Testing W2 Hardness Quenched with Parks 50 and 9% Brine Solution.

[Natlek]

Jeff , did you measure hardness from both side of test piece ? Or just from one side ?

 

[Jeff the Millwright]

Jeff , did you measure hardness from both side of test piece ? Or just from one side ?

I just test on one side of the test piece because I have to grind through the decarb layer on most of my pieces because usually right at the surface does not indicate true hardness and I can tell you it takes a long time to grind every piece down just .005" when the hardnesses are in the 60s and the tips of my fingers are already wearing out so it would take a lot more work to test both sides.

 

[Jeff the Millwright]

I have learned that several others have encountered the exact same problems as me with consistency hardening W-2 from Aldo’s so today I approached my latest test with the resignation that quite likely I am fighting a losing battle and that this steel simply is bad as some have suggested and that it is impossible to get uniformly and consistently hard results.

So I expected today would be my last test because I thought I have done just about everything I can think of and if things didn’t work today I would know one way or the other but of course my results were inconclusive and once again I feel I am one day away from really knowing if it is possible.

If you look at the following photograph I did something I haven’t done before and that is I punch marked my test pieces before I cut them up so that I would know what parts of what pieces were adjacent to what other pieces. This way I hoped I could possibly determine if I was dealing with soft spots on the bar of steel that overlapped adjacent test pieces. This way I was hoping that it would tell me if my problem was with my test pieces and the way they were treated or if my problem was with the steel and soft spots within the bar.

So once again for the eighth time I tried my 1425°F to 1450°F austentizing temperature test in 5°F intervals. And on this test I did normalize at 1650°F, 1350°F and 1250°F and I did cool to black between each temperature but I did not quench at any point during the normalizing because I suspected that was causing my problems.

If you look at my results for the first three temperatures I thought I had actually solved the riddle because my hardnesses after removing .005” of material to get through the decarb layer I was in the mid-60s. Not ideal but pretty darn close. But then I hit 1440°F and once again my numbers went for a crap but they pretty much continued to be awful for the rest of the test.

The fact they fell off and didn’t really recover properly makes me suspect that I simply dropped off the high tolerance for astentizing temperature for this steel. I know these numbers seem low but I just think my pyrometer for my oven is not calibrated properly and from my tempil stick test I assume it is out by about 30° so the 1435°F I think is around 1465°F which is about the recommended temperature.

So one of two things happened either the same bizarre all over the road inconsistencies hit me again and it’s just coincidence that they were the last three highest temperatures or I got good results and it’s just that I went beyond the austentizing temperature range of the steel. And I know it is pretty bizarre it would fall off so quickly from a 5° variation but maybe it’s possible.

But the idea that the steel has bad spots with low carbon and that’s why certain pieces didn’t harden seems unlikely because I know from my punch marking which pieces were beside each other and some of the highest were directly adjacent to some of the lowest for example the 1430°F sample was right beside the 1440°F sample and shared a lengthways edge and I tested along both these edges and on the 1430°F sample it was hard all along the edge that was cut out directly beside the 1440°F sample and I tested that edge and it was extremely soft all along. And the same thing happened with the 1435°F sample which was hard all along its edge that it shared with the 1445°F which had a soft spot along its edge.

I know none of this is conclusive but it certainly suggests that there are not soft spots on my bar because those soft spots just seem to coincidentally exactly match the shape of the test pieces I cut so this tells me that my problem is how my test pieces are being treated and not that there is something inherently wrong with portions of my stock.

So what I am saying is at this point I think there is still hope and that I tested too high a range for my autentizing temperatures and that if I test from 1435°F down in 5°F increments I may actually find a good range where everything is hard. If that doesn’t work I am done but then again I say that every day, LOL.

 

[Kentucky]

The sequence I use is 1650 for 10min, cool to black (magnetic), 1550f for 10 min, cool to black, 1450f for 10min, cool to black and quench. 1350 or 1200 will stress relieve and fine spheroidize the steel. .

This is what we had to do with his 80crv2..It would harden but not harden all the way, 40's-50's Im guessing. After running normalization cycles it came out slate white and hard as a woodpeckers lips..That was for stock removal.

