Heat treating

Phil, Thanks very much for your post it clears a lot of things up and also makes me wonder about a lot more also. 100 degrees difference! I would have never guessed the temp difference could be so large, wow. Is that at the same level or 6 inches higher or lower? Is that a common difference?
How many blades do you heat treat at the same time? I’m guessing you don’t do single blade batches. Are there a largest amount and a minimum amount of blades you’d do at the same time? I’m wondering a lot more butthats probably more than I need to know already. I personally think the heat treatment is the most important part of a knife, and the most misunderstood part too.
 
Phil (or anyone who has the technical know-how),

Maybe I'm a little off in the way that I think of things, but wouldn't it make more sense to use a medium that has a much higher specific heat than air to transfer the heat to the blade.

Aluminum, with a specific heat of 900 joules/K/g, and a melting point of 1220F, could be used for heat treating anything that you would use at 1400+ F, and with such a high specific heat, you would have very accurate control over the temperature of the molten metal.

Zinc could be used for lower temperatures, because it's molten at 787F, but it doesn't have nearly as high a specific heat, at 388 J/K/g.

Also, you would have much more mass to deal with, making things even more stable in the system. I imagine that temperatures could be kept within 5F for a decent system. Just to compare, air has a specific heat of ~1 J/K/g

The downside would be the energy required to melt your bath every time you wanted to start heat treating, and the obvious danger of having a quantity of molten metal sitting around, but if you waited until you had several blades ready to treat, you could do them in series, and not have to pay the energy costs of remelting the metal.

If I'm way off base here, then let me know.
 
the_mac said:
... use a medium that has a much higher specific heat than air to transfer the heat to the blade.
Click to expand...

Salts and molten lead are indeed used for heat treating.

Phil Wilson said:
In spite of all this if I do a temp survey inside I find some places where there is a 100 F difference within 6 inches.
Click to expand...

Have you any experience with commercial ovens? Is the variance similar?

The Rockwell certified test block is accurate to .5 pt. The machine on a good day is probably good to about the same .5 pt. So if all this adds the wrong way it is easy to be off a full pt. even with the best equipment.
Click to expand...

It depends somewhat on if the variances are random or systematic however you can reduce them by simple statistical methods of greater sampling. For example if the test block has a known tolerance of +/- 0.5 pt and you take readings on four blocks and average them your tolerance is ideally 0.25 pt with some assumptions about the nature of the block variance. The blocks are expensive though.

-Cliff
 
The largest temp differences in the furnace are near the ends where there are no elements. There is an area in the center about 20 inches long at the same eleveation where differences are 20 degrees or so (furnace is 30 inches long). This is plenty good enough for what I do. I used the 100 degrees as an example of what could happen in a large commercial furnace that heat treats 200 blades at a time. I used it also to point out that unless a temp traverse is taken in a furnace with a separate thermocouple one would not know that temps could vary so much inside. I usually do only one blade at a time.
Some times I can do 2 if the blades are small. This gives me the control I need and I only make 5 or 6 knives per month so this is not a production set up. Before I built this furnace I did some research on different heat treat mediums as suggested. Molten salt would work but I found out that at the higher temps some of these salts are toxic and corrosive. Liquid metals would work also but as mentioned the energy required to melt them would be prohibitive in a shop my size. PHIL
 
One would think that a large commercial furnace is not as accurate as the smaller versions but that is a false assumption. Many of the commercial grade furnaces (vacuum specifically) are certified to a temperature tolerance by their customers. If you are doing aircraft work 25 degrees plus or minus is the maximum allowed. With conversations with many commercial treaters over the years, plus or minus 10-15 is typical. The commercial treaters also have computer controls and multiple thermocouples. The multiple couples are to monitor any variations like the ones Phil monitors. The heat treaters to watch out for are the intermediate sized ones with little controls.

Many of the large commercial knife manufacturers use the tighter controlled vacuum furnaces for their work.
 
Satrang, very interesting. I'm more confused than I was when I started this thread. The only thing I'm fairly sure is that it's probably wrong to assume a batch or large batch heat treatment is not as good as a single or small batch. Thanks. :)
 
Satrang, Very good information. I agree the better the equipment the better the result. I’m sure the companies interested in quality work and quality control will use the best tools available. The original question was-- is there a difference between heat treat of a large batch of blades and a small batch. I don’t want to beat my own drum but one point I made was that by dealing with single blades I can certify individual hardness. If you want RC 60 on say S30V then I can tell you that it will be between 59.5 and 60.5. I can also err on the higher hardness side and not have to play it safe by shooting lower. The difference here is that the blade can be optimized to the specific steel and specific use. Please also understand I was not intending to in any way bash the factory knife industry. We now have quality knives made from the newest and best steels at a reasonable price. This is a great deal. Phil
 
To add to what Phil is saying. He is correct that an individual with good controls and a good process can get results that are more consistent than most commercial batches. It's a baby sitting level that the big furnaces just can't do. This is the main reason for the standard two point range you see in commercial heat treatment. Phil can probably hold much tighter.
 
