I read an interesting article by Gene Wengert The Wood Dr. in a furniture industry rag I receive monthly. https://www.woodworkingdigital.com/wwnet/library/item/3946503/
The gist was, if my understanding is correct, wood shrinkage, the first time it occurs, when the wool experiences a new low EMC, will always have a greater dimensional change in the loss part of the cycle (the first time only at that new EMC low), than when it returns to the original EMC. There is a significant difference between dimensional reduction moving from higher EMC to lower EMC (the first time), than the dimensional expansion, when the wood recovers its previous EMC. Repeated loss cycles to that previouis low EMC will not be as aggressive a dimensional loss as the first time the low EMC was attained.
This explains the problem with various dimension based Soundboard EMC gauging methods, including one I and others use, which is using a gauge stick to record and monitor dimensional changes in the soundboard preribbing.
The Forest Service dimensional charts we use are based on first time drying. They are not predictive of cycled expansion/shrinkage.
It also suggests to me, that, before ribbing, it makes sense to reduce EMC beyond the worst we expect to see in service, and then, at least in my own case, let it pick up moisture to 6.5% EMC before ribbing.
It's interesting to see him discuss something I have seen over and over again in any composite wood structure. That is, if the assembly makes it through the first year of cycling, it will be in a good position to survive the long haul. The first year, gremlins will mostly appear, if they are going to appear. It also may explain why some compression boards fail very early (not all, but some).
It's also why I like using old boards if that is possible. If you can assure the glue structure is excellent, the system is more stable, and behaves in a unique way, relative to new wood. However, and this is a big however, glue structure must be carefully tested and improved if necessary (almost always necessary in my experience), as the glue is a huge part of the structure. Another plus in an old board, with age, cellulose, lignin and hemicellulose degrade, and the wood becomes lighter and more resonant, and I think a little less stiff...something I have noticed repeatedly when messing with old boards. They have a resonant and relaxed sound, their juvenile cousins lack, assuming the glue joints are excellent and of course bridge terminations are new.
Maybe. But the first time the wood in a panel is dried is not when you're making a soundboard out of it. It's when the wood is milled into boards or even before. By the time you make a soundboard the wood has gone through multiple cycles sometimes over years and the effects of hysteresis are diminished.
On another point I would consider 6.5% EMC to be inadequate to avoid the panel cracking later during dry periods. To get to 6.5% EMC you're talking about 80 degrees (F) at about 35% RH. I would consider that risky and certainly inadequate for any significant compression.
I think we have different experiences. I've seen boards ribbed at 6-6.5% form cracks during periods or low humidity but not single digit. And that's in Northern California where humidity levels are generally fairly benign. My goal is to not let the board ever go under tension which means an EMC at ribbing and prior to installation of 5% in my area. I think if I were sending a board to Arizona I'd dry it down further. That is achieved by controlling the hot box RH and temp for an adequate period of time to cover the panel thickness There are charts to help with that calculation. I agree that cutoff bars seem to exacerbate the problem of cracking but not if the assembly is dried prior to installation.I always thin the panel on the bent side corresponding to about the length of the rib taper. Most at the bass end. Second most right behind the top capo section. Less behind the killer octave section. Lightly on the spine and very lightly at the belly rail. I don't see thinning as creating any problems but I do crown ribs (typically about 18M radius throughout so I'm not relying on compression totally for crown formation. I'm not a fan of very tight rib radii
Sitka panels get more thinning than white spruce and sitka also seems to have less of a tendency to form pressure ridges. With white spruce I don't thin along the spine or the belly rail. With Sitka I do and I also thin slightly under the length of the long bridge.
I have measured panel shrinkage with multiple dryings (not ribbed) and have not noticed any significant differences In the degree of shrinkage from one drying to the next. I think the initial drying of the wood may well result in more shrinkage than later drying. But by the time I get the panel the wood has been cured, meaning multiple dryings over time. I don't really worry about the hysteresis effect that you mention. So far so good.
ChrisHonestly I have no idea what you're talking about.