 

[Willie71]

This is what we had to do with his 80crv2..It would harden but not harden all the way, 40's-50's Im guessing. After running normalization cycles it came out slate white and hard as a woodpeckers lips..That was for stock removal.

The 52100 requires this, other steels have been hit or miss depending on the level of farts in the mill, or whatever other variable results in the inconsistencies between bars. I usually do 5-10 pieces at a time, depending on the size of the pieces. It doesn't take long, and I've been getting really consistent results. I haven't had a bad bar of W2 yet though.

 

[Willie71]

Jeff, how thick is this stock? 1/4" might not harden fully even with parks 50. I've seen auto hamons on 1/8" steel on occasion too. 1/8" should through harden though. No guarantees on anything thicker than that. The area above a hamon will test in the high 30s or low 40s iirc.

 

[Kentucky]

The 52100 requires this, other steels have been hit or miss depending on the level of farts in the mill, or whatever other variable results in the inconsistencies between bars. I usually do 5-10 pieces at a time, depending on the size of the pieces. It doesn't take long, and I've been getting really consistent results. I haven't had a bad bar of W2 yet though.

Yea we have had to do the same thing with some 52100. We use to use a lot of it with kevin cashens heat treat..Good stuff

 

[Jeff the Millwright]

This is what we had to do with his 80crv2..It would harden but not harden all the way, 40's-50's Im guessing. After running normalization cycles it came out slate white and hard as a woodpeckers lips..That was for stock removal.

I think you are right Kentucky that the secret to heat treating W-2 properly lies in the normalization cycles and once I figure out my best austentizing temperature I am going to play around with that little more.

 

[Jeff the Millwright]

The 52100 requires this, other steels have been hit or miss depending on the level of farts in the mill, or whatever other variable results in the inconsistencies between bars. I usually do 5-10 pieces at a time, depending on the size of the pieces. It doesn't take long, and I've been getting really consistent results. I haven't had a bad bar of W2 yet though.

I am not sure I understood you properly Willie when you described your heat treating system for W-2? I know you use 1650°F for 10 minutes and cool to black, 1550°F for 10 minutes and cool to black, 1450°F for 10 minutes cool to black and then quench but I thought from what you wrote that you only do 1350 and 1200 if you are going to do some grinding or working the steel so I didn’t bother doing the last two lowest temperatures when I did your system so do you always do the lowest two temperatures along with the higher three because that may be why your system didn’t work with me since I eliminated the lower two?

And since you haven’t had a bad bar of W-2 I guess what you are doing must be working so if you are doing all five temperatures in your normalizing I will try that once I get my austentizing temp down tomorrow.

In to answer your questions in the next post my stock is supposed to be 3/16 but it’s about .200”.

 

[Willie71]

I am not sure I understood you properly Willie when you described your heat treating system for W-2? I know you use 1650°F for 10 minutes and cool to black, 1550°F for 10 minutes and cool to black, 1450°F for 10 minutes cool to black and then quench but I thought from what you wrote that you only do 1350 and 1200 if you are going to do some grinding or working the steel so I didn’t bother doing the last two lowest temperatures when I did your system so do you always do the lowest two temperatures along with the higher three because that may be why your system didn’t work with me since I eliminated the lower two?

And since you haven’t had a bad bar of W-2 I guess what you are doing must be working so if you are doing all five temperatures in your normalizing I will try that once I get my austentizing temp down tomorrow.

In to answer your questions in the next post my stock is supposed to be 3/16 but it’s about .200”.

If the blade is profiled and shaped, I don't do the subcritical cycles. If I'm doing a kitchen knife, I'll cycle first, with a final of 1200, so it drills and grinds easily. I don't like to cycle thin blades because of the decarb and risk of overheating the thin edge. I heat treat then after grinding with no further cycling. I don't use a 1350 cycle. I will use 1400 if I'm doing a water quench though. So 1650, black, 1550, black, 1450, black (quench if desired).

If I'm doing a water quench, 1400, black, quench (I wouldn't have quenched after 1450, to clarify.)

Now I can heat treat at 1460f. If I'm doing a thin knife, 1200f for 2h, then grind, and heat treat at 1460f.

