The Acoustic Guitar Forum

Go Back   The Acoustic Guitar Forum > General Acoustic Guitar and Amplification Discussion > Build and Repair

Reply
 
Thread Tools
  #16  
Old 05-29-2012, 01:28 PM
charles Tauber charles Tauber is offline
Registered User
 
Join Date: Feb 2011
Posts: 8,381
Default

Quote:
Originally Posted by wrench68 View Post
I would make an audio recording of the noise and do spectral analysis of it to define exactly what it is you hear (or don't hear)...

Depending on who does the troubleshooting, this information may be important to find the problem quickly.
Having done it the "old-fashioned" way, I like the idea of doing a spectral analysis to aid in troubleshooting. What set-up (hardware, software) is required or do you use to perform spectral analysis?

Thanks.
Reply With Quote
  #17  
Old 05-29-2012, 01:43 PM
gitnoob gitnoob is offline
Registered User
 
Join Date: Jun 2009
Location: Emerald City
Posts: 4,327
Default

Charles, download a copy of Audacity:
http://audacity.sourceforge.net/download/

Use the "plot spectrum" feature in the "analyze" menu.

Algorithm: spectrum
Size: 16384
Function: Hanning Window
Axis: Log Frequency

OP, you can alter the bridge mass (and thus the top's resonance frequency) with bridge pins. Experiment with some brass pins. You may only need one or two.
__________________

gits: good and plenty
chops: snickers
Reply With Quote
  #18  
Old 05-29-2012, 02:08 PM
wrench68 wrench68 is offline
Registered User
 
Join Date: Dec 2011
Posts: 402
Default

Quote:
Originally Posted by CrankyChris View Post
What does this mean "some tap tone recordings are necessary"? How does one do that?

And how might I experiment with adding subtracting mass?
Tap tones are a guitar's series of natural frequecies: the main air, the top, the back, and the neck. It is possible to identify them by rapping the center of the bridge and recording the resulting sound. I like to record at the sound hole, the top at center of the lower bout, the back at the center of the lower bout, and the back at the center of the upper bout. I then use the "Analyze" function of Audacity to analyze the resulting sounds. I then use that information as a means to understand what part of a complex sound is the guitar separated from the music played on it. When analyzing a recorded piece of music, I want the guitar's natural frequencies to be very low amplitudes.

It's very easy to experiment with mass-frequency. As a general rule, adding mass lowers the natural frequency, while reducing mass increases it. When I find a guitar is producing overtones that interfere with the music (wolf notes, thuds, etc.) I tune the top and back to the nearest prime number to minimize the resonances excited by musical notes. Adjusting frequencies to other non-prime numbers will most likely just move the problem, not minimize it. Adding weight is easier than subtracting it, and easier to undo. It takes a surprisingly small amount of weight to change a top or back several Hz. This usually gets me to the nearest prime, and often solves the problem. Temporary weights are easy to fabricate. Craft putty, coins wrapped in tape, and chips of wood all work well. For semi-permanent corrections I use chips of wood and mild glue so I can undo the correction if necessary at a seasonal change. A very precise scale is essential.
Reply With Quote
  #19  
Old 05-29-2012, 02:50 PM
wrench68 wrench68 is offline
Registered User
 
Join Date: Dec 2011
Posts: 402
Default

Quote:
Originally Posted by charles Tauber View Post
Having done it the "old-fashioned" way, I like the idea of doing a spectral analysis to aid in troubleshooting. What set-up (hardware, software) is required or do you use to perform spectral analysis?

Thanks.
What Gitnoob said.

A couple of valuable features not immediately visible in Audacity are the expansion of the wave view and the EXPORT option in the analysis window.

By expanding the wave view, you are able to see the shape of the wave, where the attack of a note and other features are visible. You are also able to see beat frequencies, as shown in this very thread.

The resolution of Audacity's analysis is much greater than it is able to display on its graph. The EXPORT feature in the analysis window produces a text file showing two columns: frequency and amplitude. This is very important in dealing with resonances because they tend to have such high aplitude differentials in very short aplitude ranges that are difficult to display. Again, using the sound clip in this thread as an example, the graph shows a spike at 124 Hz that shows other features, but not clearly. The export produced a table that showed the frequency at 118 Hz was -50 db, but at 123 Hz jumped to -16 db. That wasn't easy to see on the graph. This table is .txt, so it can easily be imported to other software (Excel, LibreOffice, etc.) for better graphing.

