Tap Test after Closing Box

[Woodsy23]

Hi there,

I am on my first steel string acoustic build using "the books". It is a 12 string with the shape of a Martin D28-12, a non-live back and top bracing similar to the J45 in the book. The X-braces and sound hole braces have been beefed up for the additional force from the 12 strings.

I am at the stage where the top and back have been glued on to close the box (no bridge yet) and I carried out a tap test as suggested in 12.6 of the white book. I get frequency peak at 99 Hz and 155 Hz but also get unexpected peaks at 77.5 and 120Hz, as shown here.



Can anyone tell me the meaning of those additional peaks? Are they significant? Also, is the peak at 155 Hz unusually low?

I should mention that, before closing the box, I attached K&K pure 12 string bridge plate transducers and drilled the hole for the jack and fitted it. I'm not sure if that is significant.

The next possible step in the build is to thin the edges of the top. Is that beneficial or recommended for this X-braced 12 string?

Any advice would be appreciated.

[Trevor Gore]

I think you're getting a lot of spurious stuff in your tap recordings; typically ambient noise like your computer fan, the bar fridge, cicadas etc. etc.. Have a look at this thread which addresses some of those issues (and you may need a better mic, if you're using one built into a laptop).

One of the guys (can't remember who, now) ended up building an acoustic "nest" around his guitar using old cushions etc. to kill of the extraneous noise and got much better results. Looking at your plots, there seems to be lot's of noise and not much signal, so the root of your problems may be the same.

[kiwigeo]

Woodsy, set up while tap testing can influence results. Can you post up a pic of your set up?

[Woodsy23]

Trevor,
Your were completely correct. I hadn't checked the microphone input level. It was way too high. How stupid of me. I also had the microphone (a Shure SM58) sitting on the computer deck feeding in the computer fan vibrations.

I guess if you're feeding the sound signal directly into VA only, with no monitoring of the input level, there's nothing to tell you if the input level is OK. So there's a tip for beginners like me - monitor the sound input level by, for example, making a recording of your voice with something like Goldwave. There are, of course many ways to check the sound input level.

Below is a new frequency spectrum which looks much better (to my untrained eye). This time, I had the microphone sitting on a cushion on a separate desk. The computer fan noise in the room is still significant so there's a lot more I could do to eliminate extraneous noise but it does not seem necessary, at this stage of the build anyway.



The peak at 99 Hz is still there but the one at 155 Hz is gone and the Main Top T(1,1) is at 207 Hz - Much closer to Fig 12-49 of the white book.
Do you think it is worthwhile thinning the edges of the top to change the top frequencies? I know you say that it would make the guitar louder buy how much effect would it have on an X braced guitar?

Thanks for your help.

[kiwigeo]

Woodsy I'm using the same mic as you but I have it hanging from my workshop ceiling.

[Trevor Gore]

On X-braced guitars the bridge has more of a mass effect than a stiffness effect. So, make your bridge (if you haven't already) and stick it on with double sided tape and tap it again. For a first 12, shoot for 180Hz for the T(1,1)2 if you can get it that low without compromising the structure, otherwise settle for 190Hz. With the bridge glued on, you'll likely come out a bit higher than with just tape, but remember that you'll have a saddle and 12 pins to add on. Don't take the edges down thinner than ~2.5mm unless you really know what you're doing. Have your end point in mind early, so you can get to it without a truck load of side mass.

Let us know how you get on!

[dshaker]

One of the guys (can't remember who, now) ended up building an acoustic "nest" around his guitar using old cushions etc. to kill of the extraneous noise and got much better results. Looking at your plots, there seems to be lot's of noise and not much signal, so the root of your problems may be the same.

