Thread: Hardening of sawplates

Q: Does the Golden Era show a decrease in variability relative to the other era's?

A: Soft yes, not statistically significant yet.

Q: More saws or more tests of the saws I have?

A: More saws.

Corrected plot of relative standard deviation vs. production period.

Median Historical hardness survey of Disston saws 100416a.JPG

Here is a scatter plot reflecting the above data. Be aware that in several cases the markers overlap so that one dot may in fact represent two or more saws.

Chris calculation median standard deviation of hardness with fit line showing individual points .JPG

Originally Posted by rob streeper

Q: Does the Golden Era show a decrease in variability relative to the other era's?

A: Soft yes, not statistically significant yet.

Q: More saws or more tests of the saws I have?

A: More saws.

Corrected plot of relative standard deviation vs. production period.

Median Historical hardness survey of Disston saws 100416a.JPG

That would've been my assessment. We'd love a test with 5000 saws or so, but I can't imagine there'd be a way to coordinate that without a lot of data adjustment from various sources - and loss of quality of the results.

Be interesting with a larger sample to see if you could draw lines at different places and see anything. Interesting that the '18-'28 period seems to be a little more variable, but then after that (the later period coinciding with circular saws killing the handsaw market) sees a decrease again in deviation.

No, 5000 saws aren't needed. The numbers I have (standard deviations and differences in central tendencies) at this point indicate that N is in the mid-20's, numbers that are easily assembled.

Without reading the whole thread ( landed here by mistake) 1906 was a big year for Disston.

In 1906 Disston installed their first electric crucible. Electric crucible means no coal fire means no coal dust swirling through the air in the foundry means no coal dust landing in the crucible in varying amounts.Coal is of course compressed carbonaceous matter with variation in coking grade and trace elements. The shift to electric crucibles meant far more precise control of the carbon content of the steel produced, and should result in a more consistant product. it also means adding more carbon to the mix to compensate for lack of coal dust. Carbon content determines hardness of course.

that may explain the 1906 anomalies.

John,

Indeed, I've been thinking about that transition as well.

By making all of these measurements I'm working toward the goal of a better explanation of why the saws from the Golden Age are so highly regarded. Better methods and controls very likely are the reason. Now I have a good idea of how many saws must be tested and how many test points per saw must be taken.

Other historical events, such as the introduction of fractional horsepower electrical motors for use in portable saws (~1930's) may also play a role in that Disston may have changed or cheapened manufacturing processes to reduce the product cost in the face of competition from the new type of saw. That competition likely ultimately lead to the demise of their production of handsaws in the way that they had been accustomed.

The Post-28 saws may be representative of the austerity imposed by the 1929 stock market crash.

The relatively good variability of the WW2 era saws may be due to the expanding use of hardness testing testing methods such as the Rockwell.

Regards,
Rob

Again - not pessimistically or cynically - a note of caution about the data from 5, or 20, saws out of a production of 1000s upon 1000s.

A couple qns ...

1. What is the R^2 score from a linear fit to the scatter plot in #273 ?

2. when you say average SD and median SD, by "SD" do you mean SD of (several measurements at separate places on one saw) ?

Thanks,
Paul

Paul,

It's a polynomial fit to the current data set. How exactly it's calculated I don't know.

The standard deviation is derived from five measurements per saw. I've run some tests with higher numbers of samples per saw by my statistician and he said not to bother, five points is fine.

With the current data set, divided into three groups, Before the Golden Era, Golden Era and After the Golden Era, P = 0.055 corresponding to a 1 in 18 possibility that the observed difference is not real.

The statistician has done a power analysis and given me an estimate of the appropriate number of saws to test from each of the seven production periods I'm looking at. These numbers are estimates but past experience has shown that 99.99 times out of 100 they're correct.

Now I have to plot a strategy to acquire the saws needed.

Regards,
Rob

49 saws now. Golden era saws denoted by the yellow boxes. Three total Acme 120's included in these data.

