Official Luthiers Forum!

Just sharing something that might be obvious to the more seasoned folks, but could be helpful to folks at my end of the experience spectrum (the hack hobbyist crowd).

When cutting rosette channels for a normal steel string thin-wide-thin pattern, I have always used the compass method and drawn out the circles before drilling the hole for my router jig's center pin. Then I would line the router bit up with the lines. It worked, but there was always the opportunity for inaccuracy. Pencil lines are a bit thick for this application, and the untouched spruce between the channels can look too thin and uneven (i.e., one is wider than the other) if you are just a little bit off.

It occurred to me that, after I locate and drill for my center pin, I don't have to rely on pencil marks at all. I can unplug the router, turn it over, expose the bit, and measure the distance between the inside edge of the center pin and the inside edge of the router bit with a dial caliper.

Example for a normal thin-wide-thin pattern, assuming a 4" soundhole, a 1/4" center pin and a 1/8" end bill router bit:

The center pin is 1/4" in diameter, so the distance between the center of the pin (which is the center of every circle at issue here) and the inside edge of the pin is 1/8". The inside edge of the router bit, and the outside edge, are obviously 1/8" apart.

So, if I want a 1/8" wide channel that is 1/8" outside the soundhole, I need to measure 2" from the inside edge of the center pin to the inside edge of the bit, lock it in, and cut. If I want another 1/4" wide channel 1/8" outside of the first channel, I need to measure 2.25" from the inside edge of the center pin to the inside edge of the bit, lock it in, and cut, then measure 2.375" from the inside edge of the center pin to the inside edge of the bit, lock it in, and cut again. If I want another 1/8" channel outside of the wide channel, I need to measure 2.625" from the inside edge of the center pin to the inside edge of the bit, lock it in, and cut. Finally, to cut the soundhole, I need to measure 1.75" from the inside edge of the center pin to the inside edge of the bit, extend the bit below the thickness of the soundboard, lock it in, and cut. This gives me channels that are spaced 1/8" apart with a level of precision I could not accomplish by just eyeballing pencil lines.

It is super important to plot this out and write the distances on paper before turning on the router.

Hope this is useful.

My only suggestion to this good advice is to use the compass in addition as what I call a sanity check. When you use your method to set the center pin to bit distance and then get ready to use it, you can check against your pencil line to make sure you didn't read a digit wrong. If it is close to the line you did it right, but if it is noticeably off, it's time to recheck. I use this philosophy in nearly all measuring, because it really sucks to do something and find you are off by -exactly- an inch or mm or whatever,
Mike

Don
I have been doing a rosette this week, and used exactly the method you describe. Agree completely. Have you tried using a grammel to scribe the outer and inner edges of your channel? It helps to get a cleaner edge. I have found that the router blade is prone to tearing out a bit at the part of the circle where it is running with the grain (3 o'clock and 9 o'clock of the circle), causing fuzzy edges.

As I read your description, jumping from fractions of inches to decimals and back again, I shake my head and wonder how you Americans can possibly function in a non-metric world?
But each to their own I guess, if it gets you the same result.

Mark

Mark--

I have a gramil for cutting binding and purfling ledges, but not a circle cutter. I may look into buying or building one. If I build one, I could probably borrow the blade from the binding/purfling cutter. That's a good suggestion, as is Mike's good advice about guarding against a flub by still drawing the lines. I could probably set the blade distances the same way, with the dial caliper. Thanks!

On metric versus Imperial, I can't stick up for our continued use of Imperial measurements, but only apologize for the fact that we prefer them because we are comfortable with them. Later generations will criticize us for it, I am certain. Thank goodness we are measuring distance instead of weight or fluid volume. Stones and quarts? We might as well use Galleons, Sickles and Knuts for money.

I just used a dial caliper.

I've always used the caliper too as Don suggests. Gotta be careful when doing the math. I just work in thousandths.

As far as the imperial/metric thing goes. As an engineer I use them all. As a guitar maker I can give you the decimal equivalents for all of the common fractions down to 1/32" so no big deal - we just have to know more. Seriously, all metric would certainly be nice but not sure when that might happen over here.

_________________
Steve Smith
"Music is what feelings sound like"

As Steve mentioned its simply an added skill set -- anyone can learn to use either measuring system.

