Official Luthiers Forum!

I'm currently on my sixth acoustic guitar build and have decided to try and address the issue of intonation in a more indepth way. Up to this point, I have been following the system set out in the Kinkade book: high E compensated by 2mm, low E compensated by 6.5mm. After the instrument is strung up, I usually end up compensating the B string additionally, as it is often sharp. The problem is that, no matter how diligent I am in doing things properly, intonation ranges from acceptable to down right rotten. I don't feel like I have any real control over the outcome. Being a professional player, I have owned a number of high quality instruments and found myself constantly fighting the intonation on them as well. I understand the principles of the equal tempered system and that basically everything is slightly out of tune. I also understand the constraints imposed on a higher volume operation such as the larger guitar manufacturers in that they are somewhat limited in what they can do to each individual instrument. However, as a one at a time builder, I am prepared to spend whatever time is required on this aspect of construction. I am considering purchasing "the intonator" as sold by Stewmac. Has anyone out there had experience with this device? Another option that I am considering is splitting the saddle, as I believe Stefan Sobell does on some of his instruments. I am wondering what the pros and cons of this technique would be. The third thing I am hoping to get some feedback on is the Buzz Feiten tuning system. I have done some reading on it and some builders swear by it. Has any one out there used this system and, if so, do you consider it worth getting involved in?

Thanks!
Pete

You might try a compensated nut, I've had good success with those by Earvanna. Consistency in action on your builds will help with intonation as small changes in action ht. have great effects on intonation.

_________________
Brian

You never know what you are capable of until you actually try.

https://www.howardguitarsdelaware.com/

I suspect Trevor Gore will check in pretty soon, but in case he doesn't, I'll put in my spoke.

First: it's just about impossible to achieve 'perfect' intonation on an acoustic guitar. The fact that the top moves (which it has to or you would not hear the thing) means that at least one end of the strings is not 'fixed' as it has to be in order to work right. Temperament adds another level of complication: if you want to play in a number of keys, or use straight frets, there will always be compromises that render some intervals less 'pure' than you'd like. In short, there's not a lot you can do about some of the things that might sound 'off' to you.

The basic problem with setting up intonation is that the string tension rises when you fret them, throwing the pitch sharp. They go sharper at the first fret than you might think, because you displace them through a pretty sharp angle. As you go up the fretboard the pitch tends to get even sharper. How much it goes out depends on a lot of variables, with the nature of the particular string and the action height being big ones. Nylon strings, being more elastic than steel, tend to go less sharp, for example. A string that is fairly close to it's breaking point; a high E or wound G, will go less sharp than one that is relatively slacker. This draqs scale length into the problem, as well as tuning (play much in DADGAD?). It's possible to calculate this stuff if you have the math chops. For those of us who have not cultivated our inner Einstein, there is another way.

If you plot out the amount a string goes sharp in cents for each fret position, you get a line that starts out a bit high at the first fret, and rises as you go toward the saddle, as I said. Simplifying with desperate brevity; moving the nut toward the first fret shifts the whole line downward, while moving the saddle away from the 12th fret alters the slope. Moving the nut forward and re-tuning is the same as moving all of the frets back toward the nut, and shifts all the pitches downward by the same amount in cents. Moving the saddle away from the 12th fret adds proportionally more to the lengths of the higher notes, and thus makes them go flatter as you go up. Some combination of nut and saddle offset will get you pretty close to 'right' for any given string, but the offsets will be different for each string. The trick is to find those offsets.

As I say, it's possible to calculate this stuff, but for those of us who don't trust our math chops as much as we trust our tool chops, it's also possible to set up a rig to measure the offsets directly. The simplest setup is a beam with a flat top, that has the first, 11th and 12th frets set in at the right places. You have a movable nut and saddle, and some way of putting on a string and tuning it. It's easiest if the saddle has some sort of a pickup built in, so you can plug into a good tuner.

