A shot of a prototype knife made from a tension bar :
A piece of tension bar, mild steel, from a discarded chain link fence was used as stock for a small utility knife. The knife was formed using a 4.5" angle grinder. The bevels were refined and sharpened with a bastard file. The steel formed a burr readily during the filing which was removed with edge-into honing. The edge was sharpened to a bevel of under 25 degrees included. A piece of known steel like colded roll bar stock of a specific carbon content would have allowed a more definate benchmark, this project however was started mainly for amusement and expanded due to the performance of the blade. It was inspired by a conversation with a custom knifemaker, Muhamad Irwan, who had made such knives out of mild steel.
The knife readily cut grass, weeds and miscellaneous vegetation with the filed finish. The edge took no damage and was still aggressive after a few dozen swings. It then cut at some alders a half an inch thick and under, which it easily chopped through. A piece of an Alder was then placed on top of a 6x6, and diced up into sections, trimming off the secondary growth and then moving down the trunk. The edge was still aggressive, no excessive blunting.
The knife was then stabbed into the 6x6 a half a dozen times, no damage. It could break pieces out of the wood on shallow stabs but prying after a stab half an inch deep bent the blade readily. Attempting to restore the blade by reverse prying would bend it in another place. The knife was restored to semi-straight by cold work with a hammer. It readily split some scrap and clear sticks using a framing hammer as a mallet. However during an attempt to split a knotty small stick (under three inches) the blade actually bent down under the impacts of the hammer - the split was aborted.
The bevels were cleaned up with a one inch belt sander using a worn 100 grit AO belt. The basic geometry was not altered, just removed the dips from the initial rough grinding and evened out the edge which was wavy because of the prying. After the grinding the edge scraped shaved a little, and would smoothly slice straight down into a sheet of newsprint. The profile was full convex, very close to flat, no secondary edge bevel, and the curvature close to the edge (0.024" thick) was 12 degrees per side, and went down to 8-10 in the shoulder.
On 3/8" hemp, with the finish from the 100 grit AO belt, the tension bar knife was used on a slice with a two inch draw for a total of three runs. Each run included a complete sharpening. The edge was checked for consistency under magnification after sharpening and showed micro-teeth up to 0.15 millimeters deep. The first run had a slight burr remaining, which was removed in the second and third run through a more careful sharpening. Averaging all three runs, the blade started off requiring 22 (3) lbs. The force required to cut the hemp smoothly increased, and was 40 (3) lbs after 126 cuts. In some detail :
| # hemp cuts | force on cuts |
|---|---|
| lbs | |
| 2 | 22.0 +/- 3.3 |
| 6 | 25.0 +/- 3.8 |
| 14 | 25.0 +/- 2.0 |
| 30 | 28.0 +/- 1.6 |
| 62 | 32.0 +/- 1.1 |
| 126 | 40.0 +/- 2.6 |
The burr on the initial run was assumed initially to be the way the edge was formed as it was the first time mild steel had been sharpened in such a manner. However as the next two runs came out clean this conclusion was in error. The first run was included in the average (actually median) to set a lower bound, it just raised the variance of the results slightly. The knife was also a bit bent during the cutting due to previous prying which hampered the performance and the second and third runs were taken with thicker and more obtuse edges due to the repeated sharpening, 14 and 14.5 degrees per side respectively.
After the rope cutting, the edge showed little aggression and took 5+ centimeters of edge to slice 1/4" poly under a 1000 g load, and could not slice newsprint. Under magnification significant regions of the edge were worn completely smooth, up to millimeters in length. The remaining microteeth were vastly reduced in size, down to a third or less. The edge was also very distorted, nothing which could be seen by eye, but under magnification dents could be seen up to 0.1 millimeters to the side. After steeling the performance would increase, but only back up to ~28 lbs, not close to optimal, again indicating the effect of significant wear.
