Official Luthiers Forum!

Hello everyone! Just finished my first guitar. It didn’t collapse upon itself after stringing up, so that’s good. I do, however, have an issue with intonation. I would like some advice before I start guessing about how to adjust things and exacerbate the situation.

I tune the guitar up using a vibration tuner and open E chord sounds great. When playing a G, D, A chord the B string sounds pretty horribly sharp I think the E string is fairly sharp as well.

The saddle is already slightly compensated, but am I correct in thinking that I should compensate the saddle more—focusing the adjustment on the B string specifically? Doesn’t seem like there is much more compensation I could do there, but maybe small adjustment = big result?? No idea. I have heard of builders compensating at the nut as well but I also have no idea what added benefit/drawback that would provide. Any and all advice is truly appreciated.

TL;DR -- What is the easiest way to determine how/where you need to adjust a guitar’s intonation?

_________________
Mark Dillon \\ Dillon Instruments :::====( o || )

Judging by inharmonicity of open chords is not a reliable means of checking intonation (especially first position chords). Your case for example - in order to make your open E sound best, the minor third interval of the B against the G# would sound best about +16¢ over equal tempered tuning (a minor third is most naturally harmonious at a 6:5 ratio, which is notably wider than an equal tempered minor third). Play a G or open A however, and now the G/B strings form a major third, which is most naturally harmonious 14¢ narrower than equal tempered, or a full 30¢ off where the open E would sound most natural.

Normally most of us are conditioned to tolerate equal tempered thirds today (this hasn't always been the case), and tolerate the color so long as you don't focus on it too intently. Try to sweeten one though, and you will almost inevitably push another across a threshold of sounding especially dissonant.

So a definite "NO" to using open chords as a reference for making adjustments, unless you are very well versed in the language of temperament enough to make the smoke and mirrors of shifting things around work reliably in your favor (this can get very complicated).

Better to get a good tuner (if you have a smartphone or tablet it's hard to beat the Peterson Strobosoft app), and come up with a more reliable and objective map of where and exactly to what degree your problems lie. Tune open, note direction and degree of errors around the 12th fret area, then check errors in the first 2-3 fret region, and if you want to be thorough you can check the upper 17+ range as well. With this rough map of trends in areas, it can be much easier to help analyze the source of your issues remotely.

My guess though, is that your nut slots are probably just way too high.

_________________
Eschew obfuscation, espouse elucidation.

How much compensation did you add to the scale length at the saddle location and is this a steel string guitar?

_________________
Jim Watts
http://jameswattsguitars.com

@David: Thank you very much for the detailed response. I have essentially zero experience with this level of frequency analysis, but it gives me a place to start in my search for answers. The more I learn about building the more I realize I have yet to learn.

@Jim: Steel string guitar. Not sure how much I added. How would I measure that? It’s standard Martin scale. I used a friend’s jig to route the slot. I will take some measurements however you would like right when I get home this evening. Thank you!

_________________
Mark Dillon \\ Dillon Instruments :::====( o || )

if you measure from the nut to the 12 fret and multiply by two you'll get the scale length.
Now measure from the 12th fret to the saddle and subtract the first measurement from it and this will give you your compensation.
BTW: the B string on a steel string guitar usually needs some extra compensation.

_________________
Jim Watts
http://jameswattsguitars.com

@Jim: Thank you so much for breaking it down simply for me. I did as you requested (and verified all of these measurements 2-3 times) and this is what I came up with. Seems a fair bit off.

Measurement of Nut to 12th fret: 12 19/32” ( x 2 = 25 3/16” )

Measurement of 12th fret to the saddle:
Low E String - 12 27/32"
B String - 12 25/32"
High E String - 12 24/32"

So, using your formula I would actually need 3/16" of compensation for the B string and 5/32" for the high E. Is this even salvageable? My saddle is 1/8" wide.

_________________
Mark Dillon \\ Dillon Instruments :::====( o || )

Mark, it seems like there might be a couple of thing going on here. If it's playing sharp I'm thinking Davids spot on with high nut slots. if you numbers are right I'd think you'd be playing a little flat. I'm guessing you action is probably high too.
You might find this stewmac calculator helpful.
https://www.stewmac.com/FretCalculator
You put in your scale length (25.187) and number of frets and it spits out the fret spacing and a reasonable approximation for the compensation amount.
As David says above get you hands on a good tuner so you can dial this in. Also, get your action dialed in at the nut and saddle and check it out, you might be able to make it work.

_________________
Jim Watts
http://jameswattsguitars.com

That's a rather atypical scale length. Where did you get your board? Did you cut the fret slots yourself? Whether from specs of supplied parts or calculations done yourself, the scale length of the fret spacing should be something you can look up as reference. I must admit, the measurements provided seem a bit suspect (both the atypical scale and what seems to be gross overcompensation). I would also add that measurments in 1/32" steps is a bit course for this task. A scale measuring in .001" increments is much preferred, or at the very least carefully and accurately read .01" or 1/64" (1/32" is >.031", which is a pretty wide increment for this application). You're measuring to the center if the 12th fret, right?

And a good tuner such as Strobosoft certainly seems in order to give us some more objectively qualified observations.

_________________
Eschew obfuscation, espouse elucidation.

That stewmac link is a great resource to have. Thanks! I was actually measuring from the nut to the starting edge of the 12th fret for the scale length. Then I was measuring from the start of the 12th fret to the soundhole-side-edge of the saddle to get the compensation. I will have to completely redo those calculations and hopefully that will standardize these numbers a bit more.

