Even fretless fingerboard levelling tips

[togril]

Yes TLC1976 - I should have thought about it more before posting - that 'Zero Block' malarky is bang on. I will probably end up doing the same as I like a little height at the nut.
Bruce: How long is that zero block along Down the Fingerboard?
You crazy Americans and your crazy "thousands of an inch" measurements.

 

[megafiddle]

On a 7 1/4" radius fingerboard, the "hump" under the outboard strings is in the range of 0.010". On a 12" radius fingerboard, it's around 0.002-0.003". That doesn't sound like much, but it's enough to cause the buzz or Mwah to be uneven.

When you loosen the truss rod to add some relief, that somewhat flattens out the humps under the outboard strings. But then the center strings have more relief than the outboard strings.

In my experience, doing this "hourglass" trim is very important on a 7 1/4" (or smaller) radius fingerboard. Mildly important on a 9", noticeable on a 12", not really noticeable on 14" or flatter.

I have to respectfully disagree.

I am getting a value of only 0.002" for a standard P bass with 7.25 radius.

The calculations are quite simple. I take a cross sectional slice at three points on the neck: one at the nut, one at the last fret, and one at the fretboard midlength.

All three slices are centered on the fretboard axis. An imaginary line is then "drawn" between the two end slices. The line endpoints are situated on the fretboard right below one of the the outer strings (eg, the E string). This line represents the ideal straight fretpath beneath the string.

It is then a simple matter to calculate the midpoint of that line in 3D space - it is the mean of the two endpoints.

Finally, the distance of this midpoint from the axis is calculated, and it's difference from the radius is the size of the error (or hump).

This is the input and output of my program:


number frets ? 20

scale length ? 34.0

radius ? 7.25

string spacing at nut ? 1.25

string spacing at bridge ? 2.25


fret 20 position 23.29

midlength position 11.65

center line to string at nut 0.63

center line to string at midlength 0.80

center line to string at fret 20 0.97

fretboard height at nut 0.0000

fretboard height at midpoint 0.0020

fretboard height at fret 20 0.0000


The actual code:

//	cylerror.c    v1.0
//	calculates midlength fretpath error based on radius and string skew
//	megafiddle    May 2017

#include <stdio.h>
#include <math.h>

#define TEMPER 1.05946

int main()
{
	int nfrets;
	float scalelength, radius, nutspacing, bridgespacing;

	double skew, v1, x1, y1, z1;
	double v2, x2, y2, z2;
	double v3, x3, y3, z3;

	printf("\n\n number frets ?  ");
	scanf("%2d", &nfrets);

	printf("\n scale length ?  ");
	scanf("%6f", &scalelength);

	printf("\n radius ?  ");
	scanf("%6f", &radius);

	printf("\n string spacing at nut ?  ");
	scanf("%6f", &nutspacing);

	printf("\n string spacing at bridge ?  ");
	scanf("%6f", &bridgespacing);

	z1 = 0.0;
	z3 = scalelength - scalelength / pow(TEMPER, (float)nfrets);
	z2 = (z3 - z1) / 2;

	skew = (bridgespacing - nutspacing) / 2;
	x1 = nutspacing / 2;
	x3 = skew * (z3 / scalelength) + (nutspacing / 2);
	x2 = (x1 + x3) / 2;

	v1 = radius;
	y1 = sqrt(v1 * v1 - x1 * x1);

	v3 = radius;
	y3 = sqrt(v3 * v3 - x3 * x3);

	y2 = (y1 + y3) / 2;
	v2 = sqrt(x2 * x2 + y2 * y2);

	printf("\n\n fret %2d position\t\t\t %4.2f", nfrets, z3);
	printf("\n\n midlength position\t\t\t %4.2f", z2);
	printf("\n\n center line to string at nut\t\t %4.2f", x1);
	printf("\n\n center line to string at midlength\t %4.2f", x2);
	printf("\n\n center line to string at fret %2d\t %4.2f", nfrets, x3);
	printf("\n\n fretboard height at nut\t\t %6.4f", radius - v1);
	printf("\n\n fretboard height at midpoint\t\t %6.4f", radius - v2);
	printf("\n\n fretboard height at fret %2d\t\t %6.4f", nfrets, radius - v3);
	printf("\n\n\n");

	return 0;
}

 

[Bruce Johnson]

Hmmm....Well, the numbers I used aren't based on calculation, they are from experience directly testing with a straightedge. Take a bass fingerboard that's a straight cylindrical cut, 7 1/4" radius, adjusted dead flat down the center by straightedge. Now, move the straightedge off center and angled to match the outboard string path. The rocking of the straightedge is a lot more than 0.002". It's clearly visible, more like 0.020-0.030" at one end, which means that the hump is 0.010"-0.015". And I know from experience regularly surfacing 7 1/4" fretless fingerboards, that I have to file/sand off around 0.010" of hump to get the outboard string paths reading straight with a straightedge.

