Page 7 of 24 FirstFirst ... 2345678910111217 ... LastLast
Results 91 to 105 of 347
  1. #91
    Join Date
    Jun 2008
    Location
    Victoria, Australia
    Age
    74
    Posts
    6,132

    Default

    Quote Originally Posted by rob streeper View Post
    Now I know why Sawmill Creek has the reputation that it does. The posters there have formed a club and outsiders are not welcome. I checked the thread from yesterday and Prashun Patel, I assume one of the moderators, is going back and revising all of the posts and has deleted many. Auditing reality to make it pretty has a bad smell to me.
    Moderators sometimes do that, mostly it's to delete personal attack type post, anyway, enough of that. Those guys over there were trying to tell you something, and that is....

    The issue for me is the high hardness numbers that you are getting... the maximum hardness for 1095 as I've already posted is around 66 Rc, that's physically it!.... and that's surface hardness in the as quenched state. no tempering at all... and it will be glass hard, a file will skate off it without leaving a scratch. It's not able to be filed until you get down to maybe mid 50's, that's not just my opinion, that's just 1095 steel.

    When you report results of 62,60,62,59,64 hardness for saw plate, there is something wrong with the methods you are using.

    A simple test is to see if you can actually file that super hard saw, if it can be filed, then the hardness test must be wrong.

    Ray

    PS, The 1095 data I'm quoting comes from The ASM Heat Treaters Guide 2nd edition

    http://www.asminternational.org/web/..._p_col_count=2

  2. # ADS
    Google Adsense Advertisement
    Join Date
    Always
    Location
    Advertising world
    Age
    2010
    Posts
    Many





     
  3. #92
    Join Date
    Dec 2013
    Location
    San Antonio, Texas, USA
    Posts
    3,070

    Default

    Quote Originally Posted by RayG View Post
    Moderators sometimes do that, mostly it's to delete personal attack type post, anyway, enough of that. Those guys over there were trying to tell you something, and that is....

    The issue for me is the high hardness numbers that you are getting... the maximum hardness for 1095 as I've already posted is around 66 Rc, that's physically it!.... and that's surface hardness in the as quenched state. no tempering at all... and it will be glass hard, a file will skate off it without leaving a scratch. It's not able to be filed until you get down to maybe mid 50's, that's not just my opinion, that's just 1095 steel.

    When you report results of 62,60,62,59,64 hardness for saw plate, there is something wrong with the methods you are using.

    A simple test is to see if you can actually file that super hard saw, if it can be filed, then the hardness test must be wrong.

    Ray

    PS, The 1095 data I'm quoting comes from The ASM Heat Treaters Guide 2nd edition

    http://www.asminternational.org/web/..._p_col_count=2
    Ray,

    I measured the hardness values for the old saws under the handles. I suspect they're harder there because they have been hammered in manufacture. I'll get after this issue later today.

    Cheers,
    Rob
    Innovations are those useful things that, by dint of chance, manage to survive the stupidity and destructive tendencies inherent in human nature.

  4. #93
    Join Date
    Dec 2013
    Location
    San Antonio, Texas, USA
    Posts
    3,070

    Default

    Quote Originally Posted by BobL View Post
    I'd like to see some more extensive testing before making any firm conclusions - certainly before naming any manufacturer.
    If one of my students was doing this as an undergraduate material science project I'd say even though 5 measurements might satisfy a standard procedure I still think 5 test points per saw is on the slim side to demonstrate hardness is normally distributed, especially if normal stats are going to be applied.

    The first thing I would ask the student to do would be to demonstrate the method can generate a normal distribution on a reference material, both in several locations of the material and then using a reduced sampling regime across the surface of the material i.e. a hardness map. i.e. 20-30 measurements in a few locations and then 5 measurements at points event distributed across the material

    I would then ask the student to to the same (hardness map) of at least one saw using the same method..
    The surface finish issue can be addressed by sanding to the same level of smoothness.

    The other concern is that typical "consumer testing" issue of just testing "one" of anything can be problematic. I would suggest the student investigate at least 5 saws of the same make and model from at least one manufacturer. i.e. no map but at least 20 measurements per saw.

