Piano World Home Page Forums Our Most Popular Forums Piano Tuner-Technicians Forum How much crown is too much crown?

For 25 years my business partner and I have been recrowning original soundboards.

just for clarity:

extract with steam
steam apart into component parts
repair individual parts
repress parts into a new crown
reinstall

working with such specialized material, I think we have been suffering from an overabundance of caution, and over crowning our boards. I'm sure that it is simplistic, but, if it is clear that too little crown is no good, and some crown is better and perhapse propper, at what point does increased crown begin to be a tonal detriment?

Mechanicly, this danger zone was revealed when the plate of an A2 required much raising to bring the bearing into reasson. Still, this is a mechanical limit. What I'm curious about is the tonal implications of various crown profiles. And perhapse bearing principles that optimise each.

Thanks,
Craig

It's stiffness that's important, not necessarily crown, which is a means of increasing stiffness without adding weight. Soundboards are more or less susceptible to stiffness loss depending on how they are constructed. Unfortunately, simply recrowning a soundboard does not bring this stiffness back, as the wood fibers have already lost their resiliency. Your best bet is stiffening the board via other means, such as auxiliary riblets and/or a penetrating epoxy treatment... or replacing the board altogether.

Beethoven,
Thanks for responding. But I'm not sure we are talking about the same thing yet. I have never found old tonewood to be lacking stiffness, resiliancy, or strength. Quite the opposite, actually. And if you would like to discuss the superior qualities of aged tonewood, I would love to.
My original question, however, was about the tonal effect of having a higher or lower crown in a given piano. It would seem to me that a higher crowned board might be notably stiffer than one more shallowly crowned. And while it may be mechanicly able to support a heavier bearing, is it able to respond as well to whatever bearing it gets, either heavy or light.

Well, I'm not sure that anyone has an exact "crown value" that has any universal meaning.

For one thing, different pieces of wood respond differently meaning that different assemblies of wood respond differently. Also, crown is a result or symptom more than a specification. That is, when a soundboard is bent it does have something we call crown. However, the purpose in deflecting the board was to create stiffness, not crown. Crown may be an indicator of panel stiffness with a given material content and construction procedure but it is not some universally sought-after value.

Also, you really can't think about crown in any meaningful sense without talking about downbearing. The more downbearing you are going to put into the system, the more crown you need to bear it. (Perhaps, and with other things being equal -- which they rarely are). These two factors are interactive. It would do just as well to say that the more crown you are going to build into the system the more downbearing will be called for.

This is not to say that low crown and light bearing are necessarily identical in performance to high crown and heavy downbearing. In the same fashion, a lighter board may need more crown and a stiffer board may be in danger of getting too much crown.

It's kinda like string scaling. You control the variables of mass, tension and length to find your "sweet spot". Multiple "sweet spots are available, depending on how you manipulate the variables.

Hi Craig,

I was at Larry Buck's shop for the Bill Shull talk where you brought some bellies to "show and tell".

I find your processes very interesting, ie salvaging the existing soundboard panel and ribs for reuse.

The question you are asking though can't be answered without knowing the overall stiffness built into the system. Relative stiffness determines the impedance levels and thus tonal response of the structure as a whole.

Crown is not necessarily an indication of stiffness. A board can have plenty of unloaded crown but lack adequate stiffness. Stiffness is the result of the combination of rib beam section, compression at ribbing, panel thickness,perimeter thinning, bridge height and the wood panel's and ribs' structural numbers(MOE etc). In addition, the relative stiffness has to match the string scale.

In your situation you are cutting new ground, because no one knows what the structural numbers of an aged reconstituted panel with its inevitable age hardening and compression sets at glue joints does to the stiffness of the panel, even if ribbed with original ribbing. I would suspect, actually, but my hunch would need to be tested obviously, that the reconstituted panel, ribbed at EMC's similar to the original say 4%, would be significantly stiffer than the original.

I would think this could be a serious tonal problem if you were retaining the original string scale.

The challenge your process presents is since you don't know what the original design EMC was, and since you don't know the actual strength of the wood in its aged state, you have to take a guess guided by empirical trial and error. Trying to repeat exactly was done originally will, in my opinion not result in a board which performs as the original performed. It might work and it might not, which is the mantra of any compression based system.

I would try to load the unstrung board evenly along the bridge, measuring what deflection you get per load so you can come up with some idea of what the spring rate of the reconstituted system is. Though this info would only be useful after doing it a bunch of times and beginning to see trends in the spring rate vs tonal outcome.

This is the Achilles heal in trying to copy the board exactly as is was originally...for the reasons stated above regarding unknown stiffness, the original board is gone. Even if you are using the original board and ribs, and you are basically doing the same thing any new belly guy does, which is redesign the belly. The problem is, that since you are usign materials for which there is no data, you gotta do the empirical research yourself.

I still find it an interesting approach, but it presents some difficult design problems.

Jim Ialeggio

Jim,
Yes, I remember you. That was a fun day. Thanks for responding.

Yeah, I guess I figured that it would come down to personal research. I was just hoping that there might be some golden rule I had missed.