 

[Bruce Johnson]

Yes TLC1976 - I should have thought about it more before posting - that 'Zero Block' malarky is bang on. I will probably end up doing the same as I like a little height at the nut.
Bruce: How long is that zero block along Down the Fingerboard?
You crazy Americans and your crazy "thousands of an inch" measurements.

The zero blocks on my basses are 1/4" (okay, 6.35mm) thick or "long" along the neck centerline. The top surface is ramped so that it forms a nice edge where it meets the fingerboard. I use an aluminum nut/string spacer, and the slots in it are deep enough to ensure that the strings rest fully on the zero block.

I normally trim the zero block to 0.015"-0.020" (0.38-0.51mm) above the fingerboard surface. From experience, that's the best combination to make the open strings sound the closest to the fingered strings. Pluck the open string harder, and it will buzz. Trim it lower than that, and I find that it's too buzzy. My customers all seem to like this setting too.

We Americans like to work in thousandths of an inch, because there are more digits, and it makes the measurements look more precise.

 

[megafiddle]

Hmmm....Well, the numbers I used aren't based on calculation, they are from experience directly testing with a straightedge. Take a bass fingerboard that's a straight cylindrical cut, 7 1/4" radius, adjusted dead flat down the center by straightedge. Now, move the straightedge off center and angled to match the outboard string path. The rocking of the straightedge is a lot more than 0.002". It's clearly visible, more like 0.020-0.030" at one end, which means that the hump is 0.010"-0.015". And I know from experience regularly surfacing 7 1/4" fretless fingerboards, that I have to file/sand off around 0.010" of hump to get the outboard string paths reading straight with a straightedge.

I don't doubt your measurements at all; I'm sure they are correct.

Even 10 or 15 thousandths would still be incredibly excellant work for a piece of wood that long. A StewMac precision straightedge can be off by as much as 0.003" over the length of a 20 fret bass neck.

I'm not saying that the straightedge is a source of any error in the measurement, only that straightness is not easy to achieve, even on a precision steel tool. As long as the same straightedge is used for all of the measurements, the results are comparative and should be accurate.

To actually make the cylindrical error measurement, it would be better perhaps, to compare the surface along the centerline to a skewed path also at the centerline. In other words, the center of the straightedge would be right on the centerline, the nut end would be displaced to the side towards the G string, and the heel end would be displaced towards the E string. This needs to be done at the center, since you can't get a reference measurement parallel to the centerline near the fretboard edge. The straightedge would fall off the fretboard at the nut end.

Also, rocking the straightedge so that all the gap is at one end, can be very imprecise due to the circular surface that it is rocking over. The gap can easily be much more than twice the hump height.

It just wouldn't be surprising to me if the 0.010" was there for some reason other than cylinder geometry.

-

 

[pilotjones]

A StewMac precision straightedge can be off by as much as 0.003" over the length of a 20 fret bass neck.

If you are correct about this, I find it highly disturbing. That's not a straightedge, it's a little more than a machined piece of bar stock. This is a straightedge:
[Invalid or Expired Link Removed]

Accuracy (in): ± .0002” per foot

 

[Gilmourisgod]

It's hard to get worked up about .0001" variations on wood that's going to move in the 1/128" range anyway based on humidity. Even if you machine wood to that level of accuracy, it won't stay that way.

 

[pilotjones]

It's hard to get worked up about .0001" variations on wood that's going to move in the 1/128" range anyway based on humidity. Even if you machine wood to that level of accuracy, it won't stay that way.

I disagree, particularly with respect to a fretless, which is not going to receive a fret leveling. .007" has got to matter with respect to the "mwah" response of notes up and down the fingerboard, just as .007" difference in fret heights would be huge on a fretted. IEven if your relief may change even twice that much with the seasons, that 's what the truss rod is for, to put it back where it should be, evenly across the length. But the profile has to be right to begin with.

 

[Gilmourisgod]

The trussrod will only change the relief on the neck, is has zero effect on seasonal expansion or contraction of the wood fibers. Initial Accuracy is a worthy goal, but a Starret straightedge is an unnecessary level of accuracy for wood that will move more than the tiny variation between a $200 straightedge vs. A $30 machined aluminum level. If you can afford the Starrett, and have other uses for it that justify the cost, have at it, but the neck is going move out of that level of accuracy at the first change of season.

