July 2019 PTG Convention class: Voicing with B-72
Read the PTG Journal article here:
PTG Journal Article
Subject: Voicing with B-72
Professional Associate, AIC
Abstract: A useful method for hardening piano hammer felt can be found in a commonly used museum conservation material. Paraloid B-72 when dissolved in ethanol (190-proof grain alcohol) can be applied to the hammer using the same methods established for using other hammer hardening solutions. The principle advantages: 1) fast results, 2) readily available, 3) inexpensive, 4) less noxious odor, 5) has the desired effect of increasing power without adding stridency to the tone, and 6) can be used on hammers that have already been hardened with lacquer or other hammer hardeners.
For many technicians, B-72 can be added to their own voicing methods, since it can work in conjunction with lacquer, key-top / acetone solutions, and collodion.
Borrowing from the Conservation Lab
Paraloid B-72 (previously named Acryloid B-72) has a long list of uses around the museum conservation lab. It is an acrylic polymer that has two very useful characteristics. It is soluble in slower acting solvents like grain alcohol, and it remains somewhat flexible when dried. This makes it an ideal adhesive for bonding pottery shards, and for the consolidation of fragile wooden objects like wooden statues.
My interest in voicing piano hammers with B-72 started when I read a research paper published in the American Institute for Conservation Journal in 1994. At the time, I was struggling with hammer hardening solutions that used acrylic key-tops dissolved in acetone. It seemed that along with building tone, a nasty “zing” in the sound occurred when it dried. This article started my thinking that the fast evaporation rate of acetone was actually the problem.
The research was done by Arno Schniewig & Peter Eastman at UC, Berkeley. It was a study in which samples of wood were treated with B-72, and after drying, were examined using a scanning electron microscope. The B-72 was visible as plugs inside the wood cells, and by counting the plugged cells, they were able to plot the presence of B-72 at the surface. As they plotted the number from deeper slices, the degree of penetration could be documented.
When the test was repeated using another solvent, a big difference was observed.
When acetone was used as the solvent, the B-72 was found to be concentrated at the surface layer. But, when grain alcohol (ethanol) was the solvent, the concentration was away from the surface. This indicated that during application phase, the solution penetrated deep into the surface; but in the drying phase, the fast evaporation of the solvent created a reverse migration, drawing the consolidant back toward the evaporative surfaces.
This finding is important! It turns out that when we mix and apply a hammer hardening solution, that is not the end of the story! During the drying phase, the consolidant in the hardener can migrate along with the evaporating solvent back toward the surface. The faster the evaporation, the more the migration.
All consolidation treatments (including hammer hardening) require the same basic elements. The consolidant must start off as a solid, and when mixed with a solvent, become a low viscosity liquid. This is possible when the molecular bonds binding a solid material can be broken in the presence of a solvent. Think of the long chain molecules as beads of a necklace.
The solvent breaks the beads apart so they can freely move around as a liquid, but when the solvent flashes off, the beads are rejoined; again as a solid, but now in a different shape, (like that of a key top).
Another important factor in consolidation is the thickness of the liquid.
The degree of penetration is determined by the viscosity of the consolidant and the porosity of the surface being treated. A thick syrup of lacquer will not penetrate very far into hammer felt, but it will quickly drain through a sieve.
Some bench testing:
Flexibility of the treated felt is a necessary consideration in voicing. The hammer must be able to produce a warm tone on a soft blow, with progressively different tone possibilities on harder blows. This range of dynamic and tonal possibilities is not possible with unyielding felt. So I undertook some testing using strips of wool felt cut from the brim of a hat.
The hat felt is a uniform thickness, which made my testing repeatable. I used the arm on my digital height gauge to apply crushing force. As I slowly lowered the arm on top of the felt, I could tell when the felt stopped compressing because the weighted base began to tip. This movement was accurately detected with a dial indicator that was touching the base. As soon as the dial indicator moved, that was the maximum compression.
I measured these samples of the same thickness:
A strip of wood compressed not at all. 0%
Felt without any hardener compressed 30% .
Felt treated with B-72 in ethanol compressed 17%
Felt treated with B-72 in acetone compressed 9%
Felt treated with acrylic key top in acetone compressed 9%
This was not a scientific test, and it was not repeated enough times to be a statistical sampling. But, it was a quick effort that yielded interesting results that others may wish to pursue with further detail. It seems to confirm that the flexibility of B-72 treated felt is affected by the solvent that is used. By using acetone as the solvent, users may not be receiving the full benefits of B-72.
The first experimentation using B-72 to voice pianos was carried out while I was working at Northwestern University. With more than 200 pianos, I found many candidates and was able to follow the results over a long period of time. I eventually took it to the concert hall and piano studios to gauge the reaction from the faculty and visiting artists. I was able to interest a few trusted technicians, but we guiltily kept it as our own secret weapon, telling very few other technicians about the discovery.