The way that I've found works pretty well is to put on a string, set up the action height, and start with the nut and saddle at their 'theoretical' positions, the same distance from the 12th fret. Tune the string as accurately as you can, and then fret it at the 12th fret, and see how much it's gone sharp. _Don't_ go by the 12th fret 'harmonic', which can actually be a bit sharp from the 'correct' pitch, due to string stiffness: always go by the fretted tone. Move the saddle away from the 12th fret and continue to check the pitches, re-tuning every time. When you get the octave as close as you think you can, go up and fret the string at the first fret. That note will be sharp too, so move the nut toward the fret until you get an exact semitone on the tuner. Now the 12th fret will be out, so you go back and fix it, which throws the first fret out, and so forth. Each time you go around the adjustment gets smaller, so at some point you just declare victory, note the offsets down, and go to the next string. I like to check each string three times, and average the results: somehow, no matter how careful I am, there's a certain amount of variation in the numbers.

I use these numbers to set up the nut, and establish the angle for the saddle slot. The lower strings often need to have the nut as much as 3mm closer to the first fret than it 'should' be, so I just cut off the fretboard there. I cut the nut, and use a Dremel to move the break point for the other strings back to the correct settings.

It seems to me that top movement is more likely to change the bridge offsets than the nut, so I don't file the saddle top before hand. Instead, I make it flat across the top, with the back edge a bit rounded, and put on the strings. Once I get the action right I use the old 'string under' trick to find the correct break points. Take a length of plain string, and put it sideways across the top of the saddle under one of the strings to lift it off the flat saddle top. Tune up the string, and move the lifter around until you get the octave pitch to be right at the 12th fret (not the 'harmonic!). Do each string, remove the saddle and file the peaks.

I did this on a 12-string I had at Montreal last summer. The octaves sounded like a Wes Montgomery solo, and when I checked it on a tuner no fretted note was more than 3 cents off. It was a certain amount of work, but I thought it was worth it. More to the point, the buyer is happy....

Great reply Al.

Fred

_________________
Fred Tellier
http://www.fetellierguitars.com
Facebook page http://www.facebook.com/pages/FE-Tellier-Guitars/163451547003866

Wow! What an enlightening explanation, Al! Thank you very much for taking the time to explain it to me. Much to ponder!

Excellent reply indeed,Alan, but did you actually mean to say "move the nut towards the fret " ...?

Yes, I'm sure he did. Al may be able to edit the original post to save some confusion down the track.

And thanks Al, for saving me a lengthy explanation!

"The Book" explains this method and a few others, varying from mathematically very intense to some mathematically much simpler methods and compares their relative accuracy. The method Al describes above doesn't take into account the frequency shifts caused by body resonances. Responsive guitars, which tend to suffer most from these frequency shifts if they haven't been modally tuned, can benefit from using this method, which is described in the book, that takes this into account. It is a method that can be fairly mathematical, but you can also "eye-ball" it with a bit of experience. The method entails accurately measuring the cents error from equal temperament for all strings, all frets, then adding nut and saddle compensation for each string until the sum of the tonal errors on each string is minimised over the frets of choice, usually those used most. Probably easier to do than it sounds!

_________________
Trevor Gore, Luthier. Australian hand made acoustic guitars, classical guitars; custom guitar design and build; guitar design instruction.

http://www.goreguitars.com.au

Great advice using the fretted string for the octave rather than the harmonic.
I wish all of the experts on the web giving intonation advice knew this.
Al, thanks for the detailed explanation.

_________________
wah
Wah-wah-wah-wah
Wah

An alternative to compensating the nut is to make a compensated zero fret - I've been doing this for a few years. Mike Doolin and Greg Byers have both written about this in the past and are worth looking up.



nigel
http://www.nkforsterguitars.com/#/book/4568936095

_________________
nigel

http://www.theluthierblog.com

Very nice Nigel, how are you bending the fret wire?
And Al, that seems like a totally reasonable method, thanks for the insight.

_________________
Jim Watts
http://jameswattsguitars.com

I start out with a wide stainless steel fret, flatten it to the correct height and to give me a broad top surface to work with, file the shape into it, then go through the grades and polish. Takes about an hour and works really well.

nigel
http://www.nkforsterguitars.com/#/book/4568936095

_________________
nigel

http://www.theluthierblog.com

Thanks for the heads up on that mistake guys: I did the edit, so now it should be right.

