Cross-grain tap test advice?

[dshaker]

I've been working through my library of sound boards, tap testing them al fresco, trimming and sanding/planing them, and then doing a proper tap test with microphone, frequency analyser, and spreadsheet. It's great fun in a very guitar nerd-luthier way. It's really amazing how much variation there can be in a pile of random tops.

Getting the long grain frequency is easy. Getting the twisting mode frequency is a little harder, but still relatively easy. But getting a good cross-grain frequency seems to be tough.

Any advice on getting a good cross-grain frequency test?

[kiwigeo]

Doug can you post up a VA screen shot of one of your cross grain tests?

[Trevor Gore]

It's mainly about where you hold and tap the wood. The general rule is hold on a node of the mode you're interested in and tap on an anti-node. For the cross grain mode, it helps to tap at the same level down the panel as you're holding it. Slide your holding point down the presumed node line (22.4% in from the long edge) and keep tapping at the same level as the hold point until you get a clean signal. Shifting your hold point between the 20% line and the 25% line will help you find the most resonant spot.

I use G-tune for this type of test, as, with the spectrum display turned on, you can see which peaks are present and which peak is being read in real time. The lowest frequency is the one you want, but make sure you're not reading a long mode frequency.

[dshaker]

Martin-

I'll post something when I get a spectrum that is a little less grassy. I tried doing one this morning, but the level of background noise was just too much. And I found I had forgotten how to copy a spectrum to post. I needed to get to work so the "Ah, f#@k it. Do it later" impulse won out.

Trevor-

Thanks for the pointers. I'll try noodling around with where I am holding to get a clearer signal. I have tried using the
iPhone implementation of StobeSoft with the low freq. add-on,
Black Cat Systems' Audio Frequency Analyzer for iPhone, and
VA for Windows

My results between software systems are reasonably consistent, so I think I'm getting good data, but I am wary of the amount of interpretation I need to do to get the lowest significant peak for the cross-grain. Long grain and twisting are clearer.

I'll take a look at G-Tune and see if that and your modification of where to hold helps get a clearer graph. Thanks.

[dshaker]

I'll take a look at G-Tune and ....

I just poked around and found
1) G-Tune is no longer available other than as a demo that times out after a few minutes.
2) They are the software engineers behind Peterson's iStrobeSoft for the iPhone and a similar product for the PC. They encourage G-Tune users to migrate to the StrobeSoft products.
and
3) The iPhone version has an add-on that shows a fairly live frequency graph.

I think I'll try that.

-Doug Shaker

[JamesO]

I'm glad to have found this thread.

When I was practicing with all of my tops a few months ago, I wasn't getting any difference between the Long and Cross grain frequency results. I inched the point where I held the tops down bit by bit from the long grain until I basically got to the middle of the tops. I tested using Audacity (plot spectrum), and then iStroboSoft.

Is it possible for the two are the same?

[dshaker]

I'm glad to have found this thread.

When I was practicing with all of my tops a few months ago, I wasn't getting any difference between the Long and Cross grain frequency results. I inched the point where I held the tops down bit by bit from the long grain until I basically got to the middle of the tops. I tested using Audacity (plot spectrum), and then iStroboSoft.

Is it possible for the two are the same?

No, the cross-grain and the long-grain frequency won't be the same. What I do is test for the long-grain frequency and then, when I tap for the cross-grain, I ignore the long-grain freq that I found and look for something a bit higher than that. If I hit the soundboard five times and I see approximately the same frequency every hit or so and it is in the range of 90-150, then I assume that is the cross-grain. But the stuff that shows as 65-80 hertz is going to be the long-grain.

[JamesO]

Doug, it's nice to meet you. I'm in Fresno. I'll go grab a top and do some taps in Audacity. I'll screen shot my results and post them within an hour.

[JamesO]

I got a little thorough with this. I wanted you to be able to hear what I'm hearing as I work down the "line" between the Long and Cross nodes. Long Grain and Twisting have always been easy. It's the Cross Grain that I have trouble discerning, which is why I'm thankful this thread was here.

I took a lot of screen shots of the graphed results and recorded the clips of my taps. Hopefully I'm presenting it in a way that's easy to understand. I'll post the Twisting, Long Grain, and Cross Grain results independently of each other, and I'll start by showing you how I set things up.

Here you see the positions where I held the top as I tapped: 1-9, excluding the Twisting node, which you see as the X at top center.

IMG_4224.jpg (139.58 KiB) Viewed 26462 times

Physical data for the piece:
- Length: 559mm
- Width: 213mm
- Thickness: 4mm
- Mass: 198g

[JamesO]

Twisting has always been easy. This one is clearly 57hz.

