VA plots for tap test on plates

[seeaxe]

I have been getting into the tap testing of the plates prior to plugging frequencies into Equn 4.5-7.

FLong is fairly straightforward and am pretty sure that this is around 80Hz from the following test results. I've reproduced this a few times and things are fairly consistent.

I note that there are two small peaks before the big one at 80, one at about 58 and one at 65.
There are also two clear peaks at around 160 and 180, then a big one at 300.

As others have found Fcross is more difficult. I've read other posts about this and I think Martin said it is usually preceded by a bit of a frequency "hump" so I have interpreted the plot below to mean that Fcross is around 160Hz..............maybe

In this plot the peaks at around 50 and 80 reappear, as well as one at 90, but also a strong peak at about 120Hz. This could be Fcross and its tempting to be guided by Table 6.1. Plus 160 is 2 x 80 so the 160 peak could be another higher FL mode.

And FLC I am struggling with. When I tap the panel what I hear is very clearly a higher pitch than FL. Of all the modes, this one is the easier to hold and tap, no ambiguities or nuance, right in the middle on the edge and tap right in the corner. I get a good clear ringing tone but it doesn't sound anything like 50Hz. Sure enough when I look at the VA plot it shows this

The big peak at 220 is what I think I am hearing. Sure there is a peak at 50 but then there's also one at 58 and a bunch more above 100.

It seems that picking one of the peaks from this plot over the others, without good reason, is out of whack with the elegance of Trevor's method.

Hence I conclude I am not doing something right. I've been through the VA set up and I think I have that right. I've changed from tapping with my knuckle to an eraser on a pencil, which helped but has got me to these plots.

I also note that the Fl and Fc are on the high side which will lead to higher E values and thus thinner target thickness.

I'm wondering whether anyone has any comments on the above. Big question is am I interpreting these correctly?

By the way, I wondered if the 50hz peak is from the equipment so I checked a plot without any tapping

PS this is one of your Engelmann tops you sent me Martin - did you do tap tests for all of them?

[kiwigeo]

Hi Richard,

Here are the typical plots I get for tap tests.
Flong

Fcross

Fdiagonal

[seeaxe]

Thanks Martin.

Well it looks like i have a few VA settings wrong then. Ill go back and reread the post on Trevors site. Also that's the "capture spectrum" screen i think. I had mixed results trying to use that.

Then I'll twiddle some knobs and see what happens.

.

[seeaxe]

Some progress.. gain was too high (I found a post of mine from 3? years ago with a comment to that effect from Trevor....)

Plots are now a bit clearer but not as clear as yours Martin, so will pursue that a bit more.

Flong

Fcross

Fdiag

No tapping

Something going on a 133hz...

So the first two are Ok, I think on the basis that I'm looking for the lowest peak that I know is not something else. eg the Flong peak appears in Fcross so I know thts not Fcross, so it must be the next one.

F diag is still a puzzle so maybe I'm not tapping it right.

Would welcome any comments on this and if anyone can see anything obvious (to them not me) sing out.
Cheers

[Trevor Gore]

Fdiag should be an easy one to get, assuming you have the VA set up right. Hold the panel in the middle of the long side edge, tap lightly on a lower corner with the mic right behind the corner you tap. Often the problem is due to tapping too hard or not putting the mic where it can "hear" the vibes.

[seeaxe]

Thanks Trevor and Happy New Year. I wish you lots and lots of heavy rain, wherever you are in Oz. I have varied the tapping a bit and F diag is very clear to listen to, but harder to get a consistent plot.

More progress today.

I've noticed a few things such as, if I watch the plot build up it rarely changes much after about 5 taps, so have stopped doing more as the picture seems more confused at times if I carry on.
I have tapped some lutz spruce and some cedar with interesting results.
I have also checked my spreadsheet using the numbers in table 4.5-3 and get same to one decimal place, I'm assuming the rest of any difference is rounding errors.

One of the good things about spreadsheets is its very easy to see how sensitive the answer is to any of the input variables.
Seems F long is the most important thing to interpret correctly.
The penny also dropped that weight/density doesn't affect the predicted thickness as square root of rho is in all the terms in the equation. That's assuming I have the spreadsheet right of course.

So I keep coming back to interpreting the plots. Here are the lutz plots.
F long I picked 82.7, seems fairly obvious

F cross - I can see the peak at 80 so the next one is at 137hz.

