Official Luthiers Forum!

I've only completed 2 guitars, both Martin kits. Number 1 was a rosewood dred & number 2 was a Mahogany OM from reject parts. The rosewood dred sounds great, everybody loves it, big lush sound. The OM is just lacking. The sound isn't bad or offensive. It's very laid back & warm, not very exciting. There is no sparkle on the highs & it has weak projection.

I'd like to try to improve the tone of the OM. Here's some details on it's construction:
Mahogany back & sides
Sitka top
Black ebony bridge
Macassar fingerboard
Bone nut & saddle
Top braced to Grellier plans
Back braced per Grellier except the 2 bottom braces are not wide & low, they're narrow & tall like the top 2

I scalloped the top braces using the Grellier plans as a guide. I scalloped until I heard pitch of the top tap drop & the sound loosened a little. I think I was careful not to go too far.

The back is only 0.095" thick - that's how thick it was when I got it. From what I've read this is too thin for Mahogany. I wonder if this is dissipating some of the energy?

I also wonder if a rosewood bridge would would help transmit the high frequencies better (less damping) and also remove a little weight from the top?

Here are some pics of the bridge. I think the break angle of the strings is ok?



The next question is about the plates. It's possible that they are incorrectly tuned & may be cancelling each other?

I know that this is complicated subject with several variables & there is no silver bullet, but if anyone has any ideas on things to try or methodology to isolate the issue(s), I'd like to hear them.

Thanks,
Kevin Looker

_________________
I'm not a luthier.
I'm just a guy who builds guitars in his basement.
It's better than playing golf.

Only use finishes that are proven by other luthiers. Some finishes may dampen the sound. Just a long shot.

How about posting a tap response? That would make it a whole lot easier to figure out what's going on (or not going on).

_________________
Trevor Gore, Luthier. Australian hand made acoustic guitars, classical guitars; custom guitar design and build; guitar design instruction.

http://www.goreguitars.com.au

How long has it been strung up? I find that my guitars evolve for at least a month after stringing.

According to your comments, sounds like it's not very responsive - so may be overbuilt. One option is to go up in string size (if you're using lights, use mediums). This might make it louder....

But this is one of those topics that will have at least as many opinions as there are people in the room. Caveat to my comments is that my opinions are based on having built ten guitars, not 100.

One thing I've really focused on is to build a bit lighter each time (within reason). My first guitars were all overbuilt, and heavy. As I worked on improving my fit/finish - tighter joints, cleaner work, dry - fitting parts so that they fit as well I could possible make them fit before gluing, the tone started to improve, and they became more responsive.

the top is most of the tone. What was the top thickness and which glue did you use. Did you do any testing on the top for stiffness? Also what gauge strings you using?
One thing you may also check is often the nut causes more problems than anything. I bet you have a straight slot on the nut. It may be following the head stock angle but I learned that this needs to be curved , Think roller , as the steel strings do not bent at a given point but tend to have an arc to them.
Nut issues usually change on a fretted note . If that isn't an issue then you may need to rethink your bracing . Keep close records and with that you can start to see the cause and effect relationship.
Joint integrity is one of the most crucial parts of building.

_________________
John Hall
blues creek guitars
Authorized CF Martin Repair
Co President of ASIA
You Don't know what you don't know until you know it

I'm at work now & the guitar is at home. I'll try to record some taps this evening & post them.

The guitar has been strung up for over a year.

The finish is Deft rattle can lacquer, it's very thin, my sand through proves it.

It's strung with lights - trying mediums is a great idea to see if it's over braced.

The top is around 0.110, I'll double check this evening & it was glued with Titebond. One other detail is that the bridge plate is Maple. I didn't do any sort of deflection testing but the top was pretty stiff. There is a slight rise at the bridge so it doesn't look like it's over braced, at least not it that region.

The nut slots are straight, I never knew to file a curve. You mean a curve in the bottom surface of the slot so that the string rides over a curve as opposed to a straight ramp?

