How to test for hardness with Damascus?

[Craig B.]

Hey all,
Been awhile since I've posted on here. Have been working on a blade with some Damscus I bought from Alabama Damascus at the Tulsa Gun show last spring. It's in the oven tempering as I type.

My question is this. I have a hardness tester. Will I be able to test it with that or should I just use a file and that will suffice? I tested with an old file before tempering right out of the forge and it skated all around on the blade except right where I had ahold of it with the pliers.

I would think that with the different alloys of low and high carbon steels that a hardness tester would not be very accurate? Correct?

I've made knives for 20 years this past Feb. but I've not worked with much Damascus. So, what's the best way to test for hardness?

Thanks all!

[GHEzell]

I cannot remember exactly what alloys they use, but I believe they are all high-carbon. With carbon migration, you likely have a piece of steel with a fairly homogeneous carbon level, so yes, there should be no problem testing with a hardness tester. Hard and soft layers in damascus are pretty much a myth, unless the maker purposely does the things necessary to stop carbon migration from happening.

[adn_sun]

Check upto four times from different points with hardness meter than get your average .

[Craig B.]

Thanks guys!

[Ray Rogers]

Craig,

I believe your original assumption is as likely to be correct as not. Under some circumstances, a hardness tester could get a fairly accurate reading on damascus but under other circumstances the reading can be useless.

Damascus is made from two or more types of steel and they harden differently. Also, some damascus has many, many layers that are very fine while others have relatively few that are wide. Since a Rockwell style hardness tester works by measuring the depth of penetration it will give a false reading if the penetrator is sitting on a soft area or if it is on a thin hard area which is allowed to move because it is surrounded by soft areas.

The tester could work well enough if the two (or more) steels respond similarly to the heat treatment used (but few actually do) or if there are enough layers that the steel becomes fairly homogeneous or if there are so few layers that substantial areas of each steel is available for testing.

Taking multiple readings and averaging is good practice when testing a single steel blade but could be misleading on a damascus blade. For instance, the average of one very hard area with three significantly softer areas could yeild a very low number but the blade might perform far better than the number suggests if the hard material is well distributed along the edge.

In short, it might work well in your particular case, and then again, it might not. Lots of factors to consider. The best rules for damascus that I've run across so far is to heat treat it as you would heat treat a blade of the primary blade steel used in the damascus. That particular part of the damascus can be expected to harden almost the same as the plain blade made from it and it should perform about the same. From what I've read, decent damascus will usually perform better than the hardness test would suggest so the best thing might be to just use the blade and see how it does ......

[Craig B.]

Originally Posted by Ray Rogers

Craig,

I believe your original assumption is as likely to be correct as not. Under some circumstances, a hardness tester could get a fairly accurate reading on damascus but under other circumstances the reading can be useless.

Damascus is made from two or more types of steel and they harden differently. Also, some damascus has many, many layers that are very fine while others have relatively few that are wide. Since a Rockwell style hardness tester works by measuring the depth of penetration it will give a false reading if the penetrator is sitting on a soft area or if it is on a thin hard area which is allowed to move because it is surrounded by soft areas.

The tester could work well enough if the two (or more) steels respond similarly to the heat treatment used (but few actually do) or if there are enough layers that the steel becomes fairly homogeneous or if there are so few layers that substantial areas of each steel is available for testing.

Taking multiple readings and averaging is good practice when testing a single steel blade but could be misleading on a damascus blade. For instance, the average of one very hard area with three significantly softer areas could yeild a very low number but the blade might perform far better than the number suggests if the hard material is well distributed along the edge.

In short, it might work well in your particular case, and then again, it might not. Lots of factors to consider. The best rules for damascus that I've run across so far is to heat treat it as you would heat treat a blade of the primary blade steel used in the damascus. That particular part of the damascus can be expected to harden almost the same as the plain blade made from it and it should perform about the same. From what I've read, decent damascus will usually perform better than the hardness test would suggest so the best thing might be to just use the blade and see how it does ......

Ray, thanks for the reply. I realize that in some cases it would never worked (i.e if nickel silver is used in the mix). I also know from previous experience that sometimes checking for hardness with a worn file is not always a very accurate test. I've had blades that the file I was using would skate across that would only be 45 on c scale when tested with the tester. The reason the file skated.....it was plumb worn out and would have probably skated on wood!

By the way, Ray I'm making a liner lock with this blade with some of the information you share on your website tutorial. Thought you might like to know that. I've not made but about 3 folders, all liner locks. Trying to get more into folders and less in fixed blades as I just absolutely hate leather work!

Thank Guys! Damascus is sure pretty but a whole different cat than what I'm used to working with. Mainly I use 1095 and 440c with occasional ats-34.

Later!

[Ray Rogers]

Making folders to avoid leather work! What happens if somebody want a belt sheath for their folder?

Send me a small (under 100k) picture of your first folder and I'll post it on the tutorial with the others.