 

[pilotjones]

Hey Gilmour, help me to understand here.

The trussrod will only change the relief on the neck, is has zero effect on seasonal expansion or contraction of the wood fibers.

What do you mean here? The way the seasonal effect on the wood manifests itself /is/ by change in the relief of the neck, which is exactly what the truss rod works on. How else do you see seasonal changes having an effect? They should not be causing any other kind of bizarre small-area changes that the rod could not adjust away. I mean there's fret sprout, but that's really unrelated to the conversation here.

Initial Accuracy is a worthy goal, but a Starret straightedge is an unnecessary level of accuracy for wood that will move more than the tiny variation between a $200 straightedge vs. A $30 machined aluminum level. If you can afford the Starrett, and have other uses for it that justify the cost, have at it, but the neck is going move out of that level of accuracy at the first change of season.

The Starrett is not cheap, but the 24" sells for $90 off their site - much cheaper than I had expected.

Now I just looked up Stewmac's straightedge (the un-notched one), and it says they are +/-.0015" per foot. While a +/- statement of accuracy is really not appropriate for straightness, for reasons that are a little beyond the scope of this forum, this spec means that there is an allowable variation of .003" variation per foot, or .006" in a 24" straightedge. This is really what I would call a sad piece of hardware, as any decent machinist on a Bridgeport should be able to beat that easily, maybe in the range of .002-.003". Then they take it over to the grinder, and take it down to the range of the Starrett offering. And the Stewmac sells for $75.64! For less than $15 more, you can have a Starrett with a max total variation of .0008".

 

[Five String]

Sub'd. This is a great thread for learning how to level a fretless neck and all the variables pros and cons. I just did it myself and I was very pleased with the results.

 

[Thegrandwazoo]

Some people want the open string to sound the same as a fingered position string. This is why the crazy low, low enough for the open string to mwah against the finger board at the nut end.
I'm about to clean up my VM jazz fretless - buzzing from the markers not being flush. So a great thread.

If it's just the markers, I bet you could just radius sand with a high (like 600-800) grit paper and knock them down without removing an important amount of wood, right?

 

[Thegrandwazoo]

Your numbers sound good to me. On my fretless basses, I use a little ebony "zero block", sort of the equivalent of a zero fret. I trim the height of the zero block to 0.015" to 0.020" above the fingerboard surface. That seems about right to me, in the amount of force it takes to push the string down at the 1st position. Lower than that, and the touch is too light. I like to have some small bit of feel of when I'm pushing the string down at the 1st.

As do I, tell me more about this 'zero block'. Is it simply a tapered, radiused slice of wood glued to the fingerboard? Oh, wait... The bass would have to be built to accommodate a zero 'fret', wouldn't it? Hmm... I'm thinking about whether I have the skills to fill my nut slot and move it back a fret's width.

 

[Thegrandwazoo]

Yes TLC1976 - I should have thought about it more before posting - that 'Zero Block' malarky is bang on. I will probably end up doing the same as I like a little height at the nut.
Bruce: How long is that zero block along Down the Fingerboard?
You crazy Americans and your crazy "thousands of an inch" measurements.

As opposed to what, hundredths of a millimeter? I'm not being a smartass, I'm actually curious. I agree wholeheartedly that SAE measurements are batscat ridiculous, but I know them and they make sense to me, so I'm ruined. I studied to be a machinist before I wound up an electrician, that's even weirder. 'Half a tenth' means .00005", and when surveying or operating earthmoving equipment here, a 'tenth' is a tenth of a foot.

 

[Thegrandwazoo]

Is there any way that, when I level my fingerboard, having the nut still in place could affect things? I'm sure it won't be perfect right up to said nut, but I'm of the opinion that it would surely resolve itself by the first-fret position, right? My problem with removing the nut on my Fender is that I kinda don't have the tools to make a new nut should I tear it up getting it out. The other style, like on Gibsons and stuff with angled headstocks, I can pretty easily knock off and put back, but as you know, Fender-style instruments are different and I've not messed with one of them. Should there be a considerable 'ramp' at the nut (I don't think there will be much), I could always cut a few more thousandths in the nut slots, could I not?

 

[Thegrandwazoo]

I'm looking at you, @Bruce Johnson.

