The Testing of Steel for Leather Paring Knives

[Update 29 Feb 2016: Starrett now tells me their red stripe blades are M2]

As the result of my own experience,  feedback from some of my colleagues, and the results of scientific testing, I’ve become convinced that A2 cryro and machine hacksaw blades (M2 or M3 steel) are the best steel types for bookbinding knives.  By best, I mean they offer what I consider an optimum balance of edge retention, initial sharpness, ease of resharpening and price for a leather paring knife.

Four leather paring knives were tested by Cutlery and Allied Trades Research Association (CARTA), located in Sheffield, England. The knives were tested for Edge Angle, Hardness, Initial Cutting Performance  and Cutting Edge Retention.  The steels tested were O1, A2, Machine hacksaw blade, and T15.  All the steels rated very good to excellent, overall.  O1 is a popular, standard knife steel, which I make many styles of knives out of, and it works very well on regular, vegetable tanned goat.  Machinery’s Handbook recommends it when extreme sharpness is required. A2 is a newer, high tech steel gaining favor among woodworkers, since it holds an edge longer than O1.    Most large woodworker blade suppliers make both O1 and A2 versions of plane blades, so there is some personal preference involved- do you like to spend a slightly longer time sharpening (A2) or sharpen more often (O1).  Some type of overall time analysis might be conducted, but I can’t imagine it would be very informative given the wide range of unique variables the affect knives in use and when resharpening.  Some bookbinders also use knives made from D2- but I find the high chromium content and coarse carbide structure (up to 50 microns!) make it cut poorly, more like stainless steel than a tool steel.  Fully hardened machine hacksaw blades are traditional metal (starting mid. 20th C.?) for bookbinders to make paring knives from, I assume since it can be purchased already hardened and shaped by stock reduction. Starrett told me their “Red Stripe” blades are made from M3 steel, and I have read the English Eclipse ones are made from M2.  T15 is a super expensive, very high tech steel that outperformed all the other blades, but the inital cost to me, and the number of grinding belts it ate up and time it took when shaping would result in a $750 knife, and  I doubt anyone would purchase a $750 knife.  And you would almost require a set of diamond stones to resharpen it.

I shaped and sharpened all of these knives by hand to a 13 degree angle.  I suspect it is one of the first times that these types of steel were tested at such low angles– at typical woodworking blade is usually 25+ degrees.  The blades were shaped on a 2 x 72″ Cootie Belt Grinder, progressing through US grits of 36, 100, 220, followed by hand sharpening on 3M microfinishing films of 40, 15, 5  microns, then stropped on vegetable tanned horsebutt with a .5 micron chromium oxide and finished by stropping on the flesh side of undressed vegetable tanned calfskin.  The sharpness was tested on a pack of test cards containing 5% silica for 60 cutting cycles, 50N test load and 50mm/sec. test speed.  CATRA has invented and sells many machines for sharpening and knife testing.


Although I was attempting a 13 degree angle, the edge angled measured between 14-16 degrees for all of the knives when measured with a laser goniometer.  I think this is the result of not being able to hold the knife angle consistently enough when sharpening and the result of stropping.  There also was a general tendency for the angles to be slightly more acute (.5- 1.5 degrees less) at the sides of the cutting edge as compared with the middle.  This is most likely the result of natural hand motion when sharpening, since I usually sharpen parallel to the cutting edge. I suspect if I was in the habit of sharpening perpendicular to the cutting edge, the edge angle would be even more obtuse.


Average from three testing points near the cutting edge in Rockwell C scale.

A2 – 62

Hacksaw (M3) – 64

O1 – 64

T15 – 65


This is how “sharp” the knife is; how far it penetrated into the cards during the first three cutting strokes.

Tested in accordance with BS EN ISO  8442-5: 2004, Part 5 (Clause 3.4- Cutting Performance)

A2 – 107

Hacksaw (M3) – 107

O1 – 98

T15 – 116


This is how long the edge lasted– ie. the ability for the edge to resist wear.  This is the cumulative depth of  60 cycles of cutting the test pack.

A2 – 522

Hacksaw (M3) – 586

O1 – 395

T15 –  921


A2 is light blue

Hacksaw is green

O1 is dark blue

T15 is red

CONCLUSIONS. Initially, A2 has a slightly slower rate of dulling, which may be an advantage, but later in the test the Hacksaw slightly surpasses it.  The O1 seems to have a very constant, predictable rate of dulling- the graph is very smooth.  The T15 is incredible- even at the end it was still cutting quite deep, and each cut it still quite deep.  Since O1 is about 400, and A2 is about 600, does this mean A2 is 20% better?  Given all the complex variables in use, it is hard to accurately  observe.  As with most scientific testing, a primary conclusion is the need for more testing.  I would like to compare bevel angles at 2.5 degree increments, add a number of different types of steel, like some of the M and S tool steel classes, maybe some of the new high carbon stainless steels and instead of testing with a sample card containing silica, test actual tanned and tawed leather.  Also the pressure on the blade may be much more than is necessary (or possible) to pare leather, and the sawing motion of the testing machine is different than a more static blade motion that bookbinders use.


