The Steel Guitar Forum

Larry,
That's why I thought this would be a good project for someone with a "hobby shop" as opposed to a full blown machine shop. If you knew exactly how to make each cut, it really wouldn't take all that long to make a few sets. I'd be glad to make $30 an hour to piddle around making rollers if I had the right equipment (which I don't).

Time for a bit of lateral thinking:
Put the roller in the lathe.
Turn a small groove.
Place a piece of guitar string of the desired gauge in the groove.
Take a measurement of the roller (with the guitar string in situ) with a micrometer.
Work out what the measurements are for all 10 or 12 rollers with their respective strings to get the string height consistent.
No trigonometry needed, just a micrometer.

To all, a very informative thread. To C Dixon you answered several of my questions in one post, particularly about V groove vs round rollers.
What it the prevalence of rounded rollers, and when was this change to V grooves brought on?

Also the adjustable height roller idea has strong merits. (as well as an increased price tag.) But this very well might help intonation issues in general and playing at 1s fret.

Mr. Donald, Mr. Dixon,

Respectfully, I am still wondering how the round bottom groove vs. V bottom groove would affect sustain...

When you have that steel down there touching them strings, and the fangers of that left hand resting on them strings down in there behind that steel, I am not understanding how that would affect sustain.

All of the information provided here has been very interesting.

------------------
"Stump" Reeves
<FONT SIZE=1 COLOR="#8e236b"><p align=CENTER>[This message was edited by David Reeves on 09 March 2003 at 06:11 AM.]</p></FONT>

Sir Richard,
You come through once again, my friend
That's the most elegantly simple solution yet proposed. I'm gonna try it!

------------------
<small>Larry Bell - email: larry@larrybell.org - gigs - Home Page
2003 Fessenden S/D-12 8x8, 2000 Fessenden S-12 8x8, 1969 Emmons S-12 6x6, 1971 Dobro, Standel and Peavey Amps

Thank you, Larry.
I've always thought that I deserve a knighthood!

David,

I did not say sustain. I said,

"A nut roller groove that allows the bottom of ths string to touch the groove will not give as brilliant and/or resonant a tone as a groove that only touches the string at two points." What I should have said is,

"A roller nut that allows the bottom AND sides of the string........" Sorry bout that. This is what Sierra does. And IMO has been poor design from day one.

This is not a new theory and predates PSG by dozens of years. Eddie Alkire new this eons ago. That is why on most of his guitars, the nut is a piece of aluminum that comes almost to a point at the top. Alkire guitars have incredible sound in most cases. And I attribute the shape of the nut and bridge for a major part of this.

The ideal situation is for the string to touch the nut (be it fixed or roller) at ONLY one point. This is generally impractical. And the ideal "point" causes a "screeching zzzzip" when tuning the wound strings.

The flatter the nut, the more loss of tone and resonance you get. The sharper the nut, you get the reverse.

When you use a roller, there is a poblem IF the roller has a groove in it. And most do. The two most prevalent choices are round bottom or V groove. Most use V groove I believe. This means two points of the string touch the V. Which is NOT as good as if only one point touched the string. But is the next best thing. Round bottom, where the string touches the sides AND the bottom are unacceptable, IMO.

Actually there is no real reason for the roller to be grooved at all; IF the pull is in a straight line parallel to the length of the guitar. There is no groove in most changers. So why have one at all in the nut. If this were the case, we are back to the "one point" scenario. Which would be better. (This provided the string does NOT tend to wrap around the roller such as strings 1 and 10 on an Emmons PSG). By the way, this is the major cause of players complaining that the 1st and 10th strings do not have that zing like the other strings do on an Emmons.

Also on a side note. Experience has proven that there is much more to making gauged rollers "absolutely flat" than what has been described in this thread. "The proof of the pudding is in the eating". And NO body to date (that I know of) makes gauged roller nuts that result in the strings being 100% flat across the tops of all strings.

The one person I thought that could and would do it, turned out to be a rip-off, sadly.