As Trevor says, the more responsive the guitar is, the more trouble you're likely to have with intonation. I've noticed over the years that the best guitars often show the most 'problems' in their spectra: things like 'split peaks' on upper partials that happen when there's a strong interaction between the top motion and the string. The thing is that you often simply don't notice them unless you know they're there, because there's so much else going on in the sound. One of my 'test mules' has it's 'main top' resonance exactly on the open G string pitch (when the humidity is right), and in a spectrogram of the pluck you can see the fundamental 'coming and going' as the two polarities of the string motion beat against each other. Most people don't hear it unless I point it out to them, though; it just makes the sound 'interesting'. This is not to say you should not try to get those main resonances away from played pitches; rather it points out that, on a good instrument, you're almost bound to have some of this, just because there's so much happening. The trick is to keep it from being a problem. But, back to the subject at hand...

It's really high time for me to re-read 'the book'. I'm sure Trevor's 'better' method is less trouble than it might seem at first, and I should give it a try. OTOH, the method I outlined above works so much better than anything I've seen before that, for the moment, I'm pretty happy with it.

Alan's explanation for correcting intonation at both the nut and saddle positions to me is the best and easiest to understand that I've heard. It comes as no surprise to me,as he always seems to come through with top notch data. Stick another feather in your hat Alan..............!!!
Tom

_________________
A person who has never made a mistake has never made anything!!!

Al, any chance you can show a picture of your beam setup?

Nigel, how did you go about figuring the different levels for your zero fret?

By reading Byers and Doolin. They've both written about how to work this out.

It's like this: set up your guitar as normal, get your tuner out (mine is a Peterson strobe) and make a map of where you're at for each string at each fret. By that I mean make a note of how many cents sharp or flat you are of equal temperament for each string at each fret. From that you can calculate whether your nut and saddle should be in relation to where they are now: on a 25.4" scale, one cent equals about 0.4mm (sorry to mix imperial and metric, I'm forever doing it) So you plan your new saddle and zero fret (or nut) accordingly adding or subtracting 0.4mm for every cent out. Then I make 'em and fit 'em and check 'em. If it's correct - and you can get extremely close to perfect equal temperament, then take notes for that scale, action and string as this is quite repeatable. Strings do vary of course, so it's worth doing again if you're changing gauge or type, but this is a very effective way to do this.

Changing the zero fret or nut in accordance to what is going on in the first few frets and changing the saddle to fit in with what is going on at say the 10th and up means you usually don't have to keep going back and forth. It works. Try it.

The hardest thing is to actually accept that equal temperament is what it is - perfectly out of tune! I've set guitars up like this so many times and still find myself tying to sweeten when tuning which is crazy - I understand the reason for equal temperament but my heart is just drawn to Just temperament...which is another story entirely....So, even when it's done all you have achieved is a compromise. But a good one.

nigel
http://www.nkforsterguitars.com/

_________________
nigel

http://www.theluthierblog.com

Hey guys I graduated from the University in 1994. My head hurts.

Al I would suspect you get 3 different numbers when you are averaging your results because of the human factor of pressing a string to a fret.

Thanks for the explanation, Nigel. One more question....How much extra length do you leave on the nut end of the FB to accommodate your zero fret?

All the advice you are getting about nut detailing, getting the action consistent and knowing what string guages you want are important, BUT:
To answer your original question:
I have an acoustic where I have been very happy to have a split saddle. In addition, the grooves in the bridge for the saddle are much wider than the saddle. I place, but do not glue, shims (veneer) made of tone wood beside the saddle to take up the space. Therefore, it only takes me a moment to loosen the strings, rearrange the shims and move the saddle half forward or backwards. Both grooves are tilted.
Stewmac's gadget is a clever way to figure out where you want to cut the grooves, I've done approximately the same thing by gluing the bridge to the guitar and drilling the string holes before cutting the saddle slots, then putting temporary chunks of wire for "saddles".
PS: Few sounds on earth are louder than taking a router to a bridge glued on a nice acoustic guitar. Wear ear protection or you'll never be able to hear the guitar so good.

I use the "thin pin" on my fretting jig on the zero fret, the "thick pin" on the rest of the frets, and add about 4mm behind the zero fret for the cut off. I don't even measure it. Just eyeball it. If it's too long you can get some odd buzzes, if it's too short it can break off during fretting.