 

[Jeff the Millwright]

Thank you for clarifying that Willie but it just deepens the mystery because I did exactly that using your system; 1650°F to black, 1550 to black, 1450 to black and quenched and some of my samples weren’t even out of the 30s for hardness?

I think we have very different W-2. Tomorrow should be quite telling because I think I am closing in on a workable heat treat but it has seemed like that several times previously. I am going to dicker around a little longer and see if I can get it right but when I feel I have done everything and it still won’t work I will send you a sample if you like to see if it will work for you?

 

[Jeff the Millwright]

Today was a very good day in the shop because not only did I have consistent results and think I have a viable heat treat for W-2 I believe I also figured out what was causing all my wild variations previously.

At the top of the page you can see yesterday’s test and from it I hoped that the hardnesses fell off after 1440°F because that was too high of an astentizing temperature and wasn’t just the usual unexplained fluctuations.

So if you see the test below that I used 1435°F for the top of my range and tested down to 1410°F and lo and behold all my numbers were in an acceptable range for hardness. That was a real relief to finally see good numbers. It looks like things began to fall off again after 1435°F so it looks like that is the top of the range. Obviously 1410°F is not the bottom because those hardness is are still quite good so eventually I will have to test the lower end of the scale further to find the true center but for now I feel I can confidently make a knife by staying in the middle of this range and can hope to be in the 66 to 68 range. And obviously my oven is much warmer than the readout indicates because I am austentizing at 1418°F and others feel 1460°F is the sweet spot.

Now that I have figured out what was an acceptable austentizing temperature I did another test with three pieces. I did them all at the same temperature of 1418°F because I figure that is about the middle of the range so far even though it might be a little cooler once I do more testing on the bottom range. In this test all I did was vary my quenching during my normalizing cycles. For the first piece I quenched after the piece turned black after every temperature which were 1650°F, 1350°F and 1250°F. For the second piece I only quenched at the final temperature of 1250°F and for the third piece I did no quenching at all.

I suspected that quenching was causing all the unexplainable extremely low hardness readings I have been getting lately and this test certainly affirms that. Quenching after every step didn’t seem to make much difference either way but shockingly only quenching after the last temperature cycle gave readings 40 and below. It is hard to believe that this commonly suggested practice renders this steel so completely useless after hardening.

I have a working system now to make a good quality knife and I will probably do that now because I have certainly had my fill of testing samples but down the road I will do some more testing because I think there probably is perhaps another hardness point or two to be gained maybe. I know some people are getting 67 ½ and 68 ½ consistently but that may be because they are using different W-2 but I will try to see if I can gain something else by continuing to fiddle with my heat treating and especially my normalizing because I think maybe if I add a couple more temperatures between 1650°F and 1350°F I may gain a bit.

But anyway it looks like I have finally gotten a good heat treat for Aldo’s W-2 and I’d like to thank everybody who was so helpful with their suggestions because it was a great help. And I replied to your message Willie that since I have success now there is no point in sending you a sample but thanks for the offer and I am saying this here because I couldn’t see my reply in my outbox so I don’t know if you will receive it.

 

[samuraistuart]

Jeff, one quick note....not meaning to throw you off or anything, but I am wondering exactly what your "thermal cycle" temps are. It looks to me like you are normalizing at 1650F (good!), but then your next thermal cycle is sub-critical at 1350F, and the third is way below critical at 1250F. Those subcritical cycles aren't doing much for grain size, as you aren't even reaching austenite state. Well, technically you are reaching the very beginning of austenite at 1350F, BARELY, and not at all at 1250F. You need to have austenite to nucleate new grain growth for the next cycle.

However, in other posts it seems as if you are cycling more "correctly". Normalize at 1650, cycle down but stay at or above critical for a few cycles, as in 1550, 1475, 1450, finish with something slightly subcritical if you like as in 1350-1375. 1250F is doing nothing but stress relieving, and also beginning to slightly spheroidize the carbides.

My comment is about aus grain reduction. However, W2 has a good dose of vanadium, so aus grain should be small anyway, even if no cycling was done after 1650 normalizing. To recap, normalize hot to break up carbide networks and help to distribute carbides evenly. Thermal cycle at or above critical to cause grain "shrink" (puts more grain in a given area). Thermal cycle below critical to reduce stress and spheroidize.