20 thoughts on “The Testing of Steel for Leather Paring Knives

  1. Tom Conroy

    Machine Hacksaw Blades.

    In considering machine hacksaw blades as raw materials for paring knives, some distinctions should be made.

    Currently, most machine hacksaw blades are “bimetallic” with a soft, tough back and a hard, narrow edge. A friend of mine has hardened and tempered common machine hacksaw blades and got good knives out of them. Unfortunately, hardware-store bimetallic hand hacksaw blades are made with low-carbon steel backs and cannot be hardened into good spine lifting knives (I have tried).

    Current “allhard” machine hacksaw blades are made of high-speed steel, an alloy formulated to remain hard at red heat. Unfortunately, in the inevitable tradeoff of properties, high-speed steel will not take an edge as sharp as high-carbon steel, and it seems to me to be distinctly harder to sharpen. My experience with new British HSS machine hacksaw blades has been that they make decent, but not spectacular, lifting and paring knives.

    I too remember the high praise the older British binders had for hacksaw blades as raw materials for paring knives, but as I remember it they always specified “prewar” or at least “old” blades. I suspect that they were praising high carbon steel blades made before high speed steel became common, although it is possible that they were praising allhard HSS as opposed to bimetallic blades.

    Currently one supplier is selling paring knives reground from discarded machine hacksaw blades. The original blue paint is still partly present, so they clearly have not been hardened and tempered, but it is not so clear whether the blades are allhard HSS or common bimetallic.

  2. Jeff Peachey

    I guess I should have been clear that I had new, allhard blades tested. The only current English ones I have examined are branded “Eclipse” , but the one I had tested is the Starrett Red stripe. In my experience and according to the test results, they posses greater edge retention and initial cutting performance than O1 (a high-carbon steel). The do take longer to sharpen, though.

    I’m not sure what you mean by the paint on a hacksaw blade- all the Starret’s are painted after they are tempered and hardened–most binders shape them by stock reduction. I would like to examine an older hacksaw blade if anyone cares to trade one with me!

  3. Tom Conroy

    Sorry about the confusion of my last paragraph. I meant that the survival of the paint shows that the blades have not been hardened and tempered **by the knifemaker**, which shows that the knifemaker did not deliberately and intelligently use discard bimetallic blades. However, in the absence of a clear statement from the supplier, I wondered if the supplier said “grind down old hacksaw blades,” and the knifemaker did that with any old hacksaw blades that came his way, whether allhard or bimetallic.

  4. Tom Conroy

    Hardness of common paring knives

    Several years ago a friend of mine kindly tested the Rockwell C hardness of one of my old Barnsley paring knives, and of one of the Howill knives now being sold by Hewit. George Barnsley and Sons was an old-established maker of cobblers’ and other leatherworking tools. Their knives were good reliable working tools, holding a decent edge for a reasonable time and easy to resharpen. I have used them for most of my work for about twenty-five years. When tested, the hardness of the Barnsley was remarkably, almost shockingly, low: rC55 near the edge, down to 54 about halfway down the blade. The Howill knife, carried by Hewit after Barnsley went out of business several years ago and made as a close copy of the Barnsley, is very similar: 54 near the edge and 53 halfway back, so close as to have no real significance.

    I’m still using my Barnsleys, which I’m accustomed to and which work well. In theory I would say that the good range for hardness for leatherworking knives is, as with woodworking tools, in the neighborhood of 60-62, and in theory I despise anything even as low as 59. In practice: don’t change a winning game.

    In a less extreme way, though, I am definitely on the side of easier sharpening with the trade-off slightly shorter edge retention. I think that differences in edge retention are rarely much noticed in work, but ease of sharpening is very definitely noticed. It seems to me that most people want an edge that lasts a long time primarily because they hate and fear sharpening, not because the longer working period has an objective advantage. Then, better edge retention becomes an excuse not to sharpen and they continue to work with a half-dull knife far longer than they would if the knife were easy to sharpen. Its a psychological effect, not a steel property, but for me it speaks for a steel that offers potential maximum sharpness and ease of sharpening without compromising these away in favor of toughness, edge retention, high speed, corrosion resistance, depth of hardening (which means deeper than 1.5″) or any of the other properties which allow alloy steel to be tailored to different needs.