Let me know if you try it - post some images - that would be great.

nigel
http://www.nkforsterguitars.com/#/book/4568936095

_________________
nigel

http://www.theluthierblog.com

Alan's post is very good. If you understood it, and are curious about what the Buzz Feiten system is, consider this:
Alan says, correctly, that the harmonic is often sharper than a true octave. The lighter the strings and the shorter the scale, the more this is a problem. The Feiten system does for a guitar exactly what piano tuners have always done: put the instrument in equal temperment in the middle region where melodies are played. Then the lowest notes are adjusted so that the real overtones of the low strings tune up with the middle range. The lowest notes are stretched lower so that the (unfortunately) sharp overtones sound beautiful with the middle range. The high notes are stretched sharp so that they will be in tune with the (unfortunately sharp) overtones of the middle melodic range. Read any account of piano tuning.
Notice also that Feiten insists that his repairmen work on the nut first. Most guitars have more gross intonation issue with the nut than with the problem that Feitan addresses. Funny story: I'm from Kalamazoo, one day two guitar lovers were talking about fixing nuts in front of an old Gibson factory guy, making it sound like rocket science. The Gibson guy says "Jeez, we used to call that 'putting the nut in the right place.'"

If you want to stretch tune a guitar (similar to how a piano is tuned) you don't need the Buz Feiten system. Tune your first string using a reference (electronic tuner, fork, whatever) then put the reference away. Then tune the double octave harmonic (5th fret) on the 6th string to match the first string. The open sixth will be 2-3 cents flat compared to an accurate electronic tuner. Ignore that. Tune the rest of the strings using only octave harmonics and fretted notes. So 6th string 12th fret harmonic = 5th string 7th fret; 6th string 12th fret harmonic = 4th string 2nd fret; 6th string 5th fret harmonic = 2nd string fret 5; 6th string 5th fret harmonic = 3rd string fret 9. There you go, stretch tuned (assuming you could get decent equal temperament before you started) which is just matching lower string harmonics to upper string fundamentals.

_________________
Trevor Gore, Luthier. Australian hand made acoustic guitars, classical guitars; custom guitar design and build; guitar design instruction.

http://www.goreguitars.com.au

philosofriend wrote:
"...the harmonic is often sharper than a true octave. The lighter the strings and the shorter the scale, the more this is a problem."

From what I've seen, it's _heavier_ strings that have sharper partials. Plain string stiffness goes as the 4th power of the diameter, so heavier strings have higher stiffness all else equal, and it's the stiffness that makes the upper partials go sharp. Windings on strings are a way to add mass without adding a lot of stiffness, and they should, in theory, have more 'true' partials. In practice that doesn't always work, since the windings, and the crud that gets between them, can mess up the mass distribution.

Thanks Alan, your reply caused me to hit the books. I'm still not certain of a conclusion, but I've realized that it is complex.
On page 313 of the Dover edition of Benade's Fundamentals of Musical Acoustics, he says that long, tight, slender strings have the least problems of inharmonic overtones.
For us as guitar players though, going to a thicker string (bad for overtones) causes us to have to pull the string tighter to tune it (good for overtones). After thinking about this for a few days, I suspect that the best compromise, and the best overtones, come from medium/light acoustic guitar strings: 11 or 12 to 50 on a 25 1/2 scale steel string.
I have only two arguments to back up this hunch: (1) the reason these strings sound so good is precisely that the overtones do line up (2) it seems to me that electrics with light guages have more problems with this, Buzz Feiten is an electric player.
Thanks in advance for any further comments.

This does get complicated, that's for sure!

One test I did when I was looking at strings a few years back was to compare the overtones of an A string on my more-or-less rigid string rig, and on a guitar. On the rig the overtones went progressively sharp, as expected. The 'main air' mode of the guitar was near that A pitch, which tended to shift the fundamental a bit. As a result, the overtones up to the fourth or fifth (iirc) were actually flat. What was happening was that the pitch shift caused by top compliance meant that the string had to be a little slacker than it was 'supposed' to be to get the fundamental to be in tune on the tuner. The overtones were not effected by the top compliance, though, and showed up flat. for that string, on that note, the top compliance was helpful.

Solid body players tend to spend a lot more time up the neck than acoustic players do, which makes any intonation problems worse, and that may account for the need for Feiten nuts. A lot of them use plain thirds, which bend more than wound ones for the obvious reasons.

So when you compensate both the saddle and the nut, how does that affect playing with a capo??? My tecnical mind tells me that it´ll work very bad, but please tell me if I´m wrong.