Pages:
1
2 |
jdowning
Oud Junkie
Posts: 3485
Registered: 8-2-2006
Location: Ontario, Canada
Member Is Offline
Mood: No Mood
|
|
A Low Cost Hardness Tester for Wood
Forum member spyblaster recently questioned if maple might make a good wood for an oud fingerboard ("Question about Woods" topic). It was thought that
maple was a bit too soft to make a durable fingerboard surface - unless it was chemically hardened as is the case with some commercially available
guitars. In depth hardening of fingerboard wood requires costly specialised equipment (pressure/vacuum chambers). Discussion then centred around
whether or not a fingerboard surface might be adequately hardened to a durable condition by soaking with low viscocity penetrating chemicals such as
super glue.
I have undertaken to run some trials to test if maple might be chemically treated to increase its hardness without resort to costly equipment. This
will involve not only use of chemicals such as low viscocity cyanoacrylates but also chemicals designed to penetrate and solidify rotted wood as well
as chemicals that have been available for centuries. These trials will be undertaken later as a separate topic on the forum.
Before any meaningfull trials can take place some means of measuring relative hardness of the wood under test with reasonable reliability must be
established. Hardness testing of wood is undertaken by the industry using the Janka apparatus where a steel ball of 11.28 mm diameter is forced into
the wood under test to half its depth. The force applied (in pounds or kilogram force etc.) is then used as a measure of the material hardness.
This test is similar to those used to test metals where a constant force is applied to a metal sphere or diamond point and the diameter or geometry of
the impression measured to arrive at a number expressing the hardness of a material according to a particular test method. The hardness numbers
obtained are not absolute values but are used to compare the relative hardness of materials.
The Janka hardness numbers for wood are used, for example, by the wood flooring industry to determine the durability of materials used for flooring
against deterioration and wear from indentation loadings and abrasion.
Commercially available hardness test equipment is expensive so a low cost alternative will be described here to enable any luthier or researcher to
carry out relative hardness testing - hopefully with results that will compare reasonably well with published Janka hardness values. This is a work in
progress - so there is no guarantee, at this point in time, that it will all work out successfully in the end but no harm in trying.
|
|
jdowning
Oud Junkie
Posts: 3485
Registered: 8-2-2006
Location: Ontario, Canada
Member Is Offline
Mood: No Mood
|
|
The 'ready made' tool to be converted into a compact and portable hardness tester is an 'automatic centre punch' familiar to metal workers. It is a
spring loaded punch designed to put a small dimple in metal that acts as a centering guide for a drill in preparation for the drilling operation.
Pressing down on the punch releases a spring loaded weight that strikes the punch with a constant force - all contained within the body of the punch.
Very convenient as there is no need to strike a punch with a hammer in the usual way.
The cheapest Chinese made automatic centre punches cost between $5 and $10 - quite well made with solid brass bodies. The punch on the one that I
bought a few years ago did not have a hard enough steel point for striking metal so failed in service. The tip looks as though it was just hard chrome
plated in place of a properly hardened point - the softer steel underneath prevailed!
The design concept behind the adoption of this tool as hardness tester for wood was to regrind the tip to a 90° cone. This was done by mounting the
tip in a drill press and grinding the tip to the required shape - as it rotated in the drill - with a Dremel grinding tool.
In use the auto punch is simply held vertically and pressed into the wood surface until the punch activates - leaving a cone shaped dimple in the
wood. The harder the wood the smaller the diameter of the dimple. I tested the force exerted by the punch using a digital bathroom scale. At the point
at which the punch released its impact force the force measured on the scale was 34 pounds weight. The impact force would be somewhat greater than
this - exactly how much does not matter as long as the force remains reasonably consistent between each operation of the punch.
The reason for this particular geometry of the tip is that by simply measuring the diameter of the indentation and squaring the result (i.e. multiply
the diameter by itself), the number obtained may be used as a measure of relative hardness of woods under test.