    To demonstrate "between sample" reproducibility I'd like to see the student repeatedly measuring the reference material between each different saw measurements to show there is nothing wrong with the apparatus or method. I'd also like to see one saw measured multiple times and in between revert to measuring the standard. A more representative value of the hardness would then be the external (or average of average values)

    Although this might be difficult for you to do to further increase the confidence of measurements I'd also ask the student to perform sone repeat tests using a different apparatus. Ultimately a completely independent set of measurements would be useful.

    I appreciate the time required to perform these measurements but I think even doing some of these it is one way to help reduce bias and hence criticism of the results
    Hi again Bob,

    Since this seems to be such a divisive issue I decided I'd better do this with an adequate number of samples. Thus I did a power analysis using the results tabulated above for standard deviation and differences in the means. Here's the log file of the calculation.

    <PS logging enabled 1/13/2015 6:53:17 AM

    Version 3.1.2
    Suggested citation:
    Dupont WD, Plummer WD: 'Power and Sample Size Calculations: A Review and Computer Program', Controlled Clinical Trials 1990; 11:116-28.
    or
    Dupont WD, Plummer WD: 'Power and Sample Size Calculations for Studies Involving Linear Regression', Controlled Clinical Trials 1998; 19:589-601.

    --------------------------------------------------------
    Type of study: T-test

    Requested output: Sample size
    Design: Paired

    alpha=0.05 power=0.8 DIFF=0.9 SIGMA=1.92 M=0
    Sample size=38

    t-test confidence interval width=1.267477

    We are planning a study of a continuous response variable from matched pairs of study subjects. Prior data indicate that the difference in the response of matched pairs is normally distributed with standard deviation 1.92. If the true difference in the mean response of matched pairs is 0.9, we will need to study 38 pairs of subjects to be able to reject the null hypothesis that this response difference is zero with probability (power) 0.8. The Type I error probability associated with this test of this null hypothesis is 0.05.>

    Of course I'm not doing a clinical trial but the sample size number needed is larger than 25 so I'll do 50 measurements.
    Innovations are those useful things that, by dint of chance, manage to survive the stupidity and destructive tendencies inherent in human nature.

  5. #94
    Join Date
    Jun 2009
    Location
    Holland
    Posts
    72

    Default

    It looks like Sawmilcreek closed the thread, just when I was outside to do some tests! How dare they?

    I got my Disston D8, popped off the handle, grabbed a new sawfile (Bahco) and after 5 strokes I had filed the nick you can see in the first picture. Then I got a Stanley plane blade from a drawer (newish UK made stuff) and tried to file a nick into the edge of that one too. After 8 strokes I got only half as deep as on the Disston and more importantly, the feel was completely different. While you can feel the file really grabbing into the steel on the Disston, it feels like you are scratching on the plane blade. Also, when looking at my poor file's edge, you can clearly see the tell tale shiny spots of a worn out file.

    Let's assume the Stanley plane blade is 58 HRc (I don't believe they would go much higher) then the Disston certainly is nowhere near that value.

    Alas, I don't have any 1095 available at the moment.

    S1.jpgS2.JPG

    Just one data point of course, but it gives some color to the data.

  6. #95
    Join Date
    Aug 2009
    Location
    Armadale Perth WA
    Age
    54
    Posts
    4,524

    Default

    Regardless of the absolute numbers, any test will show that there is a difference between the 30-ish numbers and the rest.

    We should be welcoming this line of investigation ... even as a first step ... even if the method and data are not at a "peer-reviewable" level.

    It asks and starts to investigate a question that maybe hasn't been looked at before ... which is what science is for.

    As with any good experiment, there are any number of qns raised ... and it's valid, indeed essential, to question and query and improve and re-verify ... but people (elsewhere) approaching this with animosity is beyond me.

    These are real measurements ... and internally consistent. There shouldn't be any backlash against publishing them ... only what to think of and do next.

