Intonation issues - Questions

[Shane Woonton]

Hi guys,

On my latest build, a black limba/euro spruce OOO cutaway, all is going really well... until I get to intonating the saddle. I've used Trevor's simple intonation fix, and moved the nut 1.4mm towards the first fret and was planning a standard saddle intonation. It's 632.46mm scale length (24.9") and I've triple checked that all the frets are in the right place etc.

I placed the bridge as per Trevor's build book, adding 3mm of compensation to the front edge of the saddle on the centreline, using the 12th fret as the measuring point (as the nut has been moved forward). So 319.2mm (316.2mm + 3mm) from 12th fret to front edge of saddle.

The saddle is 5.5mm deep. Saddle is skewed by 2.5mm from high to low E.

Thing is, all of my 12th fret readings when trying to mark the saddle positions for each string are flat when set at the very front edge of the saddle. So I have no room to move the string positions any further forward...

The fretted 12th fret notes are flat by the following amounts from low to high:
E - 0 cents
A - 6 cents
D - 9 cents
G - 13 cents
B - 7 cents
E - 10 cents

All indications are (to me) that the saddle needs to be moved towards the nut (by at least a few mm?) but the placement should be spot on. I'm trying to understand why things are not working when the bridge has been placed "correctly". The only things that are different to my last build is the scale length (650mm on last one) and the fact that I have moved the nut toward the first fret by 1.4mm.

When I moved the nut should I have made an adjustment to the bridge placement?

Any thoughts on what is going wrong here?

Cheers,

Shane

P.S. I should add one caveat which is we had a new baby 2 weeks ago so it is feasible that sleep deprivation is a factor

[Shane Woonton]

Couple of other factors to consider:

Action is nice and low - 2.4mm low E and 1.6mm high E
Strings are Elixir custom light 11 - 52

Shane

[jeffhigh]

The short answer is you have too much compensation.
If you do nut compensation you have to take that away from what you would otherwise use at the saddle.

I don't set the bridge measurement at the centre, I use measurements to the high E and low E
Using the Stewmac calculator you would have 2.2mm compensation at the high E and 5.3 on the low E
These are to the break point of the string not to the front of the saddle

From your measurements you have 3-1.25+1.4 =3.15mm at the high e and 3+1.25+1.4 =5.65 at the low E and these are at the front of the saddle, giving you nowhere to go.
Unfortunately your nice wide saddle is wasted

In addition the very low action will be contributing to the problem.

I would try raising action to 2.8/2.0, and cutting back the break point of the nut individually
This may get you to an acceptable result otherwise you are going to have to bring the saddle slot forward.

[jeffhigh]

ps
You are not alone in this problem
I set out my bridge on my last build 4mm too far back. I blame it on sleep deprivation too.
I had to fix it by shortening the neck....

[Shane Woonton]

The short answer is you have too much compensation.
If you do nut compensation you have to take that away from what you would otherwise use at the saddle.

Thanks Jeff, that's the piece I must have missed in my reading.

I would try raising action to 2.8/2.0, and cutting back the break point of the nut individually
This may get you to an acceptable result otherwise you are going to have to bring the saddle slot forward.

Good idea. I had considered raising the action slightly but not adjusting the nut. If that fails, I have room on the bridge to move the saddle slightly forward also. Not something I want to do but will see how I go.

Cheers,

Shane

[Shane Woonton]

ps
You are not alone in this problem
I set out my bridge on my last build 4mm too far back. I blame it on sleep deprivation too.
I had to fix it by shortening the neck....

Yikes!! Shortening the neck. I have my fingers crossed it doesn't come to that.

Thanks,

Shane

[Shane Woonton]

So, I raised the action to 2.0mm on the high E and 2.8mm on the low E (0.4mm higher on each) and it had the right effect, just not quite enough of it.

The readings for the fretted 12th fret notes have moved to (from low to high):
E - ok (correct compensation is now 2mm onto the saddle)
A - ok (correct compensation is no 1mm onto the saddle)
D - 4 cents flat
G - 8 cents flat
B - 5 cents flat
E - 7 cents flat

So raising the action by 0.4mm had the effect of shifting the reading by about 5 cents on the D & G strings and by 2 - 3 cents on the B & high E. And by a little more on the low E and A which can now be compensated with the saddle in the current location.

So.... onto the nut....

[Trevor Gore]

Shane, check out Design section 5 for layout procedure, which covers the situation with and without nut compensation. Basically, if you do nut compensation there is no need to add more at the saddle, as Jeff has mentioned.

The typical numbers I have given in the text are for lights (12-53) on a 25.4" scale length with a 3mm 6th and 2mm 1st action. Your nut shift will equate to ~ 4 cents flat on all strings, the un-required saddle compensation will also contribute and the low action and custom lights will account for the remainder.

There are numerous ways you could go to get back into tune. The least painful structurally would be to cut the nut back the 1.4mm you moved it forward, so that you have an "L" sectioned nut with the string take-off in the theoretical zero fret position, use light gauge strings (12-53) and raise the action. This should get you considerably closer. Of course, it won't give you the feel you were after. To get closer to the anticipated feel, you will need to move the saddle slot, which may be possible if you have enough material in front of the saddle after filling the old saddle slot. A two part saddle might help you achieve this.

