MIMF

Many people use (for this thread let's call it) a compensated nut location by moving the nut a few hundredths of an inch or so toward the first fret. I'm not talking about adding width to the nut for each string (truly compensating the nut), just getting closer by moving the whole nut location. When doing this, is there a rule of thumb about changes you should make to your built in saddle compensation? Let's say that my starting point for saddle compensation for a given scale length is going to be 0.09" (high e) and 0.21" (low E) (knowing I am going to adjust the saddle bearing surface further once strung up), but then I also compensate the nut location by, let's say, 0.03"; Do I reduce my low and high e saddle compensation numbers by the same amount to save me some work on intonating the saddle at setup? Some other amount? I'm just trying to get into the ballpark but it would be nice to get into a smaller ballpark <g>.

Ohhhh Bryan, you just sent up a red flare for Trevor, didn't you!

Check this out: of all the descriptions of saddle and nut intonation methods I've read over the years (online, and in decades of American Lutherie articles), this one thread over at the ANZLF makes the most sense to me and seems most achievable. Actually, I printed out the first post to save. Since I've been using zero frets, I haven't tried it, but it is Trevor approved!

http://www.anzlf.com/viewtopic.php?f=33&t=6262

I too have been using zero frets but am considering getting away from them. Thanks for the link, I'm going to read it right now.

Nice link, thanks. Dougs procedure looks pretty easy to follow - and it's Trevor approved!

Doug's procedure in his post on ANZLF is brilliant, and really easy to follow. I'm a very occasional visitor here, but am extremely interested in this topic; I hope I'm not speaking out of turn. Out of purely academic interest (since Trevor has validated Doug's procedure for all practical purposes), I wonder why Doug chose to start the iteration by measuring pitch at the 5th fret, rather than a lower one.

Is it possible that starting at a lower fret would shorten the procedure (in terms of the number of iterations before reaching the practical limit of construction tolerances)?

I've built a rig that measures necessary nut and saddle compensation for strings at a particular scale length, playing action and string type (but just for a single string at a time) and my limited practical experience so far has been that the lower the fret you can start off with to measure nut compensation needed, the better the result. I find two iterations to be sufficient if you use the 2nd or 3rd fret, and while I would not claim I get the accuracy obtained by these others, these few iterations seem to be enough to make the guitar intonate very well. Why the 5th - does anyone have an idea why that might be a better choice than a lower fret?

My choice of the fifth fret was purely pragmatic. I play blues and most of the time I am fretting at the third, fifth or seventh fret. The fifth fret was in the middle of that, so I assumed that if I got the nut to fifth fret right, the errors would radiate out from there and things would be pretty much OK for most of the music I do. Most of the players I observe also spend most of their time between the nut and the ninth fret, a focus on that neighborhood was confirmed from a second direction.

But if you get better result from focusing in the third fret, I'm happy to have that bit of info. And I'm happy my procedure/explanation worked for you. Let's hear it for nut compensation! (Though my wife means something completely different by that phrase.)

Thanks!

I'm confused from the description how you are using a saddle as a stand-in nut. I think this is a case where a picture would be worth a thousand words, any chance you could post one?

Thanks!

When you are making this compensation, is it for the fact that the string at the nut is higher than it is on the frets, or some other factor?

Would there be any reason to compensate a zero fret?

A properly made nut should put the strings no higher off the fretboard than a zero fret would; in either case they need to be high enough to avoid buzzing at the first fret and back buzzes higher up. They don't need to be any higher. So I can't see where having a nut would be appreciably different from a zero fret.

At any rate, if you look at the geometry of the situation, the strings are stretched a bit more when they're fretted at the first fret than they are at the second, more there than at the third, and so on. When the string is fretted on those low frets it goes sharp: that's why you need nut compensation. How much you need will depend on a lot of factors: it can be quit a lot and it might even be none at all in special cases. It's going to be something a bit different for each string, though, just as saddle compensation is, and that's why I don't think a zero fret is likely to work as well.