    [
    One day ... for myself ... I'd like to know more about what "tension" means in the physical reality of a saw-plate. eg #12s (of which I have a couple dozen in varying ages and degrees of attrition) are dependably sonorous, and usually thin and flexible.
    ]


    Cheers,
    Paul

  7. #96
    Join Date
    Dec 2013
    Location
    San Antonio, Texas, USA
    Posts
    3,070

    Default

    Quote Originally Posted by pmcgee View Post
    Regardless of the absolute numbers, any test will show that there is a difference between the 30-ish numbers and the rest.

    We should be welcoming this line of investigation ... even as a first step ... even if the method and data are not at a "peer-reviewable" level.

    It asks and starts to investigate a question that maybe hasn't been looked at before ... which is what science is for.

    As with any good experiment, there are any number of qns raised ... and it's valid, indeed essential, to question and query and improve and re-verify ... but people (elsewhere) approaching this with animosity is beyond me.

    These are real measurements ... and internally consistent. There shouldn't be any backlash against publishing them ... only what to think of and do next.

    [
    One day ... for myself ... I'd like to know more about what "tension" means in the physical reality of a saw-plate. eg #12s (of which I have a couple dozen in varying ages and degrees of attrition) are dependably sonorous, and usually thin and flexible.
    ]


    Cheers,
    Paul
    Hi Paul,

    I'm looking into that tension issue as well. See my D8 replica posts.

    As to Sawmill Creek, well what can I say. I'd heard in several places some hints that the posters there could be difficult but I thought at least some people would appreciate the information. Some posted thanks openly but the majority sent me PM's suggesting that they didn't want to publicly thank me for my effort. I'll continue to go there and watch but I doubt I'll be posting anything more. My experience left me with a bad feeling.

    Cheers,
    Rob
    Innovations are those useful things that, by dint of chance, manage to survive the stupidity and destructive tendencies inherent in human nature.

  8. #97
    Join Date
    Dec 2013
    Location
    San Antonio, Texas, USA
    Posts
    3,070

    Default

    Quote Originally Posted by Corneel View Post
    It looks like Sawmilcreek closed the thread, just when I was outside to do some tests! How dare they?

    I got my Disston D8, popped off the handle, grabbed a new sawfile (Bahco) and after 5 strokes I had filed the nick you can see in the first picture. Then I got a Stanley plane blade from a drawer (newish UK made stuff) and tried to file a nick into the edge of that one too. After 8 strokes I got only half as deep as on the Disston and more importantly, the feel was completely different. While you can feel the file really grabbing into the steel on the Disston, it feels like you are scratching on the plane blade. Also, when looking at my poor file's edge, you can clearly see the tell tale shiny spots of a worn out file.

    Let's assume the Stanley plane blade is 58 HRc (I don't believe they would go much higher) then the Disston certainly is nowhere near that value.

    Alas, I don't have any 1095 available at the moment.

    S1.jpgS2.JPG

    Just one data point of course, but it gives some color to the data.
    Hi Corneel,

    When I was testing and posting yesterday I was also doing some heat treating and annealing. I was hardening the round floats I posted about earlier in the year and they turned out quite well. I took some readings and got back values corresponding to Rc 58. An older file I had barely burnished the teeth of the hardened float. The steel I used was O1 if I'm remembering correctly. The thickness of the stock may play some role in perceived file-ability or the alloy may just be more abrasion resistant, at this point I don't know. Experiments and posting of results will continue however, here only though.

    Cheers,
    Rob
    Innovations are those useful things that, by dint of chance, manage to survive the stupidity and destructive tendencies inherent in human nature.

  9. #98
    Join Date
    Mar 2010
    Location
    US
    Posts
    3,075

    Default

    Some of us do read more than one forum.

    There is a misinterpretation of pressing people to provide correct data with it being a personal assault. It is nothing of the sort. It's the desire to see meaningful data and not just data.

    As has been said previously in this thread, the high end of the hardness numbers provided for several western saws make no sense, and it's not worthwhile to push the data if the data is wrong. Incorrect data is worse than no data. I agree with that.

    Kees just demonstrated why those high values aren't relevant, and quickly - even with a blade that may not even be 58 hardness.