Furthermore, if the tool works, it should be possible to obtain an approximate correlation with published Janka hardness test. Note that due to the
great variability of wood as a material the published Janka values are only accurate to ±20%.
Measurement of the indentation diameter can be estimated using a scale with suitably fine graduations be it a mm scale or
a 1/60th inch draughtsman scale (and a magnifying glass to accurately read the scale) - it doesn't matter what the absolute measure is, only the
number of chosen units matter.
So, for example, the indent made in a sample of Black Walnut shown in the attached image measures 7 units units on the 60 divisions to an inch scale
or about 5.2 units on the millimeter scale. So the relative hardness number using the former scale would be 7X7 = 49 or for the latter 5.2X5.2 =
27.
For the 'rocket scientists' my calculations concerning the geometry for the punch tip will follow.
[file]25320[/file] [file]25322[/file] [file]25324[/file] [file]25326[/file]
|
|
jdowning
Oud Junkie
Posts: 3485
Registered: 8-2-2006
Location: Ontario, Canada
Member Is Offline
Mood: No Mood
|
|
The attached image is a copy the design calculations that I have used to determine the geometry of punch of the wood hardness tester - for general
information.
Nothing very complicated and not beyond basic high school geometry.
I have chosen a special case geometry for the tip of the punch - a cone whose height is the same as the radius of its base. This means that the length
of the slant of the cone can be conveniently represented as proportional to the radius of the base.
This device provides a constant load to the punch each time it is operated causing the punch to penetrate into the wood under test a certain distance
- greater for softer woods and less for harder. The resistance provided by the wood is a pressure effect. Pressure is defined as force per unit area -
so it is the surface area of the tip that determines how far the tip will be driven into the wood under test.
Hardness tests do not provide an absolute value of hardness - just a number (dependant upon the test procedure used) to allow relative hardness
comparisons to be made between materials.
For this little device it so happens that a relative hardness number (lets call it the 'Downing' hardness test number for want of a better term) is
obtained by simply measuring the diameter of the indentation (D) made by the punch tip, multiplying that figure by itself (D²) and calculating the
reciprocal 1/D² which is the relative hardness number
So on the 'Downing' hardness scale a low number will indicate a softer material than one with a high number.
Futhermore by testing known wood species it should be possible to convert 'Downing' hardness to published 'Janka' hardness - at least within the
accepted tolerance of ±20% ?
So the next step is to test a number of wood species to see if this simple little hardness tester will provide reasonably reliable results comparable
to the Janka industrial standard.
This will take a little time - more to follow.
|
|
fernandraynaud
Oud Junkie
Posts: 1865
Registered: 7-25-2009
Location: San Francisco, California
Member Is Offline
Mood: m'Oudy
|
|
We don't know how hardness measured by indentation relates to resistance to "minute (in amplitude) transverse and longitudinal string vibrations", but
this is wonderful.
|
|
freya
Oud Junkie
Posts: 276
Registered: 2-8-2004
Location: Asheville NC
Member Is Offline
Mood: Happy in the mountains of WNC
|
|
John,
I stumbled across this when I was trying to figure out various ways to test the hardness of different woods treated with Tung oil, CYA, Nitro, etc.
You may not have it in your library.
http://pac.iupac.org/publications/pac/pdf/1965/pdf/1003x0239.pdf
Best regards
|
|
jdowning
Oud Junkie
Posts: 3485
Registered: 8-2-2006
Location: Ontario, Canada
Member Is Offline
Mood: No Mood
|
|
Thanks Harry - No, I have not seen that report before. I have only quickly read the first few pages and it all looks clearly written, very relevant
and useful particularly in defining the problem . Will run a hard copy off and hopefully read it this evening.
The subject matter appears to focus on the problems of testing thin film coatings over harder or softer substrates that is certainly of interest when
it comes to whether or not fingerboard coatings have a significant value in preventing or slowing down wear due to string impingement under pressure.