I had another thought on Rockwell testing damascus. If it is true that some damascus can be accurately tested and some cannot then how can we tell which is which? It occurs to me that the answer to that may be found in the averaging that adn_sun suggested. It is common practice with Rockwell testing to measure a piece of steel 4 times, discard the lowest result, and average the remaining three. This works with a piece of homogeneous steel because we can reasonably expect the 4 tests to yield very similar results, say within 1.5 Rockwell points of each other (hopefully, within 1 point ). So, if you used that method on your damascus and measured 4 points scattered across a large area of your blade (not just in a half square inch, for instance) and you got at least three results that were withing 1.5 Rockwell point spread then an average would be as meaningful as it would be on a plain steel blade. If, on the other hand, the readings varied greatly ( say 5 or more points) then you have one of the mixes that will not produced meaningful results from Rockwell testing ....

[Craig B.]

Originally Posted by Ray Rogers

Making folders to avoid leather work! What happens if somebody want a belt sheath for their folder?

LOL, then I'll have to bite the bullet and make one. I still make sheaths, just don't like too. Had a customer the other day who wanted me to rehandle his schrade sharpfinger with buffalo horn and make a sheath for it. I told him the price and he called me back and said he'd found a sheath to fit it and just to rehandle the knife. Man, I sure was glad!

Thanks!

[adn_sun]

Taking multiple readings and averaging is best in 288 and above layer count .
After this long discussion i had experimented on six blades with same heat treatment and same layer count .
1. test five times each blade .
2. leave the two lesser results .
3. get an average of three higher figures .

i had got some very good results with the difference upto 0.35 to 0.84 HRC between 6 blades i think it is best .

Ed please check it.
_________________________
info@pakhandicrafts.com

[Craig B.]

Tested the blade with the tester just for kicks and grins. It averaged out to 60c, so not bad. Would've been much better if I would've not inadvertantly ground on the lock at the wrong time with the wrong belt

Oh well, maybe I can fix it with a new liner for the lock

[Ray Rogers]

Craig,

The blade 'averaged' 60? Doesn't that mean that some of it was above 60? That might be a bit high for simple carbon steels . Might be worth trying a flex test (brass rod) on the edge to see it it chips out.

Adn_sun,

Sounds like you might be on to something there but be cautious with your method. If you test 5 times and discard two then you are discarding 40% of your results and that is significant. The more tests you do, the more numbers you will get in every possible range so if you discard a high enough percentage what is left will always be the higher numbers which, when averaged, will give a desirable looking value that really doesn't say anything realistic about the blade. To take an extreme example, suppose you test a blade 100 times and discard the lowest 75, averaging the top 25. If the tests gave you numbers ranging from ranging from 30 to 60 such a process could give you a result in the high 50s when the actual blade might have more areas in the 40s that high 50s.

The purpose of discarding 1 in 4 is to allow for the 'flier' reading that so often happens with Rockwell testers because of their mechanical nature. The process only works if at least 3 of the 4 tests are very close together numerically (the 4th may or may not be close). So, if the tests were, for instance, 59.4, 58.7, 60.1, 55 then dropping the 55 and averaging the others yields a reasonable result. But, if the tests were something like 45, 58, 53, 48 the the result would not be meaningful because the range of those results exceeds the operational accuracy of a calibrated Rockwell tester......

[Craig B.]

Ray, it read 60, 61,60,59. The 59 was in the aread where the pin would have rested when closed so it would be ground out anyways. So the average was 60 along the rest of the blade.

I have an Ames handheld tester which I clamp in a vise by the handle when taking readings. It has a little site glass bubble to get your reading and sometimes if I don't have my glasses it get a little harder to read it precise. It is pretty accurate though as I have a test block and it is almost always dead on when testing. Just need to remember I'm starting down the other slope and do need to wear glasses occasionally to see where I'm going!

I had already decided to give the blade another tempering cycle which should get it in the 58-59 range if I can salvage the whole knife from my mistake. I won't know until I get time to get back out to the shop to make another lock for it.

If not, I'll start all over again (next time with different materials)

Thanks for all the replies!

[adn_sun]

Ray the results were not as differ as you had given example they were in the range of 0.5 to 1.5 HRC .

[Ray Rogers]

I'm glad to hear that the results were closer than in my example. But, 1.5 Rockwell points is still a fair amount of variance. If you received one such number out of 5 then everything is probably OK but if there were two you may have a problem.

Anyway, I haven't done very well at getting explaining the important part of what I'm trying to say. Another way to say it might go like this: if you test 4 times and discard one test then average the other 3 you are doing things the way the industry does and your results will be meaningful when compared to other Rockwell values on other blades. If you test 5 times and discard 2 you have discarded 40% of your tests rather than 25% as is the norm. This is enough to skew the result so that a comparison to other Rockwell values on other blades is no longer accurate (maybe not by a lot but possibly enough to create a false advantage for your measurements). In short, if you want to compare apples to apples the same test criteria must be used .....

[adn_sun]

"if you want to compare apples to apples the same test criteria must be used ....."

it sounds good