I think that was me. In order to get a decent plot I needed to cover the disk side of my laptop with two pillows, turn off the fluorescent lights (which was a lot; I don't use incandescents unless I have to), and put my mike in the middle of a pile of pillows, with an open area pointing at the guitar. Those things cleaned up my plots a lot, though there was still a lot of grass. Then I made a hammer out of an eraser and a brass rod and I practiced hitting the guitar cleanly, without a bounce/second hit.

My wife thought I was insane, sitting in the dark with pillows on top of my computer, hitting my carefully made guitar box with a "special" hammer.

-Doug Shaker

[Trevor Gore]

Thanks for posting that again, Doug.

I couldn't find the original post. Maybe it's not even on this forum!

[Woodsy23]

Trevor,
As suggested, I attached the almost complete bridge with double stick tape and retested. The bridge was complete except for tapering the thickness at the back edge so I figured that the mass of the wood not yet removed in the tapering would be about equivalent to the mass of the 12 bridge pins (6 grams) and nut. Anyway, the frequency spectrum is below. The T(1,1)1 has dropped to 97.5 Hz and the T(1,1)2 is at 188.5 Hz.



I would appreciate your suggestions about where to go from here.

Am I correct in believing the T(1,1)1 is getting too low - too close to the G note at 98.0 Hz? For the T(1,1)2, is it feasible to get it down to 180 Hz or is it better to try to get it to 190 Hz, as you suggest for dreadnaughts in 3.2.2.4 of the white book?

Thanks

[Trevor Gore]

Woodsy,

It's hard to give you any more advice than I have already or is in the book. Check out Fig 4.6-3 in the black book and look at the SS X-brace part of the histogram. The chart is for glued on bridges. See how that compares with what you're getting and then make a call. You want to shoot for 180 or 190Hz. You may need to alter the bridge mass, possibly by using wood of a lower density to make sure you get up to 190Hz. You have a lot more information then I have so you are best placed to make the call of whether to go up or down. A well done 180Hz 12 string will be a total killer though.

You could also mass load the guitar in increments whilst tap testing to get an idea of its sensitivity to added mass, which will help your decision making.

I'd say don't worry too much about the T(1,1)1 at the moment. If it ends up too close to G and you can hear it, leave it a good while to see where it goes. If it doesn't move, a small increase in sound hole diameter will fix things. It's a very easy adjustment to do with a sanding block curved to the current sound hole diameter and then a bit of shellac wiped around the hole to seal the end grain.

[Woodsy23]

In order to get a decent plot I needed to cover the disk side of my laptop with two pillows, turn off the fluorescent lights (which was a lot; I don't use incandescents unless I have to), and put my mike in the middle of a pile of pillows, with an open area pointing at the guitar.

As you can see from the previous posts, following Trevor's suggestion, my problem of spurious resonance peaks was solved by correctly setting the microphone input level and placing the microphone on a separate desk (from the computer desk) on a cushion. Earlier today I did a spectrum test with the microphone sitting on the computer desk again (as in my original test). The results were still OK with only a couple of small spurious peaks. This suggests to me that the main problem I had was the microphone input level. Set really high, it was picking up the smallest sounds. My speculation is that the much louder tapping sounds were being clipped (kept to a fixed level) so the background low level sounds ended up high relative to the tapping sounds and, hence, showed up as additional spurious peaks. The computer "hears" the background noise as nearly as loud as the tap noise.

If the frequency spectrum obtained by VA is showing similar results to my original one, I think it would be a good idea to make a recording of the tapping (using any one of a myriad of programs available) to check that the sound being received by VA is similar to what your ear is hearing.

[Woodsy23]

Trevor,

OK. I'll do some more tests.

My blackwood (690 kg/m3) bridge currently weighs in at 27g and after tapering the thickness at the back edge will probably get down to about 24 grams. I'm not sure I can easily get much lower with a 12 string bridge with a curved back edge (like in fig 20-3 of white book). My bridge has a plan area of 5750 mm2. So I'm tempted to shoot for 180 Hz, particularly if you think that could produce a "killer" guitar, but I'll check the effects of mass loading before deciding.

