Fret Layout Accuracy?

[Will Reyer]

So, here's my question first, then the longer babble:

How accurately do frets need to be placed to expect an acoustic guitar to play in tune? What sort of allowable tolerance in positioning is possible and still make it work?

I make fingerboard blanks just over a quarter inch thick, with parallel sides just over 2-1/2”, then slot them on a table saw sled using an aluminum template notched for frets at a 25.4” scale.

I went to a machinist's supply and bought a metal-slitting saw .032” thick and 5” in diameter, for $35. I had my machinist turn a bushing to adapt the 1” arbor to the required 5/8” needed on my 10” table saw. Fret-slotting blades from luthier's supplys are close to 3 times more expensive. The 5” diameter blade protrudes more than the necessary distance above my Powermatic's table. These blades come in various slotting widths. I could have bought one .023” thick if I wanted to pound in frets, but I do Don Teeter's method pushing them into epoxy with my fingers.

I had enough fret wire from a local guitar shop, now defunct, to make the first two guitars, and the barbs on the tangs were .032” wide. Then I ran out of wire and bought a pound of Stew Mac #1048 as the dimensions appeared the same, except the tang barbs on the Stew Mac measure .029”.

This means that if I put fret tangs .029” wide in a fret slot .032” wide that they can be mislocated by
.0015”, one and a half thousandths. I don't lose any sleep whatsoever over this. My #4 guitar, built like this, intonates just fine.

The end of the aluminum guide is not entirely square with the edge, making the spacing for the first fret from the nut problematic. I cut the end purposely long, and disc sand down to dimension after slotting. This is hard to hit exactly. With a 25.4” scale, the center of the first fret should be 1.426” from the close edge of the nut. I stick a .032 feeler gauge in the first fret slot and caliper from the nut end of the fingerboard to the far side of the feeler gauge, making the required distance 1.426” plus half of the slot width, .016”, yielding a caliper dimension of 1.442”.

How far can I be off from this dimension without having intonation problems?

[ken cierp]

but I do Don Teeter's method pushing them into epoxy with my fingers.

Really? As compared to tapping them in with a hammer Don's bound finger board re-fret method is an incredibly messy time consuming process.

The high level of accuracy regarding intonation and play-ability is what makes a hand made guitar a musical tool.

[John Parchem]

I read about Don Teeter's method but never tried it. If you have a saw sharpener service around you you can have them reset the kerf to where it can be pushed in but still center correctly.

[Will Reyer]

I guess either my communication or my comprehension is going the way of my hearing and my hairline. I'll try again:

How far from the mathematically calculated, three decimal places positions for fret locations can the frets actually be mislocated before it appreciably affects the perception of pitch in fretted notes?

I don't care about Teeter's method for bound fingerboards, and I could just as well make narrower slots and hammer in frets.

What I'm doing now, using tang barbs .029" wide in .032" slots means that each individual fret could be out of position by .0015". The worst case scenario then, engineering-wise would be to have fret one .0015" out of the mathematical location toward the nut and then fret two be .0015" out of position toward the bridge, making the spacing between them the .003" max mislocation. Except that each fret would still be only .0015" out of position, playing notes fretted there.

I maintain that I could probably not hear the difference between a fret at the correct location or out of position by .0015". I've seen photos of people doing fret layouts using steel rules. If a fret was misplaced by 1/32" would it sound sharp or flat? Is 1/64" close enough to not notice misplaced frets? Does it require accuracy to thousandths of an inch? If so, how many?

I'm more concerned about sanding the nut end of the fret board down to provide the proper distance from the nut to the center of the first fret. I assume if frets could be misplaced by .0015" that the distance between the center of fret one and the near side of the nut could be off by that much. How much more could it be off, being too short or too long?

[John Parchem]

My guess is that people make bigger errors than 1.5 thousands when re-crowning frets.

[Will Reyer]

O.K., so for guitar #5 I bought a new machinist's slitting saw with an .023" kerf. I looked on-line, actually they were maybe $22 each but the minimum order was for $25 so I just bought two. Still lots cheaper than one from either luthier's supply. These last two were Czech-made, work just fine.

Also borrowed Kjell's fret radiusing tool and rolled out 16" radiused fret wire sufficient to utilize on #5, completed, and #6 still in progress. Tapped in frets for #5 with a little plastic faced hammer not used for other purposes. Worked just fine, and simpler for radiused fretboards than my previous flat ones that utilized the push-into-epoxy-with-fingers Teeter method.

Also had my machinist friend make me a new notched piece of aluminum to use with my tablesaw sled for kerfing fretboards, 25.4" scale, but this one has a slot for a zero fret also, something the last one lacked, and I'm going to try the zero fret method for #6.

But I'm back to my original question: How misplaced can a fret be and still play in tune? Or, how far off could the first fret be from the nut and still sound right?

Several books show people laying out fret spacings with steel rules calibrated in sixty-fourths. One sixteenth of an inch is decimal .0625". I used to do mechanical design; you can probably ignore the fourth decimal place, a ten-thousandth of an inch is pretty small. That makes a thirty-second of an inch equal .031" and a sixty-fourth equal .015".

It would seem then that if you can lay out fret spacings with a scale calibrated in sixty-fourths, that fret location within .015" is good enough. Except I just bought from Elderly the Antes plan for a tenor ukulele, and his instructions say "The careful builder will hold fret spacing closer than plus/minus .003".

So ignoring compensation at the saddle for now, I guess I'm back at my original question, how close do you have to locate frets and still be playable in tune? What's your method, what tolerances to you try to hold?

[Dave Bagwill]

I bought a template to use when I'm slotting my own - from KMG, I'm sure others sell them too. That's been close enough that no human ear has been able to notice anything 'off', though the dog next door howls when I play or sing.

To my utter, complete, and unredeemable shame, when I order FB material, I usually pay for the slotting service.