It appears very simple. But try it on a lathe. The results do not bear this out. The reason for this is, trig IS necessary and SO is a bit of calculus. The angle of the groove versus the cross section of the string gauge, plus the the radii of the roller cause a mathematical variable that is NOT easy to solve, WHEN you are talking about 10 or 12 different gauges.

If it was easy, I am certain the manufactuters would have long since made it work. The facts of the matter is this has NOT been the case. If you doubt this, simply call a builder, tell them what gauges you have and ask them to make you a set of gauged roller nuts. THEN, when you get them, see for yourself.

Remember, unless the top of ALL strings are dead level (at the nut), you ARE going to get string rattle with light bar pressure at the first fret. That bubba is the "eating"

I am sorry to say the results have been painfully sad. Thus the reason for this thread, and my (and others) 100% futility in trying to get a set that were completely satisfactory.

carl

Mr. Dixon,
God Bless Your Heart,

I went back and re-read your post... y'know, my wife always tells me that I don't listen... I guess that I don't read right neither....

------------------
"Stump" Reeves

Richard Burton,

I agree with the process you outlined for making a set of rollers "from scratch". That's basically what I did when I reworked the set on the Zum 2000. If memory serves me I only had to modify strings 7 thru 12. With the proper cutter it wouldn't be any big deal, really. And the right tool isn't hard to obtain, we just don't happen to have one, guess we should buy one, huh?? That will fix my problems, the rest of you guys are on your own !!! (ha ha)

I suppose the manufacturers don't fix this because it would involve having maybe 7 different rollers as opposed to 3, and on 10 strings it isn't as much of an issue. But I have yet to sit down at a 12 that didn't have this problem initially. It probably doesn't bother a lot of people, but it sure irritates me. I don't like to apply a lot of bar pressure. Of course we re-worked Jeff Newman's guitar last year and it wasn't perfect either, but doesn't seem to hurt his playing. Personally, I need all the help I can get.

Thanks to all for the good ideas and participation in this thread.


------------------

Doug Earnest
The only Zum Keyless U12, Fender Cyber Twin

Mr.Dixon,
The keyless Zum pulls the strings thru the tuner in a straight line. I think I'll make a set of rollers for it that don't have any grooves, and vary the diameter to fix the string height. This would be a good test of the single point vs. V notch theory in terms of tone, don't you think? I'm pretty sure that on the other Zum the bass strings are sitting on the bottom of the string notch (not necessarily a V) and at about the time I made the switch it developed this GREAT BIG throaty growl. It sounded good previously, but is exceptional now. Even Bruce says so and he's darn picky (that's good). I'm pretty sure there's a correlation in there.

------------------

Doug Earnest
The only Zum Keyless U12, Fender Cyber Twin

FWIW Bruce Zumsteg makes only two different groove depths on his nut rollers. I was surprised to learn that on 10 string E9, the two deep ones are used not on strings 9 and 10, but on strings 1 and 2. The difference in depth is much greater than the difference in string diameters for these high E9 strings (not so for low strings on C6 or extended E9). Apparently this helps the tone of the high G# string. It works well.
Thus trial and error may work better (sound better) than a theoretical calculation that puts all the string tops at the same height.
And the calculation is well within the capability of Jethro Bodine if the string sits in a groove whose V shape has a right angle at the bottom.

I just keep coming up with ideas that spawn a half dozen more questions.

But, the tops of the strings don't HAVE to be the same height to be PLANE. There's no problem with the big ones being higher than the little ones as long as the top of the strings are in a straight line. That line can go downhill and still achieve our 'no rattle' / less bar pressure requirement.

------------------
<small>Larry Bell - email: larry@larrybell.org - gigs - Home Page
2003 Fessenden S/D-12 8x8, 2000 Fessenden S-12 8x8, 1969 Emmons S-12 6x6, 1971 Dobro, Standel and Peavey Amps<FONT SIZE=1 COLOR="#8e236b"><p align=CENTER><FONT SIZE=1 COLOR="#8e236b"><p align=CENTER>[This message was edited by Larry Bell on 09 March 2003 at 10:41 AM.]</p></FONT>

<SMALL>But, the tops of the strings don't HAVE to be the same height to be PLANE. There's no problem with the big ones being higher than the little ones as long as the top of the strings are in a straight line</SMALL>

Right, but that would be the case only if the string diameters changed in a linear fashion, but they don't. The lowest string is fatter than that, etc.
The Zum system works well on the high strings, but works poorly on the low strings. On my 12 string E9, puts the deep grooves on strings 1,2,11,12. I still get rattle on most of the wound strings except 12.