I presume by mass loading you mean side mass (not adding mass to the bridge). I have mass support blocks glued in so I can do the mass loading.

Is it realistic to expect a drop of 2 to 3Hz (or more) by thinning the edges of the top to the 2.5mm you suggested? That's going to be one of my key decisions - to thin or not to thin.

Thanks again.

[Trevor Gore]

I presume by mass loading you mean side mass ...

Yes.

Is it realistic to expect a drop of 2 to 3Hz (or more) by thinning the edges of the top to the 2.5mm you suggested? That's going to be one of my key decisions - to thin or not to thin.

Depends how close to the linings you took the bracing and how thin you scalloped it to. But in general, I've not had any trouble dropping 2-3 Hz and more.

Take it down gently, keep tapping and as you're getting close leave it a while to settle before moving on. Remember also that you have clean-up after binding and finish sanding to do.

[Woodsy23]

Hi Trevor,

Today, I was able to make 4 steel blocks to mass load the sides and I have some more spectrum tests with two different masses.

I first added 394 grams total (2 steel blocks each side) and obtained the following result. The T(1,1)2 is at 181.5 Hz, a drop of 7 Hz or 3.7%. Figure 2.3-13 would suggest a drop of 3.1% for 400 gm mass loading.

I then removed two of the weights, leaving 202 grams total, and obtained the following results. The T(1,1)2 is at 183.3 Hz, a drop of 5.2 Hz or 2.7%. Figure 2.3-13 would suggest a drop of 1.5% for 400 gm mass loading.

My 2 results are not as linear as the plot in Figure 2.3-13 but I am very happy with the results. The mass loading is more effective that I had hoped (based on Figure 2.3-13). They give me confidence to aim for a T(1,1)2 of 180 Hz. With edge thinning and a mass of a little over 200 grams, I hope I can get close to that value.

I may have to get used to the idea of a heavy acoustic guitar but having a heavy body may be a good thing. With the weight of 12 tuners at the peghead, the guitar may feel more balanced.

I will now proceed to do the edge thinning, following your advice and recommendations. I'll let you know how I go.

Thanks again.

[Trevor Gore]

I may have to get used to the idea of a heavy acoustic guitar...

Well, it will be lighter than a Les Paul or a Smallman classical...

[Woodsy23]

OK. I have done some edge thinning this weekend and achieved pretty close to 2.5mm with the top sanded to remove all tool marks and grain tear out. I have repeated the tap tests with 394gm, 202gm and no side mass loading. Generally I'm getting a reduction of around 2.0 to 3.0 Hz as a result of the edge thinning. With 394 gm of mass the T(1,1)2 is 179.0 Hz (181.5 Hz before thinning). With 202gm of mass it is 183.5 Hz (183.5 Hz before thinning). With no side mass, T(1,1)2 reduced to 185.5 Hz (188.5 Hz before thinning). The result with 202gm before thinning (183.5 Hz) may have been an anomaly as I obtained a higher value of 185 Hz after a first stage thinning to 2.8mm.

Don't take the edges down thinner than ~2.5mm unless you really know what you're doing. Have your end point in mind early, so you can get to it without a truck load of side mass

Is this because it would be too risky for a beginner who would not be able to tell (by the sound) when it is getting too thin (resulting in a guitar that "will sound somewhat flabby")?
If I ended up having to use 400 grams of side mass, is that considered to be "a truck load" of mass?

My inclination at this stage is to stay where I am with the edge thinning (at 2.5mm). Hopefully, I will still be able to end up with the T(1,1)2 at close to 180 Hz by using a side mass of 400 grams or less. I note from 22.3.1.4 of the white book that I could also add a little bit of mass at the bridge although there is already the additional mass of the K&K pure 12 string bridge plate transducer so I don't want to add too much.

[Trevor Gore]

That sounds all good, Woodsy.