 

[Jeff the Millwright]

Jeff, one quick note....not meaning to throw you off or anything, but I am wondering exactly what your "thermal cycle" temps are. It looks to me like you are normalizing at 1650F (good!), but then your next thermal cycle is sub-critical at 1350F, and the third is way below critical at 1250F. Those subcritical cycles aren't doing much for grain size, as you aren't even reaching austenite state. Well, technically you are reaching the very beginning of austenite at 1350F, BARELY, and not at all at 1250F. You need to have austenite to nucleate new grain growth for the next cycle.

However, in other posts it seems as if you are cycling more "correctly". Normalize at 1650, cycle down but stay at or above critical for a few cycles, as in 1550, 1475, 1450, finish with something slightly subcritical if you like as in 1350-1375. 1250F is doing nothing but stress relieving, and also beginning to slightly spheroidize the carbides.

My comment is about aus grain reduction. However, W2 has a good dose of vanadium, so aus grain should be small anyway, even if no cycling was done after 1650 normalizing. To recap, normalize hot to break up carbide networks and help to distribute carbides evenly. Thermal cycle at or above critical to cause grain "shrink" (puts more grain in a given area). Thermal cycle below critical to reduce stress and spheroidize.

I appreciate your posting Stuart because I was thinking of doing more experimenting with my normalizing temperatures and you have definitely pointed me in the right direction. And yes you are right my normalizing temperatures that were successful for me were 1650°F, 1350°F and finally 1250°F. I believe I wrote these down from one of Stacy’s postings.

And Willie did suggest also that I use more cycles above critical so I did do an experiment using his suggestion 1650°F, 1550°F and finally 1450°F. But when I did the test with these temperatures I also quenched after the final temperature when the sample piece cooled to black. I got dismal results and I assumed it was because these temperatures were different from Stacy’s but what I didn’t know at the time was that it was the quenching that was causing all my problems so I blamed the temperatures instead incorrectly. And that is why I reverted back to my 1650°F, 1350°F and 1250°F cycles.

But after I knew that my problem was quenching during my normalizing cycles I suspected that I could gain a little by trying higher cycling temperatures again and thank you because your posting verifies that’s a good idea like Willie also suggested and really explains why so I appreciate that and I will be trying more cycles above critical.

And when I do that experiment I think I will try two samples and do one sample with all above critical temperatures and a second one where I will include a couple of sub-critical temperatures to see which yields the best results with no quenching.

And actually I might even experiment later with quenching because it is truly mystifying at this point. When I quench only on my last cycle after it cools to black that causes a catastrophic failure in hardening but quenching after every cycle has actually increased hardness on a few samples so I may try eventually not quenching at all and quenching after every cycle also.

 

[Jeff the Millwright]

Again today my tests had unexpected results. These tests again affirm that W-2 is an especially finicky steel when it comes to heat treating and validates the practice suggested by many much more experienced heat treaters than me of the importance of doing your own tests because once again today conventional wisdom was shown to be inapplicable to this steel.

It was suggested to me by samuraistuart in a sincere attempt to help me that I wasn’t cycling “correctly”. He wrote, “Normalize at 1650, cycle down but stay at or above critical for a few cycles, as in 1550, 1475, 1450, finish with something slightly subcritical if you like as in 1350-1375. 1250F is doing nothing but stress relieving, and also beginning to slightly spheroidize the carbides.”

I had been using Stacy’s cycling temperatures of 1650°F, 1350°F and 1250°F. Stuart’s suggestion made sense because I was normalizing to break up the coarse Spheroidized carbides so today I tried two different heat treating cycles. I had one where I had three above critical temperatures where I used 1650°F, 1550°F and 1450°F and for my second heat treating I used those three above critical temperatures +2 below critical temperatures in addition being 1350°F and 1250°F.

In the past it seemed that not quenching during normalizing gave me the best results but also one test gave me good results when I quenched after each temperature cycle so I also did two samples in each normalizing test one without quenching and one with quenching after each temperature. So I had four samples in total; one sample went through five cycles without quenching, one went through five cycles with quenching at every temperature, one went through three above critical cycles without quenching, one went through three cycles above critical with quenching.

Here are the results.

