Interesting intonation problems with two small high-performance guitars I just completed

Talk about musical instrument construction, setup and repair.

Moderators: kiwigeo, Jeremy D

Post Reply
User avatar
stepheneb
Wandoo
Posts: 14
Joined: Wed Jul 26, 2017 3:19 am
Location: Shutesbury, MA USA

Interesting intonation problems with two small high-performance guitars I just completed

Post by stepheneb » Wed Nov 21, 2018 3:26 pm

I've recently finished two small high-performance steel-string guitars inspired by the 1918 Martin 1-18 plans Ted Nelson created for GAL.

By "finished" I mean in the last couple of weeks I've finally got the bridges glued on and have done some initial work on action and intonation. The guitars of course are still changing and settling in.

I created these two guitars over a four year period meeting once a week with small group of folks all building similar guitars inspired by the 1918 Martin 1-18 plans. We were mentored by John Fabel who has experience building and repairing guitars and violins. I was also inspired by Trevor Gore's work and book: Contemporary Acoustic Guitar Design and Build.

FYI: lots more info, pictures and data on this page:
https://learnmaketeachshare.org/instrum ... itars.html

The current problem:

The tops are lightweight and efficient (top panel weights: 135 and 150 grams, bridges about 15 grams) and the main monopole resonance in the top as well as the main air resonance appear to have enough power to shift the pitch measured over a larger range of nearby individual scale note than I expected.

A few more details below (lots more in the page linked above):

I was surprised at how much the Helmholtz and T(1,1)2 resonances affected the intonation in the scale tones. Multiple scale tones below a resonance were flattened while those above were sharpened.

While I can certainly still improve the intonation the large effect of the scale tones coupling with the existing resonances will make it difficult to get the intonation I hoped for.

Reflecting on this intonation issue I suspect the coupling of scale tones with resonances is probably inherently more pronounced when a top is very light and responsive and designed to work well with the smaller amounts of energy available in light and very light strings.

A couple of steps I plan to try first to lessen the intonation errors before working on correcting them by making further changes in nut and saddle offsets.

1. Lower the fret height. I used EVO gold FW43080 fret wire and with the light and extra-light strings I designed these for I don't need all 0.043" of the fret height. While examining the intonation errors in the first few frets on E2 I noticed I didn't need to press as hard as I normally do to get a solid stop on the fret and that pressing harder increased the sharpness of the intonation error.

2. Double check the neck relief using the appropriate strings for each guitar and use the truss rod to set it to 0.1 mm.

3. Lower the action.

After I'm satisfied with these steps then finish making changes to saddle and nut offsets to correct intonation over the scale tones I care most about as much as is practical.

Any other ideas, tips, suggestions are greatly appreciated.

Thanks, Stephen
Stephen Bannasch

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

Re: Interesting intonation problems with two small high-performance guitars I just completed

Post by kiwigeo » Wed Nov 21, 2018 3:59 pm

Thanks for that Stephen. Very interesting stuff which might take a while for me to digest. Trevor Gore will no doubt have some useful comments.
Martin

jeffhigh
Blackwood
Posts: 1536
Joined: Thu Feb 07, 2008 5:50 am
Location: Caves Beach, NSW
Contact:

Re: Interesting intonation problems with two small high-performance guitars I just completed

Post by jeffhigh » Wed Nov 21, 2018 5:06 pm

A few comments,

When testing for body, top and back, forget about plugging the soundhole, it's the open soundhole frequencies and only from tapping at the bridge that matter and these are the ones which can put off notes nearby.

In the Gore method, the plugged soundhole results are only used for monopole mobility calculations.

The expectation is that only the nearest note will be affected and that it will be repulsed and I don't really see that happening, some of your notes above the monopole frequencies seem to be flat and those below sharp.