_________________
Ken Cierp

http://www.kennethmichaelguitars.com/

When going through end grain I discovered it really helps to make sure you are doing a climb cut on the outer edge which means rotating the router in different directions for the inner and outer border of the rosette.

Good tip, I'll try it as a backup to my usual method and maybe my usual method will become a backup to your method.

_________________
It's not what you don't know that hurts you, it's what you do know that's wrong.

Don I think that we could say that guitar building is avoiding doing anything stupid while waiting for that next ah ha moment...

One of my ah ha moments had to do with dish sanding the rim. I would install my reversed kerfed linings before going to the dish.... We can't write dumb arse on this forum can we... One day it occurred to me that if I dish sand prior to lining installation and then carefully install the linings only say 1/32" proud of the sanded rim the next session in the dish was only a few minutes instead of an hour or more... Duh.....

I do my initial measurement like Don describes in his first post, it's also a good idea to cut you channel into some mdf or something to check what your diameters are, as you want to be sure before you commit to a top.

_________________
Jim Watts
http://jameswattsguitars.com

Yup. No matter what process one may employ, testing on scrap is an invaluable step.

_________________
George :-)

_________________

For sure -- we now have the luxury of CNC, but when using a circle cutting router set up we had a test plate right next to the project sound board -- an over-sized rosette channel will make you very sad.

_________________
Ken Cierp

http://www.kennethmichaelguitars.com/

Yep, got a CNC machine that took over a year to build and hundreds of hours to learn. Now I can cut rosettes like I used to with a home made circle cutting jig! Progress


Sent from my iPhone using Tapatalk

"It is super important to plot this out and write the distances on paper before turning on the router."

It is also important to always measure to the right edge of the bit - another reason drawing a few lines on the top is not a bad idea.

I have a piece of ply, into which I've already cut the rosette channels. I set the router bit in those channels and I'm ready to go.

No measuring, except for the first one in the ply.

Steve

I have one of these, http://www.leevalley.com/en/Wood/page.aspx?p=69874&cat=1,250,43298,43314,69873 and i bought an extra item "B", from which I removed the center point and inserted a 1/8" dowel pin. I use this setup to pre-cut my rosette channels, then use my router with a Wells/Karol style jig. Pretty easy to measure the radius with calipers on the modified Compass Center and the Slicing Blade. You can also use the Groove Blades for cutting individual channels. Great tool.

Bob

Definitely good advice added regarding the need to test on something other than the top, and the convenience of creating a gauge out of scrap so you only have to measure once for a particular rosette style. In fact, I was already following these suggestions (in a way). My first test rosettes using the dial caliper were into UHMWPE molds, rather than tops. The molds let me build the rosette away from the guitar, glue the lines together in the molds, use the molds as gauges (I also measured with the caliper and drew lines with a compass--triple checked!), and thereby have an easier time when I glue the lines into the top.

A few things about the UHMWPE mold idea (thanks to posts from OLF member RNRoberts for the idea):

1/2" thick kitchen cutting boards are a great inexpensive source for the plastic.

You will see suggestions about forming rosettes in a mold, shooting them with CA such that it wicks down, then pop the rosette out of the mold after the glue sets. I did all of that, only with thinned hot hide glue instead of CA. It worked great! For general purpose hot hide glue, I normally use a water to granule ratio of 1.9 to 1 by weight, but to get it a bit thinner for good wicking, I went with about 2.1 to 1 for this task. After it dried, everything just came right out of the molds, easy as pie. I had to scrape a bit of the dried glue off the edges, but since it is fully dried hot hide glue, it is easy to scrape. A little touch up sanding to the rosette pieces and they are a perfect fit in the tops. Working with 1 piece per channel is a lot easier at glue time than separate multiple lines.

Sharp bits and a wash coat of shellac on the tops the day before help keep fuzz and tear-out at bay.

The logic behind my being more persnickety about these rosettes than I have been in the past: I made the lines and herringbone myself out of dyed veneers instead of using commercially available herringbone and lines. The dyed veneer is more fragile and prone to breaking than the fiber that goes into the commercial stuff. Bending lines on the hot pipe and constructing the rosettes outside the top seemed a safer way to go.

Lots of extra work, but worth it.