Two different issues being addressed. One is drying the panel, the other is thinning the panel,Controlling the drying of the panel is being way overanalyzed. Put the panel in a hotbox controlled for temp and RH to achieve the emc you want, simple process. If you are working in a high humidity environment have a way to determine if they panel is changing when out of the box. If it is, put it back in the box. No brainer. I find the Nossaman gauge does that very well.
I don't know anyone drying panels to 3%. The panel will crown at 4%, 5% or 6% if the final ambient humidity is higher than that level. If there are differences in the crowning due to panel differences or dry down levels that can be taken care of when setting the bearing since soundboards are non linear springs. Want it stiffer? Set more downbearing or adjust later. Adjustable perimeter bolts help there. If you dry down the panel to higher emc levels you might have problems with cracking or inadequate crown during period of low humidity. Choose your poison.Re panel thinning, I don't see any reason to further restrict movement of the panel at the rim if the rib scale is adequate to support the structure and create adequate stiffness. I find panel thinning a benefit especially in the bass and high treble, and not thinning the panel robs the piano of some power while not necessarily enhancing sustain. Sitka panels, I find, need more aggressive thinning than white spruce. YMMV but I don't see slight variations to be problematic. They just create slightly different outcomes.
All these superlatives about what works and what "can't possibly" are so much hot air.
ChrisI didn't ask about how you get the board 3-5 lbs lighter in the post you responded to. I take you at your word however you do it. I just don't think it makes sense to just remove mass for its own sake. The rib scale determines the mass mostly. The panel, other than some slight variations in thinning, is more of a set quantity. I determine the rib scale by what is required in terms of support not targeting some amount of weight. If I wanted less mass in the ribs I would make them taller rather than wider since that allows you to reduce the volume of the rib without sacrificing strength or support
As far as cutoff bars, when I install them, which is not all the time, they are simple straight cutoff bars in the bass corner, only enough to get the longest ribs to under 1 meter (39"). That usually only affects 3-4 ribs not unlike what you would see in a Steinway D. I'm doing a 6'4" Knabe right now with a small but curved cutoff bar that was part of the original design. It too only affects a few ribs and brings them down to a similar maximum length. Seems to solidify the mid tenor some. I'm not convinced by very large cutoff bars though they do make the working part of the assembly lighter if that's the goal.
I like Jim‘s idea of sailing when the humidity gets too high except for the fact that I’m prone to getting seasick
"Good, better, best; never let it rest, 'til the good is better and better best!"
"Providing quality service for the world's pianos"
George W.R. "Bill" Davis, RPT
The Piano Place GA
2315 Rocky Mountain Rd NE
Marietta GA 30066
I always had a deep love for music and the instruments that make music. I built my first instrument back in 1978, It was a type of harpsichord called a Pentagonal Virginal. I came across a little booklet called, "Italian Harpsichord Building in the 17th and 18th Centuries" by John D. Shortridge. It gave just enough details for me to draw up the plans. I already had 4 years of mechanical and architectural drafting experience because in High School drafting was my favorite class.
From there I built several more virginals, harpsichords, clavichords and Appalachian Dulcimers.
In the 90's I made the transition to piano rebuilding. After several years of learning the basics of piano technology, I moved on to the more advanced aspects such as pinblocks and soundboards.
There were many experiments along the way. One such early experiment was to freeze a piano, (excluding the action, plate or strings). Sounds silly, but I read that violin tops are sometimes frozen to manipulate and lower the woods vibrating frequency. The idea of this is to end up with a violin with a warmer tone. Freeze to Warm, hmmm? Freezing a piano though had the opposite effect. The vibrating frequency went higher in pitch. This is because during the freezing process, a violin top is normally removed from its support, and a piano soundboard is not when freezing. Anyways, it was just fun to look at everyone's faces rolling a piano into a sub zero meat locker.
After years of rebuilding I wanted to move on to Piano Design and manufacture my own pianos. In 2010, I designed and built a 7' tall Upright piano with a repetition lever in the action. I called it the Mammoth VCG.