It would appear that even among experts in this specialised field there is much disagreement.
In order to test thin films some kind of 'microtesting' apparatus is required. If an indentation hardness method is used the depth of intentation
would need to be quite a bit less than the film thickness I would imagine - so I am interested to learn about the test procedures and apparatus
recommended in the paper.
Clearly my simplified hardness testing tool will (hopefully) provide useful measurable relative results or even results that can be correlated with
the Janka tests. This test procedure is of course a macro test with maximum indentations of perhaps a mm or two in depth - so will really only be
practical (hopefully) in testing wood samples of a cm or so in thickness - including chemically treated samples have been impregnated through and
through. But at least this will be a start on the right path.
Fernandraynaud - I have to suspect that a very hard wood will have a much longer service life than a softer wood - perhaps an 'indefininite' service
for most users. I have my doubts about the long term usefulness of a maintenance dependant thin coating in preventing wear. However, full impregnation
of softer woods to make them harder and more durable seems a more practical solution (allowing use of cheaper, more readily available hardwoods)
particularly if this can be readily undertaken at reasonable cost by a luthier in his own workshop.
The little tester is now complete and ready for use so will try to find time to run a few preliminary tests today on known species of hardwoods that I
have to hand (Eastern White Cedar - relatively soft to Brazilian Rosewood - relatively hard) just to get some idea if the tester is a working
proposition.
Should be interesting.
|
|
jdowning
Oud Junkie
Posts: 3485
Registered: 8-2-2006
Location: Ontario, Canada
Member Is Offline
Mood: No Mood
|
|
A few brief trials with the tester were made this afternoon - mainly to evaluate the potential range of the tester and to identify potential problems
with its use.
It was decided to restrict hardness testing to only hardwood samples. The softest of these to hand is so called Yellow Poplar (not a Poplar at all but
genus tulepifera - a large beautiful tree native to the Southern U.S.) and the hardest, Brazilian Rosewood. These woods have a published average Janka
hardness (±20%) of 540 lb-ft for the former and 2720 lb-ft for Brazilian Rosewood.
The respective indentation diameters produced by the hardness tester measured 9 units and 4.5 units. Squaring the indentation diameters gives a
'Downing' hardness value - according to this specific (and crude) test method - of 81 and 20 respectively. The ratio of 'hardness' value - soft to
hard wood is, therefore, 81 /20 (= X 4.05) - so if the average published Janka hardness for Yellow Poplar (540 lb-ft) is taken as the base point, the
relative value for the Brazilian Rosewood works out to be 540 X 4.05 = 2187. Close but a bit on the light side perhaps? - until one considers the
practical 20% tolerance limit either way of wood samples tested under the Janka test that results in a range for Brazilian Rosewood of between 432 X
4.05 (= 1750 lb-ft) and 648 X 4.05 (= 2624 lb-ft) calculated from the Yellow Poplar base. This falls nicely within the published range of Janka
hardness values for Brazilian Rosewood of 2176 lb-ft to 3264 lb-ft. calculated from the average published Janka value of 2720 lb-ft.
So, in this example the calculated equivalent hardness value comes reasonably close to the average Janka hardness value for this material.
Nevertheless, some improvement in the test equipment or procedure is indicated in order to provide a closer corellation - if at all possible.
There were some problems noted during the test that may predudice the relative comparative accuracy and hence repeatability of the results. The
conical punch created compressive/shear forces that caused surface fractures in both soft and hard woods under test. Waxing the punch tip helped to
produce a more clearly defined indentation.
1) the ground finish of the punch tip is still quite rough - which might affect reliability of the measured values due to excessive friction(?).
Action: the tip will be abrasive polished to a mirror finish.
2) Measurement of indentation diameters using a finely divided ruler is awkward and not very accurate in assessing and correcting any assymmetry in
the diameter measurement An alternative simple but more accurate tool (a comparator) will be made consisting of holes of accurate diameter drilled
into a sheet of thin clear plastic. In use, the comparator will be placed over an indentation to determine its average diameter.
