Cliff Stamp
BANNED
- Joined
- Oct 5, 1998
- Messages
- 17,562
Recently I noted a simple model to apply to edge retention and how to use and interpet it to comment on the relative performance of steels using a large comparision of several S30V and ZDP-189 blades :
http://www.bladeforums.com/forums/showthread.php?t=416896
Recently I set out to benchmark a U2 in SGPS at 62 HRC and chose the closest knife in geometry a Jess Horn in ZDP-189 at 65 HRC. Both knives have high hollow grinds on 0.095" thick steel. The U2 was reground by Krein and the edge is sharpened flat to the stone at 3.9 degrees per side. The Jess Horn is also sharpened in the same manner but since the grind is sabre/hollow it is significantly more obtuse at 7.9 degrees per side.
The first trial run was using a new cord which is very thin, only about a mm or so wide. A 50 g section of mild steel bar provided the tension. 20 m of 1/8" ridged cardboard was cut through 20 cm sections. The cutting was done on a slice through 3 cm of edge with the sharpness measured on the middle two cm. The 20 m was cut in 12 intervals with each each point consisting of an average of four sharpness measurements and five runs were made with the U2 and four with the Jess Horn. The results :
This is the kind of rough trial result that never actually leaves the lab. There are a couple of interesting trends but the graph is too noisy to tell if they are real or just random. The cord wasn't that consistent, repeat cuts at the same point could generate 0.75 cm and then 1.25 cm. There was far too much tension being used. The cord basically busted under the weight and tended to get tore apart rather than cleanly cut. So drop the weight back to 30 grams and take eight measurements of sharpness at each point. I ran 5 trials with the U2 and 6 with the Horn. I split the run up into 15 intervals to give more data to allow a better curve fit. Now at each point there was 40-48 measurements of sharpness for a total of 600-720 points in total. This gives :
The U2 gives better edge retention. To determine how much do a simple transform, take the ratios of the two data sets after subtracting the initial sharpness. Fit this curve to the same power law model. The coefficients of this curve then show directly how much the edge retention of one knife is ahead of the other. The transform :
The results :
The power law fit shows the power is barely different from zero which you would expect as these two steels should blunt in the same manner. The multiplier coefficient showing that the Jess Horn is blunting at about 75 (8) % as fast as the U2. Now this raises some interesting questions because the Jess Horn compared well to number of S30V blades and looking at SGPS it is essentially S30V with a lot less vanadium and a little more moly. Thus you would expect a greater hardening responce (vanadium ties up a lot of carbon) and a greater secondary hardening but the wear resistance should still likely favor S30V. I would be really curious if SGPS was tempered high.
Now is the difference seen here due to the hardness of the S30V vs SGPS? Or is it the difference in primary edge angle. If so how does the reduced edge angle have this effect because all the blunting is constrained to the 15 degree microbevels (checked under magnification) and the initial sharpness is near identical for both so it doesn't seem to effect the measure of sharpness. It could be because the lower edge angle decreases the lateral forces around the edge which I can see because there would be a significant reduction in localized distortion.
The above proposes the following work would be of interest. Take a few blades and run them through the above work with standard micro-bevels and 5,10,15,20 degree primary bevels, or as fine as increments as desired and see if there is a definate correlation. One other thing which comes to mind is that the tension used could effect the results. Consider a 4.5 tpi vs 11 tpi saw ran both very light and very hard cutting woods and the performance ratio is massively different. This would then logically imply that slicing with a very coarse micro-tooth pattern would favor a high cutting load and low cutting loads would favor finer finishes.
One thing I would note is that I really don't expect this kind of numerical detail from a general hobbiest looking at edge retention. However from makers/manufacturers who are seriously trying to produce meaningful data and make definate statemens about steel performance, the above is an illustration of how to do it using standard statistical methods. Anyone who wants to do the analysis but isn't sure of the math drop me an email and I'll either do it for you or give you the necessary code to do it yourself. There are freeware programs on any platform which can do this very quickly.
