How to test sharpness in digits.

[nozh2002]

I make some device last night which show how many force you should put to cut thread:

It shows 50 gramms for shaving sharp knife and 200 gramms for dull blade. Now I can run one or other tests to dull the blade by cutting wood or rope or something else and check differences in sharpness in gramms.

This is a bulat knife which I like to test (bulat by Ivan Kirpichev, Tver, Russia).

What do you gentelmens think?

Thanks, Vassili.

 

[T. Erdelyi]

I like that, now I gotta make one.

 

[WadeF]

Where do I get a scale that measures in grams? Is that scale much more sensative than a postage scale?

Also for us to compare our results between us we'd all have to use the same kind of string.

 

[neosporin]

Now you just need a conraption to hold the knife and you're golden

 

[nozh2002]

Where do I get a scale that measures in grams? Is that scale much more sensative than a postage scale?

Also for us to compare our results between us we'd all have to use the same kind of string.

This scale I still from my wife in the kitchen.
Of course it depends on string properties etc - results can not be absolute and meaningfull only to show dynamic of sharpness during the test - like sahrpeness droped from one mark to another after cutting 10 cm of Jacarta rope...

So it is more for comparison at one testing session. My number will be different then your numbers but if I test different blade with same rope myself it may be usefull.

I fixed string with scotchtape.

This thing show so far good consistency.

Thanks, Vassili.

 

[Mirror-Saw]

Wade, you could get a scale that measures grams pretty easily, just check out some bong shops

 

[Laceration]

Seems like a less sophisiticated version of what Cliff Stamp uses. I never thought to use a kitchen scale like that though. Good idea.

 

[HoB]

Seem to me, at least in principle, identical to how Cliff Stamp tests for sharpness. I think its a pretty good test of sharpness. Since we don't really go into contest with each other, a standard thread isn't really required. I use this test more as a threshold test. I have a type of thread that I like and a fixed weight. If the knife cuts the thread, I declare it as sharp enough for me, if not, back to the stone. Works quite well. Dependened on your skill and expectation you can set the threshold arbitrary high or low. What ever suits your needs and purposes.

What I never liked about the test, is that is uses thread. The fibers are cut sequentially and the fibers are so small that even a dull knife cut rip through them one by one. I always felt that that would distored the test. I talked to Cliff about that and he had the idea of using music cassette tape, which I think should work better. But Cliff also mentioned that the required force to cut through the tape is quite high. Also a tape has the advantage of testing a longer edge length for sharpness than a thin thread.

 

[Gollnick]

This type of a test is one way to test relative sharpness. That can be good an useful. But, like any measurment scheme, it has its limitations. As long as you know and accept those limitations, it can be very useful.

Consider, for example, two knives sharpened such that they measure a certain sharpness on this device, perhaps 50g. Both knives are then used to, for example, chop through identical 2x4s. Both are then remeasured using the proposed method and the first measures 150g while the second measures 250g. It is now reasonable to say that the second knife demonstrated significantly poorer edge retention in this test. That is a way in which this type of method can be very useful and can bring some quantitative objectivity to such testing.

 

[Cliff Stamp]

The fibers are cut sequentially and the fibers are so small that even a dull knife cut rip through them one by one.

I tried mono fishing line and the results were no different, forces were different, but nothing different in behavior, I don't think there is a distortion by some kind of partial fractionation of the filaments.

As for the method in general, it is very sensitive if you use light thread, I use baisting thread which is "wafer thin", even the sharpest blades I have seen (commercial shaving disposable razor blades) require significant force to cut it, a blade that is any way dull will show a force ~10x optimal (i.e. you can see the edge has dulled visibly).

If you use thicker cord you are measuring cutting ability not just sharpness and thus the results will be effected by the geometry of the blade, as you proceed to thinner string you focus in on pure sharpness. The precision also increases rapidly for the same reason, you can get to within 5% easily if you average a few trials.

The biggest problem is that it will focus the sharpness test on a very small portion of a knife, so if you are actually going to compare two knives you better take a lot of readings on each blade to insure that your comparison isn't distorted by spot effects as blades dull randomally. I take about a dozen readings per inch usually.

It is really sensitive with the light thread, you can for example readily measure the difference between shaving sharp and hair above the skin sharp, and will show the difference in edge loss after just cutting a few pieces of rope even with the better steels.

Note however that this is one one of the measures of sharpness which is push cutting, you also have to measure the other which is slicing. A blade can be sharp in one of them and not the other, you can't truely optomize both at the same time. If you are doing a lot of slicing for example this is a poor way to measure sharpness as it tests push cutting and you are doing slicing for the blunting work.