    As far as sawmakers making saws with various saw plates, it's not new news that some of the european manufacturers, and I'm sure western, too, do not use 1095 in their saws, and they do not harden 1070 or whatever other substitute is used to a high level (Certainly not the 48-52 range that 1095 is commonly found). One would expect if a boutique maker of saws doesn't specify the composition of the saw plate, you don't really have a reason to expect that a saw is 1095 and thus that it will be the same hardness as 1095 (be it a supply house that provides 48 hardness or 52, whatever it may be, either will make a good saw that is at least the equal of vintage backsaws from any era).

    That said, 30 hardness is not acceptable, but providing that information without confirming that it's consistent and naming the maker doesn't do anything but indict all boutique makers who haven't been confirmed to not be that maker, and that is unfair. Especially when you are promoting your own saws and saw parts.

    I have no objection to the data, but it has to be useful for something, and in order for it to be useful, it needs to be correct. Especially when another much more experienced professional toolmaker provided you with the actual results from strikes he made on saws with a much more sophisticated device (a versitron). What he provides as seeing saws between 40 and 45 hardness is data, and relevant. To suggest no other data has been provided is not accurate. It is less data, and less resolution, but much higher precision. With the direction that the aspersions are being cast, precision is awfully important.

    I commend ray for directing you with exactly the same thoughts as I had, he is more mannerly than I am. It is not by chance that the most experienced folks are telling you that you have a problem to correct and the least experienced are cheering you on, correct or not.

  10. #99
    Join Date
    Aug 2009
    Location
    Armadale Perth WA
    Age
    54
    Posts
    4,524

    Default

    Quote Originally Posted by D.W. View Post
    Some of us do read more than one forum.
    There's a good chance every reader here reads more than one forum.
    It doesn't require reference to any other forum to discuss data and points that were presented here.

    It's the desire to see meaningful data and not just data.
    I would almost say *all* data is meaningful. This data clearly is.
    At the very least it demonstrates at least separate groups in the recently manufactured items.
    Rob could simply have posted ... "I have a number of custom saws ... and one of them doesn't perform."
    That is opinion. So I'd say the call should be for "meaningful opinion, not just opinion".

    There are various ways you go about trying to quantify performance or quality ... usage tests like Derek has pursued eg.
    This is a perfectly reasonable step to take if the equipment is available.

    In the first instance, the exact numbers do not matter ... even though there was a calibration stage.

    If these numbers are "wrong" then the "real" ones are going to be a linear or near-linear or monotonic or non-linear function of the readings ... and for a virtual certainty ... the scores around 30 are going to remain clearly separate from the scores around 50-60.

    This is not some offence against science ... this is exactly how it works. A first test is a first test.


    As has been said previously in this thread, the high end of the hardness numbers provided for several western saws make no sense, and it's not worthwhile to push the data if the data is wrong. Incorrect data is worse than no data.
    That is false. If the data is precise, then accuracy can be corrected.
    It depends on what you are measuring and in what sense it is supposedly "incorrect".

    A difference in means doesn't rely on accuracy ... precision is sufficient.
    Medical risk statistics mostly never know what absolute level is observed ... it's proportionality that is investigated.
    And most of all, we perform useful rough surveys all the time.

    If I sort through my handsaws, some of them ring and some of them "thunk".
    When you buy wood at a chain hardware store you sort through boards into "looks alright" and "not now, not ever".
    It's useful at that level, and it is of course useful at this more explicitly quantified level.

    If there's then more to investigate, then encourage more investigation ... and constructively suggest aspects that don't gel or could bear fruitful investigation.

    This is a consistently supportive forum and a real good in the online world.
    We don't do a bunch of negativity for very long around here.


    That said, 30 hardness is not acceptable, but providing that information without confirming that it's consistent and naming the maker doesn't do anything but indict all boutique makers who haven't been confirmed to not be that maker, and that is unfair. Especially when you are promoting your own saws and saw parts.
    Again I disagree. People will report on their purchases in the general areas all the time. "This is great", "This is a dog". There's nothing different going on here other than it is quantified. Plus this is a niche section of these forums. I doubt if there are thousands of trooping onlookers coming to carefully analyse the saw-making forum. This is where this should be posted, among the people who are interested in the details.