Audio clip on SoundCloud

Twisting.png (150.93 KiB) Viewed 26462 times

[JamesO]

Long has also been straightforward. I get a clear tap, and clear results on the graph. I'm confident saying the long grain on this is 47hz. In the recording, the tapping results seem best in positions 1-5. Positions 1-3 located right around where they should be as seen in the book (Build pg. 6-7).

Long Grain Audio clip on SoundCloud

Results of the long grain taps in positions 1-8

Long 1.png (148.92 KiB) Viewed 26462 times

Long 2.png (151.06 KiB) Viewed 26462 times

Long 3.png (153.51 KiB) Viewed 26462 times

Long 4.png (146.61 KiB) Viewed 26462 times

Long 5.png (148.04 KiB) Viewed 26462 times

Long 6.png (156.48 KiB) Viewed 26462 times

Long 7.png (150.85 KiB) Viewed 26462 times

Long 8.png (153.15 KiB) Viewed 26462 times

[JamesO]

Cross is where I run into trouble. There isn't an obvious peak that jumps out as the leader. It seems like it could be 50-ish, 125-ish, or 175-ish. The clearest audio results were when I held it in positions 5 and 6, which correspond best with where that node should be as pictured in the book (Build pg. 6-7).

Cross Grain Audio Clip on SoundCloud

Results of the cross grain taps in positions 1-8

Cross 1.png (135.2 KiB) Viewed 26462 times

Cross 2.png (151.26 KiB) Viewed 26462 times

Cross 3.png (167.04 KiB) Viewed 26462 times

Cross 4.png (162.69 KiB) Viewed 26462 times

Cross 5.png (134.87 KiB) Viewed 26462 times

Cross 5.png (134.87 KiB) Viewed 26462 times

Cross 6.png (158.3 KiB) Viewed 26462 times

Cross 7.png (127.06 KiB) Viewed 26462 times

Cross 8.png (130.16 KiB) Viewed 26462 times

[kiwigeo]

I'm glad to have found this thread.

When I was practicing with all of my tops a few months ago, I wasn't getting any difference between the Long and Cross grain frequency results. I inched the point where I held the tops down bit by bit from the long grain until I basically got to the middle of the tops. I tested using Audacity (plot spectrum), and then iStroboSoft.

Is it possible for the two are the same?

James,

What wood are you testing? Ive got a library of tap tests on Swiss, Engleman and Sitka tops in my stash so should be able to pull up a screenshot of what you should be looking for. Also what hardware are you using? (ie mic, pre-amp? etc).



My set up consists of following:

1. VA running on Windows laptop. Software setup is as per Gore and Gilet's books.
2. A Shure SM47 mic.
3. a pre-amp to give signal a bit of boost.
4. A mallet consisting of an eraser stuck on the end of a piece of dowel.

[kiwigeo]

Here are some VA plots for tap testing I recently did on a top set......I think it was Swiss Spruce.

Along grain

Across grain

Diagonal

[Woodsy23]

James,

I haven't got the books with me at the moment so I can't check but my recollection is that the point where you hold the board to generate the cross grain vibration is further from the edge of the board than indicated by your marks. I may be wrong but it's worth checking the books.

[JamesO]

James,

What wood are you testing? Ive got a library of tap tests on Swiss, Engleman and Sitka tops in my stash so should be able to pull up a screenshot of what you should be looking for. Also what hardware are you using? (ie mic, pre-amp? etc).

Thanks for the help, guys. This is redwood, and I'm using:

1) Apogee MiC 96K
2) Into a MacBook Pro
3) Using Audacity
4) Setup to use an 11025 Hz sampling rate, size set to 16384, and Axis set to "Log Frequency"
5) I tap with the fleshy part of my index finger

I'd like to know how to interpret the graphed results. How do I know which peak is the winner?

[JamesO]

James,

I haven't got the books with me at the moment so I can't check but my recollection is that the point where you hold the board to generate the cross grain vibration is further from the edge of the board than indicated by your marks. I may be wrong but it's worth checking the books.

Hey Woodsy. I had Build open when I did these today. I marked the X's where the top became the most resonant and rang like a bell. In the book, Cross is further down the board, but I think it looks like it's in line with the Long?

[kiwigeo]

Thanks for the help, guys. This is redwood, and I'm using:

1) Apogee MiC 96K
2) Into a MacBook Pro
3) Using Audacity
4) Setup to use an 11025 Hz sampling rate, size set to 16384, and Axis set to "Log Frequency"
5) I tap with the fleshy part of my index finger

I'd like to know how to interpret the graphed results. How do I know which peak is the winner?