If you want to measure slicing aggression, push down on the blade with a set amount of force and see how much of it you have to draw along the cord to cut it. This is effected by grit finish just like the push cuts are but in the opposite way generally, but there are limits.

As well do *NOT* use one sample of the cord for all your testing, you have no way to note then that any work you have with other rolls is the same. The best way is to buy a bunch and randomally sample from it. Otherwise use a stock roll and test other rolls against it on a benchmark blade and adjust your calibration baseline accordingly.

I started doing it when Phil Wilson noted he did all his rope cutting on a scale so that he could measure the blunting in a quantitative way and not just stop by feel or looking at the rope. He learned this from Goddard. I moved from the rope to the thread and really light cord for as noted in the above to focus in on sharpness.

Have fun with the loops, I have an edge trial on carpet scheduled for tomorrow which will use 4 blades, I plan to do at least 10 stages depending on how long the blades last, which means 480 loops of thread to be cut in total, that is about an hour of loop tying and cutting on its own. That is the worse part of it, cutting the rope under tension is *WAY* faster.

-Cliff

 

[nozh2002]

I got today "Grandma's Best 100% Mercerized Cotton crochet thread Style 295 21 White 10 size" it takes 100 gr for just sharpened knife and 400 gr (at the end of the scale) for dull knife (fortunately I found one).

I found that measure blade at several points actually show the state of the blade better then if you test sharpness with newspaper. I found dull portion in the middle of the blade this way, but it cuts newspaper perfectly.

Regarding to fibers I noticed that at this sizes fibers actually flatten and spread on the edge evenly so should not be much differense with solid string, of course it will not be the case for thick rope.

I'll continue my observation, but my point is to came up with some kind of measurement which anybody can maintain easely at home to be able to test knives himself. Some basic methodic which will be accepted by everybody and can be executed by everybody - then we will have much more test results from variety of different tests runed by anybody.

Thanks, Vassili.

P.S.
I should say that it is not my idea, but my Russian friend Nikolay who actiually former experimental phisicist - those guys always make some equipment to run their experiments from whatever they have around and he suggested me to make this "device".

 

[fulloflead]

What I never liked about the test, is that is uses thread. The fibers are cut sequentially and the fibers are so small that even a dull knife cut rip through them one by one. I always felt that that would distored the test. I talked to Cliff about that and he had the idea of using music cassette tape, which I think should work better. But Cliff also mentioned that the required force to cut through the tape is quite high. Also a tape has the advantage of testing a longer edge length for sharpness than a thin thread.

How about something cylindrical like a drinking straw? It might even allow you to use a slicing motion.

You need something under the scale pointer that will get pushed down and stay in place so it will mark the heaviest point (peak pressure) throughout the motion. (Hard to explain.)

I like the idea overall. It'd be neat to create some sort of standard system that we all could buy and set up. Then we could test various things like razor blades to create benchmarks.

 

[Megalobyte]

And here I thought with the title: "How to test sharpness in digits" that cutting off your digits/fingers would be involved.

"My knife is 8 fingers sharp..."

 

[JDBLADE]

I have found Kangaroo Lace a good medium because you can use it for the push cut and the slicing cut plus you can also us a 45 degree angle and with all cuts you have a more visible cut for examination then with string or cotton.

 

[Planterz]

And here I thought with the title: "How to test sharpness in digits" that cutting off your digits/fingers would be involved.

"My knife is 8 fingers sharp..."

That's what I was thinking.

"I closed the Spyderco Manix on my hand and cut 3 fingers off. The Strider only cut two off...."

 

[HoB]

I tried mono fishing line and the results were no different, forces were different, but nothing different in behavior, I don't think there is a distortion by some kind of partial fractionation of the filaments.

That is excellent information. Thank you!

How about something cylindrical like a drinking straw? It might even allow you to use a slicing motion.

I like the idea, but it would be difficult to keep the "straw" from deforming which would exaggerate the results. A dull blade will deform the straw more before the straw gets cut and because of the deformation will increase the surface presented to the edge, which in turn would require more pressure and would lead to greater deformation. It could help to distinguish between two blades that are very close in sharpness though. But as Cliff pointed out, the tread cutting test does already very well in distinguishing between different sharpness levels.

 

[nozh2002]

I did some testing - in general I tested methodology.