3) To reduce material distortions due to the impact forces applied by the tool, tests will be made to evaluate any improvements that might result in
reducing the impact force odelivered by the punch main spring (i.e. use a less 'powerful' spring).
So far so good!
[file]25352[/file] [file]25354[/file] [file]25356[/file] [file]25358[/file] [file]25360[/file]
|
|
fernandraynaud
Oud Junkie
Posts: 1865
Registered: 7-25-2009
Location: San Francisco, California
Member Is Offline
Mood: m'Oudy
|
|
Looking at this paper we can see the enormous complexity of the issues.
In the case of a coated fingerboard, the entire objective is to extend the interval between leveling/planing operations, as these are not trivial. The
resulting lowering of the fingerboard calls for adjustments to the bridge, nut and neck, that are especially difficult if the instrument is as simple
as an oud. Where labor costs are high, this is a real problem.
Replenishing or reapplying a "sacrificial" coating that wears is much easier. It is in practice literally trivial. But hardening the substrate itself
seems like a desirable alternative. Can low-viscosity superglue make ANY difference? See how it soaks into rosewood?
http://www.youtube.com/watch?v=b4XL4i4-3D8&feature=youtube_gdat...
Any measurements, such as John's, on surface-treated woods would be a great start.
|
|
jdowning
Oud Junkie
Posts: 3485
Registered: 8-2-2006
Location: Ontario, Canada
Member Is Offline
Mood: No Mood
|
|
I have had time to read through the paper 'Hardess Testing of Organic Coatings' and find it refreshing in the clarity of the writing and presentation
of the material for a scientific subject - very easy to read and understand even for those with a basic scientific training I would imagine. It is a
compilation of articles covering various aspects of testing paint surface coatings - prepared by applied chemists who are specialists in this field -
and includes a discussion about the mechanical aspects of surface coatings, problems of testing and description of several manual testing methods. An
excellent and thought provoking paper that I shall retain for future reference as these trials progress.
Unfortunately the video link provided by fernandraynaud does not show how deeply thin super glue penetrates into the rosewood - it shows the fluid
spreading laterally along the exposed longitudinal surface cells of the wood (?) which may mean only a very superficial, microscopic penetration (one
wood cell diameter?). The superglue is also seen wicking into the saw cuts provided for the fret markers - but that condition will not apply for any
surface treated wood penetration test - unless, of course the wood is first prepared with a multitude of very fine, closely spaced deep cuts (to the
required depth of penetration) prior to application of the glue.
As previously mentioned, I shall be testing the possibilities of being able (or not) to deeply penetrate and harden wood by surface application using
simple apparatus and a variety of chemicals - including superglue. Depth of penetration in each case will be determined by sectioning each sample
followed by microscopic examination.
However, all that is to be the subject matter of a separate topic on this forum that I shall initiate when I can get around to it.
First, a simple method of of measuring relative hardness must be devised and developed - which is the objective of this thread.
|
|
freya
Oud Junkie
Posts: 276
Registered: 2-8-2004
Location: Asheville NC
Member Is Offline
Mood: Happy in the mountains of WNC
|
|
Gentlemen,
I have used "water thin" superglue in numerous experiments. It's "surface tension" (I am assume this is the mechanism but just guessing really) is
considerable less than water, i.e. where water will bead up on a surface or only penetrate slightly, the thin water glue will "vanish" into the wood
in huge volumes - it has to be going somewhere. That said, I have not cut cross sections to see the actual penetration depth. This is worth a try at
least on spruce or cedar where I use it the most either to repair cracks or harden the surface (mostly in the case of cedar). BTW, spraying water thin
CYA is the nastiest operation I've ever done any I am unlikely to repeat it unless someone donates a full HAZMAT suit to me.