Check nut slot depth, If this is too high it can give the sharp notes you are getting in some locations

Also could be the strings, silk and steel? perhaps uneven windings

blackalex1952
Blackwood
Posts: 776
Joined: Tue Aug 27, 2013 6:36 pm
Location: North East Victoria

Re: Interesting intonation problems with two small high-performance guitars I just completed

Post by blackalex1952 » Wed Nov 21, 2018 5:42 pm

How high are your frets and what method did you use for marking out and slotting them. How consistent is your fretting technique? Are you getting consistent errors or are the results being affected by how hard you press and whether or not you are accurately fretting just behind the fret with the lightest pressure to achieve a buzz free note.
"Everything I say on the topic is based solely upon inexperience and assumption!"

blackalex1952
Blackwood
Posts: 776
Joined: Tue Aug 27, 2013 6:36 pm
Location: North East Victoria

Re: Interesting intonation problems with two small high-performance guitars I just completed

Post by blackalex1952 » Wed Nov 21, 2018 5:59 pm

Just thinking on my feet here- some instruments end up with quite a powerful 'marimba' mode caused by the neck and body resonating at a strong frequency below the T1,1. This can cause a dip in the response of T1,1 and T1,2 -I have recently read about this. I hold the instrument where the headstock joins the neck, damping the strings with my fingers and heel tap around where brace one sits. If you use the free software Visual Analyser you may see the marimba bar frequency frequency. Then look at the Visual Analyser read for the taps at the bridge in the normal way. See if the dip in the response is there- the neck resonance can be shifted by changing mass at the headstock
"Everything I say on the topic is based solely upon inexperience and assumption!"

User avatar
stepheneb
Wandoo
Posts: 14
Joined: Wed Jul 26, 2017 3:19 am
Location: Shutesbury, MA USA

Re: Interesting intonation problems with two small high-performance guitars I just completed

Post by stepheneb » Wed Nov 21, 2018 6:41 pm

Hi Jeff,

The graphs and intonation error tables are all from the White Pine guitar using Martin M540 phosphor bronze light strings.

With the sound hole open (without any side mass weights) I measure the Helmholtz at 115 and the monopole at 213.

In the intonation error table for no side mass weights I see this amazing shift from D3:5-7 where the notes are getting progressively flatter: -11, -13, and -18 cents and the next three scale tones: d3:8-10: +9, +5, and -4 cents.

It seems like something is coupling with a frequency close to 218 Hz (which is 5 Hz above the measured monopole) and pushing scales tones below that flatter and shifting ones above sharper.

I also marked D3:7 which I measured at 217.7 Hz with an "unstable pitch when tuning" background color (yellow). There are certain notes that when playing them I can't get them to settle at the pitch I'm tuning for in a stable fashion. Getting as close as I can they start sharp and within about 0.5 s have switched to flat. A tiny bit tighter on the tuner and they are just sharp, a tiny bit looser and they stay flat. My interpretation is that they are coupling with a similar resonance.

I wonder if I have some small systematic error in my spectral measurement and the monopole isn't 213 Hz but more like 218 Hz??

This gets more confusing however with the data from the measurements taken with the two 300 gram side mass weights attached. In this case with the sound hole open the monopole measurement has dropped to 201 Hz but the intonation error inflection point is seemingly still sitting in between D3:5-7 and D3:8-10.

D3:5-7: -8, -5, -9
D3:8-10: 13, 9, 10

I'll check the slot depth on the nut.


Hi Alex,

My frets are nominally about 0.043" high. The neck and fret installation went well so I didn't need to do much leveling. After leveling I drew black marks on the fret tops and then rounded the frets until the tiniest thread of that black line was left visible.

Am planning to re-do the intonation table with an alternate fretting technique -- only press hard enough to get a clear note. Previously I was pressing down with the amount of force I normally use when fretting.

I don't think the frets need to be as high as they are to get good clear stops on the strings so I'll measure their actual height and consider lowering them a bit.

I think the frets are quite similarly set. I've already set aside one saddle because I set the action so low that I started getting buzzing on multiple frets. The saddle in now is only about 0.75 mm taller.

I used a stewmac jig to cut the fret slots and their 25" slotted fret scale ruler to set the intervals. I double checked the position of the fret slots with a ruler with 0.01" gradations and a powerful magnifying glass and they were correct.