  5. Jeff Peachey Post author

    Another aspect to consider is the type of sharpening system you are using. Various steels seem to work better or worse with different methods- oilstones, waterstones, ceramic stones, lapping compounds, abrasive papers, microfinishing films, diamond stones and compounds and various water/ ceramic stone combinations. I have a D2 pocketknife which takes forever to sharpen with microfinishing films- diamond works great, however. But A2 and hacksaw sharpen quite quickly on microfinishing films, but I wouldn’t recommend cheaper (Sun or King) waterstones.

  6. Jeff Peachey Post author

    In rereading this post, since I have linked it to the knife catalog, it appears I missed one comment Tom Conroy made, “Unfortunately, in the inevitable tradeoff of properties, high-speed steel will not take an edge as sharp as high-carbon steel.”

    The chart above clearly shows that in my testing, the M3 hacksaw blades had the highest ‘initial cutting performance’ ie.sharpness of all the steels tested.

  7. Ian Burns

    What a great article thank you. Im currently testing D2 and RWL-34 to replace my O1 knifes and your comments on D2 were interesting.

  8. Eric J

    Jeff, were your mentioned angles per side, or included angle? T15 is a very hard grade of HSS, so it makes sense to me that it would hold an edge the longest.

  9. Eric J

    Gotcha, thanks. I recently started working on a straight razor made from M2, looking to hit around 16° – 18° included with that. I have done some testing and found that using anything other than diamond or CBN abrasives leaves quite a lot of exposed/jutting carbides and torn out carbides. The particular steel I’m using contains carbides that are about 4 – 6 microns in diameter.

  10. Eric J

    Oh, and in my testing, the HSS is easily able to take as keen an edge as plain carbon steel at 18° included angle. The only thing you have to watch out for is carbide tear-out at the edge.

  11. Jeff Peachey Post author

    Thanks for the information. A couple of questions: How did you do your initial cutting performance testing? What other abrasives did you test (and in what substrate?) How did you measure the micron size of the carbides?

    I’ve occasionally encountered this tear out, but much less frequently than with A2.

    I think M2 would make a great straight razor, because of the fact it seems to strop back into shape (I usually use CrO2 and horsebutt) quicker than other steels I’ve used, though this is pretty subjective.

  12. Eric J

    Hey Jeff,

    I was able to measure the carbide size using a modified USB high magnification microscope. Once the scope is calibrated to something of known dimension it can be used to measure via software by clicking on endpoints of an object (i.e. diameter of a spherical object). I am currently testing for keenness using bounced light at a very narrow angle from the apex. I’ll try to include a link to some of my testing here:

    I’ve tried several natural stones commonly used for razors as well as synthetics, and diamond. Diamond of course doing the best as far as edge cleanliness.

  13. Eric J

    Thanks. If you check back there in a day or two I can put up a test shot or two for you if you’d like. I could baseline it with an 8k diamond stone and hit it on a CrO strop and then bare leather so you can get an idea of what you’re doing to your knife at the apex. Or you could outline your full sharpening routine and I could do my best to use the same progression (depends if I have all the stuff you used – I probably do or at least close).

  14. Jeff Peachey Post author

    Sure, I use 80-40-15-5 (sometimes 1) Microfinishing film, then .5 CrO2 and bare leather, all free hand side to side. CATRA measured the edge after all of this with a laser and it usually was around 15 degrees, which works great for leather and the microbevel easy to maintain in practice.

    I suspect there are many variables when freehand sharpening with different techniques and pressure, etc….

  15. Eric J

    Yup, all sorts of variables. Ideally, if using lapping film (mostly aluminum oxide abrasive on these) you’ll want to keep pressure as light as possible to prevent carbide tear-out. I would think about moving to diamond stones and diamond or CBN strop paste if you want to get even a little more edge life – but you will still get pretty good edges with what you’re using now.

    For the edge shown in the latest set of scope shots (at the link I posted earlier) I ran a quick progression using the DMT EEF (3 micron) plate, then switched to 3 micron and 1 micron AlOx lapping film in turn, before stropping on CrOx for 30 laps and 60 more on plain leather. It showed some tendency to cut around the carbides and leave them exposed in even that small amount of steel removal with the AlOx abrasives.

    Another site you might find interesting is my friend Todd’s site – he has access to a Scanning Electron Microscope and is very interested in this stuff too! Check it out here:

  16. Jeff Peachey Post author

    Thanks for the info, if you post the images for comparison online somewhere let me know.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s