Randolph Rhett wrote:I'm confused from the description how you are using a saddle as a stand-in nut. I think this is a case where a picture would be worth a thousand words, any chance you could post one?

Randolph-

I don't have a nut-less guitar in process which I could use as a model but let me try to explain more clearly.

The idea of the feeler gauges and the saddle is to simulate a nut, but in a way that I can slide the nut back and forth to test different positions. The feeler gauges make a platform, the thickness of the fretboard, where the nut would be. The saddle on its side puts the strings at the approximate heights of the strings above the fretboard in the completed nut. You could use anything - I just happened to have an unused saddle that, flipped on its side, was about the height I wanted the strings to be above the fretboard. If it wasn't the right height, I think I would have notched it (or shimmed it) to bring the strings down (or up) to the right height.

I hope that is clearer.

Chris Garland wrote:I wonder why Doug chose to start the iteration by measuring pitch at the 5th fret, rather than a lower one.

When I do this "properly" I run an optimisation routine that covers all the frets I want to incorporate (usually 0-12 on strings 4, 5 and 6 and 0-18 on strings 1-3; no point in forcing a sub-optimal result because of frets you never play on). The reason Doug's method works so well, I think, is because if you had to pick just one or two positions to get get right, you'd pick the full, 3/4 string and 1/2 string lengths, which are frets zero (nut), 5 and 12. If those are right, the others can't get very far out.

Was going to mention that a couple properly compensated Moscow Mules with obviate any nut compensation. But I suspect you knew that.

One thing that does happen is that, in general, using more compensation at the nut means that you need less at the saddle. Obviously you could use too much at the nut end, and mess up the lower notes, but it opens up the possibility that moving the zero fret in a bit would at least improve the intonation in the lower positions, and still allow you to compensate the saddle to get the higher position notes close to right. In theory I suppose there's one 'correct' compensation for any given situation, but in practice there seem to be a range of options that all work reasonably well, even though they're slightly different.

I'm using a zero fret and medium frets. If I lightly fret each note, they are all in tune, according to several different regular old guitar tuners. If I press down hard, I can get them to go about 50% sharp. It's easier to get the first few frets to go sharp than down around 12. I'm thinking it's because there is more room between the frets to get more of the string mashed in there. Don't really know.
Anyway, it seems like it isn't the first little bit of stretching to get to the fret that makes it go sharp, but the extra pressure afterwords. Logically it is a factor, but my tuners can't measure it, and I can't hear it.
It's pretty hard in a playing situation to know if I'm using more pressure than I need. I suppose a more talented musician could actually hear and adjust for minor discrepancies.
It seems like how tall your frets are would have some effect on how much compensation you would need, assuming that a taller fret has more opportunity for "mashing". Also, playing style might have to be considered.

I find that with 0.5mm compensation at the nut, the saddle compensation is less dramatic; I file the individual nut slots back if necessary. It leads to less compromise for where the intonation is proper. With many production guitars, even with careful setup, the 1st to 3rd frets are sharp unless the saddle is overcompensated, giving flat notes (medium pressure) up past the octave. With the nut compensated, I can get the intonation close enough that the biggest factor by far is how hard the player frets and strums.

Of course, string guage, prefered string height, and being light or heavy handed are factors. With heavy strings, low action, and a light touch, compensation is a much smaller concern.

Lots of great information on nut/ saddle compensation here.

http://www.setitupbetter.com/

I've built a rig that I use for each guitar I build, to test the amount of nut and saddle compensation needed for the exact string type, scale length and playing action that the guitar will have (I build individually to order, and am still a relative beginner at the craft). I find that I usually need to set the nut forward by 2.5mm or so for one or both of the bass strings, and some string slots are filed back from this position as not all the strings need this much nut compensation. I have also noted a definite impact on the saddle compensation: less is needed than with no nut compensation. I have not tried to analyze this properly, but typically saddle compensation seems to be only about half of what I would use without compensating the nut.