During this time my wife and I decided to move out of the city and into the country. We bought a fixer upper on some property in East Tennessee. I spent the next five years fixing up the place, building a shop, clearing land, doing piano work part time, and remarketing my business.
Eventually, the pianos started coming in on a regular basis, and I was putting in new soundboards one after another. Thanks to the higher output, I was fortunate to be able to compare well known brands to many obscure brands. I began noticing that some soundboards sound so much better than others do. So I had to learn why!
The first thing I did was to collect data on every soundboard I got my hands on. I then created charts and spreadsheets to be able to analyze the data. I spent two years building up a database that now has the info of over a hundred piano soundboards. The first thing I learned is that no two soundboards are the same. I once had 3 Baldwin R's side by side. They were all good sounding pianos but one in particular was much better than the rest. It was more responsive with a richer voice. The word to describe it would be quality. Quality is such an interesting word because there is no real definition, but we all know it when we see or hear it.
It wasn't the action either because I could switch them between the three.
So after years of scientific and empirical study, hundreds of experiments, hundreds of pianos, here is the improvements on a compression crown soundboard that I incorporate to greatly enhance the sound.
I Use the scalloping to control the deflection values on each end.
Pushing on the bridge location, the deflections on each side of the rib is made to match. This makes the bridge the center of amplitude (maximizing bridge movement).
The rib scale is engineered with the proper safety factor built in. With an emphasis of NOT over-engineering nor under-engineering. When I studied the 100 soundboards from different makers, it became evident of what was normal, and also, what was not enough or too much regarding mass.
Maximum rib length should not exceed 45" as the Moment of Torque becomes too high with proven failure from fatigue.
The rib scales are further designed to be even, so that the load is distributed evenly amongst them and create smooth stress curves.
The Section Modulus must match the length curve. Basically, the longer a rib span is, the larger the rib should be.
The panel is thinned in such a way, so as to maintain that the rib scallop deflections stay in balance with the others.
The Strength to Weight ratio is increased. The boards maintain their Strength value but weight is reduced. This is possible because the bridge is off center and creates an uneven work load across the rib length, so unnecessary wood can be removed from the long side.
The bridge cap must have a Janka hardness of 1800 psi to avoid false beats and have a cleaner tone.
I often slightly lengthen the string scales and string segments. 53mm for .031 has become my standard.
A few months ago I created a new downbearing setting method for setting a light, even downbearing. Insuring that a board is not overloaded and choked. Although it adds much accuracy to the process, its very time consuming.
Last but not least, yesterday,I just created a new stringing procedure for installing new strings on a newly reshaped capo, that avoids unnecessary in the wrong place grooving. Hated that.-chris
I like Jim's idea of sailing when the humidity gets too high except for the fact that I'm prone to getting seasick
ChrisI don't make ribs taller, shorter and narrower to reduce weight. I said one could. Please read more carefully.
You have frequently commented that you reduce weight by 3-5 lbs as SOP. I don't worry about the weight of the assembly. I chose panel thickness based on common standards, usually about 9 mm. I thin the panels as I have described. I design the rib scale to provide the required support based on the string scale. The weight falls where it falls. 3-5 lbs reduction is equivalent to a couple of entire ribs. I don't get why you would feel compelled to simply reduce weight. But to each their own (how's that for pronoun gender neutrality 😉 I'm getting woke).
Since we haven't ever compared assemblies in terms of weight or anything else I wouldn't draw too many conclusions about what you think our differences are.BTW on panel drying we haven't mentioned drying down the panel to the ribbing EMC before you cut the panel to fit the rim. That's pretty important too. It can shrink a lot. So in addition to the panel, or flitches, being cured before you even get them, the panel is dried down prior to fitting to the case and then dried again prior to ribbing and yet again prior to installation.
This is a complicated question.