F diag - hard to be sure, the first peak of any significance is 51, I dont know what the one at 60 is. It looks nothing like your plot Martin.

Using these numbers though, I get an E long of 10.95 Gpa and a target thickness of 2.3mm. Bear in mind this is a classical so I have used f = 60.

For the Engelmann I started with, I get an E long of 11.1 and a target thickness of 2.3mm

So then I tried some WRC I got from Stewmac, nice glassy tone.
F long - At first I picked the big peak at 78 and ignored the 35 hz assuming that was flc

F cross. I could see what I had assumed was Flong at 78 so took the next peak as F cross

F diag - I just took the lowest obvious peak at 27.6

So you can probably guess where this went. I ended up with an E long of 25GPa and a top thickness of 1.6mm. Unlikely to be correct I thought.
I went back and revised all the assumptions for Fl and Fc one peak down ward and came out with a finished thickness of 3.1 which makes more sense.

I've got to sort out the drum sander before I do much more as most of the tops I have need some work to get them to tappable condition.

Cheers and thanks for any comments. Hope everyone is safe and sound over there.

[kiwigeo]

Yes I use the capture screen.

Thanks Martin.

Well it looks like i have a few VA settings wrong then. Ill go back and reread the post on Trevors site. Also that's the "capture spectrum" screen i think. I had mixed results trying to use that.

Then I'll twiddle some knobs and see what happens.

.

[seeaxe]

I decide to haul out my entire stash of sound boards and prepare them for tap testing as the idea is to find the best of the bunch. Some had some saw marks and some had not been sanded at all, however they are now. I also made sure all the ends were trimmed square and sides were parallel. I then went through and tap tested them all again. While things were a little clearer there was still a few tap tests that didn't really make sense as you can see from the results.

SUMMARY.JPG (36.23 KiB) Viewed 19451 times

LS002 is clearly wrong. I could not get a low Fdiag peak no matter what I did, whereas it had been quite easy to spot on the others. There is a pink stripe down the middle of these panels and I guess its something to do with that. If I used the same F diag as he other Lutz set I get a more normal result.

As I understand it, if all these tops were built to the Books, then the lightest would be the most responsive. The person I am building for wants a spruce top as he already has quite a nice cedar top. So using this as a guide to which is the better set it looks like LS001 is a clear winner.

I might use the WRC- C on one of the others.

[Trevor Gore]

Were any of the panels not flat enough?

[seeaxe]

Nope they were all flat.

[johnparchem]

I would do what you did and put an average value or 1 for the 6.9 for the G for plate ls002. Your results will be a lot close to correct and make for a better comparison. With that low cross grain stiffness the fdiag might be really low.

[seeaxe]

Thanks John.
Having identified the better tops i have now started on the sides as i want to be in a position to use the top or back as soon as its built. Auckland has permanently high humidity so i don't like leaving the plates hanging around for too long.

If you don't mind me asking, for the brace locations on your classical build, did you follow the falcate plan in the books (even though its for a SS) or develop your own??

Cheers
Richard

[johnparchem]

Thanks John.
Having identified the better tops i have now started on the sides as i want to be in a position to use the top or back as soon as its built. Auckland has permanently high humidity so i don't like leaving the plates hanging around for too long.

If you don't mind me asking, for the brace locations on your classical build, did you follow the falcate plan in the books (even though its for a SS) or develop your own??

Cheers
Richard

The pattern I use is very close to the steel string from the book. I use the same forms to make the braces but I shift them a bit to line up where I want them on the bridge. The angle on my sound hole patch is a bit different to take into account the way I laid out the primary braces. I ended up with taller falcate braces than the steel string. To get the top resonance up to 190 or 200 Hz with a top that is significantly thinner than a steel string needs the stiffness from somewhere. I end up with 10 mm or 10.5 mm tall primary falcate braces depending on the target resonance I am trying to hit. THis surprised me because they are much taller than my fan brace height for a traditional classical. But leaving out the lower transverse brace lengthens the brace span. There is a cubed relationship between the length of a brace and its stiffness.

I have thought of adding an open lower transverse brace to make it more like a classical guitar. In that case the braces could be a lot lower as the effective length of the brace is shortened. But I really like the performance of the guitar without the lower transverse brace. My guitars are loud, responsive and have a lot of sustain.

I have build blogs of almost all my guitars. You get a pretty good view of my tops. http://harvestmoonguitars.com/classicalguitars.htm

[seeaxe]

Thanks John, I will go look at your blogs.