As a reference, a SOG SEAL with a 600 grit DMT finish took ~40 lbs to make the first cuts on the hemp, and after only 14 slices on the hemp the force required exceeded 50 lbs. The tension bar knife thus showed much better cutting efficiency and edge lifetime. This shows quite clearly the massive influence that geometry and grit choice for sharpening can make on cutting efficient and edge holding.
For use on cardboard the edge was polished using waterstones and finished with 0.5 micron chroimum/aluminum oxide loaded leather and a final micro bevel with a fine ceramic rod set at 22 degrees. The knife easily shaved and push cut newsprint. It was then used to cut 1/8" thick, ridged cardboard, with full slice over the three inch blade. It cut six meters of cardboard before it would rip instead of cut. It was then aligned using a smooth steel from Razor Edge, and would cut another meter of cardboard before ripping it. The knife was then again steeled and another meter of cardboard cut before the blade would again tear instead of cut. The knife was brought back to a fine shaving finish with three passes per side on a fine ceramic rod.
In the kitchen, it was used to make a salad, with a Japanese styled utility knife as a reference. On mushrooms there was no advantage to the Japanese knife as both knives were razor sharp and mushrooms are soft so they are not influenced significantly by blade geometry as they can't generate binding forces, similar for green onions and chicken pieces. A tomato was also cleanly cut by the mild steel blade, but the blade was too thick to make extremely thin slices. On the celery the thickness of the tension bar knife could again be felt as the celery was rigid enough to induce enough wedging to be readily noticed. No significant blunting was induced. It was also used to cut up a couple of turnip slices into small strips, and then peel six unwashed potatoes. It handled well for the potato peeling, but was too thick for the turnip slices. After the cutting it was still shaving sharp, no real effect on the edge.
For wood working, the knife was used to hack down an Alder about one inch thick. The stick was bent, and given a few pops with the knife. The blade easily lopped off the branches with wrist flicks, removed the bark, and pointed the end, which was cracked off and repointed. About 100 slices were used for the pointing in total. A 1.5" thick piece of seasoned pine (two years old), was also pointed with knots removed by cutting straight down to section them and then chiseling them off. About fifty cuts were made to shape the point, more force here than on the Alders. The knife was less than 100% on shaving, but could still pop down dandelions, however could not cut individual stalks of grass. A couple of passes on a smooth steel and it was back to 100%.
It was also used to cut down four Alders about two inches thick. Various techniques were used, the optimal being to bend the wood under tension and cut straight through with the knife which required from 15-25 slices. The knife was used to limb out the sticks which were from 7-9 feet in length, and had 2-3 major branches from 0.5 to 0.75 inch thick, which could be used to frame out the sides of a shelter. The many smaller branches could then be used to weave in the gaps. The blade remained sharp enough to pop off the small branches, 1/16" thick to the end. After the chopping, the knife still had the ability to cut the heads off of grass stalks with a wrist pop. Later that night it was used to cut paper and tape for packaging, just a few pieces of each. Steeling made little difference to the edge however five passes per side on a fine ceramic rod and it was back to smoothly shaving sharp again.
The knife was batoned through the same piece of seasoned pine. Starting with low force from the baton (18 oz ball pien hammer) but towards the end the spine was getting flattened from the impacts, and the piece of wood broke when the first cut was attempted on the back side after cutting half way through on the front. The knife was then used to make a large notch through a pressure treated and well seasoned, 6x6. The notch was two inches deep and four inches wide through one corner. The edge while now blunted, was free of visible damage, and nothing could be felt by thumbnail. The edge was still sharp enough to pop dandelions heads. It needed 10 passes per side on the ceramic rod to be restored to 100%.
Using the piece of seasoned wood earlier batoned in half, the knife made one teaspoon of fine scrapings for fire starting. The edge suffered no damage and was restored to shaving with a couple of passes on a smooth steel. Some hairs were missing on a shave, so a couple of passes on the ceramic rod were used to set it back to 100% shaving.
On hard plastic, the knife cut the top off a pop bottle, made multiple cuts down the sides, and trimmed off the bottom section which is very thick and hard plastic. No visible damage to the edge, no reflection of light.