    All this says to customers is that one individual saw out of four individual saws seems not to be up to functional expectations. It might be a maker, it might be a batch, it might be one-off. It might be unintentional. So customers learn to ask more questions, which they already had to learn anyway to know if they want a rip or xcut or hybrid saw. What pitch? What handle wood? So now they ask also about steel hardness. That's better for everyone.

  11. #100
    Join Date
    Aug 2009
    Location
    Armadale Perth WA
    Age
    54
    Posts
    4,524

    Default

    Rob - do you know this forum? http://woodworkerszone.com/igits/forumdisplay.php?f=63

    They're still american ... I don't suppose there would be more in the way of welcome of the data ... but they are keen on their saws.

    Cheers,
    Paul

  12. #101
    Join Date
    Mar 2010
    Location
    US
    Posts
    3,075

    Default

    Paul, when you say the data is precise, I think you're getting resolution and precision twisted around.

    Resolution implies showing a lot of detail, precise implies correct. There is a lot of resolution here, but the precision is not there for some of the measurements. The trouble is resolution implies or at least is perceived as precision by a lot of people.

    Hardness spec is less important for people to know as a saw buyer than is the type of steel. 1095 is sold 48 to 52 hardness that I have seen. Anything in that range makes a very fine saw, and to be honest, I would prefer something at the low end of that because it's easier on files and barely wears any more poorly in wood. And as Ray suggested earlier, just because a saw is struck somewhere and it reports a number doesn't mean that the same number will carry over to the tooth line.

    The saw that is 30 hardness has a problem of some sort. If it's intentional from a seller, then it should be known. I had two "deer brand" or something similar dovetail saws that said "made in germany" on the plates. They were cheap, and I doubt they were harder than that. I gave one to my FIL and threw the other one away. There were other aspects about the saw that were undesirable other than the very soft plate. Now, those are something like $12 saws, so one should probably expect that.

    It occurs to me that if a beginner is to be recommended something, its that they should ask if the saw they're buying is 1095. If someone said "no, but it's 1080 and it's hardened to 48 hardness".

    If someone were to buy a 1095 saw that was soft, then they'd need to take it up with the maker to get their saw replaced. George Wilson often mentions that a lot of the tradesmen at Williamsburg were using Pax saws (I think it was pax) and that the saws were too soft and filed too easily. They were not made of 1095.

    Data is fine if it's correct, but as a user and sometimes maker of tools, I would lean much more on someone experienced telling me "yes, it's good" or "no, it's not".

    I didn't have the sense to do that when I was a beginner because I didn't understand the context of the tools in work. I wanted something that sounded good on a spec sheet. I suppose that's a market that can be satisfied, the catalog spec sheet hardness.

    I did relay on another forum that a friend (different person with access to a versitron and an expert to operate it for us) and I had a whole glom of chisels tested, and one of the things we bought was a japanese iyoroi mortise chisel to compare to an LN mortise chisel. Of course, we concluded the japanese chisel would be better because it said it was 64 hardness, and the LN said 60-62 or something. They both tested the same (strikes averaged 61). I was very disappointed to find that out and considered the chisel junk at the time because it didn't meet the spec sheet. When we divvied the items, I took the LN chisel and my friend took the japanese chisel. I know enough now to know that at 61 hardness, it's probably a pretty good chisel. I also know that the japanese chisels that are hardened way up into the mid 60s, and truly are what they claim can be much harder to sharpen, especially on a natural stone. If the chisels were legitimately labeled based on the numbers I mentioned, every beginner would say "one is 65 and one is 61...65 is bigger, that's better". Most beginners would also have trouble properly sharpening the chisels that are 65 hardness. The same thing would occur with saw plates. "that one's 50 and this one is 55...55 is better". We know enough to know that a 55 hardness plate is trouble for files, but I'll bet a beginner would go for it every time, bigger number. An experienced user (especially one who uses tools often enough to need to sharpen them with frequency) would go for the one at 50.

    That's sort of two different points. I never would've posted about any of this on any forum had the strikes suggesting vintage saw hardness of 58 or 61 been omitted, though, or corrected before presenting the data.

  13. #102
    Join Date
    May 2008
    Location
    Australia
    Posts
    2,357

    Default

    DW's comments have covered my own personal concerns with the general direction of this post. By raising concerns from a 1 off saw made by Maker 3, and then leaving the conjecture open as to who that may be, it unfairly raises a cloud of doubt on the work of all boutique saw makers.