Another detail I forget to mention is that the braces go all the way through the sides. There may be little gaps between the tops of the braces and the sides - possible energy loss?

Kevin Looker

_________________
I'm not a luthier.
I'm just a guy who builds guitars in his basement.
It's better than playing golf.

Kevin, I'm not much of an expert on this but I own a couple of old Martins from the "overbuilt" 70's era that have had some aftermarket hot rodding by a well known guy who does this sort of thing. The D18 had the bridge plate replaced and some moderate scalloping, the D12-28 had a new lighter b/p installed and a lot of work done at the bridge. Both had neck resets at the time which gave much better saddle height and angles. When the D18 returned home after having all the work done my wife's comment was "you are playing louder tonight". I didn't think I was - same strings, pick, sloppy technique.

I'll also add that my small body guitar builds (OM, 000 and 0 sizes) have all been very heavily scalloped and the are all loud and balanced across the strings. They may blow up some day, but in the mean time.....

edit to add, looking at your pics the saddle doesn't seem all that tall. How high are the strings off the top, what is the action and neck angle?

Kevin: Some of my beliefs about responsiveness considering the posts that have been presented.
1. If you have built large guitars and start building smaller guitars it is very easy to over build.
2. As John Hall mentioned,joint integrity is key.
3. I don't buy the HHG is the only way to go. Nothing wrong with Titebond.
4. Weight of the completed top plate is also key. There is only so much energy in the strings.A smaller lighter bridge of a wood not as dense as ebony is a big step in the correct direction. Your maple plate is OK as long as it's not to big. Any thing over 1.625 in width is getting big for me.
5. Don't couple any braces below the sound hole into the lining except the X if you want to use that for registeration. Then taper the lower ends until there is just enough left to do that job.
6. Thin the edge of the lower bout about 1/4 of the original thickness tapering in about 2" from the outer edge.
Tom
Get yourself a gram scale,check and compare the weights of your tops and components. If you have not done that it can be eye opening.

_________________
A person who has never made a mistake has never made anything!!!

Also, an OM is different from a dred. Just in case you are comparing the two.

Can you check the bridge rotation before and after stringing up? If the top is too thick or overbraced you probably won't have very little bridge rotation. From memory, seems 2-3 deg is about right.

_________________
Formerly known as Adaboy.......

I spent some time with the guitar tonight.

I bought a set of medium strings to try.

As I loosened the lights, I paid attention to what the bridge & top did. They moved more than I thought they would but I didn't try to measure it. The bridge rotated & the top sunk down a bit as the tension was relieved, not the easiest thing to measure with just a ruler. I'll break out the dial indicators next time.

I then went to pull the saddle & I noticed that it had some play in the slot. I could make it rock slightly. I measured the thickness of the saddle with my calipers & found that there was 0.006" of slop between the saddle & the slot.

I made a new saddle that fit better and restrung the lights so that I would only change one thing at a time.

The sound is definitely louder but the strings are shot. The sound is very bassy & mid-rangy, I'll put a new set on tomorrow & see how they sound.

Kevin Looker

_________________
I'm not a luthier.
I'm just a guy who builds guitars in his basement.
It's better than playing golf.

FWIW, here is my point of view on your problem.


I think you are on the thick side, specially if you feel the top was pretty stiff to start with. My tops on OMs are usually around .090 for sitka, depending on stiffness of course, but never .110''.


How thick are the braces when they go through the sides? If they are 1/4'' or more, they are killing a good part of the sound. Try to make them no more than 3/32''.

I understand I'm not much of a help for this perticular guitar, but it might help for the next ones...

_________________
Alain Moisan
Former full time builder of Acoustics, Classicals and Flamencos.
(Now building just for fun!)

Working on the bridge/saddle did help things a little but the overall tone of the guitar is still lacking, IMO.

I did some tapping & made some recordings.

Here are some spectrum plots, if they are of any use. I can also post sound files once I figure out how.

I recorded with the mic placed directly in front of the sound hole.