 

[Turnaround]

There have been some wonderfully detailed explanations here, and some of them are complex enough that it's a little difficult to get your head around it all.

But it's not that simple. There are a few more things that go into the equation. First let me say that what I am about to demonstrate is not really practical, but it does explain some of the complexities involved. The first is that in order to have a equal distance from the strings to the surface of the fretboard, the plane of the strings cannot be on the same radius as the surface of the board, that is to say if your fingerboard radius is 16 inches, your strings will not be on a 16 inch radius . It's easiest to see this with the radius curves exaggerated, so we get this:

View attachment 2815959
Curves A and B have the same radius, and the distance between each pair of coloured dots is the same. But we measure the distance from string to fingerboard on a line perpendicular to the fretboard, like the distance from the yellow dot to the A curve at point x. You can see that it is less than half of the distance from A to B. In order to maintain a constant distance from the curve B to the curve A we need to have two different radii, like this:

View attachment 2815979

So if the distance between curve A and curve B is n, then the radius of curve B is radiusA+n.

As discussed earlier, since the strings are placed narrower at the nut than at the bridge, and are arched from side to side, the string plane formed is the surface of a cone. And we can create a matching cone-shaped surface at the fingerboard to provide equal distances from string to fingerboard. But we actually don't want that. We want the strings to be closer to the fingerboard at the nut than at the other end of the fingerboard. We can do that and maintain string-to-fingerboard distances at each position in two ways; by the hourglass manipulation described by Bruce, or by setting the parameters of our sting plane radii to radiusA+n along the entire length of the fingerboard, where n will vary depending on the distance from the nut. If we don't vary the "n" part of the radius we end up with the opposite problem at the nut than the first curve shown above, like this:

View attachment 2816070
Now the blue and yellow points are much further apart than the red and green points. So we must vary the factor "n" along the length of the conical surface. This sounds more difficult than it actually is so long as we maintain a constant taper in the conical structure.

OK, now we've got that sorted, all should be good. Well, not yet. We don't actually want each string to be the same distance from the fingerboard, we want the heavier strings to have a little more clearance since they vibrate in a larger arc. But before we get carried away with the geometry of this skewing, we have bigger problems to address. So far we have been talking about static geometry. But once we pluck a string we are introducing movement into the equation and that means reworking things for a dynamic environment.
String vibration is complex with waves moving side to side, up and down and along the length of the string. And we are trying to adjust for these complex movements with a fixed fingerboard with a conical surface (or modified conical surface as described by Bruce Johnson) that is skewed in relation to the conical plane of the strings in stasis. Or are we? The fingerboard is vibrating too. You can feel the vibration. Vibration is movement, so we should account for that in our dynamic model to achieve an ideal fingerboard geometry.

This all gets past the realm of practicality pretty quickly. Even if we could work out all of this stuff to the finest degree and were able to manufacture such a fingerboard, we would lose most of the finer points if we made the board from wood. Wood is itself not a static material - it responds to changes in humidity by changing its dimensions. In practical terms we are best to pursue a course as Bruce outlined of creating a longitudinal flat surface along each string path, then adding a bit of relief, usually by string tension adding deflection to the fingerboard under the control of the truss rod. And with careful manufacturing, that bit of relief will be close to the precise curve we would like to optimize dynamic string clearance. Of course that's only possible if we are using high quality wood with even density throughout and straight growth patterns, cut to minimize short grain runout which will affect stiffness, the curve of flexure under string tension, etc. Such wood tends to be expensive. Such care in manufacture tends to get expensive. It's not just about having the parameters worked out, it's about the manufacturing complexity, the quality of materials used, the cost effectiveness and the compromises one is willing to accept.

 

[Gilmourisgod]

Dang! I’m still trying to learn how to make a consistently straight, single radius board! Cones beyond me at this point, but it’s a good argument for a radius jig. Gonna have to read this one a few times to digest

 

[rwkeating]

This information will be very useful and inspiring as I continue to make flat (non-radiused) fingerboards . Just joking. Thanks for the information!

 

[Turnaround]

This information will be very useful and inspiring as I continue to make flat (non-radiused) fingerboards . Just joking. Thanks for the information!

You are welcome. And please feel free to forget everything I said so long as you continue to make flat boards. To be fair, I was pushing the point anyway with my examples. There are two ways (perhaps more) to avoid all of the problems I outlined. First is to have a flat board. The second is to have a board with the spacing at the nut and the bridge the same. The further you deviate from those two conditions, the more the parameters I outlined become relevant.

Yah, I know you were just joking. But your joke is relevant.