Beyond a PC and free software, the hardware is still pretty cheap. I use very inexpensive input devices - a cheap Radio Shack audio microphone and a Korg CM100 contact microphone. Remember a mic detects vibration of air, where the contact mic detects mechanical vibration. Audio mics are often weighted for a particular purpose, but are still still usable for relative applications. When I need something more absolute, I use a Radio Shack mic capsule that is flat 30 Hz to 30 kHz.

You already have a PC. You probably already have a mic. Audacity is free. The learning curve will be pretty short given your engineering background. I think this setup could be useful to you very inexpensively and very soon.

I can't emphasize this enough, though. You will be amazed that so very much of what you see on those graphs will so much agree with your ears.
Reply With Quote
  #20  
Old 05-30-2012, 09:13 AM
CrankyChris CrankyChris is offline
Registered User
 
Join Date: May 2011
Posts: 458
Default

Update:

I took it to an "Authorized Martin" repair person. It should still be under warranty.

I don't have a lot of faith b/c before he even heard the buzz he started taking things apart. So he really has no idea of what the problem is. I also mentioned the resonant frequency/system b/f# stuff...he said "huh?" So I'm sure in 3 days, he'll say it's fixed and I'll go pick it up, and the buzz will still be there and there will be a couple of new dings on it.

I'll follow up with results. I'm going to try to add mass after I get it back. This is pretty interesting either way!
Reply With Quote
  #21  
Old 05-30-2012, 11:23 AM
wrench68 wrench68 is offline
Registered User
 
Join Date: Dec 2011
Posts: 402
Default

Quote:
Originally Posted by CrankyChris View Post
Update:

I took it to an "Authorized Martin" repair person. It should still be under warranty.

I don't have a lot of faith b/c before he even heard the buzz he started taking things apart. So he really has no idea of what the problem is. I also mentioned the resonant frequency/system b/f# stuff...he said "huh?" So I'm sure in 3 days, he'll say it's fixed and I'll go pick it up, and the buzz will still be there and there will be a couple of new dings on it.

I'll follow up with results. I'm going to try to add mass after I get it back. This is pretty interesting either way!
Given your interest and commitment to seeing this issue through, I will explain a little more about what I saw in the analysis of your sound clip. Very few people approach the sound of a guitar this way, but to me, it is a useful approach to better understand what I hear.

I loaded the sound clip into Audacity and expanded the wave view to get an initial look at the waveform. I immediately saw that the section with the repeated B notes showed a beat - another wave with a different frequency traveling along the main wave with their peaks coinciding 12 times over a 1 second period. So when Audacity did the FFT and displayed the result, I looked for some signals with a 12 Hz differential. Your B note was in tune at 124 Hz, and there was a strong F# at 372 Hz. 124 x 3 = 372, so F#4 is the 3rd harmonic of B2. This explains how F# always shows up with B, and it is normal. The problem appeared at 120 Hz. It is not a musical note, so this signal must be coming from the guitar or some external interference. 372 Hz (F#4) - 360 Hz (3rd harmonic of 120 Hz) = 12 Hz. Voila! The 120 Hz signal is likely the problem, and therefore the signal whose amplitude we need to lower to make the 12 Hz beat go away.

The next question to answer is what is causing the 120 Hz signal? This is how we will use the tap tones of the guitar. 120 Hz can be a subharmonic of the back, a subharmonic of the neck, the top (I've never measured a Martin M top, although 120 seems a little low), this could even be electrical interference from your TV (unlikely, but possible). Don't laugh too hard at that; about a year ago I diagnosed a home recording problem whose root cause turned out to be a fluorescent lamp sitting on the same surface as the microphone. 120 Hz may even just be a resonant frequency tuned into the top. Whatever it is, we can find it.

If your Martin Repair shop is able to resolve the problem, please post a follow up sound clip. I would find it interesting to analyze the resulting sound of the repair.

Last edited by wrench68; 05-30-2012 at 01:41 PM.
Reply With Quote
  #22  
Old 06-24-2022, 01:02 PM
Tapping Tapping is offline
Registered User
 
Join Date: Jan 2017
Location: Miami/Ft. Lauderdale
Posts: 31
Default

I was having a similar problem. So glad we have this forum. I have a very loud C# and F# with some strange high overtones that my K&K- like UltraTonic hears it. I don't have bridge pins on my Breedlove to add weight.

Sound:
https://soundcloud.com/matthew-levy-...tthew-20200624

Freq Analysis
https://www.dropbox.com/s/28knvqk2jg...06.24.jpg?dl=0

What to do? Thoughts?
__________________
Matthew
Ed Gerhard Signature Masterclass Breedlove
Strandberg Metal NX

Youtube
Spotify
Tidal
ResultsManagementGroup and deepwork.coach

Last edited by Tapping; 06-24-2022 at 04:25 PM.
Reply With Quote
  #23  
Old 06-24-2022, 01:47 PM
Alan Carruth Alan Carruth is offline
Registered User
 
Join Date: Feb 2011
Posts: 4,180
Default

Software for sound analysis is easy to come by and use these days. I've used some old shareware and freeware for years; so old that it won't run on 'modern' operating systems. A friend wrote a nice free app for Android devices called 'Luthier Lab' that has a useful spectrum analysis module, as well as a signal generator that can be used to help find resonances by driving them. You'll need some sort of amplifier for that, of course.
Reply With Quote
  #24  
Old 06-24-2022, 02:30 PM
Tapping Tapping is offline
Registered User
 
Join Date: Jan 2017
Location: Miami/Ft. Lauderdale
Posts: 31
Default

Quote:
Originally Posted by Alan Carruth View Post
Software for sound analysis is easy to come by and use these days. I've used some old shareware and freeware for years; so old that it won't run on 'modern' operating systems. A friend wrote a nice free app for Android devices called 'Luthier Lab' that has a useful spectrum analysis module, as well as a signal generator that can be used to help find resonances by driving them. You'll need some sort of amplifier for that, of course.
Thank you, Alan. I have audacity and put the analysis above. I can also create the resonance and even drive it in to feedback at 130 or 180hz no problem and I know where the top vibrates. I just don't know how to fix it, even it out. Does anyone just glue a thin piece of spruce underneath on the resonant spot, mine is 1.5 in southwest of the bridge corner. What to do?
__________________
Matthew
Ed Gerhard Signature Masterclass Breedlove
Strandberg Metal NX

Youtube
Spotify
Tidal
ResultsManagementGroup and deepwork.coach

Last edited by Tapping; 06-24-2022 at 04:27 PM.
Reply With Quote
  #25  
Old 06-25-2022, 05:59 AM
Tapping Tapping is offline
Registered User
 
Join Date: Jan 2017
Location: Miami/Ft. Lauderdale
Posts: 31
Default

Thank you Alan. I just read a bunch of your posts including the one about gluing an electricians nut as a temp fix. Much appreciation.

Since I have a Breedlove with a thinnish top and a bridge truss, can I adjust the bridge truss to raise the resonant frequency? Also scale length is 25.75" at low E. and I'm using 13-56 strings.

Wrench68, I'm curious what you see in my spec analysis above.
__________________
Matthew
Ed Gerhard Signature Masterclass Breedlove
Strandberg Metal NX

Youtube
Spotify
Tidal
ResultsManagementGroup and deepwork.coach

Last edited by Tapping; 06-25-2022 at 10:26 AM.
Reply With Quote
  #26  
Old 06-25-2022, 04:54 PM
Alan Carruth Alan Carruth is offline
Registered User
 
Join Date: Feb 2011
Posts: 4,180
Default

130 Hz is high for the 'main air' pitch, but not unheard of. We call it an 'air' pitch but it's really the lower peak in a 'bass reflex couple' between the Helmholtz air resonance and the loudspeaker-like 'main top' resonance. Both are moving at both pitches. 180 Hz is a bit low for the 'top' pitch, but, again, not unheard of. To isolate them you can cover the sound hole with something fairly stiff, such as heavy card stock. This will enable you to find the 'top' (and 'back') pitches without the influence of the air pressure changes in the box. The pitches will be different in this 'uncoupled' condition, of course.

The Helmholtz frequency is largely set by the size and location of the sound hole. Isolating that involves loading the box so that it can't move; one researcher buried guitars in sand with only the sound hole open. The 'assembled' 'main air' pitch will be lower than the isolated Helmholtz frequency, and the 'top' pitch higher with the coupling than without. If the 'main back' pitch is close to the 'top' pitch that gets into the act too, and usually drops the 'air' pitch some due to the added mass.

All of this assumes that the issue is with the fundamental of the note, which may not be the case. Try recording the actual played note and looking at the spectrum. I've seen cases where the real 'wolf' was on one of the partials of the string. The tell-tale here is that that partial will be split: there will be two peaks in the spectrum. This happens when the top is moving enough so that the end of the string is not actually stationary when it's moving 'vertically' with respect to the plane of the top, but it is when it's moving 'horizontally'. Since there is normally motion in both directions the string actually produces two pitches at the same time, and the buzz is the difference frequency between them. If the problem is at, say, the second partial, then that peak, and probably the fourth partial peak, will be split.

These things can be hard to track down, in part because you can have more than one problem at or near the same pitch, and they effect each other. You can take some comfort in the thought that these things are much more common on better' instruments. Once you figure out where the problem is it's often reasonably easy to fix it, but hunting wolves can be hard work.
Reply With Quote
  #27  
Old 06-25-2022, 06:19 PM
Tapping Tapping is offline
Registered User
 
Join Date: Jan 2017
Location: Miami/Ft. Lauderdale
Posts: 31
Default

So generous, Alan. Alan and all you guys. I'm so excited to be taught by you here. What an education on overtones, harmonics, noise, analysis and tools. <3 It is a hand made Breedlove Masterclass, so decent, and fairly large, 17in bout.

Above was with a pickup and amplification as I'm also fitting a new pickup which is currently extenuating the C# problem.

So below is natural, right into a USB mic to Audacity.

Slow 6 string Strum - 12th Harmonic then Open 4th as C#, F# on 6th
Audio

Analysis Strum

Analysis C# This is my loud overtones

Analysis F# This one is my feedback frequency.

Closing the hole and tapping around on the top I get mostly 150hz and a couple pockets of loud 130 hz with many overtones and a few spots at 180hz.

The back is all 180hz. Sides are 130, 150 and 180hz at the most hollow places.

So, I guess this means that the thole thing is really buzzing at these frequencies. Hence the reason so spend on dense hard woods for back and sides, duh. OK, I'm learning.

So what does this mean? Is there anything I can/should address other than EQ? Can this be improved? Can a slight tightening of the bridge truss help?
__________________
Matthew
Ed Gerhard Signature Masterclass Breedlove
Strandberg Metal NX

Youtube
Spotify
Tidal
ResultsManagementGroup and deepwork.coach

Last edited by Tapping; 06-26-2022 at 10:13 AM.
Reply With Quote
  #28  
Old 06-26-2022, 02:41 PM
Alan Carruth Alan Carruth is offline
Registered User
 
Join Date: Feb 2011
Posts: 4,180
Default

I doubt that adjusting the truss rod would help; it's function is not to alter resonances, although it might do that anyway.

I don't have experience with Breedloves (with the JDL bridge system, right?), so I can't say much directly.

It would help to isolate the vibrating areas; to actually see the resonant patterns. This can be done, up to a point, using 'Chladni patterns'. Basically, you drive the instrument at a resonant pitch, and sprinkle some dust (sawdust and aluminum glitter are popular) on the surface you want to look at. The dust jumps off moving areas ('antinodes') and gathers along the 'node' lines in between. It can take a fair amount of power (30 W or so) to drive things hard enough to form patterns at frequencies as high as 500 Hz, but you can usually see the 'main' top and back modes, and the 'dipoles' and maybe even 'tripoles'. The signal generator in Luthier Lab is good for that (that's why he included it), but you'll need a power amp and a speaker that can handle it. Use pure tones; sine waves. Wear hearing protectors!
Reply With Quote
  #29  
Old 06-26-2022, 03:00 PM
Tapping Tapping is offline
Registered User
 
Join Date: Jan 2017
Location: Miami/Ft. Lauderdale
Posts: 31
Default

Quote:
Originally Posted by Alan Carruth View Post
I doubt that adjusting the truss rod would help...JDL bridge system, right?...
Yes a bridge truss, like JDL, not a truss rod, to stabilized the bridge for a thin top. Anyone have experience with that in this kind of a situation?
__________________
Matthew
Ed Gerhard Signature Masterclass Breedlove
Strandberg Metal NX

Youtube
Spotify
Tidal
ResultsManagementGroup and deepwork.coach

Last edited by Tapping; 06-26-2022 at 03:21 PM.
Reply With Quote
Reply

  The Acoustic Guitar Forum > General Acoustic Guitar and Amplification Discussion > Build and Repair

Thread Tools





All times are GMT -6. The time now is 09:03 AM.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.
Copyright ©2000 - 2022, The Acoustic Guitar Forum
vB Ad Management by =RedTyger=