Almost every result was unexpected. Firstly only doing three above critical cycles whether I quenched are not gave terrible results not even out of the 40s for hardness. This absolutely is against conventional thought because these cycles are the ones that are supposed to break up the course spheroidize carbides and it was thought this is the part of the normalizing that leads to greater hardness when in fact seems to have done very little.

And both the samples whether they were quenched or not where sub-critical cycles were added they were significantly harder even though this is only supposed to “stress relieving, and also beginning to slightly spheroidize the carbides”. So this test basically showed the complete opposite of what I expected and that is that adding additional higher than critical temperature cycling doesn’t aid in hardening and without the addition of subcritical cycling temperatures they don’t do anything to harden.

But the other surprise was what quenching and not quenching did. I suspected that quenching after every temperature might be beneficial because it had been once in the past and by doing that and cycling all five temperatures I attain the same hardness as I had previously with my other successful normalizing procedure so this wasn’t a big surprise but affirmed what I suspected. But what was a total surprise was that not quenching using the five different cycling temperatures did not give me good hardening again because not quenching did give me good hardness results when I used it with just three cycles of 1650°F 1350°F and 1250°F.

No wonder I had so much trouble getting consistent results because this steel reacts in very unexpected ways. Why is it that when I use five cycles and don’t quench I didn’t get good hardness but when I used three temperatures with one being above critical and two subcritical without quenching I get good results?

So right now I have two ways to get between 66 and 67. I can use five temperatures, three above critical and two below and quench after each cycle or use three temperatures one above critical and two below and not quench at all. This makes no sense but this is the reality.

So I do have consistent ways to do heat treating with W-2 but just out of curiosity I think I have to do a test only doing subcritical temperatures normalizing. It makes sense at this point that it really is just the subcritical temperatures that are doing the work because doing just above critical does nothing and doing above critical and below critical works so it only makes sense to do an only below critical because it seems like that is what is working? And I think I will do with and without quenching just for the fun of it.

And Stuart I do appreciate the input and I am not disputing your reasoning or knowledge and I am sure with other steels it does apply so I am just reporting my results and I can’t explain why I am getting them.

 

[me2]

temperatures below critical temperature aren't normalizing. They are mostly stress relief or sub-critical annealing, depending the beginning state of the steel. Just a quick note for anyone looking into it in heat treating texts or such, as normalizing has a fairly specific meaning.

 

[samuraistuart]

Jeff, this W2 has been talked about and cussed about by a few in my circle of knife making buddies. There are issues with it, no doubt about it. The normalizing and thermal cycling I lined out is exactly what I use, and I am getting excellent results, every time. The only problem I have had personally with this W2 is the ATP 641 anti decarb compound leaves little "divots" in the steel that I have to sand out. No other steel I have used does this. I have no idea why. I have heard there is some flux or something in the ATP that causes this, but not with any steel other than this W2. I've been lucky, others have had a heck of a time hardening the W2, especially the thicker stuff like 3/16". I use the 1/8" and always have bevels cut before HT, edge at .020". Comes out of P50 harder than woodpecker's lips with grain like a file or drill bit. I use the exact same thermal treatment on 52100, as per Kevin Cashen's suggestions, especially dealing with the heavily spheroidized 52100.

Off the top of my head, I can say the more you cycle W2, or quench it, (or any steel), it will decrease it's hardenability. That is to say, the more you cycle/quench it, the faster your quench needs to be to harden it. W2 is already extremely shallow hardening. It *may* be possible that 5 cycles is too much, reducing hardenability to the point P50 struggles.

You might try this: Normalize 1650F, thermal cycle 1475F, thermal cycle 1450F, thermal cycle 1400F. Quench at 1460-1475 (re-calibrate those numbers to match your thermocouple that's off). Or 1650F, then 1450F twice. Do that 1 normalizing, do the 2 or 3 thermal cycles at the most. Again....anything under 1350 is not doing much for grain refinement, that is spheroidizing/stress relief temps. However, if you dont get good results, I would stick with using what you did get results with.

 

[Jeff the Millwright]

temperatures below critical temperature aren't normalizing. They are mostly stress relief or sub-critical annealing, depending the beginning state of the steel. Just a quick note for anyone looking into it in heat treating texts or such, as normalizing has a fairly specific meaning.