This is the frequency spectrum plot of tapping the top with the back free to resonate also. The strings were removed.

This is the top tapped with the back damped by placing it on a blanket.

This is the back tapped with the top damped.

Is this information helpful at all?

Thanks,
Kevin Looker

_________________
I'm not a luthier.
I'm just a guy who builds guitars in his basement.
It's better than playing golf.

Kevin, most people tend to overbuild and overbrace their first guitars. Don't worry about it. My first few flattops sound very weak and toneless to me -- although other people like them a lot, I have been spoiled by the tone of the instruments I built after I improved my skills a bit. lI would recommend John Mayes' DVDs on voicing the top; they are excellent. Once you watch them a few times and think carefully about what he has said, it's hard to make a bad sounding guitar.

IMHO, things like the finish, glue choice, etc, might or might not make a small difference in tone. (I personally do not believe it makes a meaningful difference given the variability in more significant factors like voicing the top, but others take a different view.) But in any event, those things won't make an otherwise well-built guitar sound bad. At most, they make a slight difference in tone. If you are unhappy with the sound, I would bet steaks to cheeseburgers that your world will change once you improve your voicing skills.

BTW, I don't think .095 is too thin for mahogany on the back. Nothing wrong with a little thicker either, but if your top were well voiced, a .095" mahogany back would not water down the sound meaningfully.

The top's too loose right around the bridge and too stiff around the edges. Get the set up right with a proper fitting saddle (you already have?) and nice curved nut slots. Lighter gauge strings may actually help a little too by not binding up the area of the top around the bridge. A Musser brace/PMTE may help, as well as shaving down the lower tone bar if you can reach it. (I assume the plan you used uses 2 tone bars)

_________________
I didn't mean to say it, but I meant what I said.
http://www.brackettinstruments.com/

Try fretting it at . . . say fret 5. Soon eliminate any Nut slots.

Kevin, Could you retest the tap freqs with the mic not in front of the soundhole? That position mostly picks up the 'main air', and if too close to the hole, the mic interferes with that resonance. And in Audacity, in the "Size" box, select 16384 instead of 1024 -- that will give a fine resolution to the plot. More tips on tap testing here: viewtopic.php?f=10101&t=31330

_________________
David Malicky

Thanks for all the replies.

Once my power comes back on I'll measure some more taps.

I've actually tried to shave down the ends of the X-braces but it's going to take some customized tools to get down there.

Kevin Looker

_________________
I'm not a luthier.
I'm just a guy who builds guitars in his basement.
It's better than playing golf.

Do you have a photo of the bracing? If not... get yourself a camera and start taking them every time. Very useful to look through all your past guitars when deciding how to carve the braces on the next one.

How tall are the lower legs of the X where they notch into the linings?

I'd probably begin with sanding it thinner around the perimeter (with strings on, so you can hear the difference immediately), since it sounds like your top is on the thick side, and that's the area where stiffness is most defined by the plate itself. You can sand the central region too, but it's hard to tell how thin it's getting and how even it is, so you might have better luck shaving braces there.

Of course, any sanding will mess up the finish... but even if you just give it a quick and dirty touch up, a rough looking, but good sounding guitar will get a lot more playtime than a shiny, bad sounding one

Oh, and .095" sounds plenty thick to me for a mahogany back on an OM, unless it's exceptionally floppy.

Well, that's a pretty strange looking set of tap responses. Not knowing exactly how you produced them (mic position, mic type/frequecny response, tap procedure etc.) makes it fairly hard to do a diagnosis, but this is how I'd interpret those sorts of charts if I'd made them using my normal technique.

What I would normally expect to see for a top tap response is a pretty high-Q peak around 100Hz, which is the main air resonance, another peak around 200Hz, which is the main top resonance and a further peak at ~240Hz which is what you get if you have a live back; and those are the first three significant peaks.