-Cliff
http://www.bladeforums.com/forums/showthread.php?t=416896
Recently I set out to benchmark a U2 in SGPS at 62 HRC and chose the closest knife in geometry a Jess Horn in ZDP-189 at 65 HRC. Both knives have high hollow grinds on 0.095" thick steel. The U2 was reground by Krein and the edge is sharpened flat to the stone at 3.9 degrees per side. The Jess Horn is also sharpened in the same manner but since the grind is sabre/hollow it is significantly more obtuse at 7.9 degrees per side.
The first trial run was using a new cord which is very thin, only about a mm or so wide. A 50 g section of mild steel bar provided the tension. 20 m of 1/8" ridged cardboard was cut through 20 cm sections. The cutting was done on a slice through 3 cm of edge with the sharpness measured on the middle two cm. The 20 m was cut in 12 intervals with each each point consisting of an average of four sharpness measurements and five runs were made with the U2 and four with the Jess Horn. The results :
This is the kind of rough trial result that never actually leaves the lab. There are a couple of interesting trends but the graph is too noisy to tell if they are real or just random. The cord wasn't that consistent, repeat cuts at the same point could generate 0.75 cm and then 1.25 cm. There was far too much tension being used. The cord basically busted under the weight and tended to get tore apart rather than cleanly cut. So drop the weight back to 30 grams and take eight measurements of sharpness at each point. I ran 5 trials with the U2 and 6 with the Horn. I split the run up into 15 intervals to give more data to allow a better curve fit. Now at each point there was 40-48 measurements of sharpness for a total of 600-720 points in total. This gives :
The U2 gives better edge retention. To determine how much do a simple transform, take the ratios of the two data sets after subtracting the initial sharpness. Fit this curve to the same power law model. The coefficients of this curve then show directly how much the edge retention of one knife is ahead of the other. The transform :
Code:
sr(x)=rr1*x**(rr2)
fit sr(x) 'card' every ::1::15 using:(($4-27.5)/($2-27.5)):(($5/$4)**2+($3/$2)**2)**0.5*($4/$2) via 'params\rr.prm'The results :
The power law fit shows the power is barely different from zero which you would expect as these two steels should blunt in the same manner. The multiplier coefficient showing that the Jess Horn is blunting at about 75 (8) % as fast as the U2. Now this raises some interesting questions because the Jess Horn compared well to number of S30V blades and looking at SGPS it is essentially S30V with a lot less vanadium and a little more moly. Thus you would expect a greater hardening responce (vanadium ties up a lot of carbon) and a greater secondary hardening but the wear resistance should still likely favor S30V. I would be really curious if SGPS was tempered high.
Now is the difference seen here due to the hardness of the S30V vs SGPS? Or is it the difference in primary edge angle. If so how does the reduced edge angle have this effect because all the blunting is constrained to the 15 degree microbevels (checked under magnification) and the initial sharpness is near identical for both so it doesn't seem to effect the measure of sharpness. It could be because the lower edge angle decreases the lateral forces around the edge which I can see because there would be a significant reduction in localized distortion.
The above proposes the following work would be of interest. Take a few blades and run them through the above work with standard micro-bevels and 5,10,15,20 degree primary bevels, or as fine as increments as desired and see if there is a definate correlation. One other thing which comes to mind is that the tension used could effect the results. Consider a 4.5 tpi vs 11 tpi saw ran both very light and very hard cutting woods and the performance ratio is massively different. This would then logically imply that slicing with a very coarse micro-tooth pattern would favor a high cutting load and low cutting loads would favor finer finishes.
One thing I would note is that I really don't expect this kind of numerical detail from a general hobbiest looking at edge retention. However from makers/manufacturers who are seriously trying to produce meaningful data and make definate statemens about steel performance, the above is an illustration of how to do it using standard statistical methods. Anyone who wants to do the analysis but isn't sure of the math drop me an email and I'll either do it for you or give you the necessary code to do it yourself. There are freeware programs on any platform which can do this very quickly.
-Cliff