    Stewie;

  14. #103
    Join Date
    Dec 2013
    Location
    San Antonio, Texas, USA
    Posts
    3,070

    Default More data.

    Okay, levity and insults aside I've done the experiment I discussed with BobL.

    Abstract:

    To briefly describe the experiment. I took some cut off pieces of 0.035" 1095 steel sheet that I recently received from Zoro Tools. They have recently been stocking a very wide variety of 1095 sizes - pretty cheap too but I doubt that they will deliver to Australia for less than an arm and a leg. I left one piece as produced or 'neat', I took a second larger piece, marked it into two sections. I beat on one section with a hammer and then ground both sections as I do my saw blades. I then measured hardness of each piece/section 50 times on the Rockwell N30 scale. The result was, 1095 was as hard as everyone expects it to be. Ground 1095 gave softer hardness readings because the surface was rougher, which causes readings of hardness to have negative deviations. Hammered and ground 1095 was harder than both neat 1095 and ground 1095. Conclusion - work hardening is real and it is important in handsaw manufacture.
    Given that I have gotten a lot of criticism on this issue I'm going to make this a long post to explain as much as I can think to at the moment so please bear with me.

    Materials and Tools:

    Software:

    Sample size calculator: http://biostat.mc.vanderbilt.edu/wik...owerSampleSize


    Hardware:

    Two pieces of 1095. Here they are.

    Test articles.jpg


    These are 1095 having a thickness of 0.035" that come from the same roll of steel that I'm using for my D-8 and half back projects. The piece on the left is divided into two sections by lightly running an angle grinder cutting wheel down the face of the plate. This produces a virtually indelible mark dividing the plate. The half of the plate to the right has a clipped corner. I did this so that I could identify which side of the plate was hammered. I hammered the side with the clip. I call these 'test articles'.

    anvil.jpg


    100 lb. Peter Wright anvil. A very sweet tool to have.

    hammer sample before grinding.jpg


    A 5 lb sawyers hammer. Neat design. Hold it in the right hand and point one face down to hammer side to side and flip it around and use the other face to hammer top to bottom.

    Destroyer of false notions.jpg


    I have come to think of this as "the slayer of unsupportable assumptions", otherwise known as an Ames 2-S superficial hardness tester for Rockwell N and T scale testing.


    Methods:

    Beating of steel sample was accomplished by applying 100 replicate strokes of the hammer from an elevation of approximately 100 mm above the upward facing face of the test article. Look at the picture to get an idea if you need to.

    about 4 inches drop.jpg


    An aside on how to take the measurements. The first task in setting up the tester is to back the indenter off so that it's hanging in mid air and not touching anything. Then you rotate the bezel of the indicator so that the needle rests exactly over the dot on the indicator face that is to the left of the marking 'set'.

    bezel set.jpg


    In this photograph the needle is set a little askew of the dot to allow you to see both the needle tip and the dot. In practice thee needle tip and dot are aligned one over the other.

    After aligning the dot and the needle one carefully winds in the indenter screw and places the test article between the indenter tip and the anvil in the left of this picture.
    While supporting the test piece until it is trapped between the indenter and the anvil carefully rotate the anvil adjusting nut until the needle moves over the set marking on the bezel as so.


    First step.jpg

    At this point move the indicator scale such that the protruding pin rests against the top of the Lucite indicator lens like this.

    Set the indicator such that the pin touches the lucite lens.jpg


    Now carefully increase the indenter load until the needle winds up to the desired loading, in this case 45kg. Be careful as the response can be non linear and it is easy to overshoot the loading.

    Crank the wheel until the indicator reads 45.jpg


    Now rotate the anvil adjusting screw in the opposite direction until the needle comes back to the 'set' point. Then look at the indicator window and read the N45 hardness.

    close up it's 60.jpg


    Here the reading is ~58. Parallax errors creep in so somewhat careful.

