Chisel Test 2 – The Test

This is what I did to test chisels that are shown in the videos on page 1. Again, every time I talk about those videos, I have to apologize for how hard they are to get information from and when there are six relatively good chisels, how little you can actually do to ethically say “this one is good and the others are terrible and nobody could work with them”. We didn’t include chisels like that. The most highly alloyed in the whole test is 52100, which is similar to O1 in terms of alloying mass, though a little different in flavor. Everything else, though not tested XRF, is by feel, more plain. I can vouch for 26c3, but the India stone communicates how plain the two Chinese chisels are and how plain the Ricther is. All three are steel with additions for properties, and not additions for wear or to impart really high toughness like 52100.

But regardless of that, here are the details:

Sharpening

1. Ground flat bevel at 23 degrees on adhesive backed paper wetted to prevent any heat issues
2. Honed at a very slight lift on fine india and “Dan’s hard” a less than ultrafine stone, but still fine
3. Buffed tangentially three passes for each chisel, which protects the edge a little bit but is not the full unicorn treatment.

From a later trial, this is much more protection than something like a 25 degree angle, but less than 33 or 34 degrees.

The Test

• 30 ounce verawood mallet wrapped with horse butt.
• Red oak, flat sawn , so quartered orientation at the edge. Each chisel mallets two separate blocks of wood 3/4″ thick x 1 1/4″ long by 1 3/8″ deep. The wood is marked every 1/8th inch to make sure the malleted leaves removed are even.
• Malleting is done briskly, but it is not intentionally harsh and the chisel is not allowed to bounce or wobble while being struck.
• The work is straight in, with the waste side forward and bevel of the chisel forward.
• The first cut in each leaf is the same corner of a chisel embedded. Meaning 3/4ths of the width is chopped off of the test piece, and then the last quarter is chopped off with the same side that was already in the cut. This is important to identify corner issues that may not show up in a test where corners are never in the cut.
• If a chisel is wider than needed to perform the prior bullet point, the excess width is not contacting wood anywhere in the test. that means the same working length of the edge on each chisel should do the same amount of work the same way.
• Two separate cells are chosen and each chisel in the “core four” works one cell, an then the others do, and then this is repeated once more

That’s it. I strike all of the chisels the same, as you’ll end up doing if you’re a hand woodworker and with experience. Work is in rhythm and much the same within a task.

Red oak isn’t terribly hard on edges, but it’s not easy on edges like cherry or mahogany. It makes a good test medium because it’s hardwood and not something abrasive but not hard or hard to the point that the stock is rarely used. For example, this test done with gombeira or katalox would’ve provided damage at even normal setups, but it’s not useful in general to do that because people hand working significant amounts of either of those are rare. Damage that occurs relatively so in katalox isn’t necessarily predictive of damage that will happen in cherry or soft maple. Even in relative terms.

A Control Check

Richter and 100cr-v Chinese chisels were purchased after the “core four” were tested, with the Richter test being done soon after, but the latter being done long enough later that the sharpening feel wasn’t fresh in my mind.

The Richter test is worse than the others notably, at least in the more aggressive setup, but the results are in line with how easy a section of edge was punched out, as well as what appears to be slightly enlarged grain on a relative basis.

Because the Richter results were so disparate, I honed the 52100 chisel and the Richter chisel in a honing guide to an angle that is 34-ish on an old microebevel contraption that I have from long before 2010. Nearly all of the gap between the two chisels disappeared at that. For this kind of work, that sort of edge treatment isn’t unreasonable, though it was less easy through the wood than the original test edge. Very little damage occurred in either chisel in that case.

I see this as confirmatory, both in confirming that reasonable setup efforts negate needing to pursue an ideal, as well as confirming that when absolute damage levels change, so too can relative differences.

By this, I mean that when you see someone online performing any test, even if it’s in a machine, and you see large differences in results, you should request that the person figure out what setup is needed for the performance gap to close. if it never does, one chisel is better than the other. if it closes with a few degrees of adjustment, you may still prefer the chisel that didn’t “win” the initial test. A real-life example of this is my start at mortising plane bodies. I wanted a chisel that would hold up the best and bought an expensive japanese chisel (Imai). The Imai chisel didn’t hold up very well in actual work compared to a stanley or PS&W chisel that just had a couple of extra degrees of bevel. The Imai chisel would’ve won a chisel test pretty easily where a fixed angle of 30 degrees was used just because it’s harder. It took far less long to do the work from one plane to the next with the stanley and PS&W chisels because while they didn’t hold up at a lower angle, they held up fine at an angle that the Imai chisel also needed to hold up mortising beech. And took much less time to sharpen. That actual act was probably the first step toward unicorn, just modifying the tip of a chisel a little to see what would happen and finding out how little it took to stop the damage.