By 2013, I felt confident enough to reveal this technique to a few more technicians. But as word spread, an important detail was changed. Technicians made the assumption that B-72 was to be mixed in acetone, with no mention of the original ethanol solutions. So, in 2018, I presented a class at the PTG Convention with my colleague, Dan Levitan. This class was very well attended and it is to be be repeated at the 2019 convention. Neither Dan or myself have ever used the acetone version, but many that have tried it claim good success.
Researcher and author David Koenig has done limited analysis on samples that I recorded using a digital recorder. Although not conclusive, one of the samples produced a result that is quite evident. In this sample using note c-5 on a Steinway grand, the before and after spectrograms detected an increase in the power and duration of the 2nd and 3rd partial, with a decrease in the power and duration of the first partial.
Four steps to trying this yourself:
Finding a small supply of B-72 should be easy. Do a web search of “museum conservation supplies”, or a search for “Paraloid B-72”. Here are two links to suppliers in the USA:
Finding high proof grain alcohol will be difficult, but some national-chain liquor stores in the USA sell “190-proof Everclear” on their website for shipping or pickup at the local store. (or in Europe the equivalent is “Navimer).
190-proof grain alcohol is best, but 150-proof will work. Alternatively, denatured alcohol can be found at the paint store and can be used as an alternative. The unknown percentage of methanol, formaldehyde, and other dangerous chemicals make this less safe to handle or breath.
3. Suggested mixing method:
Cut a 5” square of white bed sheet and place 10g of B-72 pellets in the middle. Use a length of string to tie the edges of the cloth into a ball, like a tea bag.
In a lidded glass jar , add 100ml of ethanol (190 proof grain alcohol) or denatured alcohol if nothing else is available.
Suspend the bag of B-72 above the bottom and tightly close the lid to trap the string.
Over the next several days, shake the bottle. When the bag is empty, it is ready for the next step.
Pass the liquid through a paper coffee filter and store in another container. Divide the mixture into two lidded glass jars; one for full strength, and one reduced to a medium strength by adding 50ml of ethanol.
4. Application using a needle oiler or .25ml pipette:
Apply 3 or 4 drops of medium strength on both sides of the crown, avoiding the strike point.
Apply several drops of full strength to the staple area.
Results should be evident in 10 minutes. Re-treat hammers as needed, (increasing the drying time with repeat application).
Apply directly to strike point when needed in upper treble, or just on two of three string grooves for better soft-pedal voicing.
The use of acetone may deprive the user of the full tonal benefits, without any real difference in the drying time. For those who use acetone, beware that re-treating an area may have a reverse effect, since the acetone will dissolve the previous layer. This will not happen if alcohol is the solvent.
Other uses for B-72
Before restringing a piano with an original pin block, I use a strong mixture of B-72 to flood the tuning pin holes. This consolidates the wood and results in a tight, smooth turning tuning pin.
Note: If cracks are present, the b-72 solution will leak from one tuning pin hole to the next while you are filling them. If this happens, the B-72 will not helpful. Instead, consider a low viscosity epoxy such as Epo-tek 301.
B-72 is also an excellent adhesive for gluing stringing cloth to the iron frame. A thick mixture will result in a viscous gel that can be painted on the plate without affecting the bronzing. it has a quick grab and makes clamping unnecessary. It is also useful to tack aliquots to the plate while stringing.
B-72 can also be added to French Polish to improve the handling characteristics and aid the the build rate.
And finally, the floppy brim of my old hat got a revival when I spritzed it with B-72 and clamped it flat. Now it’s my favorite hat again!
Can piano hammer leather be treated with hardening solution?
Preliminary testing by John Watson, Williamsburg. 2019-04-25
INTRODUCTION: While hardening solutions have been used for many years in adjusting the voicing of modern piano hammers, their use has not to my knowledge been successful on leather hammer covers in pre-1850 pianos. In fact, the appropriateness doing so has been unequivocally rejected by one of the most knowledgeable and respected restorer/makers.1 Nevertheless, the great difficulty of finding adequate leathers is reason to re- consider the issue and test new approaches from time to time. This paper reports on new experimentation on that front. The task is to approximate as closely as possible the nature of historical piano hammers using leather bulked with a resin for control of the leather’s tone-producing characteristics. Note that piano hammers typically were (and remain today) covered with layers in a way that make them harder at lower layers.2
MATERIALS: The leather I used in this testing is thick Elk chamois leather from the German firm Friedrich Herzog & Co. It looks similar to the period leather often seen in early 19th century English square pianos but when untreated is too soft and produces a muffled tone. The hardening resin I tested is Paraloid B-72, a thermoplastic resin from Rohm & Haas. This resin has long been used by conservators because of its solubility in several solvents and its stability over time.