Don't take the edges down thinner than ~2.5mm unless you really know what you're doing. Have your end point in mind early, so you can get to it without a truck load of side mass

Is this because it would be too risky for a beginner who would not be able to tell (by the sound) when it is getting too thin (resulting in a guitar that "will sound somewhat flabby")?

There's a number of different reasons. Stiffness is proportional to thickness cubed, so removal of 0.1mm at 2.5mm can start having serious effects with rapid changes. It gets more difficult to maintain a good taper in thickness (because the wood flexes away under your tools) and with binding clean-up etc. it is very easy to end up thinner than you wanted. With a 12 string you'd more likely run into structural problems before you run into acoustic problems so the "flabby" sound is less of a risk than the top pulling up/off.

If I ended up having to use 400 grams of side mass, is that considered to be "a truck load" of mass?

Depends who you're building for! It wouldn't bother me, but the guitar would feel noticeably heavy to a lot of punters who still equate heavy with bad. 400grm gives me >5 Hz shift in the T(1,1)2, (and still proves to be independent of body size/shape) so is enough to solve most over-coupling problems.

My inclination at this stage is to stay where I am with the edge thinning (at 2.5mm).

As I was reading through your post I was thinking exactly the same!

[kiwigeo]

400gram weights shouldn't be a problem...just supply a free set of barbells with the guitar and the new owner shouldnt have any problems lifting the instrument.

[Woodsy23]

Trevor,

I must be the slowest guitar builder around this forum. It has taken me since January to complete the guitar. However, the build seems to have worked out reasonably well.

I previously wrote:

With 394 gm of mass the T(1,1)2 is 179.0 Hz (181.5 Hz before thinning). With 202gm of mass it is 183.5 Hz (183.5 Hz before thinning). With no side mass, T(1,1)2 reduced to 185.5 Hz (188.5 Hz before thinning).

After I glued the bridge on and installed the bridge pins and unintonated saddle, the T(1,1)2 with the full 394 gm side mass increased to 183.5 Hz which is a little higher than I had hoped. The T(1,1)1 is 96.5 Hz. I saw your suggestion in another thread to test with the strings installed, which I did. The T(1,1)2 reduced to 181.5 Hz which is an improvement.

Is that close enough to the target 180 Hz? I could add a little more side mass to lower it more. It's already a relatively heavy guitar but largely because of the wiight of the all the tuners.

[Trevor Gore]

181.5 Hz should be fine. I doubt you'll be hearing wolves!

[Dave M]

Woodsy, Trevor

For someone exploring this area for the first time this was a really useful thread. You have been exploring exactly the sort of things we beginners are stumbling over.

So, having cracked getting the spectra and looking at how to manage the resonant frequencies I wonder what happened next.

Did you get the frequencies where you wanted them Woodsy and more importantly what did it sound like?
Was it extra heavy...

Any info would be welcome. Dave

[Woodsy23]

It has been 10 months since i finished this guitar. I built a top deflection measuring rig (similar to Trevor's and Jeff's in this thread viewtopic.php?f=33&t=5464&hilit=deflection) to test my second guitar, a falcate steel string to the design in the book, so I tested the 12 string at the same time. The top stiffness I measured was 62 N/mm and the uncoupled top frequency I measured is 174 Hz. By my calculation, that gives a monopole mobility of 17.6x10^-3 s/kg.

Did you get the frequencies where you wanted them Woodsy and more importantly what did it sound like? Was it extra heavy...

Sorry I didn't answer your questions last year. The air and main top frequencies of 96.5 Hz and 181.5 Hz are just a little above the ideal values. The guitar sounds good to me but, as a hobby builder, I don't get to hear many acoustic guitars, except on commercial/studio recordings so its hard for me to appraise the sound objectively. It is a heavy guitar as far as acoustics go but, as I said in previous posts, the side weights help to balance the heavy headstock (with 12 individual tuners).