You may have missed my point, Doug
Maybe all I did was state the obvious but I'm pretty good at that.

I suggested that the grooves could either be cut such that the height of the top of each string is the same OR such that they graduated SMOOTHLY (you cut the grooves so the height decreases evenly down to the smaller ones). I still think it's easier just to put the string in the groove and adjust the grooves for all rollers to make the same distance from the bottom of the roller to the top of the string. That's clearly do-able AND I believe it can be done by either varying the groove depth or the roller diameter.

------------------
<small>Larry Bell - email: larry@larrybell.org - gigs - Home Page
2003 Fessenden S/D-12 8x8, 2000 Fessenden S-12 8x8, 1969 Emmons S-12 6x6, 1971 Dobro, Standel and Peavey Amps<FONT SIZE=1 COLOR="#8e236b"><p align=CENTER>[This message was edited by Larry Bell on 09 March 2003 at 10:49 AM.]</p></FONT>

Larry is correct.

If I implied that the tops of the strings had to be parallel with the neck I apologize. As long as the tops are in a straight line as Larry pointed out is fine.

And yes, it would be much easier to calculate the size of the nut roller IF the string did not fit into a groove. In fact it would be better in every way as far as I am concerned.

There MIGHT be a slight problem on "keyless" guitars in that the string may tend to shift away from the center of the roller according to the wrap around the locking screw. Not sure. But all in all, I think it would still be better than a groove,

carl

When I made the rollers for Joaquin's guitar, I cut them so that the top of the string would be level with the top of the roller, the idea was that the bar could slide right over the nut and not make a 'clunck' or a buzz, which of course was less of an issue for him than it was for me. The math was simply, circle in triangle. The height of the triangle was the depth of the cut. To be practical, I didn't cut them full depth.

If you assume a right angle at the bottom of the V, and string top flush with roller top,
then from the center of the string to the bottom of the V is the hypotenuse of a right triangle whose other 2 sides are the string radius r, in other words r*sqrt(2). To the top of the string add r again. Total depth = r*(1+sqrt(2)), or 1.212 times the string diameter.

Mr. Dixon, Earnest: You may have overseen that I posted the formula for any V angle yesterday. It is, in fact, very simple.

But it is not simple to find someone to make these things.

Larry, I believe you and I have the same conclusions, just stated in different ways. The only thing we are after is to get the tops of the strings in the same plane, it really doesn't make much difference how you get there, diameter of roller or depth of notch.

Earnest B., you are right in that Bruce only uses 2 size rollers on the E9 neck. However the bottom two strings on the C6 neck are a different size yet, we actually have three different sizes of rollers at the Zum Werks.( I am the part time assistant chief hole driller and student guitar assembler at Zum Steel). With maybe one more size of roller, they would be good enough on a U12. It doesn't take much adjustment to make one heck of an improvement.

With some reluctance, I am going to add my experience and suggestions. While the formula and methodologies outlined, so far, may be simple for some to understand, they seem too complicated for me. My technique should require little, if any, exacting measurements - if one starts with existing rollers and some manner of mounting them to a device such that they can be acurately turned against a cutting tool.

Let me first say that I have NOT yet performed this operation on rollers but the manner in which my lap steel nuts are formed gives me confidence in the expected results.

I'll try to describe that technique.

Beginning with flat 3/8in brass bar stock that has been cut to the appropriate height and length, the desired profile is hand filed. I built a jig to firmly hold the bar sitting on its bottom edge. The jig also is used to guide the file so the cutting strokes are straight and, initially, are parallel to the work surface. On the bridge side of the nut, the jig has a raised surface to hold the file at the finish height of the nut. On the key side, is another, lower, surface which provides a 'stop' so that, as the file is tipped during the cutting stroke, a consistant profile is achieved.