WRT fingerboards, I use it on fingerboard grooves in softer fingerboards mixed with (typically) very fine rosewood dust - for repairs of other
people's instruments mostly. A note of caution - the water thin CYA polymerizes so quickly in fine dust that it gets hot enough to smoke - and those
are some nasty fumes.
For the fingerboards I make from scratch, I'll use Honduran rosewood preferentially and EIR when I don't have any thicknessed Honduran around. I
suppose this stuff will wear eventually but probably not before I'm too old to remember my name.
Best regards
|
|
jdowning
Oud Junkie
Posts: 3485
Registered: 8-2-2006
Location: Ontario, Canada
Member Is Offline
Mood: No Mood
|
|
Continuing with the development of the hardness tester.
The punch tip has been polished to a mirror finish with a Dremel tool - the tip being rotated in a drill chuck. This will help minimise friction
effects that may, in turn, affect tip penetration under constant load (and hence measured diameter of indent).
I am now measuring indent diameter using a metric dial caliper - with scale divisions of 0.02 mm - as the simplest and most accurate way to go. The
positioning of the caliper tips must be viewed under magnification for accuracy - a 10X hand lens will do the job but better still and more convenient
is to use a low power microscope like this one that I picked up at a yard sale some years ago for a few dollars.
Retesting with selected hardwood samples showed an improvement in the clarity of the indents under the tester impact load but for the softer woods
(e.g. Soft Maple) the edges of the indents were slightly rounded and less well defined and for the harder more brittle woods (e.g rosewoods) the edges
of the indent were partially obscured by surface fractures. So the next step will be to reduce the strength of the main spring of the punch to
(hopefully) minimise these effects and improve accuracy of measurement of the indent diameter.
More to follow
|
|
jdowning
Oud Junkie
Posts: 3485
Registered: 8-2-2006
Location: Ontario, Canada
Member Is Offline
Mood: No Mood
|
|
The hardness of wood measured by indentation has three values - side grain radial, side grain tangential and end grain - as illustrated in the
attached sketch. The highest value of hardness by the indentation method - in general but not always - is end grain followed by radial side grain
followed by tangential side grain (quarter cut).
The Janka test measures side grain radial and side grain tangential and averages the result. Published values are within the range of ±20% due to the
variable properties of wood of the same species dependant upon environmental conditions. There is said to be little difference between the Janka
radial and tangential values (±10%).
In fact the indent hardness values taken at all three impact locations are corellated - the cone shape of the punch creating resultant vertical and
horizontal forces that travel along both longitudinal, transverse and radial direction of the wood grain.
For these tests using the conical punch tool - the most brittle and hardest woods like East Indian Rosewood tend to fracture at the surface under
impact loadings leaving a less than clear indent diameter to measure. Reducing the impact load may improve matters but first it was decided to run a
test on end grain indentation to see what happens.
All of the end grain indents were clear and sharp (even for the softer species like Black Ash) allowing a more precise measurement of indent diameter
and more accurate relative comparison of the results with Janka published values.
[file]25428[/file] [file]25430[/file] [file]25432[/file] [file]25434[/file] [file]25436[/file]
|
|
jdowning
Oud Junkie
Posts: 3485
Registered: 8-2-2006
Location: Ontario, Canada
Member Is Offline
Mood: No Mood
|
|
I have just noticed that for some reason I did not post an image - 9 days ago - of my calculations for the geometry of the punch tip. That post has
now been edited to include the calculations.
Here is another image of the missing calcs.
|
|
jdowning
Oud Junkie
Posts: 3485
Registered: 8-2-2006
Location: Ontario, Canada
Member Is Offline
Mood: No Mood
|
|
I have been testing a variety of end grain wood samples to determine and compare the D² values representative of relative hardness using this
tool.
The main spring of the tester has now been replaced with a less powerful spring to reduce the impact loading and crushing damage for the harder more
brittle woods.