PROBLEM: Reverse migration occurs when using resin solutions in the consolidation of porous materials. After the leather is thoroughly impregnated with the solution, the resin follows the solvent to the surface where the solvent evaporates, leaving the resin more heavily concentrated at the surface, giving the leather in effect a hard exoskeleton. Some resin does remain on the interior as air begins to reach back into the porous substrate causing some evaporation to occur throughout the substrate. Due to reverse migration, hardening solutions applied to a piano hammer results in hammers that are hard on the outside and softer on the inside, causing an unwanted “zing” when the hammer is in contact with a vibrating string. This would be a principal reason why consolidants have been unsuccessful for leather hammer covering.
THE HYPOTHESIS: If thick leather can be consolidated with a solution of B-72, allowed to dry on a plate of glass then skived (split) approximately in half, the result will be two pieces of leather, each relatively soft on one side and harder on the other side. If glued to the hammer head with the soft side facing out, the hammer will have the conventional construction with a soft outside and harder inside and will produce the required tonal affects. The concentration of the resin, the thickness of the leather, and the relative hardness of the top, bottom, and inside of the leather are all variables that can be adjusted to achieve a musical result.
1 See Christopher Clarke, “Fortepiano Hammers; A Field Report” in Keyboard Instruments – Flexibility of Sound and Expression (Proceedings of the harmoniques International Congress, Lausanne, edited by Thomas Steiner and published 2002 by Peter Long. pp. 225-258) Clark’s comment about hardening solutions is on p. 247, “Hardening solutions have absolutely no place in our work.”
2 For an explanation the elastic properties of hammers and their dynamic hardness increasing with velocity of the blow, see Clark op cit. p.225-228. Clark’s article provides an excellent summary of piano hammer variables and how they affect tone.
TEST 1: A sample of the leather was vacuum impregnated with green dye in acetone and allowed to dry.
Result: The test indicated areas of the leather collected different concentrations of dye (in decreasing level of concentration)
- Upper edges (heaviest concentration)
- Top and side surfaces
- Bottom surface
- Interior (lightest concentration)
TEST 2: Three leather samples were vacuum impregnated with different concentrations of resin
- 1% B-72 w/v in acetone
- 2% B-72 w/v in acetone
- 10% B-72 w/v in acetoneResult: I judged the 1% and 2% solutions to have too little affect. The 10% solution resulted in considerable stiffening of the leather with the differences of hardness as predicted by the dye in Test 1
TEST 3: Various methods of processing the leather treated with 10% B-72 and trial on piano hammers, always orienting the harder surface to the inside of the hammer
- Fairly extreme bending to loosen the stiffened leather, then skiving the leather approximately in half provided two options. The lower half had less resin as predicted by test 1.3 and 1.4. It was too soft for the hammers.
- The top half of the leather was then tried on the hammers. It resulted in considerably better tone.
- Another sample of the leather treated with the same 10% solution but with slightly less bending gave a little more firmness to the leather and leaving some hammers slightly too bright but receptive to final voicing by manipulating the leather on the hammer. Voicing responds to needling with very fine needle and to hot ironing.
Figure 1. Green dye migrated to the surfaces during drying. The left side of the sample is cut off to show interior
Figure 2. After impregnation in this vacuum chamber, the leather strips dry on plates of glass, encouraging most evaporation to be from the top.
Figure 3. Leather is split after the resin treatment. This exposes the softer inside which is best for the outside of the hammers.
The preliminary results of this testing suggest the hypothesis is correct. It is possible to cover piano hammers with leather, the hardness of which can be controlled by the use of B-72 in acetone. The reverse migration of the resin can be used to advantage in producing leather that is softer on the outside and harder on the inside of the hammer if the leather is skived after impregnation and attached with the hard side down. Several variables can be adjusted to control the effects: the concentration of the resin, the amount of bending the stiffened leather, the thickness to which the leather is skived, the choice of the top layer or bottom layer of leather (the top having a heavier concentration of resin at its top surface than the bottom layer at its bottom). Other resins, solvents, or leathers were not tested. It is unlikely this system can be successful when the leather is thin and not to be skived to thickness after impregnation.
Do hammers constructed this way have the elastic characteristics needed to respond to the pianist’s touch for tone colors from soft to loud? First trails in the hands of fortepianist Andrew Willis suggest a resounding yes. The piano used in this testing is by Loud & Brothers, Philadelphia, 1826, and is owned by Preservation Greensboro (NC) and exhibited at Governor Moorehead’s Blandwood Mansion.