To check work on bones, a chicken was separated and cut into chunks for stock. The legs and wings were first removed and disjointed. The blade cut through the flesh, fat and tendons and other connective tissues readily. The cutting ability was high as soft materials only require a high sharpness, and the blade was shaving sharp. No loss of aggression was noted after the cutting, and much more was done that necessary to actually prepare the chicken, the breasts were cut into chunks and so on. The rest of the chicken was broken up into fist size pieces breaking the joints. The knife was then used to cut the small rib bones, through the spine across on a slant, and then to the heavier bones, some of which required body weight pressure on the blade and a back and forth rocking motion to shear through the bones, about six times or so. The edge was still free of damage, it would scrape a few hairs on one side, none on the other. After two passes per side on a smooth steel, the blade was sharper than when initially finished on the ceramic rod, so little or wear had taken place.
Cutting light metals : the knife was stabbed into a soup can lid (0.008" thick) and then worked across it to the edge. This was repeated making a plus sign which was pried open. The edge was now reflecting light along the contact areas. The can was washed out and the knife stabbed into the side which was of similar thickness but ribbed and again worked back and forth to make a slit up to the top lid. This didn't significantly effect the blade further. The lid of the can was formed by a double layer of much heavier metal, 0.038" thick. The knife had a *lot* of difficulty in trying to cut this metal on a push or a pull, a serious attempt was made with both methods, getting only to 3/4 of the way through. The blade was then inserted under the cut with the blade resting on the opposite side of the top of the can to serve as a leverage point, the knife was then able to cut through the lid by pressing heavily down on the handle. This push induced a dent in the edge about 0.5 cm long and the blade was 0.015" thick at the back of the dent. The pull induced far less damage, about 0.2 to 0.3 millimeters deep, and less than a millimeters in width. The rest of the blade also suffered some impaction, as the knife slipped along it on both cuts after failing to go through the lid, the impaction was less than 0.1 millimeters in depth.
As a reference, the metal cutting was repeated by a "stainless steel made in Japan" kitchen knife, 0.035" thick, with an edge grind of 9 degrees per side with a 22 degree per side micro bevel with a fine ceramic rod. The kitchen knife cut through the lid on the same type of can with no visible damage. It made a cut up the side with no further problems and was thin enough to just be pulled through no need to be worked from side to side. Again a leverage technique was used to cut through the lid which induced a small bend in the edge, 0.1 millimeters to the side about one millimeter long. As a check the side cut was repeated, and an attempt made to cut through the lid on a pull, this failed, though a push succeeded. The blade however was so thin and weak that it was flexing significantly on both attempts. The pull induced small chips about 0.1 millimeters deep, the push which was much heavier and made a chip about 0.5 millimeters deep and twice as long. The push chip was large enough to be visible by eye, the knife was 0.005" thick at the back of the chip. The push made a larger chip mainly due to a flex in the knife which caused it to bend during the cut and thus induced a lateral snap on the edge - the knife was simply too thin and didn't have enough stiffness to remain straight. The knife was sharpened in under a minute, 1000 and 4000 grit waterstones followed by a 5x5 sharpening on a fine diamond rod which left an edge which could cleanly slice newsprint.
Due to the massive difference in blade geometries, the two trials can't be completely contrasted based on steel differences alone. The thinner blade generally went through the metal easier however due to flexing it took much more strain when trying to cut through the can lid. Thus it can be concluded with little surprise that it is a much harder and stronger steel.
Geometry defines the limits of cutting performance, steel defines the limits of functional geometry which are influenced strongly by user skill level and physical ability. The above work is mainly to set a very low end benchmark and show that even in extreme cases, the difference in performance between steels can be overcome by geometry and optimal grit selection. Addition work was later done with another tension bar knife of similar style.
Comments cna be sent to : cliffstamp[REMOVE]@cutleryscience.com or by posting in the following thread :
| Updated : | Jul 10 : 2006 |
| Originally written : | Jul 14 : 2003 |