Another option is Jeff's solution, which is not as difficult as it may sound, depending on the sort of neck joint you have. With a bolt on neck it is relatively easy to pare material off the shoulders of the heel, thus bringing the frets closer to the saddle rather than bringing the saddle closer to the frets. And, of course, there's fiddling with a bit of everything, grabbing a millimetre here and there, until you've made up the numbers.

I hope it works out for you.

[simonm]

The readings for the fretted 12th fret notes have moved to (from low to high):
E - ok (correct compensation is now 2mm onto the saddle)
A - ok (correct compensation is no 1mm onto the saddle)
D - 4 cents flat
G - 8 cents flat
B - 5 cents flat
E - 7 cents flat

I will admit straight off that I know nothing about intonation. So my question is whether most or any musicians will actually notice this "deviation"? At what point does an intonation variance start to become noticeable to a human as opposed to a computer? 1 cent? 10 cents? It depends?

[Trevor Gore]

At what point does an intonation variance start to become noticeable to a human as opposed to a computer?

It varies greatly by human. It is most noticeable to most people when playing intervals or when playing with other instruments. For example, equal temperament 3rd are already ~14 cents sharp compared to Just temperament, which sounds a lot more "in tune". So adding more error, say another 6 cents on top of the 14 cents gets you out to 20 cents which most people will recognise as an out of tune interval. I have one particular customer who can routinely pick a 2 cents error across intervals. Absolute pitch errors on single notes are much harder to spot and will mostly go unnoticed. Some people struggle with guitars because they always sound out of tune. Other people hear the out of tune as part of the endearing character of the guitar. Mostly, those who have played a really in-tune guitar that's been properly intonated never want to go back to a standard instrument. There's also a bunch of side effects: because the note harmonics line up better across intervals, guitars sound louder and clearer with better note separation and articulation, which is seldom a bad thing.

[Nick]

I will admit straight off that I know nothing about intonation. So my question is whether most or any musicians will actually notice this "deviation"? At what point does an intonation variance start to become noticeable to a human as opposed to a computer? 1 cent? 10 cents? It depends?

A good musician will pick it straight off & lets face it, if we want to sell our work why not spend the little extra time involved and get it right? Even a poor muso will know something's wrong, maybe just not what, when they play chords. Also it may not be so apparent to a " sit at home" strummer when it's just him & the guitar but when you play alongside other muso's that have properly intonated instruments it will be obvious something ain't right I'm picking.

[kiwigeo]

Its an interesting discussion. Last year I did a right to left conversion on an Ibanez 6 steel string. Because the thing had been strung for a south paw the intonation was absolutely crap. The owner however, a professional musician had been happily playing the thing and not noticing the serious issues with the instrument's intonation. Guess it proves you can be tone deaf and still make a living out of playing music.

[Nick]

Guess it proves you can be tone deaf and still make a living out of playing music.

Had a bass player like that once His bass played in tune but his ears didn't, I'd have to spend most of the gig yelling out the chord changes because he couldn't hear when he'd hit the wrong root note (could play a whole run or arppegio a semitone or whole note out & not hear anything was wrong) Got really awkward if I was doing vocal harmonies, I'd have to turn away from the mic inbetween lines to call the changes!

[Shane Woonton]

Thanks Trevor, I'll do a bit more reading and some careful calculations before proceeding. I have some ideas about getting the extra movement I need at the bridge end. If necessary it will also be simple enough to remove the 1.4mm compensation I added at the nut.

It's always a shame to get this far into a build and find you made a mistake but this one has been a really good learning experience. Having read the literature on intonation, it's not until it goes wrong and you have to think through "why" that it begins to sink in properly.

I'll also be taking some careful measurements off my other build which was easy to intonate but was slightly different - 650mm scale vs 632mm and 0.5mm higher action. But all else was the same so it will be interesting to see how much different the compensation is for each once I get it sorted.

Cheers,

Shane

[Shane Woonton]

The good news is my self-imposed problem is fixed!

I thought some of you might like to see how I resolved the problem.

I had used a 5.5mm bridge so had room to reduce the width of the saddle slot using a piece of ebony binding cut to fit, glued in place and sanded.

I then removed some of the saddle so that the bridge was shifted forward by 2.2mm.

I also removed the 1.4mm compensation I had built into the nut by filing a ramp into it so that the bottom of the string slots was in the correct place.

The two actions combined allowed me to file the saddle such that the fretted 12th fret notes were exactly where they should be!

All in all a valuable learning experience - don't try to figure out intonation when you've had 3 weeks of interrupted sleep! - and the end result sounds beautiful, even if I have lost the 1.4mm of nut compensation I originally planned.

I also managed to keep the action to 2.4mm low E and 1.9 high E so I'm pretty happy with the outcome.

Thanks again for the advice!

Cheers,

Shane

[Trevor Gore]

Nice save, Shane!