    Now, that I've described how the meter works I go back to the experiment. 100 strokes with the hammer and the test article looks like this.


    close up of hammered sample.jpg


    After very light grinding, just enough to remove the bluing, it looks like this.

    hammered sample after grinding.jpg


    Then take the indicated N45 reading, look at the hardness conversion table provided with the meter.

    convert the N reading to C using the chart provide.jpg


    Scan down to the 58'ish part of the table, scan across and get the Rockwell C reading.

    roll down the list until you 58 and read across.jpg

    Thus the result for the pictured shot is Rc ~53. Now for some measurements. Can't hold both my phone and the tester at the same time so bear with me while I test and write down data.

    Beginning again holding the tester the right way, i.e. by the handle. Then 50 total measurements as described above. I put this picture in especially for Ray as he wants to see raw data.

    Write down the result.jpg


    Data table with first 50 data points.

    Data data data.jpg


    Now for another hundred.

    My helper.

    Abdul looks on.jpg


    After doing all of this here is a pic of the ground test article to prove to the best of my ability that I actually did the measurements.

    pricked plate.jpg



    And here's the other.

    Here's the other.jpg

    Results:



    Conclusions:


    First for a few QC checks. Here is a power analysis of the comparison of 'neat' to 'ground'

    PS logging enabled 1/13/2015 11:30:22 PM
    Version 3.1.2

    Suggested citation:
    Dupont WD, Plummer WD: 'Power and Sample Size Calculations: A Review and Computer Program', Controlled Clinical Trials 1990; 11:116-28.
    or
    Dupont WD, Plummer WD: 'Power and Sample Size Calculations for Studies Involving Linear Regression', Controlled Clinical Trials 1998; 19:589-601.

    --------------------------------------------------------
    Type of study: T-test
    Requested output: Sample size
    Design: Paired
    alpha=0.05 power=0.8 DIFF=1.18 SIGMA=2.68 M=0
    Sample size=42
    t-test confidence interval width=1.660548

    We are planning a study of a continuous response variable from matched pairs of study subjects. Prior data indicate that the difference in the response of matched pairs is normally distributed with standard deviation 2.68. If the true difference in the mean response of matched pairs is 1.18, we will need to study 42 pairs of subjects to be able to reject the null hypothesis that this response difference is zero with probability (power) 0.8. The Type I error probability associated with this test of this null hypothesis is 0.05.


    Okay, so with 50 points I've exceeded 42.

    Comparing the data for 'neat' to '1095 hammered and ground' I get these results.

    Type of study: T-test
    Requested output: Sample size
    Design: Paired
    alpha=0.05 power=0.8 DIFF=3.78 SIGMA=2.3 M=0
    Sample size=5
    t-test confidence interval width=5.64008

    We are planning a study of a continuous response variable from matched pairs of study subjects. Prior data indicate that the difference in the response of matched pairs is normally distributed with standard deviation 2.3. If the true difference in the mean response of matched pairs is 3.78, we will need to study 5 pairs of subjects to be able to reject the null hypothesis that this response difference is zero with probability (power) 0.8. The Type I error probability associated with this test of this null hypothesis is 0.05.


    Okay, even better because of the smaller standard deviation in the populations and the larger difference between the means. 5 samples are all that is needed for this comparison.

    Now for the comparison of 'ground' to 'hammered and ground'.

    Type of study: T-test
    Requested output: Sample size
    Design: Paired
    alpha=0.05 power=0.8 DIFF=4.96 SIGMA=2.63 M=0
    Sample size=4
    t-test confidence interval width=7.505376


    We are planning a study of a continuous response variable from matched pairs of study subjects. Prior data indicate that the difference in the response of matched pairs is normally distributed with standard deviation 2.63. If the true difference in the mean response of matched pairs is 4.96, we will need to study 4 pairs of subjects to be able to reject the null hypothesis that this response difference is zero with probability (power) 0.8. The Type I error probability associated with this test of this null hypothesis is 0.05.

    In this comparison, again because of the larger difference between the means despite the somewhat larger standard deviation, the required number of paired samples is only 4.

    Thus with 50 specimens I have produced statistically significant data consistent with the expectations of clinical trial designs as mandated by the United States Food and Drug Administration, i.e. life and death situations.