Regarding marimba mode: that's interesting and I'll investigate it. If you see my response to Jeff earlier in this message some of the intonation errors, for example D3:5-7 and D3:8-10 seemed to indicate a string coupling with a resonance around 218 Hz which is 5 Hz higher than the T(1,1)2 resonance I measured at 213 Hz.
Stephen Bannasch

User avatar
stepheneb
Wandoo
Posts: 14
Joined: Wed Jul 26, 2017 3:19 am
Location: Shutesbury, MA USA

Re: Interesting intonation problems with two small high-performance guitars I just completed

Post by stepheneb » Thu Nov 22, 2018 6:50 am

I added two graphs to my article (and attached images to this post) showing intonation errors for the guitar with the White Pine top and Martin 540 phosphor bronze light strings. The first is for the case with no attached side mass weights. The second is for the case with 300 grams of attached side mass weights.

The X axis is frequency over the whole range of scale tones from E2:0 to E4:12. This presentation of the data make it easier for me to compare intonation errors on different strings in overlapping pitch ranges.

This is just a new visualization of the same data I collected earlier. I haven't yet made any changes to the guitar or recollected intonation data by fretting lightly.

Looks like the monopole T(1,1)2 resonance is affecting multiple scale tones on strings 5, 4, and 3.
wp-0g-intonation-errors.png
white pine top, Martin 540 light strings 0 gram side mass weights
wp-0g-intonation-errors.png (24.3 KiB) Viewed 17846 times

wp-300g-intonation-errors.png
white pine top, Martin 540 light strings 300 gram side mass weights
wp-300g-intonation-errors.png (23.41 KiB) Viewed 17846 times
I find this presentation where fretted notes on different strings with the same pitch are at the same location on the frequency axis much easier to reason about than the same data in a table form.
Stephen Bannasch

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

Re: Interesting intonation problems with two small high-performance guitars I just completed

Post by Trevor Gore » Thu Nov 22, 2018 10:22 am

I've just had a quick read of this so may not have picked up on all the nuances. However, it is possible that your tuner (measuring string pitch) is triggering on the second harmonic of the note rather than the fundamental. That would be quite normal on a small guitar typically plucked at L/4 from the bridge (L=string length). Hence the coupling frequency causing the intonation problems would be at ~ 430Hz. Changing the frequency of the T(1,1)2 by adding side mass as you have done should move the intonation problem by one fret, though it's not clear from what say that this happened. So go looking for a top mode at ~430Hz (or other low order "harmonics" of ~ 215Hz) and you may find your problem there.

User avatar
stepheneb
Wandoo
Posts: 14
Joined: Wed Jul 26, 2017 3:19 am
Location: Shutesbury, MA USA

Re: Interesting intonation problems with two small high-performance guitars I just completed

Post by stepheneb » Thu Nov 22, 2018 10:27 am

Responding to earlier questions:

Nut slot depth: I checked the nut slots and the strings are sitting about half way into the filed slots.

Action at fret 1 when fretting between frets 2 and 3: strings 6, 4, 3, 2, and 1 have about 0.25 mm gap to top of fret 1. string 5 has about 0.1 mm gap.

Overall action: the action at the 12th fret/body join from string 6 to 1: 2.5, 2.9, 3.0, 2.6, 2.1, 1.9 mm

Current fret height: measuring center of fret between strings 3 and 4 -- height varies between 1.1 and 1.2 mm. averaging 1.14 mm (0.045").
Stephen Bannasch

User avatar
stepheneb
Wandoo
Posts: 14
Joined: Wed Jul 26, 2017 3:19 am
Location: Shutesbury, MA USA

Re: Interesting intonation problems with two small high-performance guitars I just completed

Post by stepheneb » Thu Nov 22, 2018 10:45 am

Without changing the guitar I collected the same intonation data using the lightest practical fretting pressure to get a clear note (presumably simulating what the effect might be if I lowered the fret height bit).
Intonation-errors-300-gram-side-mass-weights-light-fretting.png
Intonation-errors-300-gram-side-mass-weights-light-fretting.png (23.87 KiB) Viewed 17823 times
While many of the intonation errors are less there are still similar patterns of errors.

I then collected the same data however this time I worked to mute the resonances in the top lower bout by covering it with a rubbery friction fabric and piling 2300 grams of unused side mass weights on top.

I was especially interested in seeing what strings 6, 5, and 4 looked like around the T(1,1)2 resonance.
Intonation-errors-300-gram-side-mass-weights-top-muted-light-fretting.png
Intonation-errors-300-gram-side-mass-weights-top-muted-light-fretting.png (23.09 KiB) Viewed 17823 times
The T(1,1)2 resonance and presumably many more were reduced but not eliminated. The large flattening below 210 Hz and sharpening above have been reduced.