Historical hardness survey of Disston saws 110616a_Page_1.jpg

Historical hardness survey of Disston saws 110616a_Page_2.jpg

Historical hardness survey of Disston saws 110616a_Page_3.jpg

At this point the comparisons of the eight periods, identified by medallion and marking types, show only that the difference between the 1850 saws and the Golden Era has achieved statistical significance with P < 0.01 (Kruskall-Wallis H test). However, comparison of the Pre-Golden era saws, i.e. all saws made prior to 1896, to the Golden era saws reveals a significant difference, P = 0.05. Comparison of the Post-Golden era saws to the Golden era saws is also significant, P = 0.04.

The Acme 120's in these plots are denoted by purple datapoints.

Historical hardness survey of Disston saws 110816a.jpg

Disston lied about the hardness of the Acme 120 saws for nearly 20 years. The two earlier Acme 120's in my collection, each from a different 19th century production period, are no harder than any other saw of their times. The oldest 120 in my collection is from 1877-1878. The 120 saws were introduced in the Disston catalog of 1876. Only the most recent saw, bearing an 1896-1917 medallion, is harder than the average. None of the 19th century taper-ground backsaws I've tested to date is exceptionally hard either.

Did the advent of the first standardized hardness testing instrumentation, Brinnell in 1899/1900, have anything to do with this change? Other methods such as Rockwell (1914), Vickers (1924) and Knoop (1939) came around too late to be drivers of the change. Was Disston afraid of being caught or was this a product of standardized hardness testing? I wonder if the same holds true for the No. 77 saws?

BTW, Paul, the only entry, wins half-credit, i.e. 5 screws of his choice, for correctly identifying one of the two saws in the challenge I opened above.

Originally Posted by rob streeper

BTW, Paul, the only entry, wins half-credit, i.e. 5 screws of his choice, for correctly identifying one of the two saws in the challenge I opened above.

Rob

Does that mean I get 2.5 screws? Can I say that ? Actually I would treasure any number as they are beautifully made.

Congratulations on such a a daunting task you have undertaken, which I imagine will be ongoing at least for your own benefit as you accrue more saws.

The results I think reflect the development in technology with regards testing and also the increased consistency with the introduction of the electric furnace. Before that time much of the processes would have been hit and miss. I have always felt it was a little telling that in the Disston company's own words, only the best blades were reserved for the No.12. This, to my mind, automatically indicated that not all saw plates came out of the furnace the same and again after tempering not all plates were the same.

The process of selection too was almost certainly based on the experience of a handful of individuals: Most likely a good guide but by no means definitive.

Even the metallurgy must have varied in the early days.

So the so-called "golden era" emerged. I am not quite sure of the exact period this represents, although I would expect 1906 to 1928 particularly in the case of Disston, but it would have applied to the other major manufacturers too. Simonds had their electric furnace around the same time and had ceased hand saw production by 1926. Interestingly, Simonds changed their logo some time before their 1907 catalogue. I have interpreted it as a time of change, but that is my own supposition.

Atkins were getting their steel from the UK and you have to suppose similar advancements were occurring there too. I am not sure about the other major player, GH Bishop.

In any event, thank you for a most informative and exacting thread.

Regards
Paul

Paul,

Five screws with nuts of your choice, any combination.

I have more saws to add to the dataset, probably this weekend. Then I'll pass the data on to the statistician to see in a more definitive way where things stand.

Cheers,
Rob

There are two No. 12's in the data, both denoted by blue diamonds. The No. 12's have lower than average variance in their hardness, considerably less than the 120's, but there are only two of them at this point so it's premature to draw conclusions.

Historical hardness survey of Disston saws 110916a.jpg

58 saws now.

Historical hardness survey of Disston saws 111216a_Page_1.jpg

Historical hardness survey of Disston saws 111216a_Page_2.jpg

Historical hardness survey of Disston saws 111216a_Page_3.jpg