One the one hand, I am of the opinion, that we need to be able to "tune", correct or tweak a board, before stringing, as after stringing its it too expensive to do any serious tweaks. I feel strongly that any board will benefit from pre-stringing tweaks, because fabrication idiosyncrasies will be unique to each and every individual fabricated board regardless of design or construction technique.The thought behind the Chladni patterns is, that we need a tool to "read" the fabricated assembly, pre-stringing. This concept makes sense to me...not necessarily the Chladni test in particular, but the utility of getting a coarse "read" of the board before it's too late. I have thought about the utility of Chladni patterns, messed with it some, and determined to go another direction. The main reason being, that Chladni patterns can be wildly misleading, given the loads we impose on the structure.
Case in point...all manufacturers are inconsistent in their build results. Steinway is famous for this inconsistency, but all manufacturers (and rebuilders) experience it. My current project is an another wildly expensive "S" piano, concert grand. The board was a dog right off the line. The piano never sold. My customer ended up with it after a difficult and unhappy legal proceeding with a dealer. Said "S" company makes a big deal in their marketing, as to how they read the Chladni patters and hand plane the entire board according to their proprietary procedures. I mapped the board thickness/thinning profile, in the piano (I am not replacing the board) and spent a fair amount of time ascertaining exactly what failed. Note, all glue joints are sound, there are no cracks, and compression ridges are present and reasonable. My read is that, in following their Chladni tests, they got the board to respond to a Chladni pattern, and in the process destroyed the board. The rib structure is fine, and I have proved this to my satisfaction. However, the board is thinned everywhere, not just at the near rim, but all the way to the bridge. One would not have just thinned a board like this, this aggressively, unless they were following some direction indicated by real time tests of some sort.
I think learning to read an unstrung board is essential for determining whether the board is too close to the non-functional edge of too flexible in some area or in the rib scale, or in some other parameter. My own process, at this point, which I am still experimenting with, is to remove bridge pins, and with a double thickness of back rail felt on the cap, strike the felt, at each rib location (only on the bridge not on the panel), with a small ball peen hammer. Then, listen to the full envelope of the board response following an impact, and compare it to its neighbors. In this case a number of issues become obvious, even to a casual listener.
1- one rib #8 (note 36-ish) produces an absolute thud, compared to ribs on either side which are producing an organized tone for at least part of the sound envelope. The thud is obvious and matches the location of a serious wolf area prior to de-stringing. This area, by the way, is an area which I have noticed this deficiency, almost universally, on many long pianos.
2- ribs 3-7, and 9 responded to the strike with an organized tone, followed by break- up of the tone. Again, a fairly obvious aural effect.
3- ribs 10-12 (notorious 4th-6th octave) no sustain of any tone to speak of. This matches the treble complete lack of sustain previous to de-stringing.Combining that aural evidence, with thickness mapping, and visual inspection of system's glue joints, I formulated several hypotheses. I use mass loads and springs to both diagnose, and test hypotheses. Then, regarding ribs 10-12, which were the worst area of deficiency, applied a 3mm veneer to the acute bent area of the near rim out to where the board was 8mm minimum. Grain at 10 degree from the original. Graduated thinning, starting at the perimeter near the bridge. Ribs 10-12 now produce a well defined sustain at the strike, which is similar in envelope to the effect ribs 3-7 and 9 produce. The veneer is almost all under the plate, and minimally visible. They all still tend to break up at the end of the envelope, but exhibit an ability to sustain tone for some amount of time. The end of the envelope matches end of envelope break up of ribs further down the scale.
After some further mass loaded diagnostic tests at the rib tapers, I will be adding strips of spruce along the ribs at the rib tapers ribs 10-12 on the other side of the bridge at the cut off bar. After doing that I will assess whether ribs 3-7 & 9 need the strips of spruce, or whether loading will increase impedance enough to control the end of envelope breakup. I know that mass loads at the rib tapers correct the breakup, so I could wait to after stringing to do that, but would prefer to add the rib taper support with spruce, if it doesn't look too weird. My feeling is, that I want the unloaded board to have enough impedance at the near-rim, to sustain without breakup, under zero load, as a safety factor.