I appreciate the clarification about the textbook meaning of normalizing but the confounding aspect of my tests are that the generally accepted rules don’t seem to be applying. Normalizing as you stated is heating a metal beyond critical in several cycles and that is used to normalize the steel in preparation for hardening and that sub-critical cycles are for stress relief and annealing but my tests don’t follow those rules.

My last test shows that with W-2 with what is considered normalizing as in heating this steel to beyond critical and cycling a few temperatures does not properly prepare the steel for hardening because just doing beyond critical cycles without sub critical temperatures does not result in a steel that is ready for hardening. In my tests the sub critical is what allows the steel to be fully hardened and whether that is in conjunction with cycling with temperatures above critical or independently has yet to be seen.

So what I am saying is that my tests show that sub critical cycling has a much more integral role in normalizing than just stress relieving and annealing because without these sub critical temperature cycles the above critical cycles do not properly prepare the steel for full hardening.

My results are so anti- to conventional wisdom I understand people are doubtful about my results and if I weren’t doing the test myself I probably would doubt it but these are my results.

 

[samuraistuart]

Jeff, like Me2 mentioned, normalizing technically is one cycle, not several. He is just trying to separate the correct terminology for future reference by others. Normalizing is done usually at least 100-150F above critical, one cycle, with an air cool. It is done to break up carbides that might have clumped together during previous forging/rolling, and to distribute them evenly. It is a one time deal. No need to "normalize" more than once, if it was done correctly.

Then we do "thermal cycles". This is really more about aus grain refinement than anything, putting more aus grain in a given area/volume, hence making them smaller. This is done usually around the critical temp, slightly above it, at it, and slightly below it. I suppose that ANY heat operation could "technically" be called a "thermal cycle", but for us knife makers, we understand that term to be limited to cycles at/around critical for aus grain refinement. After "thermal cycling" has been done, if no further machining needs to be done on the work piece, go straight to hardening. I asked Kevin what the best condition a steel could be in for hardening, he said either fine pearlite or upper bainite. Not a spheroidized structure, not a martensitic structure. That means that after the air cool is done on your final thermal cycle, you should have fine pearlite, one of the best matrices to harden from. I won't get into bainite, as that has a specific HT to develop.

Below "thermal cycle" temps is annealing temps ~1300s. For high carbon steels it is usually always best to spheroidize anneal if any further machining needs to be done. There will be no change over to austenite at these temps (below 1350f), so technically not much in the way of grain refinement here, mainly to soften things up. The mill that serves Aldo uses heavy spheroidizing on some of the steels, like 52100, and it is done at these temps for extended periods of time, hours upon hours from what I understand, with a very very slow drop in temp from sub critical on down to 900F or so. This causes large spheroidized carbides. Butter soft on tooling, but not so great to harden from.

Below annealing temps is stress relief temps ~1200s. Interestingly, if you take a martensitic blade (as quenched from critical) and "temper" at ~1200-1250F for a couple hours, you will fine spheroidize the steel for excellent machinability. This is also an excellent state to harden from, as it is extremely fine spheroidized, and does not need normalizing or extensive soak times to put the carbides back into solution.


Below ~1200s you get into the secondary hardening temps of tempering, post quench of course.

Your "assumption" that the subcritical temps you are using is what gets your W2 sample to harden properly is not correct at all. AAMOF, you do NOT need to do ANY subcritical temps to get a steel to harden after normalizing. Your test may seem to support your theory, but it is not correct by any means. Not meaning to hammer at you, just saying. Your 3rd sentence in the post above....again, no, subcritical cycling has nothing at all to do with normalizing. It is a completely separate operation for completely different purposes.

 

[Willie71]

Normalizing is typically above ac3, as I understand it. The other 2 cycles are typically ac2, and ac1 for grain refinement.

I'm suspecting the cycling is leaving your 0.200" steel so shallow hardening p50 won't do it. Grind a piece to 3/16", and try the cycling Stuart and I have recommended. If it doesn't through harden at this thickness, there is a problem with this steel. I've done pieces in 1/4", and can't get consistent readings on the tang, but even on the angled edge, get over Rc66, accepting tester error on non parallel stock.