The first phase of the operation creates an even profile accross the entire length of the nut. Next, the nut is released, moved 3/8in (my string spacing) and resecured. HERE'S THE TRICK. Before re-clamping, a couple short pieces of string are placed under the nut so that it is raised by that amount from the previous filing. The jig keeps the file always in the same cutting pattern. The result is a 'graduated' profile on the nut conforming to the size of the individual strings. No fancy tools or machining skill required.

That works fine for lap steel nuts. A similar setup should work for cutting grooves in rollers. Presume we start with a roller properly mounted to some device which holds it in a fixed position while acurately spinning it. (Like a jeweler's lathe.) Next, presume a cutting tool mounted to a holding device that centers the cutter laterally to the roller but can be moved toward or away from the roller but not side to side. And finally, a movable 'stop' that can be secured between the cutter holder and roller holder.

With a CADD program I have determined that, using a 30 degree 'V', the depth of cut should be 1.5 times the diameter of a particular string. That would result in the top of the string to be flush with the edge of the roller.

First, using feeler gauges, select one that is 1.5 times a string diameter. Position the gauge and 'stop' between the cutter holder and roller. Move the assembly so the cutter point just touches the roller and secure the 'stop'. Removing the gauge will now allow the cutter to be advanced toward the roller and cut a groove of a depth equal to the feeler gauge. Start the roller spinning and do it.

If it is desired to have the top of the string(s) above the roller, simply subtract that amount when selecting the feeler gauge.

And, Carl, I totally agree with you about an adustable nut for each individual string. Even before reading your posts, I prototyped something quite similar to what you've described. So far, there's been no machine shop interested in producing such an item - even without pricing having been discussed. Making you a set has been in the back of my mind but coming up with a retrofit to your guitar would be problematic.

--Everett

Boy, sometimes we just seem to trip over a dollar to pick up a dime.

------------------
"Stump" Reeves

Wow, when I started this thread I had no idea this problem was bothering so many people. I thought this was just an oversight by Fessenden and that all I had to do was find another manufacturer's graduated rollers for replacement. I still hold out hope for that. There are several other threads on this from the not too distant past if you search on "guaged rollers." You will find some of the same suggestions (from some of the same people) as in this thread. Some guy named Justin offered to make sets of guaged rollers and took several people's money ($30 for a standard 10-string set, $40 for a custom set), but it is not clear if everyone or anyone got them. Here are some of my thoughts.

If you put a straight edge across the tops of the strings at the nut and sight from the end of the key head you will see a huge curved gap. As Earnest says, it is not linear, it curves up steeply at the last few wound strings, especially on 12-strings, and that is the big problem. What is tolerable on a 10-string, becomes really bad on a 12-string. If you have a 12-string, put the straight edge only on the first 10-strings, then on all 12. The gap doubles.

That we pay thousands of dollars for these instruments and still have these kinds of problems seems ridiculous. But most pedal steels are made by players and machinists, and few if any have been rigorously designed by mechanical engineers. Also, the instrument has developed by adding on additional strings and pulling mechanisms, and the new additions have created problems that time has not yet solved (e.g., the extra strings on 12- and 14-string necks). After all, pianoes, horns and orchestral strings developed over hundreds of years.

Although some manufacturers may actually have individually graduated rollers, most (if they address the problem at all) simply have 2 or 3 different sized rollers. Emmons uses two (at least on my '80s P/P 12-string, don't know what the new LaGrandes use), and that seems to work well enough for 10-strings, at least for playing at the first fret, although that slide over the nut with the bar may still rattle. But two sizes doesn't work for 12-string, you need at least 3. Zum uses three sizes. Which may be good enough for a 12-string. We just need to get some measurements to see which brands of rollers can be switched easily. I took measurements of my Fessy and Emmons P/P, but left them at home. I remember they both had 1/8" width rollers, but the axles were different sizes. I would like to get the measurements for the Zum rollers and axles. Probably we can just mix and match without having to do any machining.