What I'm seeing in your top response is a low-Q peak ~ 75Hz (which is very low frequency for a main air resonance on a 000 sized guitar and also recognising you're using a log frequency scale) another peak around 190Hz, which is probably the main top resonance, and nothing that seems to correlate with back activity. If the main air resonance is that low, it can only be because you've occluded the soundhole (unlikely) or either one or both of the top and back panels is very soft (i.e. not very stiff, very low resonant frequency).

Now, if the peak at ~190Hz is the main top peak (and let's assume that, as it is about the right place and nothing you've said about the bracing etc. would make me think otherwise), that would mean that the back has to be very soft for the main air resonance to be as low as it seems to be. If we look at your back tap, there's a peak around 100Hz with the next significant one around 350Hz. If you made this chart with the guitar face down on a towel and the mic above the back, I would not expect to see much of a peak corresponding to the main air resonance, but you still have a peak at ~90Hz. If we look at the top response with the back damped, there is no real peak in the 90-100Hz range. All this seems to be pointing at a back main resonance tuned to around 90-100Hz, which is really low, but fits with the logic of a low main air resonance.

Having a back that soft (tuned that low) would be sufficient to explain the symptoms you describe.

Problem is, I can't imagine how you'd get a back tuned that low unless the back braces are detached or the mahogany panel was really floppy before you braced it.

Does this diagnosis pass a reality test? Well, no. The charts aren't like anything I've seen before (and I've seen a lot of charts!) so I would suspect the data gathering before anything else (if you made these charts using the built in mic of a lap-top computer, all bets are off!), but you get the picture about how I'd go about analysing them. If you're convinced that the data is pretty good, the next test would be to see if the peaks you have identified correspond with the Chladni pattern frequencies for the modes you think they are.

_________________
Trevor Gore, Luthier. Australian hand made acoustic guitars, classical guitars; custom guitar design and build; guitar design instruction.

http://www.goreguitars.com.au

Power is still out here - using phone.

All measurements taken with mic directly in front of sound hole less than one inch from top.

Mic is ok, integrated stereo mic on Sony PCM-M10 recorder.

Will record new taps using David's procedure unless there are some other ways to try?

Thanks,
Kevin Looker

_________________
I'm not a luthier.
I'm just a guy who builds guitars in his basement.
It's better than playing golf.

That could explain a lot. I usually have the mic about one metre away. I use a Shure PG57 plugged straight into my laptop with Visual Analyser as the app.
More info and discussion here. Hope you get your power back soon.

_________________
Trevor Gore, Luthier. Australian hand made acoustic guitars, classical guitars; custom guitar design and build; guitar design instruction.

http://www.goreguitars.com.au

If it's lacking highs,
with good bass and mids,
I would think it's under built.

Is your phone an iPhone? There is an app called iAnalyzer for 10$, or iAnalyzer lite which is free, which gives you handheld FFT convenience. I tested it against my studios' signal generator and it was within .1 hz, which I think should be close enough. Then, (and pardon me if you know this) you can get a screen capture by hitting the home button and the little square one on the top at the same time.

Trevor, The low Q was because the FFT size was only 1024 or 2048 samples -- Audacity's default sampling frequency is 44 kHz.

All, The resolution of an FFT plot is related to both the FFT sample size (1024, 2048... 16384) and the sampling frequency. (In Audacity, sampling frequency is set by Edit > Preferences... > Quality > Default Sample Rate). I'm not an EE, but here's my understanding:
For the finest resolution without a phenomenon called aliasing, set FFT sample size = (1/2) * (Sampling Frequency). Technically, that sample size is the Nyquist frequency.
The factor of 1/2 could be as low as ~1/3 and still get decent resolution and pretty sharp peaks for guitar purposes (in Audacity, 16384 / 44100 Hz).
If the FFT size is too high relative to the sampling frequency, aliasing will produce artificially noisy FFT plots. E.g., a 16384 size at 22 kHz sampling will produce many sharp peaks but most of them are aliased.
Maybe someone better versed in EE can check if the above is right.

_________________
David Malicky