The measured static load prior to release of the spring is now 7 lbs compared to the original 34 lbs. This makes the tool easier to handle with one
hand but the diameter of the indent is subsequently reduced making precise measurement of the indent diameter more critical.
It has, therefore, been decided to modify the test procedure by producing each indent with 10 successive impacts of the tester. This results in a
larger diameter indent - the diameter increasing with each impact.
More results to follow.
|
|
fernandraynaud
Oud Junkie
Posts: 1865
Registered: 7-25-2009
Location: San Francisco, California
Member Is Offline
Mood: m'Oudy
|
|
John, any chance you can standardize the image size, and/or include a ruler, so we can get a sense just by looking?
|
|
jdowning
Oud Junkie
Posts: 3485
Registered: 8-2-2006
Location: Ontario, Canada
Member Is Offline
Mood: No Mood
|
|
The images posted so far are just for general information - indents to show smoothness or surface damage for example. Detailed dimensions and other
data will be made available as trials proceed.
Otherwise is there any need for me to standardise image size? I always make image size the maximum acceptable in order to provide best resolution and
clarity of detail - this particularly applies to cropped images that may not fit the standard VGA 640 X480 format.
|
|
jdowning
Oud Junkie
Posts: 3485
Registered: 8-2-2006
Location: Ontario, Canada
Member Is Offline
Mood: No Mood
|
|
Theoretically for the Janka and this hardness testing tool - both using an indent method - the Janka hardness value should be proportional to 1/D² so
plotting Janka hardness values against 1/D² should be a straight line. If this is the case then it might be possible to 'calibrate' the hardness
tester from the published Janka hardness values?
However, there are a lot of potential variables - friction, shear forces and the highly variable nature of wood itself that may influence the
viability of such a theoretical relationship.
Time for an evaluation test.
[file]25519[/file]
|
|
jdowning
Oud Junkie
Posts: 3485
Registered: 8-2-2006
Location: Ontario, Canada
Member Is Offline
Mood: No Mood
|
|
The first preliminary attempt to 'calibrate' the hardness tester using published Janka hardness values is less than satisfactory.
The attached plot of published Janka hardness values against the hardness tester hardness values (1/D²) - using 10 impacts per indent reading with
the low tension main spring - gives only a vague straight line 'through the origin' relationship even if an envelope of ±20% of the published values
is assumed.
The problem is that without access to a Janka hardness tester to enable direct comparison of results it is impossible to know exactly where within the
envelope of maximum and minimum Janka hardness values each wood sample under test might be.
Time to rethink the test procedure.
|
|
jdowning
Oud Junkie
Posts: 3485
Registered: 8-2-2006
Location: Ontario, Canada
Member Is Offline
Mood: No Mood
|
|
Inventor of the Janka hardness test for wood, Austrian, Gabriel Janka - found that hardness is approximately proportional to the density of a wood.
Further research found a relationship between Janka hardness and specific gravity (density of wood compared to the density of water) that may be
expressed as a power formula.
The attached graph is derived from that power formula for North American hardwoods (at 12% moisture content) provided by the American Forest Products
Laboratory (Research note FPL-RN-0303) and relates Janka hardness values to specific gravity of wood species (S.G. varies widely for each wood
species).
So, if the specific gravity of my wood test specimens can be measured then this might be a way to more accurately correlate my hardness tester values
(1/D²) to Janka hardness values - which might be quite useful?
So - next - how to easily measure the specific gravity of my wood specimens.
|
|
jdowning
Oud Junkie
Posts: 3485
Registered: 8-2-2006
Location: Ontario, Canada
Member Is Offline
Mood: No Mood
|
|
Note that these days the current Janka published values are only for side hardness (average of radial and tangential grain readings). End grain values
are apparently no longer available. However, the U.S. Forest Products Laboratories did at one time (1980) gave Janka end grain hardness derived from a
specific gravity power formula - calculated values that were about 27% greater than the calculated side hardness values.