As far as machining goes, here are some thoughts. Simply changing the radius of the entire roller would seem to work. But the radius effects tone, as Bobbye Seymour will tell you about the changer. A single point of contact is best, but if grooves are needed (as they are for key heads) then the two point V is best. Grabbing the string with a U requires more precision than is found with string guage tolerances, and doesn't allow for small changes in string guages (which are common among steelers). Most roller grooves seem to be a U, but except for the smallest strings, the strings don't fit all the way in the groove and instead sit at the top of the U in a two-point contact. Applying a V formula to this situation would seem to be tricky, unless one machines the entire U into a V. Because the walls of the U are parallel, simply machining down the radius of the entire roller (as Larry suggested) would work, if radius is not important for tone or string breakage (most strings break at the changer, not the nut).

The formulas are nice, but most of us would have no way to implement the results of the formula with precision machinary. But just spinning the roller against a triangle file, as suggested by some, would probably be fairly simple. One could do this off the guitar and use the string to judge when it was right. You could just measure, or even eyeball, until by trial and error you get the string to the right height above the roller (or flush with the top if that is your goal).

Machining will be a last resort for me if I can't find some ready made rollers from one or more manufacturers that I can mix and match. Toward that end, we should all post the measurements of our rollers here. You need the roller width, diameter, groove width, and axle diameter - oh yeah, and the brand and approximate year for your guitar. I am particularly interested in the Zums, because they already have 3 groove sizes. Also, MSA and Sierra have long histories with 12-strings and already put guaged rollers on some of their models, so they are also of major interest for mixing and matching. If people who have these don't post measurements here, I suppose one of us could give the manufacturers a call.

Finally, a couple of other issues were brought up. Earnest pointed out that one manufacturer used guaged rollers to intentionally place the 3rd string (E9) higher than the 1st and 2nd strings (or rather to place the 1st and 2nd strings lower) in order to improve tone for the 3rd string with increased bar pressure. Now that Jagwire is making a 0.0115 string, that solution no longer seems needed. These strings balance the 3rd string tone with the others, and they last a little longer before breaking.

As far as individually adjustable string heights, this seems like overkill and would seem likely to adversely affect tone. Such adjustments are needed for fretted guitars to get the action just right. We don't need that for steel guitars. We just need to get all the strings on a linear plane. That's a one shot design correction for a manufacturer, and if it were done right then, we wouldn't need to ever adjust it again. If someone wanted to put on a radically different string guage, they might have to buy a new roller for that one string. Big deal. It's only a problem when different guaged rollers are not available because manufacturers skimped on this important design feature. If you're buying rollers by the hundreds or thousands, it would seem to be a really trivial additional expense to get them with several different sized grooves.

So for the simplest solution, let's all post our roller measurements.

<FONT SIZE=1 COLOR="#8e236b"><p align=CENTER>[This message was edited by David Doggett on 10 March 2003 at 12:32 PM.]</p></FONT>

David,

You said,

"Wow, when I started this thread I had no idea this problem was bothering so many people."

And indeed you are correct. Having been in the design and troubleshooting business for most of my life, I can tell you that what you said is MORE the case than not.

RCA once proved that for every percentage point of complaints from their customers, there was a 57.8% factor that could easily be added to that complaint. Meaning that only a few notice and/or complain.

But it does NOT mean the others are all satisfied. Most simply live with a problem until it is solved. Also, some feel somehow it is a reflection on "their" judgement if they state that there is something wrong with a product they have purchased.

The good part is. It is the complainers that bring about improvements. If it were not for that, there is little likelyhood man would have advanced much past the stone age.

Thank God we don't still have "outhouses",

carl

Okay, so, Carl, let's get out of the outhouse. Measure the rollers on your guitar(s) and post them here. If others post their measurements, I'll put them all together in a table so we can see what the mix and match possibilities are. We need the roller width and diameter, the groove width(s), the axle diameter, and the make and model of the guitar. Who knows, maybe everything we need is already out there somewhere, and we don't have to do any machining ourselves.