The fretboard was pre-slotted when I bought it from a sort-of-local gentleman, but I haven’t been happy with the quality of some of the other materials I purchased from him so at this point I’m hoping the fretboard was done accurately. The Cocobolo rosewood back & side set I purchased ended up warping/doming tremendously after I brought it home; so much so that I didn’t feel comfortable using it at all for my first build.

What do you recommend for an accurate ruler / measuring device? I will purchase that tuner app this evening.

_________________
Mark Dillon \\ Dillon Instruments :::====( o || )

Measurement from the nut to the middle of the 12th fret is 12 21/32"—which gives me a scale length of 25 5/16". That’s at least a bit closer to standard, correct? I apologize but I don’t currently have access a ruler that shows anything greater than 32nds. Also, the budget for Luthiery comes after the Budget for Baby’s Diapers, Food, Gas, Car Repairs, etc. I’ve got a 6" caliper, but that’s about it for now other than the tape measure. Sorry!

Measurements from the middle of the 12th fret to the string break on the saddle:

B String: 12 24/32"
High E String: 12 23/32"

Which gives me the following for compensation:

B String: 3/32"
High E String: 1/16"

I truly appreciate all the help. It means a lot. What do you recommend based on these numbers?

_________________
Mark Dillon \\ Dillon Instruments :::====( o || )

Those numbers actually look quite good and typical, at least in the ballpark of average. We'll have to wait for tuner results, but my guess will be that your intonation is pretty close in the frets 5 and up, but most notably sharp the closer you get to the nut. If this is the case, it would essentially confirm that your nut slots are too high (by far the most common cause of bad sounding first position chords).

_________________
Eschew obfuscation, espouse elucidation.

Lay your guitar down on a bench where you can get some good back lighting, and a good close view from the side of the neck at the nut. Lightly press down the a string before the first fret with your left index finger. Then with the middle or ring finger of your right hand, press the string down just past the third fret. Now get a real close look from a shallow angle across the board and good back lighting to the neck, and with the free index on your right hand tap the string between the second and third fret. If your frets are truly leveled you will hear a slight 'plink', and see the tiniest gap of light appear and disappear between the string and second fret as you tap and release (you should have to be viewing from just the right angle with just the right light to see this).

This is your reference gauge - learn it well.

Now to check your nut height, with your right index press the string down past the second fret. With your left index, tap between the nut and first fret. The 'plink' should be the same, and the tiny spec of light appearing and disappearing between the fret peak and string should be almost as hard to see.

This is not an easy skill to master (viewing and gauging accurately by eye takes a lot of practice), but gross discrepancies you should be able to spot. And at the risk of sounding presumptuous, the mini-tutorial I just laid out is a real golden nugget if you want to learn setups, a skill that even many tenured professionals never master, and a factor of fine tuning that often separates the "meh" setups from the "on my God I've never played a guitar that felt like this before" ones. It will take repeated practice to learn, longer to master, and you will mess up a lot of nuts by cutting too low in the process (when the 'plink' disappears). It also has very significant influence on first position intonation. I credit broad lack of skill and awareness in this aspect of setup with the misguided development of systems like Earvanna (grossly) overcompensated nuts.

Of course from your decription of the questionable quality of the board, all this advice may be for naught, and you could just have a majorly screwed up fretboard. Still worth checking at least in this case, and definitely worth learning overall. If the board spacing is reasonably accurate, your saddle measurements appear decent, and should not be affecting gross errors in first position chord intonation. High nut slots certainly would.

_________________
Eschew obfuscation, espouse elucidation.

Just to make sure, you're not suggesting that an E chord would sound better with the B 16 cents sharp, right? That would be seriously dissonant with the fundamental. The way you "fix" that chord, is by flatting the tempered G#.

Right - didn't mean to imply that the b should be sharper, but simply that the g#/b interval is preferred to be wider. In the E chord the 2nd/3rd string conflict prefers to be widened by flattening the 3rd, while in G and A shapes the natural tendency would be to narrow their relation by flattening the 2nd

_________________
Eschew obfuscation, espouse elucidation.

This should be enshrined in the tutorial forum. Great thread

@David: I can’t thank you enough for your thorough and easy to follow recommendations. I finally got an evening to spend some time examining this and your diagnosis was nothing less than perfect. The frets tested as being completely level but I was getting way too much space between the fret and the string when running through the test of the nut on the B & high E strings.

I didn’t want to mess up this nut [because I don’t have a source for another nut locally] so I was very slow about removing material, but after 3 or 4 rounds of filing, tuning back up, then checking again the problem is gone!

This makes for a memorable day, as this is the first moment I have felt 100% proud of the guitar I made—and that is thanks in large part to you. Prior to this I was always slightly fearful this was a sign of a larger issue.

The guitar will be played in front of others for the first time tomorrow morning at our local church service, as I’m part of the music team there. I couldn’t be happier! My sincere thanks for all of your help.

_________________
Mark Dillon \\ Dillon Instruments :::====( o || )

That's great to hear it worked out so well! I had a hunch (because this is an extremely common issue), but was still a bit worried it could have been a more serious issue with board/nut/saddle layout. Congratulations!

_________________
Eschew obfuscation, espouse elucidation.