I am quite surprised that the intonation errors in string 6 around 110 Hz seem to have almost disappeared.

Here's what the actual top muting attempt looks like.
wp-muted-top-with-fabric-and-weights.jpg
wp-muted-top-with-fabric-and-weights.jpg (52.81 KiB) Viewed 17823 times
Stephen Bannasch

User avatar
stepheneb
Wandoo
Posts: 14
Joined: Wed Jul 26, 2017 3:19 am
Location: Shutesbury, MA USA

Re: Interesting intonation problems with two small high-performance guitars I just completed

Post by stepheneb » Thu Nov 22, 2018 3:31 pm

Trevor,

That's an interesting suggestion to look for low-order harmonics of a root around 215 Hz. I'll grab some loose black tea and get out the low-frequency speaker and amp combo I use for Chladni patterns and investigate.

I can also measure the pitch of the scale tones where I'm getting large intonation errors with Audacity and check the peaks in the spectrum.

I also wrote a special pitch-detect plugin for Audacity and I'll try that also.
Stephen Bannasch

User avatar
stepheneb
Wandoo
Posts: 14
Joined: Wed Jul 26, 2017 3:19 am
Location: Shutesbury, MA USA

Re: Interesting intonation problems with two small high-performance guitars I just completed

Post by stepheneb » Fri Nov 23, 2018 6:56 pm

I re-collected all the scale tone intonation errors by collecting audio samples of each scale tone in Audacity. I actually had to do this twice, once at 8000 Hz to get high-resolution spectral data and again at 22050 Hz to get better resolution data for the pitch detect plugin to process.

I used my extended version of the Pitch Detect plugin which allows me to additionally specify an offset time to start sampling. My settings were a 0.25 s offset followed by an 0.5 s sampling period.

New intonation error chart as well as spectral data for scale tones with larger intonation issues.

https://learnmaketeachshare.org/instrum ... he-problem
Stephen Bannasch


User avatar
stepheneb
Wandoo
Posts: 14
Joined: Wed Jul 26, 2017 3:19 am
Location: Shutesbury, MA USA

Re: Interesting intonation problems with two small high-performance guitars I just completed

Post by stepheneb » Sun Nov 25, 2018 8:09 am

I made a few changes and collected more intonation data and tried to force some visible chladni patterns on the top.

Changes:

1. Reduced the side mass weights from 300 to 200 grams.
2. Took 0.1 mm off of frets (about 1.15 to 1.05 mm) and re-crowned.
3. Lowered nut 0.1 mm.
4. Lowered saddle 0.2 mm.

Changes in measurement procedure:

1. Collecting 12 scale tones and the 12th fret harmonic with Audacity and later generating pitches using Pitch Detect plugin,
2. Now muting other strings during collection.

Now the maximum flat intonation error is visible on the 5 and 4 stings and is deepest around 185 Hz.
Intonation-errors-200-gram-side-mass-weights-audacity-pitch-detect.png
Intonation-errors-200-gram-side-mass-weights-audacity-pitch-detect.png (23.66 KiB) Viewed 17671 times
I put a loudspeaker over the sound hole (strings under tension and tuned), sprinkled black tea leaves and swept through a range of frequencies looking for resonance patterns. Didn't get very many but there was a strong resonant responses at 353, 366, and 369 Hz which is just about exactly twice the deepest intonation flattening at 185 Hz ... hmmm.

Here's a picture of what I was able to collect.
chladni-20181124-catalog.jpg
chladni-20181124-catalog.jpg (72.82 KiB) Viewed 17671 times
Stephen Bannasch

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

Re: Interesting intonation problems with two small high-performance guitars I just completed

Post by kiwigeo » Sun Nov 25, 2018 9:29 am

stepheneb wrote:
Sun Nov 25, 2018 8:09 am


I put a loudspeaker over the sound hole (strings under tension and tuned), sprinkled black tea leaves and swept through a range of frequencies looking for resonance patterns. Didn't get very many but there was a strong resonant responses at 353, 366, and 369 Hz which is just about exactly twice the deepest intonation flattening at 185 Hz ... hmmm.
Hi Stephen,

When doing Chladni testing I generally hold the speaker over the bridge or lower bout area...depending on where the antinode(s) is/are. I also find that the higher frequency modes are usually never as distinct as those at lower frequencies.

