Official Luthiers Forum!

What are your favorite methods of intonating a bone saddle? I have a new guitar to setup, so I thought I'd see what you all like to do.

I've had good success with the method outlined in Trevor Gore's book. He slides a B string over the flat top of a saddle blank until the open string and 12 harmonic match pitch. Then he marks the saddle with a pencil to note where the string was so he can file it appropriately.

Your method of marking and filing the saddle to the appropriate point is what I'm most interested in.

I use Trevor's method and it works well for me too. Looked at the StewMac intonator but decided I would just stay with what I was doing.

_________________
Steve Smith
"Music is what feelings sound like"

James: I use the same method that Trevor recommends but got it quite a while ago in the Don Teeter books. I use a sharp lead pencil and mark on the saddle either side of the string plus front and back of the B string hockey stick. After doing all strings I work the saddle so the peak is at the center of the fore and aft lines. I try to blend the space between contact areas to make things look as pleasing as possible
Tom

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

Yeah --- Don Teeter published his saddle marking method back in 1974 -- perhaps revolutionary at the time, for sure copied and practiced by many of us today. There are threads of Don Teeter's genius throughout our craft.

_________________
Ken Cierp

http://www.kennethmichaelguitars.com/

To set compensation, wouldn't you use open and 12th fretted to match? I'm not being argumentative, just missing something, as it would seem that open and 12th harmonic would match with zero compensation. Maybe I'm not interpreting what you said correctly,
Mike

Yeah forget about the 12th harmonic just open and 12th fretted

What Tom Said. Some times I don't even bother. I find that taking 1MM off the top end of the fret board gets things just so darn close.

I've been adhering to the Teeter method since the late 70's.

_________________
"Act your age, not your shoe size" - Prince

One of the simplest ways I find is to use a dummy saddle during initial setup to reference against. I keep a drawer full of saddles of various thickness and radii, beveled fully toward the fromt if made specifically as a test saddle, though a shaped and compensated saddle can work fine as well. Using shims to simulate final height, you can check the intonation with the dummy saddle and determine how far you wish to change compensation with some simple math.

For most common guitar scale lengths, the constant averages around .007" for every cent you wish to affect change. If a B string is 8¢ sharp with a full forward bevel test saddle, then shape the replacement around .056" back. If the low E is 5¢ flat with a prior generic compensated saddle, shape the new one about .035" further forward than the test saddle.

You can get more mathematically precise by using the actual formulas with 2^(1/1200) and scale length, but in terms of realizable benefit, the .007"/¢ constant works quite reliably.

_________________
Eschew obfuscation, espouse elucidation.

The actual formulas? Sorry if everybody knows this already, but I'm not sure what they are.

And, your examples seem to indicate that the simple version of the formula is .007 x ¢, and not .007/¢?

Not trying to pick, just understand.

Thanks

I read that as moving the breaking point of the saddle .007" equals 1¢ in change of pitch. So the / doesn't mean "divided" rather "per" in this situation.

Thanks, guys! Very helpful. I really like the formula, David. Seems like a great constant.

D'oh!

"/" = "per" not "divided by"

Yes, of course. I misread that.

Yeah, sorry about that - "/" was intended as "per" rather than "over".

As to the formula - one equal tempered semitone is based on the 12th root of 2. One cent is one hundredth of a semitone, and therefore based on the 1200th root of 2.

So say we have a 25" scale, tune the open string true, and the 12th fret reads 5¢ sharp. To lower that 12th fret by 5 cents, we need to multiply the speaking length of that note (let's say it's 12.6" in this case, or half the scale length plus .100" existing compensation) by the 1200th root of 2 to the 5th power (5¢) for our new length. Subtract the original tested speaking length from this result, and you arrive at the amount of change you will have to move the saddle back. If it were flat then we would divide by the 1200th root of 2 to the 5th power rather than multiply.

So ((2^(1/1200)^¢ * speaking length) - speaking length = amount of change from tested position to flatten. To sharpen, replace the * with a / (actually meaning divide here ).

On a 24" scale this comes to seven thousandths per cent at the 12th, or on a 26" scale closer to seven and a half. How much better will perfect math get you? Well let's take an extreme example and say you had a 25.5" scale with the low E testing 15 cents sharp. Use our approximate .007" constant and we end up moving it .105" back. Do the actual formula full out and it turns out we should actually move back .111" - which amounts to just under 1¢ difference at the far ranges of error. So in the range we're typically dealing with it is perfectly suitable to use a simple rounded constant of .007" per cent, flattening or sharpening, and not worry about the math.

There are lots of perfectly fine ways to do this, some working better than others depending on how you're set up and what you're used to. For me, it doesn't get much simpler than taking a tuner reading of how far I'm off, then multiplying that by 7 to tell me how far to move the saddle peak in thousandths of an inch.

_________________
Eschew obfuscation, espouse elucidation.

Very useful thread. Let me reiterate: Very. Useful. Thread.

_________________
George :-)

I muffed it! It's harmonic at the 12th, fretted at the 12th.

My roommate, who can do math, was kind enough to write out calculator instructions on how to perform the equation David Collins speaks about.

As in, type this number, press the button that looks like this, add it to this other number.

By some small coincidence, I ran into a hard print copy this morning of exactly that formula.

I will post a picture of the paper as clear as I can.

Practically, after using the formula ad nauseum, I found that using the same saddle shape, located in the bridge the same way every time, yielded accurate results that did not require individual guitar to guitar intonation.

Harmonic at the 12th may not always be EXACTLY the octave of the open string due to string inharmonicity
Best practice is to use open and fretted 12th with an accurate tuner.
Harmonic and fretted 12th is what we used to do by ear before good tuners.

You'd want to know which partial your tuner was triggering on, though!

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

http://www.goreguitars.com.au

Yep, once it's right, it's generally stays that way for similar guitars.

Whilst we're at it, no one has mentioned Frank Ford's simple method yet.

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

http://www.goreguitars.com.au

Frank Ford method is giving about double as Denis calculation. Or I am missing something?

Good catch - tell you what, put my advice on hold for the moment. To be honest it's been years since I thought about this in any depth or crunched numbers manually, as I have a spreadsheet I made a long time ago where now I just punch in numbers on my bench computer and it gives me accurate results. I was spewing out that info above from memory, and am away from my computer now, but let me think it over and model it later and I'll get back with you. It's quite possible I gave some bad numbers there.

_________________
Eschew obfuscation, espouse elucidation.

Going from memory and not doing the detail checking (always dangerous, but I'm pushed for time at the minute) Frank's argument goes something like this:

1) Tune the string
2) Measure the cents error at fret 12. Assume for the moment it is sharp.

Then, if you just compensated for that error by adding extra length to correct the fretted note, the open string would now be flat because of the extra length. So the string is re-tuned, which leaves the fretted note sharp again. To get it right, both the open string and the fretted note need to be in tune which means adding twice the compensation that was first thought of.

Clear as mud?

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

http://www.goreguitars.com.au

Yes, Ed's and David's formulas appear (to me) to basically agree with each other, but don't agree with Frank Ford's formula.

As far as I can tell.

(As a side note, Frank gives the example of a scale length of 25.25, then a first fret distance of 1.43, but 1.43 the first fret distance for a 25.5 scale. The 25.25 is obviously a typo.)

So anyway, Ed's/David's formulas for 8 cents over on 25.5 = 0.059.

Frank's formula for 8 cents over on 25.5 = 0.114.

So, I'm not sure what's what.

Which is not unusual for me.