I think your hunch is right. I don't have a lot of experience with new soundboards, but I would say that they are significantly stiffer. The old wood is some pretty hard stuff.

Why should this added stifness prove a problem in association with the original string scale? This is a real rise in impedence. There is no added mass,and perhapse even less, as old wood tends to be lighter. Why woudn't this add up to a more responsive board?

I like the Idea about loading the unstrung boards with weights to produce some resistance curves. We have a few pianos at that stage, so the comparisons should be interesting. I guess its time to start taking notes.

Thanks

With new boards that are installed at low moisture, the boom after a light hit on the middle of the bridge, will let you know when the board is opening up. String the piano just after the volume and sustain start to diminish. If your lucky you'll catch the sweet spot of the crown.

OK, I don't do belly work. I would like to understand more about the subject, though. What the heck, I'll just come out and say it. I think there is a lot of snake oil being sold. Can anyone answer me this: How is the stiffness of a soundboard computed or measured and in what units, and then how do you use this to determine the appropriate scaling? I suspect the answer will boil down to "whatever works!"

Del has written tomes on this on this very forum over the years.

The short answer is, it depends on whether you are compression crowning or rib crowning.

With compression crowning it is somewhat of a crap shoot. You do what seems to work more than not. Sometimes you get a really good board and lot of times you get a honking beast. THe big problem being, the same problem Craig has to deal with...that is, each panel will have its own stiffness profile, depending on EMC at ribbing, board thickness, and wood strength values. Though wood strength values are quantified with data, the data is an average, and the board you actually use could be close to or well removed from that average.

Del's, Ron Nossaman's and Ron Over's work all take the design of the belly's stiffness to a more quantifiable level by designing a rib scale using standard engineering beam deflection values amongst other parameters to predict and trend the known loads imposed by the strings at a given design downbearing. As well, they minimize the effect of the panel regarding stiffness, by ribbing the panel much closer to ambient EMC's. My own work is derivative of the work of these chaps, but is taking their work and going with it in my own direction.

Though the above rib crowned and supported system tries to approach this stiffness design from a more quantifiable position, the entire assembly is obviously more complex than the beam and load formulas. The beam formulas, at least in my take, are used as a means to analyze with quantifiable data observed statistical trends in the belly as a hugely complex whole.

As far as your question of what units, I'm using spring rate (deflection/inch at a given load)during the spread sheet design phase, then emprically quantifiying the "as built" spring rate of the finished loaded board. That's why I suggested to Craig that he come up with a way to load the board before stringing, to see what spring rate he had achieved. At that point, he as well as the rest of us, have to figure out, by observing the tonal response at various parts of the scale at various spring rates, what "as built" stiffness we are looking for.

Jim Ialeggio

Mechanical Impedance is a function of the board's inertia, mass and elasticity(stiffness).

You've not added mass, maybe even a wee bit lighter. But both of our hunches point to a wood panel that has a greater stiffness to mass ratio than the new wood had. I think crown as a design feature somewhat confuses this issue. By taking the aged density of the panel and drying it down to original EMC's before ribbing, since the board is denser than originally (probably), but the same mass (close to I think),I think you are compressing the system into one that is stiffer than the original...hence the off-the-chart unloaded crowns. But the crowns don't create the stiffness in and of themselves...I think in this case, they are more a symptom than the cause.

As I think about this, my first attempts at figuring this out would be to empirically test out the effect of less crown, meaning ribbing at EMC's higher than the original EMC's.


Yeah...I built a go bar deck using spring loaded go bars specifically to come up with this data on my own Rib crowned and supported boards. By the way, credit for the spring loaded go bar deck goes to Jude Revely of Absolute Piano for the idea, which I am happily running with.

Jim Ialeggio

Thanks Jim!

I really want to get this down to "brass tacks", so I need some clarification.

First:



You mean you would measure the stiffness differently depending on the construction? Wouldn't the result be prejudiced?

Second:



Good! A true engineering unit.

BUT Third:



Now things gets foggy, as they always seems to. How are you observing the "tonal response"? Is there a measurement or just a subjective observation?

This was referring how the stiffness is computed, not measured. My take, measuring "as built" spring rate would be measured identically for both construction methods, as I described to Craig.

Computing...well there really is no way to "compute" a compression crowned board. Its an gut level guess informed by lots of experience and lots of unsuccessful bellies. Sometimes it works and sometimes it doesn't.

Rib crowned systems give you beam formulas which don't, for sure, tell the whole story, but if you set them up appropriately, you can use them to help see numerical statistical trends in the larger more complex structure. At least that's my take. Its not 1+1=2 simple algorythms, but it gives you a way to look at relative deflections and spring rates, and to draw numerical and/or visual trends.

The key word is trends, in my view, quantified trends, that is.


What'a ya mean "just a subjective observation". This is music ya know...geez. If you like the sound you're there, if you don't, you're not.

I said ..."as built" stiffness we are looking for"... "we" being the editorial we in my shop and we as my clients and myself in the context of our professional relationship. Not we as the human race. There is no holy grail sound this is all honing in on. Being a subjective enterprise, some folks will "look" for a particular kind of sound, and some folks will look for another type of sound. But if your trends produce so and so sound consistently, and you are looking to add another dimension to the sound, you tweak your numbers or some parameter of the structure, rather than do the same thing you always do and hope for a different outcome.

The key in the rib crowned and supported design process, is to amass a certain amount of data which correlates to actual achieved piano sounds, with actual spring rates and beam dimensions,and overall statistical trends. Then change 1 parameter at a time to see either if the effect is what you are looking for, or in the direction you are looking to go, or not. Either way you learn something,as its empirical and the numbers, being numbers are repeatable.

Jim Ialeggio

Wow,
A lots been said while I was away plying the trade. More than I can respond to, but I want to touch on some particular points.

Keith,
When you say that crown is not a universally sough after value, are you refering to a bellying theory that does not include crown in some fashion at all? I'd be curious if you were.

Its interesting that you should bring up the difference between a light and a heavy board. The latest board is a Chickering 109-C. Compared to a Steinway this board seems almost delicate. A thin panel, light, generously spaced ribs, with no long flat tapers, they just notch into the first half inch of the rim. This board took on a very nice crown. Because of its lighter construction, the crown has a lot of spring to it. Do you think a high compliance board like this would respond better to a light bearing scheme or or a moderate? There is no danger of pressing it flat.

Dave B.
Are you saying that if you can string a board at a particular point in its post installation re climatization, then you can "engrain" a favorable behavior into the wood?

Jim,
Concerning the Off the wall crowns: We have not been getting too much crown because of something the wood has been doing. We have been getting too much crown because we have been puting too much in. We use a combination of compression crowning and custom curved caulding. We have been getting the crown we have been after, we were just after too much. In the last couple of boards we have reduced the curve in the cauls to lessen the final crown, and that has been successful. We still dry the board to form internal compression to complete the crown.

And do you really have to tell me that I cant avoid having redesigned any given board because of changes in the properties of the wood? That kind of stings. Redesign usually involves a discarding of original principles. I'm trying to preserve something, and keep my own footprint to a minimum. Can we say that I'm trying to "backward-engineer" the old boards? Still, I can understand what you are saying about stiffness.

I'd like to get a look at one of thos spring loaded go bars. I met jude at that same seminar, and he gave me a quick tour of his shop. Gotta love that climate control room. Its time I built myself one.

Thanks for all the input.

As you might have guessed, it's not a philosophy I agree with. Most of the rebuilt pianos I've seen with original soundboards were woefully inadequate from a performance and aesthetic perspective. While that isn't necessarily a cause and effect, it's a striking correlation. I'm not saying that there aren't situations where retaining the board isn't valid, or even a requirement, but even in those cases, I am satisfied with other methods of restoring the boards. Perhaps if I am ever in your area, I'll give you a chance to convince me otherwise.

Jim, again, thanks for the reply:



Of course there is always a difference between “as designed” and “as built.” It is interesting that you see a larger difference for compression crowned boards. All else being equal, that would be reason enough to prefer rib crowned boards.



Hmmm….

Can you see my point about it boiling down to whatever works? And if compression crowned boards are less predictable to make are you sure that this does not affect how you perceive the sound?



Ok, are there some rules of thumb, for what can be measured, that you have found do work? I imagine something like: “Given an as built spring deflection rate of 0.1 inch per 100 pounds, an unloaded crown of 0.4 inch, and a total string tension of 37,500 lbs; the plate should be set so as to result in a loaded crown of 0.2 inch and a downbearing of 0.3 degrees. But the average 150 lb per string tension for a spring deflection rate of 0.1 inch per 100 pounds is too great. It would result in a “pinched” sound. It is better to reduce the string tension a bit and lower the plate in order to keep adequate downbearing. ” I’m not sure if these numbers would describe a real life scenario, but I think the math works out with itself.

As a start, one needs to remember that the piano soundboard must respond adequately to low frequencies as well as high frequencies. Soundboard response and thus its stiffness in the treble will be different than response and stiffness in the bass. As well, differing frequency requirements graduate high to low and all the way in between the two. The soundboard must be able, in the appropriate region of the board, to respond with freqeuncy appropriate amplitudes.

Overall string loads vary with different pianos and makers, with low string loads being in the 500lb range and high loads in the 700lb range.

Looks like you have math. Get hold of a string scale spreadsheet (or write one in excel), get some sample scales and analyze them from this load perspective. Work the numbers, look for trends and then beat up on an experiment piano, preferably a grand in my opinion. This is what everyone who is building a good belly has done to some degree...work the numbers, look for trends, then woodshed it...using the boo-boos as high value information gathering. Boo-boos are good!

Jim Ialeggio

Thanks, Jim:

I would vary much like to get into all aspects of rebuilding, including belly work. It is just not in the cards in the foreseeable future. One thing I had consider, if the time ever actually came, is to have a way to actually measure the s/b impedence at different frequencies. You know, measure how much energy a s/b actually absorbs from a transducer.