Looking at some of the frequency plots you did earlier on in the thread I'm pretty sure I saw a peak around 430Hz. It was the test you did with soundhole open and no weights.
canvas.png
canvas.png (38.64 KiB) Viewed 17660 times
Martin

User avatar
stepheneb
Wandoo
Posts: 14
Joined: Wed Jul 26, 2017 3:19 am
Location: Shutesbury, MA USA

Re: Interesting intonation problems with two small high-performance guitars I just completed

Post by stepheneb » Sun Nov 25, 2018 10:08 am

An interesting aspect of collecting the intonation data in Audacity is that it's very easy to correlate the much sharper initial drop in the shape of the amplitude decay curve with notes that seem "dull" and more "thunky" when I play them compared to nearby notes.

I already knew the most likely reason is coupling with a resonance that matches the scale tone -- but I find I understand the connection at a deeper level when I can combine multiple senses and representations.

The notes pitched around G3, G#3, A3, and A#3 all have sharper decays than nearby notes.
wp-A2-D3-G3-amplitude-envelopes-20181124.jpg
wp-A2-D3-G3-amplitude-envelopes-20181124.jpg (24.2 KiB) Viewed 17659 times
Intonation waveform data from A2, D3, and G3. Tracks are lined up so similar scale tones are vertically aligned.
Stephen Bannasch

User avatar
stepheneb
Wandoo
Posts: 14
Joined: Wed Jul 26, 2017 3:19 am
Location: Shutesbury, MA USA

Re: Interesting intonation problems with two small high-performance guitars I just completed

Post by stepheneb » Sun Nov 25, 2018 10:19 am

Hi Martin,

I started by holding the speaker just as you describe, as close as practical to lower bout and over the bridge. I had a hard time getting the tea leaves to migrate effectively. By resting the speaker on the strings over the sound hole it seemed to work a bit better.

I was also under a family constraint to get this over as quickly as possible. I was in my basement workshop, with my good hearing protectors and the volume all the way up. The folks upstairs thought it was unreasonably loud!

I'd like to explore this much further when no-one else is home.

In that test with no side mass weights there was a set of four resonances showing up: 375, 399, 435, and one with a bit less energy at 482 Hz.

The first three peaks were also easily visible and somewhat dropped in pitch in similar spectral data collected when 300 gram side weights were installed: 363, 379, and 401 Hz.

I'm going to replicate that test with the 200 gram side weights installed.
Stephen Bannasch

User avatar
stepheneb
Wandoo
Posts: 14
Joined: Wed Jul 26, 2017 3:19 am
Location: Shutesbury, MA USA

Re: Interesting intonation problems with two small high-performance guitars I just completed

Post by stepheneb » Sun Nov 25, 2018 12:08 pm

Frequency response for the White Pine top with 200 gram side weights.
frequency-response-WhitePine-top-20181124-bridge.png
frequency-response-WhitePine-top-20181124-bridge.png (31.16 KiB) Viewed 17655 times
Stephen Bannasch

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

Re: Interesting intonation problems with two small high-performance guitars I just completed

Post by kiwigeo » Sun Nov 25, 2018 4:53 pm

stepheneb wrote:
Sun Nov 25, 2018 10:19 am

I started by holding the speaker just as you describe, as close as practical to lower bout and over the bridge. I had a hard time getting the tea leaves to migrate effectively. By resting the speaker on the strings over the sound hole it seemed to work a bit better.

I was also under a family constraint to get this over as quickly as possible. I was in my basement workshop, with my good hearing protectors and the volume all the way up. The folks upstairs thought it was unreasonably loud!
I use a 50 watt Bradley signal generator and a 4" speaker. I find that with the higher freq modes I really have to wind the signal volume up to close to full tilt to get the top excited. I also use poppy seeds...they seem to work a bit better than tea leaves, glitter and other visual indicators. Exciting the top entails a bit of hovering around with the speaker until I find a "sweet spot".....the exact spot will differ with each mode.
Martin

Post Reply

Who is online

Users browsing this forum: No registered users and 83 guests