Although this end grain hardness data is now out of date it should still be useful for the purposes of these trials where the hardness tester is being
used on end grain specimens.
The graph of Janka hardness/specific gravity previously posted has now been revised to include the end grain (1980) data. Note that the Specific
Gravity (G12) is defined as oven dry weight (grams) of a sample based on its volume (cubic cm) at 12% moisture content (M.C.). The
gram/centimeter/second standard is convenient to use here as 1 cc water weighs 1 gram - so that the wood density measured in gm/cc is the same as the
specific gravity (e.g density of sample = 0.6 gm/cc , S.G. = 0.6).
|
|
jdowning
Oud Junkie
Posts: 3485
Registered: 8-2-2006
Location: Ontario, Canada
Member Is Offline
Mood: No Mood
|
|
The specific gravity of the wood samples has been determined by direct measurement and by the so called 'flotation' method.
Density is a property of a material defined as the mass per unit volume (for simplicity refer to mass as weight).
Specific Gravity (SG) is the density of a material compared to the density of water (at 4°C). Specific Gravity is, therefore, a dimensionless number
expressed as a decimal.
The measurement of the density of wood (or its SG) is complicated because wood - being of cellular structure - is not uniform. Unless being perfectly
dry (oven dry) its cells contain water (moisture content - or MC) as well as (for certain species) other materials such as oils , salts etc that add
weight and so affect density.
The standard condition for measuring the density of wood is that the volume of a sample is measured at 12% MC (i.e. an average air dried MC) but that
the weight of the sample is measured when oven dried.
For the purposes of these trials - not conducted to laboratory standard accuracy - the density and hence specific gravity has been determined from dry
samples that are probably around 8% MC or less.
So for the direct measurement procedure each sample has been measured dimensionally (in centimetres) as accurately as possible (length (L) with steel
ruler and width (W) and thickness (T) with dial calipers). The volume is then calculated as L x W x T. The weight is measured on an accurate scale.
For the preliminary trials the weight of each sample was measured on a digital kitchen scale accurate to only 1 gram - so not particularly precise.
For future trials a more accurate digital scale will be used. These are available at very reasonable cost - as pocket sized scales used for weighing
jewellery - around $5 to $15 on EBay dependant on accuracy. The scale I have ordered measures to 0.01 gram - good enough for these trials.
More to follow with the flotation method.
[file]25617[/file]
|
|
jdowning
Oud Junkie
Posts: 3485
Registered: 8-2-2006
Location: Ontario, Canada
Member Is Offline
Mood: No Mood
|
|
The size range of the samples under test vary in length from 13 cm to 22 cm, uniform cross section from 1.3 x 1.3 cm to 3 x 1.6 cm and weights from 16
grams to 48 grams.
The precision digital scale will enable improved accuracy and allow use of smaller samples
The second way to determine SG without costly laboratory equipment is by the so called 'flotation' method. Here the samples (of uniform cross section)
are floated (vertically) in water in a tall glass container (here I am using a lager glass for convenience). The sample is slowly lowered into the
water until it floats and the sample marked at the water level.
The total length AB) of the sample and the immersed length BC) is then measured (in cms). The specific gravity of the sample is the ratio of BC/AB. So
for the black walnut example illustrated in the attached images AB = 184 cm and BC = 129 giving a specific gravity = 0.7.
In order to guide each sample in the vertical position allow it to slide freely between the fingers or in a wire guide.
Note that this method can only be used for woods of specific gravity less than 1.0. For 'heavier than water' woods such as my samples of ebony and
brazilian rosewood, specific gravity must be calculated by measurement of dimensions and weight using method #1.
The results from methods #1 and #2 were consistent in agreement - although not to the highest degree of accuracy. Note, however, that published
results of wood species specific gravity are mean values of - at times - widely ranging min/max values - so even the above methods can give more
meaningful results than reference to published values.
