if that's any consolation.

[blackalex1952]

Thanks Trevor!I am reluctant to shave the braces to lower those resonances because, apart from the wolf notes, the guitar sounds great, and the bracing is pretty light. I have been mucking around with ladder braced Selmeroid builds for some time now, but this is my first X braced guitar...so risky to hack into it at this stage when the guy I made it for is keen to have it. I am grateful for your input and your books! A fine investment! But a lot to get my head around.
I almost doubled the side mass for only a small drop in frequency, seems contrary to fig 2.3-15 build volume. Also, there seems to be a discrepancy between the Chladni frequencies and the VA taps-the Chladnis resonate the most violently a few Hz lower than the VA taps. I have tested VA with A440Hz and it is accurate...I have also noticed that, even with the side mass added, the Chladni nodes for the main top don't seem to move all that much towards the edges of the soundboard...maybe the edges are a little stiff? Edge thinning a hassle as I had lacquered the top before resonance tuning.Next build I wont apply a finish to the top until I have tuned the box up! At this stage, I guess I'll lie awake at night thinking about the next step!!!!

[blackalex1952]

"Wolves are never very far away on high mobility guitars."
A good luthier is surrounded by wolves!

[Trevor Gore]

I almost doubled the side mass for only a small drop in frequency, seems contrary to fig 2.3-15 build volume.

Everything I've built since I originally drew Fig 2.3-15 has (surprisingly) followed the gradient of that line. That's about 8 different body shapes and sizes, classical and steel string. I didn't expect that, but that's how it is. The extra mass has to be seriously attached, though. Normally, I'm not trying to move more than 5Hz. Over large shifts (>10Hz), I think the line is more of a curve.

Also, there seems to be a discrepancy between the Chladni frequencies and the VA taps-the Chladnis resonate the most violently a few Hz lower than the VA taps.

Can be dependent on how the guitar is held in each case. I've been able to get them to agree very closely.

I have tested VA with A440Hz and it is accurate...

On some computers spectrum measurement (tap) and count (frequency meter) can give different results. Also, you need to calibrate the sig gen. If using VA, I've found the sig gen (wave) and frequency meter to agree, but be different from spectrum. Using VA 9 on a high spec. HP laptop with Win7 they are all very close to each other and very accurate. Was not so on my previous Toshiba running XP.

[kiwigeo]

Also, there seems to be a discrepancy between the Chladni frequencies and the VA taps-the Chladnis resonate the most violently a few Hz lower than the VA taps.

I've never had this issue. I'm using a Bradley 50 watt tone generator and poppy seeds for my Chladni testing and the frequency at which the seeds dance always coincides with the peaks from the VA tap testing +/- 2Hz.