! sharpen that bulat knife and measure sharpness with this gadget.
I did this around 1/4, 1/2 and 3/4 of blade

After diamond rods on Spyderco Sharpmaker (not shaving):
1/4 220 180 200 180 200 196
1/2 220 180 200 200 220 204
3/4 160 200 220 170 150 180
Average 193

Medium rods (unexpectedly it starts shaving after this):
1/4 100 140 140 120 130 126
1/2 130 130 120 140 120 128
3/4 120 140 130 140 130 132
Average 129

Fine Rods (hair popping sharp):
1/4 100 120 115 140 140 110 90 140 80 70 110
1/2 120 140 90 90 110 100 80 70 100 100
3/4 90 100 105 130 100 105
Average 105

Ultra Fine Rods (sharp!!!)
1/4 70 80 150 70 80 90
1/2 70 100 90 100 90 90
3/4 90 100 105 130 100 70 80 90 140 130 107
Average 96

I did additional checks when I feel suspitiont (too different results etc), but all data is here

So sharpness dynamic is 193 - 129 - 105 - 96. I forgot all what I learn in about mthamatical statistic etc, but it seems to me that even for simple average number we have consistent behaviour.

It actually shows difference between Fine and UltraFine which was hard to recognize for sure before.

Next day I deside to make intensive testing of this methodology. I run 30 measurement for each part of the blade:

1/4

90, 80, 140, 90, 70, 140, 80, 130, 70, 80
70, 100, 80, 90, 170, 120, 110, 120, 90, 50
70, 80, 90, 60, 50, 70, 70, 100, 80, 80
Average 91

1/2

120, 70, 100, 80, 90, 90, 110, 90, 70, 90
80, 90, 80, 70, 70, 100, 150,120, 70, 110
70, 70, 70, 170, 90, 70, 110, 70,90, 100
Average 92

3/4

100, 90, 90, 80, 70, 80, 50, 60, 80, 50
50, 60, 50, 90, 90, 70, 120, 70, 90, 90
80, 60, 70, 80, 50, 80, 80, 100, 70, 60
Average 75

More tests done - better average number. Sometimes I may notice fluctuation - like 170 gramm, which is way unexpected (thread is just stronger there or something), it should be some formula from mathematical statistic to take care of this, but I forgot, may be someone can halp me with this. And last test as day before kind of different - may be I just tiered, or it somehow depends on temperature or humidity, I made it later at night then first two.

I think 10 measurement will be enough for simple testing, even 5.

I am very happy because this is what I looking for - objective, consistant testing method availabale for everyone.

This is video how I do this

http://playground.sun.com/~vasya/thread.mpg

I encurage everybody to try this. Now you may came up with you own test - cutting wood, rope, brass, frozen chicken etc and measure sharpness and compare it between different knives.

Thanks, Vassili.

 

[fulloflead]

suspitiont

I had to look that up! It even confused dictionary.com

Did you mean:

1 entry found for suspicioned.
sus·pi·cion ** (*P*)**Pronunciation Key**(s-spshn)
n.

1. The act of suspecting something, especially something wrong, on little evidence or without proof.
2. The condition of being suspected, especially of wrongdoing.
3. A state of uncertainty; doubt. See Synonyms at uncertainty.
4. A minute amount or slight indication; a trace.

Great post! I can just see you suffering over that project like a mad scientist. You have the same affliction many of us do.

I would LOVE to see that same test done [hint hint] where you take several of the same knife and sharpened it at different angles.

I would also love to see it after the edge is stropped on a loaded leather strop to see what the MEASURABLE results. I've preferred the strop for years and I've seen how one can improve an edge significantly, but I'd love to see the numbers.

Nice job. You are a contributing asset to the forum.

 

[Cliff Stamp]

Fine Rods (hair popping sharp)

104 +/- 4

Ultra Fine Rods (sharp!!!)

97 +/- 5

The numbers after the +/- are standard deviations in the means. These are how you would expect the numbers to vary if you repeat the testing.

Specifically 95% of the time you should get a value within double them so you are looking at :

96 - 112 : for the fine rods

87 - 107 : for the ultra fine rods

I would argue these are not significantly different.

... it should be some formula from mathematical statistic to take care of this

Use medians instead of averages, they ignore outliers. With averages you want to calculate the standard deviations as well as it indicates the consistency of the sharpening.

I think 10 measurement will be enough for simple testing, even 5.

If the blade is consistently sharpened yes, however try some production knives or knives used and blunted, there the sharpness can vary significantly along the edge.

If you pick up a Ti-83+, cheap on ebay, these calculations are very easy.

-Cliff

 

[nozh2002]

This is why I post numbers here. So you may interpret it right.

Cliff can you give me exact formula. I am not only forget what I accidently picked up in my university years, but also I don't know terminology in English.

Can you applay same to last 90 measurements? I think thay show better sharpness after Ultra Fine.

Thanks, Vassili.