    The mean 45N hardness values for the three test articles are 55.61 for the neat 1095, 54.43 for the ground 1095 and 59.39 for the ground and hammered 1095. The statistical significance of inter-article Students t statistics are all below 0.05 suggesting that the differences in hardness are real and significant. Comparing neat 1095 to ground 1095 we see that grinding decreases the apparent hardness by 1.18 N45 points. Comparing ground 1095 to hammered and ground 1095 it can be seen that hammering has increased the apparent hardness by 4.96 N45 points. Thus, assuming that grinding had little appreciable effect on the actual hardness of the test articles we can posit that the actual hardness of the ground and hammered 1095 is the sum of the apparent hardness plus the 1.18 N45 offset caused by grinding, or a corrected N45 hardness of 60.57. Referring again to the Ames scale correlation table the final, corrected Rc hardness of the hammered and ground specimen is estimated to be between Rc54 and Rc55. It is notable that this result correlates perfectly, within the limitations of the analytical method applied here, with the readings obtained under the handles of the older Disston saws leading the author to conclude that it is highly likely that the steel under the handles of the Disston saws tested was impact hardened in manufacturing with exception of the softer Number 240 metal cutting saw.

    Appendix:

    Here's the Excel worksheet if anybody want's to work with it.

    Attached Images Attached Images
    Innovations are those useful things that, by dint of chance, manage to survive the stupidity and destructive tendencies inherent in human nature.

  15. #104
    Join Date
    Dec 2007
    Location
    Melbourne
    Posts
    3,277

    Default

    Quote Originally Posted by planemaker View Post
    DW's comments have covered my own personal concerns with the general direction of this post. By raising concerns from a 1 off saw made by Maker 3, and then leaving the conjecture open as to who that may be, it unfairly raises a cloud of doubt on the work of all boutique saw makers.

    Stewie;

    I agree 95% as I did say earlier.
    The remaining 5% - Only we know it wasn't one of your saws because he said it was poor finished and fitting in other areas, there by removing the possibility of it being one of your meticulous saws
    We know it wasn't one of IanW's saw because I don't believe he charges what their worth so it wouldn't have been hundreds of US dollars.


    Everybody else is a potential suspect..
    …..Live a Quiet Life & Work with your Hands

  16. #105
    Join Date
    Jun 2008
    Location
    Victoria, Australia
    Age
    74
    Posts
    6,132

    Default

    Hi Rob,

    I'm not seeing any of the attachments, something about invalid attachment?

    "Invalid Attachment specified. If you followed a valid link, please notify the administrator"


    On the subject of who's saw used the soft steel, I'd argue that it's better just to be upfront and say who's saw it was. You are certain to get criticized either way, so this is the least painful way forward, also, I'd be totally open and upfront about the fact that you are considering starting a saw making business. People will be suspicious of your motives if you aren't totally open and honest.

    I believe and understand that you just want to make a better saw, I suspect that there will likely be others won't see it quite the same way as I do.

    Ray

    PS.. On the subject of saw smithing, I have a range of different sized dogs head hammers that are used for saw doctoring work, also a nice big proper saw anvil, and a few other saw doctoring hammers, I am missing the twist pein hammer... but I know a guy who has a couple.. one day I'll make him an offer he can't refuse..

    The dogs head is what you need for the hammering you are doing.

Page 7 of 24 FirstFirst ... 2345678910111217 ... LastLast

Similar Threads

  1. Hardening & Tempering
    By Dovetail in forum HAND TOOLS - UNPOWERED
    Replies: 6
    Last Post: 8th February 2014, 11:07 AM
  2. Case hardening
    By Pete F in forum METALWORK FORUM
    Replies: 11
    Last Post: 18th November 2011, 10:05 PM
  3. Induction hardening
    By morrisman in forum METALWORK FORUM
    Replies: 5
    Last Post: 2nd October 2011, 09:59 AM
  4. brass hardening
    By Eldanos of KDM in forum HAND TOOLS - UNPOWERED
    Replies: 14
    Last Post: 8th July 2010, 12:56 PM
  5. Timber hardening
    By boris in forum FINISHING
    Replies: 1
    Last Post: 31st January 2004, 11:01 AM

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •