Double Peak Coupled Freq Resp ?

You can ask questions here about Trevor and Gerard's exciting new book on Luthiery.

Moderators: kiwigeo, Jeremy D

Post Reply
Deems Davis
Sassafras
Posts: 47
Joined: Mon Jan 18, 2016 2:26 pm
Location: Payson, Arizona
Contact:

Double Peak Coupled Freq Resp ?

Post by Deems Davis » Tue Jan 19, 2016 2:11 pm

I recently finished my 1st build following as closely as my developing skills will allow the principle and practices in Trevor's book/s. I'll attempt to add a pic or two to the gallery.

I'm attaching a copy of the VA results from the completed guitar and would appreciate any help interpreting them.

1. What would make the main Air Freq so much larger than the Top Main? I've determined that I've likely made the sound hole too large, but I don't think that's what's contributing to the amplitude.

2. It appears that the coupled Freq has 2 peaks ?? Am I reading this correctly? and if so is there a likely cause or remedy that can be applied so that the 2 converge into one.

Thanks for any feedback

Deems
VA Results guitar #1.JPG
VA Results guitar #1.JPG (48.79 KiB) Viewed 17370 times

User avatar
kiwigeo
Admin
Posts: 10582
Joined: Sat Sep 29, 2007 5:57 pm
Location: Adelaide, Sth Australia

Re: Double Peak Coupled Freq Resp ?

Post by kiwigeo » Tue Jan 19, 2016 7:35 pm

What's the scaling on the Y axis.......? I think this is why the T(1,1)1 peak amplitude looks so high. If you're using VA then plot with Y axis using dB and x axis using Hertz. Note that you shouldn't read too much into the amplitude of the peaks as you're tapping the top manually with a rubber tipped mallet and striking force for each stroke wont be exactly the same.
Martin

Deems Davis
Sassafras
Posts: 47
Joined: Mon Jan 18, 2016 2:26 pm
Location: Payson, Arizona
Contact:

Re: Double Peak Coupled Freq Resp ?

Post by Deems Davis » Wed Jan 20, 2016 2:00 am

Thanks for the tip Martin, any thoughts on the double peaks for the T(1,1)3?

User avatar
Trevor Gore
Blackwood
Posts: 1605
Joined: Mon Jun 20, 2011 8:11 pm

Re: Double Peak Coupled Freq Resp ?

Post by Trevor Gore » Wed Jan 20, 2016 9:06 am

Deems, see if you can post a shot directly out of VA rather than via Excel. To convert from the .wmf file type (which the forum doesn't support) to .jpg (which it does) use a program such as Irfanview.

User avatar
kiwigeo
Admin
Posts: 10582
Joined: Sat Sep 29, 2007 5:57 pm
Location: Adelaide, Sth Australia

Re: Double Peak Coupled Freq Resp ?

Post by kiwigeo » Wed Jan 20, 2016 9:10 am

Deems Davis wrote:Thanks for the tip Martin, any thoughts on the double peaks for the T(1,1)3?
Trevor will have some ideas. Chladni testing would enable you to pin down whether only one of the peaks is your T(1,1)3 or both are.
Martin

Deems Davis
Sassafras
Posts: 47
Joined: Mon Jan 18, 2016 2:26 pm
Location: Payson, Arizona
Contact:

Re: Double Peak Coupled Freq Resp ?

Post by Deems Davis » Wed Jan 20, 2016 12:34 pm

Trevor Gore wrote:Deems, see if you can post a shot directly out of VA rather than via Excel. To convert from the .wmf file type (which the forum doesn't support) to .jpg (which it does) use a program such as Irfanview.
Here's one I just took direct from VA with the instrument strung up.
Attachments
Capture#1.JPG
Capture#1.JPG (50.21 KiB) Viewed 17331 times

User avatar
Trevor Gore
Blackwood
Posts: 1605
Joined: Mon Jun 20, 2011 8:11 pm

Re: Double Peak Coupled Freq Resp ?

Post by Trevor Gore » Wed Jan 20, 2016 12:49 pm

It still looks like you have a linear Y-axis scaling. Choose log. (dB) as Martin suggested (check box in the bottom right of the VA window) and post another pic. Keep the linear X-axis scaling.

Deems Davis
Sassafras
Posts: 47
Joined: Mon Jan 18, 2016 2:26 pm
Location: Payson, Arizona
Contact:

Re: Double Peak Coupled Freq Resp ?

Post by Deems Davis » Wed Jan 20, 2016 2:47 pm

Trevor Gore wrote:It still looks like you have a linear Y-axis scaling. Choose log. (dB) as Martin suggested (check box in the bottom right of the VA window) and post another pic. Keep the linear X-axis scaling.
hmmmm made the changes and repeated a 10 tap average, the results don't look anything like what I was getting before??? I must be missing something.
Attachments
Capture LogY1.JPG
CaptureLogY2.JPG
CaptureLogY2.JPG (36.98 KiB) Viewed 17319 times

User avatar
Trevor Gore
Blackwood
Posts: 1605
Joined: Mon Jun 20, 2011 8:11 pm

Re: Double Peak Coupled Freq Resp ?

Post by Trevor Gore » Wed Jan 20, 2016 5:27 pm

Now it looks like a signal level problem.

In the lower of your pics, move the Y-axis slider to one mark down from the top of the scale. If you then can't see the signal displayed, mouse over the Y-axis scale and move it around until you can. Where it says "Average", select 1. On the settings menu, go to the "Capture Scope/Spectrum" tab and set "Capture Spectrum" to 10 buffers. Then see what you get. You want just one tap captured per buffer, so take care with how you time your taps. If you still don't get something that looks more sensible, try reducing the input level from your mic. If that doesn't work, try increasing it. Post your best looking effort and how you got it.

Deems Davis
Sassafras
Posts: 47
Joined: Mon Jan 18, 2016 2:26 pm
Location: Payson, Arizona
Contact:

Re: Double Peak Coupled Freq Resp ?

Post by Deems Davis » Wed Jan 20, 2016 8:38 pm

Trevor Gore wrote:Now it looks like a signal level problem.

Post your best looking effort and how you got it.
Here's another sample. Settings are as noted in the snips below Capture Spectrum buffers = 10.

Input gain adjusted from ho to low ( didn't seem to have significant effect). 10 taps on center of bridge using dowel with eraser. held aprox 2 ft from microphone. Mic is a USB " Blue" brand. After last tap click on Capture Spectrum button, wait for graph to build, zoom in on Freq's 0 - 350 hz and snip.

The graph appear to me to be VERY 'noisy' I'm wondering if this might be due to the mic I'm using.
Attachments
Capture LogY3.JPG
Capture LogY3.JPG (31.03 KiB) Viewed 17308 times
CaptureLogY2.JPG
CaptureLogY2.JPG (34.03 KiB) Viewed 17308 times
Capture LogY1.JPG

Dave M
Blackwood
Posts: 595
Joined: Tue Jul 15, 2014 6:44 am
Location: Somerset UK

Re: Double Peak Coupled Freq Resp ?

Post by Dave M » Wed Jan 20, 2016 10:36 pm

Ah ha.

I went through this, it was infuriating. Click on capture THEN start tapping at just over 1 sec intervals. When done VA will give you a new window with the captured spectrum and a phase spectrum. You can get rid of the latter by unclicking in view. Then select your region with the mouse. Also you can expand either axis to give clearer view.

I prepared these before I saw all the other stuff from Martin and Trevor so I'll throw them up anyway. The first two are with y axis set to linear, one with mic pointing generally at the top about half a metre away. The second closer and pointing directly at the soundhole. You can see how this sort of thing affects the result. The third is with y set to log and shows the sort of spectrum you would hope to see. (this is a classical with falcate bracing.

The images are in reverse order - still haven't really cracked adding pics to posts.

Cheers
Attachments
02 log y axis.JPG
02 log y axis.JPG (79.61 KiB) Viewed 17301 times
02 linear y 2.JPG
02 linear y 2.JPG (61.86 KiB) Viewed 17301 times
02 linear y.JPG
02 linear y.JPG (75.33 KiB) Viewed 17301 times
------------------
Dave

Dave M
Blackwood
Posts: 595
Joined: Tue Jul 15, 2014 6:44 am
Location: Somerset UK

Re: Double Peak Coupled Freq Resp ?

Post by Dave M » Wed Jan 20, 2016 10:37 pm

Oh and have you got your fft and sampling sizes right?
FFT at 16384 and
Sampling at 11025 Hz
------------------
Dave

User avatar
Trevor Gore
Blackwood
Posts: 1605
Joined: Mon Jun 20, 2011 8:11 pm

Re: Double Peak Coupled Freq Resp ?

Post by Trevor Gore » Thu Jan 21, 2016 12:05 am

Deems Davis wrote:After last tap click on Capture Spectrum button, wait for graph to build, zoom in on Freq's 0 - 350 hz and snip.
Dave M wrote:Click on capture THEN start tapping at just over 1 sec intervals. When done VA will give you a new window with the captured spectrum and a phase spectrum. You can get rid of the latter by unclicking in view. Then select your region with the mouse. Also you can expand either axis to give clearer view.
Yep. Has to be like Dave says, otherwise you are recording nothing, which is what your signal looks like. When you click Capture Spectrum, you will see the display flicker about every 1.5 seconds (if you have the FFT and sampling sizes right (see Appendix 1, Design) and Dave's post above). You need one (and only one) tap per buffer, i.e. about every 1.5 secs. If the display doesn't refresh every 1.5 secs, you should see the blue capture spectrum bar increment every 1.5 secs. After 10 buffers (15 secs) the capture window should pop up with a display, hopefully one that makes some sense!
Dave M wrote: The images are in reverse order - still haven't really cracked adding pics to posts.
Click on where you want the image in the message (insertion point), then click "place in line". Re-select an insertion point each time you want to hit "place in line".

Deems Davis
Sassafras
Posts: 47
Joined: Mon Jan 18, 2016 2:26 pm
Location: Payson, Arizona
Contact:

Re: Double Peak Coupled Freq Resp ?

Post by Deems Davis » Thu Jan 21, 2016 9:45 am

Trevor Gore wrote:
Deems Davis wrote:After last tap click on Capture Spectrum button, wait for graph to build, zoom in on Freq's 0 - 350 hz and snip.
Dave M wrote:Click on capture THEN start tapping at just over 1 sec intervals. When done VA will give you a new window with the captured spectrum and a phase spectrum. You can get rid of the latter by unclicking in view. Then select your region with the mouse. Also you can expand either axis to give clearer view.
Yep. Has to be like Dave says, otherwise you are recording nothing, which is what your signal looks like. When you click Capture Spectrum, you will see the display flicker about every 1.5 seconds (if you have the FFT and sampling sizes right (see Appendix 1, Design) and Dave's post above). You need one (and only one) tap per buffer, i.e. about every 1.5 secs. If the display doesn't refresh every 1.5 secs, you should see the blue capture spectrum bar increment every 1.5 secs. After 10 buffers (15 secs) the capture window should pop up with a display, hopefully one that makes some sense!
Dave M wrote:
Dave, Trevor... Thanks for walking me through this. I'm consistently getting results like this: BTW what is the rationale for the preference to create the y axis using LOG? Now that I have some credible figures I need to digest these and see what it means to my build. At 1st blush T(1,1)2 @ 163 seems a bit on the low side.

Deems
Attachments
Capture LogY1.JPG

User avatar
Trevor Gore
Blackwood
Posts: 1605
Joined: Mon Jun 20, 2011 8:11 pm

Re: Double Peak Coupled Freq Resp ?

Post by Trevor Gore » Thu Jan 21, 2016 10:10 am

Well done, Deems. That looks like credible data, at least.
Deems Davis wrote:BTW what is the rationale for the preference to create the y axis using LOG?
Basically, because we hear in log. The range of loudness over which we can hear is massive, so best expressed logarithmicaly. More here. (See the last table in the piece).
Deems Davis wrote:Now that I have some credible figures I need to digest these and see what it means to my build. At 1st blush T(1,1)2 @ 163 seems a bit on the low side.
Yes. The normal ranges are in the book, of course. Tell us a bit more about the type of guitar and its construction.

Deems Davis
Sassafras
Posts: 47
Joined: Mon Jan 18, 2016 2:26 pm
Location: Payson, Arizona
Contact:

Re: Double Peak Coupled Freq Resp ?

Post by Deems Davis » Thu Jan 21, 2016 10:44 am

Deems Davis wrote:Now that I have some credible figures I need to digest these and see what it means to my build. At 1st blush T(1,1)2 @ 163 seems a bit on the low side.
Yes. The normal ranges are in the book, of course. ://en. Tell us a bit more about the type of guitar and its construction.[/quote]

About the guitar:

Scale Length 632 mm
Lower Bout Width: 381
Upper Bout Width: 288mm
Body Length: 493mm
Body Depth @ 12th Fret = 100mm
Body Depth @ tail: = 115mm
Engleman Spruce top
EIRW back and sides
Ebony Binding with Maple purfling,
Stained Glass Rossette ala Jason Kostal
Live back
Walnut CF laminated Bridge (Although the CF layer in on the bottom vs the top of the blank)
Falcate braces - stika spruce (although I fear they too low and the stock I had at the time didn't meet the height sugestions. CF reinforced with 6K tow (I couldn't find any reference in the books, but I see from the forum that 3k seems to be what is optimum.)
CF reinforced bridge, but the bridge is over sized because I predrilled pin holes and when mounting the top didn't pay close enough attention the their location to I needed to widen it to accompany the Saddle.
Nut and Saddle compensation using the quick and dirty method.
The neck is BOBO and is carved on a CNC router that I built a couple of years ago (The learning curve on that and the CAD/CAM software is equal to digesting 'the books'. I built a CAD model of the guitar using Solidworks thinking I could use the CNC to produce the bulk of the pieces. I've found that other than the neck and the Bridge it's easier to produce most manually.

I put up a few pics in the Gallery, viewtopic.php?f=25&t=7187
its a Flourentine Cutaway because I REALLY like the way they look..

User avatar
Trevor Gore
Blackwood
Posts: 1605
Joined: Mon Jun 20, 2011 8:11 pm

Re: Double Peak Coupled Freq Resp ?

Post by Trevor Gore » Thu Jan 21, 2016 3:45 pm

Deems Davis wrote: I put up a few pics in the Gallery, viewtopic.php?f=25&t=7187
Oops, missed that. Nice work!

A lot of guys seem to have opted for 9mm high braces on their first guitars to get where I get with 7 hi. If you were lower than 7mm with a larger bridge, that would explain the low T(1,1)2. Once I've thicknessed the top to spec, I take very little more off it. With clean-up and a couple of re-finishes you likely got rid of a lot more top material than I do, and that would also help explain the low T(1,1)2 (and it being a bit bass heavy). To raise the T(1,1)2 will be pretty hard - loosing mass or increasing stiffness - neither easy at this stage. Lighter gauge basses may help balance it a bit more to your liking.

Enjoy building the next one. I'm sure you'll get it closer to where you want it!

Post Reply

Who is online

Users browsing this forum: No registered users and 34 guests