Also, diagnostic rib mass loads, at the rib flare on ribs 3-7, correct the thud at rib 8 by about 80%, so that is another argument for adding the spruce strips spine side, ribs 3-7
The protocol development is still in process, but I do think any board would benefit from pre-stringing real time tests of some sort. Not Chladni for me, but some sort of indication of what how the structure will perform when strung.By the way, mass loads applied at the rib taper, have a very different effect than the coarse application of mass under the bridge, in terms of damping. I still would like to avoid them if possible, but I have not found a piano that did not benefit in some way from targeted precision minor mass loading, at some point in the scale. Reason being, we are not perfect and no board is perfect.
JimI don't disagree that having a pre-installation method of assessing the assembly would be a nice thing. However, I'm not sure it's really possible at this point. I think the only sure fire method is to build a whole bunch of soundboards and track the changes that you make, preferably one change at a time, so you can limit the number of variables and isolate what affects what. I'm not sure there are any shortcuts.
I think there are many manufacturers who produce very consistent results. I don't think it's as much of a crapshoot as you suggest. Steinway has some different outcomes but I'm not sure what the reason is for that. Having taken apart a lot of Steinway pianos I can see that rib scales, panel thickness and thinning, grain density all vary. I don't know what kind of consistent control they have over EMC during the ribbing process. But many manufacturers produce very consistent results: Yamaha, Kawai, Bosendorfer, Bechstein, Fazioli to name a few. Hamburg Steinway I find more consistent than NY Steinway and there are others that do a very good job of consistent results IMO. For rebuilders who have more limited production capacity it's more difficult to assessBut what I find in many of these discussions are claims about finding the holy grail that just aren't convincing to me. Even a blind squirrel finds a nut sometimes. But I think we need to be careful about the conclusions we draw as to what was responsible for positive or negative outcomes unless we really take pains to control the variables one at a time and see consistent results over many iterations.
3- ribs 10-12 (notorious 4th-6th octave) no sustain of any tone to speak of. This matches the treble complete lack of sustain previous to de-stringing.Combining that aural evidence, with thickness mapping, and visual inspection of system's glue joints, I formulated several hypotheses. Using mass loads to both diagnose, and test hypotheses. Then, regarding ribs 10-12, which were the worst area of deficiency, applied a 3mm veneer to the acute bent area of the near rim out to where the board was 8mm minimum. Grain at 10 degree from the original. Graduated thinning, starting at the perimeter near the bridge. Ribs 10-12 now produce a well defined sustain at the strike, which is similar in envelope to the effect ribs 3-7 and 9 produce. The veneer is almost all under the plate, and minimally visible. They all still tend to break up at the end of the envelope, but exhibit an ability to sustain tone for some amount of time. The end of the envelope matches end of envelope break up of ribs further down the scale.
After some further mass loaded diagnostic tests at the rib tapers, I will be adding strips of spruce along the ribs at the rib tapers ribs 10-12 on the other side of the bridge at the cut off bar. After doing that I will assess whether ribs 3-7 & 9 need the strips of spruce, or whether loading will increase impedance enough to control the end of envelope breakup. I know that mass loads at the rib tapers correct the breakup, so I could wait to after stringing to do that, but would prefer to add the rib taper support with spruce, if it doesn't look too weird.
Quit being such an ass and find a new translator while you're at it.
Changing one variable at a time is basic science if you want to know the effects of that change. Independent builders have a harder time doing that compared to manufactures who make hundreds or thousands of pianos a year. That's just a fact. The independent builders I know spent years experimenting with different iterations, not all successful, and for many independents that's burdensome.
I have no problem with Jim and I doubt he needs you to translate. I find Jim's questions compelling and well thought out and I engage him in these discussions because of that and because they raise other questions for me. Your arguments, on the other hand, I find to be mostly shameless self promotion masquerading as junk science.
I am hardly a contrarian. My approaches are fairly conventionally at this point and that after much experimenting with new ideas and approaches whose outcomes I didn't particularly care for.
But please, review your rules governing logical fallacies. To misstate my questions and arguments and then attack them based on your own convenient interpretation falls clearly in that category.