James,

I was using VA running via Parallels on a MacBook Air but I had big issues getting a big enough signal into the computer. The problem is the Mac "Mic" input is really a line input and I had to run the mic via a pre-amp to get enough signal. I ended up buying a cheap windows laptop and running VA native...alot easier moving the data files around. Your Apogee mic is a USB mic right? Even when I ran my Shure via a USB converter/preamp and into the Macs USB port I still found the signal needed boost.

With VA I set the frequency scale to a linear scale rather than log scale. The frequencies of interest are within a restricted range (<300Hz) so no real advantage using a log scale and the linear scale expands things out a bit.

Instead of your finger try a mallet.....I use an eraser stuck on a piece of dowel. Ive rounded one end of the eraser and left the other end square so Ive got a variety of striking profiles. I strike the top and then let the eraser bounce freely off the strike point.

How do you know which peak is the one? You know roughly where the peak should be (see G and G for typical values) and you're looking for the first clear peak....as in the plots I posted up. Ive done over 100 spruce tops (Sitka, Lutz, Swiss and Engleman) and the visual pattern of the peaks is very consistent. For example the diagonal peak is always a very abrupt sharp peak and the cross grain peak is always preceded by a broad "hump".

[kiwigeo]

James,

I haven't got the books with me at the moment so I can't check but my recollection is that the point where you hold the board to generate the cross grain vibration is further from the edge of the board than indicated by your marks. I may be wrong but it's worth checking the books.

Ok for the diagonal test but you definitely need to move in a bit further for the cross grain test. On your plots it looks like your diagonal is showing up at around 50Hz which confirms youre holding too close to the long edge.

Start about 1/3 of the way across the board and then work back to the edge until you get a nice peak.

[kiwigeo]

Cross is where I run into trouble. There isn't an obvious peak that jumps out as the leader. It seems like it could be 50-ish, 125-ish, or 175-ish. The clearest audio results were when I held it in positions 5 and 6, which correspond best with where that node should be as pictured in the book (Build pg. 6-7).

James

There seems to be alot of noise below 50Hz and alot of it is of higher SPL than the tap test peaks.

What sort of plot do you get if you just record with mic open and without tapping the wood? You're in the US arent you? If you were in Australia Id get you to check that the peak at 50Hz isnt AC hum....I pick it up through my pre-amp if its too close to the power pack of the laptop.

[JamesO]

Thanks for all of this, Martin

I've had a busy day today, so my apologies for not responding sooner. I'm finishing up a draft of a paper I'm writing with a professor and need to be done tomorrow morning.

The Apogee MiC I'm using is a condensor mic with a built-in preamp. It's runs to the computer's USB port. I had the gain set right below clipping, so I think the signal is healthy. It sounds like my next practical step is to make the "mallet" with an eraser and some dowel. Maybe I can even get that done tomorrow. Then I'll take some more plots.

I tried to hold the top right where the stars are in the book. For twisting, that was right along the edge. I wish I could see the graph "live." I have to record the taps, then plot the graph in Audacity. You're telling me to keep moving my fingers ("pinch point" you could say) around until that peak in the target region is the highest?

[Woodsy23]

James,

I was going to look at the books last night but forgot. My recollection is that the location to hold the board (the node) is a point at about 23% of the length from the end, for the longitudinal stiffness, and 23% of the width from the side for the cross stiffness. These are the points in the natural vibration shape where there is no movement (the node) so the restraint your fingers are providing does not stop the vibration.

Here is a diagram of the vibration shapes of a free bar. The dark blue is the first vibration mode (what you are tyring to measure). You can see that, in this mode, the point that doesn't move is at about 23% of the length from the end.

When trying to measure the cross grain frequency, the trick is to hold the board at a position along its length that eliminates (or at least reduces) the long grain vibration. I think that is at the mid length of the board but check the book.

So, if you have a board that is 650mm long and 200mm wide, you should hold it between your fingers at a point half way along (325mm) and 23% of the width (46mm) from the edge.

I would normally check my advice carefully against the book before posting but I am at work and don't have access. I hope I am correct.

[Trevor Gore]

I tried to hold the top right where the stars are in the book...

I hope that's just a typo. Hold on the dots, tap on the stars...

[Woodsy23]

James,

After checking the Black book, I see that my advice was slightly in error. For measuring cross grain frequency the position along the board to hold is not half way but, as shown in Trevor's Fig 4.5-3, its about 3/10 along the board. My figure of 23% was close. It's 22.4%, as shown in Fig 1.6-1. Sorry.

[JamesO]

I tried to hold the top right where the stars are in the book...

I hope that's just a typo. Hold on the dots, tap on the stars...

Ha! Yes, typo. Sorry. I got them mixed up. I held at the dot (node), tapped at the star (anti-node).