empirical compensation

The distance from the nut (or zero fret in my builds) is 'almost' the scale length. It is actually a hair longer, an additional distance that is called "compensation". The compensation is there because as strings are fretted they are stretched slightly which makes the fretted note a bit sharp, so the bridge is moved back a bit so that the note sounds in tune. The amount of compensation depends on a number of factors, including: the particular brand of strings, the diameter of the strings, the distance of the strings above the fret (the 'action'), and the scale length. Given all of this it is is very difficult, if not impossible, to predict the correct compensation for a instrument in advance.

One of the instruments I am building in this round is a guitalele with a 19 inch scale length. Being a small guitar it is getting strung with classical guitar strings. Now, what the heck sort of compensation should I build into the bridge?

To answer this question (I have had the same question on other instruments in the past) I built a simple setup that allows me to set the bridge position exactly, getting the compensation as an empirical process. Since strings are generally shipped pretty long, and in this case are very long for the 19" scale instrument, I made an 'end block' to hold one end of the string and a block of tuners to hold the other end of the string. The string itself is stretched over the instrument, resting on the 'bridge' and the zero fret, as would a normal string. However, because the string is just resting on the instrument, there is no string tension force on the instrument itself and I can move the bridge around. With a small chromatic tuner resting on the instrument I can set the bridge position so that the note fretted at the 12'th fret is exactly an octave above the open note. I can do this with the brand of strings that will eventually be put on the instrument.

I use small 'chips' of bridge material sitting in the saddle slot so that I can easily set the action to a reasonable height (action goes into compensation) and there is room next to the chip to drill small positioning holes for nails down through the saddle slot so I can put the bridge on in just the right place.

I may be accused of quibbling (who, ME?) but I would assert that the distance from the nut (or zero fret) to the saddle IS the scale length. Try deciding on the scale length to which you will build, then work out the fret positions according to the "Rule of Eighteen". That is, use 18 as your divisor in arriving at the fret positions.

It worked for centuries, and then someone decided that 17.81715 (or whatever) was more accurate, so long as you then extend the scale length to provide compensation - which is not required if you use 18.

John Colter. (Climbs down off hobby-horse)

its always 2-3mm (tenor ukes of 17" scale are 2mm, guitars with a 25.5" are 3mm -[That is, the top E string])

Jupiteruke, I did pretty much the same thing on the ukes I built and it worked well. I used the finished, compensated saddle for the test, too. My tenor ended up needing more than 3 mm, for some reason, but intonation is right on.

To add to the picture/confusion, the scale length and the particular set of strings seem to make a rather large difference. On the guitalele, scale length 19", strung with Thomastik-Infeld classical guitar strings there was a large difference in the compensation needed between the 1'st string (E) and the 6'th string (low E). One typically sees the 'crooked' saddle on steel string guitars, and the same effect is required here. I am also building a parlor guitar, scale length 20.75 inches, strung with Savarez New Cristal high tension classical guitar strings, and there is almost no difference in the compensation required between the 1'st and 6'th strings. A straight across classical saddle works just fine. A very large difference between the two environments, so much so that I re-ran the test 3 times on each instrument, always with the same result. Just like the building of instruments in general, what matters? Everything.

jupiteruke said:

On the guitalele, scale length 19", strung with Thomastik-Infeld classical guitar strings there was a large difference in the compensation needed between the 1'st string (E) and the 6'th string (low E). One typically sees the 'crooked' saddle on steel string guitars, and the same effect is required here. I am also building a parlor guitar, scale length 20.75 inches, strung with Savarez New Cristal high tension classical guitar strings, and there is almost no difference in the compensation required between the 1'st and 6'th strings. A straight across classical saddle works just fine.

Click to expand...

That is a surprising result. I guess we better not decide to try other strings later.

Your problem string based.
You need to use the same strings for such tests. Jumping from Thomastik and Savarez certainly will yield different results.

Beau Hannam Ukuleles said:

Your problem string based.
You need to use the same strings for such tests. Jumping from Thomastik and Savarez certainly will yield different results.

Click to expand...

Isn't that exactly the point? Beau, perhaps you missed that these are different instruments. The 19" with Thomastik strings and a lot of compensation is my commission. I have these strings on a factory Cordoba with a straight saddle, and I can attest it does indeed have some intonation issues. Not terrible but noticeable. I anticipate Jon's build will be better in every respect, with intonation being but one aspect.

My question is what happens if I decide to put different strings on it later? From this discussion, I would expect the Savarez set to intonate "poorly" on my instrument after it has been setup for the Thomastik, but how "poorly"? Probably "not terrible but noticeable" in the opposite direction. I guess that's just the nature of the beast.

Opps- i must have missed the original question.
If a bridge/saddle/intonation is place to accommodate a particular type of string, then it can certainly happen that the intonation is out if you change string brand/type/tension/etc, or change the action.
Some strings are strange.
It shouldn't be radical difference- when i set up an instrument, I get the intonation to within 5 cents.