Note that the samples were all air dried to 8% MC or less and that no attempt was made to determine oven dry weight or volume at 12% MC of the samples
in accordance with industrial practice - not necessary given the practical level of accuracy possible by the above methods that give results that are
close enough for the purposes of these trials.
It can be seen that measuring SG alone can give a reasonable estimate of the Janka wood hardness without need for a hardness tester (and it will not
always agree with the mean published values of hardness).
Next to see if there is a possible correlation of Janka hardness (determined from measured specific gravity) and the indent diameter measurements
obtained from the hardness tester.
[file]25620[/file] [file]25622[/file] [file]25624[/file] [file]25626[/file] [file]25628[/file]
|
|
jdowning
Oud Junkie
Posts: 3485
Registered: 8-2-2006
Location: Ontario, Canada
Member Is Offline
Mood: No Mood
|
|
The attached plot is the result of a preliminary test to determine if there is a correlation between Janka hardess values (calculated from measured
specific gravity) and hardness values (1/D²) of the 90° cone tester.
There is some scatter of the data points taken from a small number of individual wood species (3 indents per sample) so more data will be required for
a more accurate assessment. Nevertheless, I have plotted a predicted straight line through the origin.
Due to friction forces and the wedging action of the cone there may be distortions of the indent affecting the accuracy in measuring indent diameters
- under estimating D for hard brittle materials (such as Ebony) and over estimating diameter for softer woods.
Also, as the test is very localised, more readings than three per sample will be required to obtain a more accurate assessment of a mean hardness
value per sample. This may be more significant for ring porous hardwoods (like Ash) than diffuse porous woods (like Maple).
However, as the purpose of developing this hardness tester is to have a tool to objectively assess 'before and after' hardness of a wood after
chemical impregnation - the correlation (if any) with the Janka test values is not really of importance (although of interest).
Next - the calibration curve for the 90° cone hardness tester.
[file]25635[/file] [file]25637[/file]
|
|
jdowning
Oud Junkie
Posts: 3485
Registered: 8-2-2006
Location: Ontario, Canada
Member Is Offline
Mood: No Mood
|
|
Of passing interest to note that researchers of the US Forest Products Lab. have modified the Janka test to measure the load required to impress the
ball of the tester to a distance less than the full depth. The measured 'hardess module' directly correlates to the Janka values by a factor X5.4.
Analagous to the 90° cone tester perhaps - constant load and depth equal to D/2?
Attachment: Hardness Modulus.pdf (350kB)
This file has been downloaded 669 times
|
|
jdowning
Oud Junkie
Posts: 3485
Registered: 8-2-2006
Location: Ontario, Canada
Member Is Offline
Mood: No Mood
|
|
Although Janka hardness values are today expressed in pounds force (or the metric equivalent) - the force (or weight) required to drive a steel ball
of precisely 11.28 mm diameter to half its depth into a wood sample, Janka apparently originally represented his hardness values as a pressure (i.e.
force/surface area of indent).
The modified Janka test procedure described in the paper previously posted does not force the Janka ball tip to half depth but measures the indent
depth for a given load - much the same principle as for the 90° cone tip tester. This way somewhat thinner wood samples may be tested.
For comparison, the attached plot of indent depth against indent surface area for each tip geometry has been prepared. From this it can be seen that
the Janka relationship is (ideally) a convenient straight line which may have been the reason that Janka chose a ball tip rather than a cone or other
tip geometry for his hardness tester.
The cone tip relationship on the other hand is a more complicated curve.
On the other hand the cone tip does not seem to require the same minimum indent depth as the Janka ball (0.1 inch or 2.5 mm according to the US
researchers).
In the #2 trials of the 90° cone tester previously reported the measured indent diameters ranged from 3.63 mm for the softest woods to 1.43 mm for
the hardest samples. So equivalent indent depths ranged from about 0.7 mm to 1.8 mm (D/2).
[file]25659[/file]
|
|
Pages:
1
2 |