Satrang said:
If you are doing aircraft work 25 degrees plus or minus is the maximum allowed. With conversations with many commercial treaters over the years, plus or minus 10-15 is typical.
Click to expand...

I wasn't that familiar with high precision temperature sensors as my background is more in pressure but that seemed exceptionally low to me so I did some checking. Even the better ones I found such as the Omegaclad XL probes have a much larger tolerance even on the special high precision ones. So there would be a greater deviation in the measured values from the thermocouples alone. In general, most people are way overconfident in terms of precision aside from the NIST guys anyway.

I also doubt that the heat treating tolerances used on production knives are similar to the tolerances on aircraft and similar. Alvin has tested a number of commercial power hacksaw blades which were production hardened obviously and doing this is actually beyond the abilities of most knifemakers because the temperatures are too high, Bos won't harden them for example. The variance Alvin found was higher than what Phil has listed for his blades. If you check around you can find reports of HRC testing on production blades in general and it is higher still than what Alvin found for the hacksaw blades.

One thing to note is in regards to what Phil said about hardness variation, if you check the numbers and the blades are spec'ed at 59/60 HRC they dip down to 55 HRC, but they never go up to 65 HRC, they always vary on the side of being softer as Phil noted. I don't think this is as much as it is the way they aim as much as anything you do less than ideal tends to soften the blades such as waiting before tempering, improper agitation, etc. . It takes a dedicated effort to get the high alloy blades hard, oil quenches, cold, really high soaks, etc. .

-Cliff
 
Cliff, talk to Stack Metallurgical in Portland or Braddock in Florida. They do quite a bit of production blades (and aircraft material). Alvin's testing on power hacksaw blades is only relevant to how they were heat treated. More information on how they are heat treated (not commercial vacuum) would be needed. Bos doesn't like to go to high hardening temperatures since it degrades the heating elements in his furnaces at a higher rate.
 
Satrang said:
...talk to ...
Click to expand...

You generally need to do a lot more than just ask someone unless you are really naive or want to proceed on faith. If you really want an independent confirmation you need the specific data sheets on the measuring devices and an intimate knowledge of their stability and the method of data collection and how all of this is effected by exactly how they are used. You basically have to do it yourself. None of this is trivial. The focus of my PhD was on exactly that topic in relationship to data collection in CIA. Do some research in a field and become familar with the limitations on measurement and then ask around and see what you hear, it won't be the same thing.

If you really want to know the details go to conferences with the NIST guys and see how they judge variances. These are the guys that will do things for example like assemble equipment multiple times and rerun data to see how the assembly/installation effects the outcome. Thus take all the thermocouples out of the furnaces and replace them with new ones in slightly difference locations and do all of this multiple times and then see how the variances change. This seems pretty extreme right - try it and see what happens. It is very difficult to obtain 0.5% max scale tolerances for the reasons noted as the measuring devices themselves usually have higher tolerances even when new and ran perfectly optimal.

In order to ignore the effect of the thermocouples they would need to be still lower or otherwise as noted they would dominate the variances at that order. I sent off a bunch of requests for thermocouple data sheets and tolerances out of curiousity, as well I emailed for information on furnaces used by knifemakers requesting for specific data tolerances. Most of the data reflects much higher numbers unless you look at long scale mean/error which is *very* different than single sample error.

Alvin's testing on power hacksaw blades is only relevant to how they were heat treated.
Click to expand...

This is the same fallacy as noted in the above. When you measure variance you see the total effect of everything which could induce a change. The steel will vary simply due to composition alone. You can for example check standard ASM tests for specific tool steel grades from different manufacturers and see an effect of the different composition as they are not integer exact in specification, they all have tolerances and will range in that tolerance. Harden 440C for example which hits the low end of the carbon and the high end of the chroimum vs a piece which hits the high end of the carbon and the low end of the chromium. Just look at the effect this makes on the tie line position alone. Then on top of this you have the extent of how they were made, how they were forged to shape, etc. . Alvin has also done hardness testing on other tools as well, taps for example and you will see similar variances. In general production knives are greater still.

-Cliff
 
As an aside, in some years thermocouples may be obsolete in terms of accuracy as infrared spectroscopy sensors are furthur developed for this application. There are some people here in Toronto doing this for combustion monitoriong for power stations in particular.
 
In general measurement advances very rapidly, it isn't as extreme as computing power which can make a purchase go from best to not even worth considering in a year, but still several years out of date can put you far behind. Some of the high alloy steels have very short soak times at the very high temperatures. I would wonder how much longer it takes for the piece fo come to equilibrium after the very thin edge has reached critical temperature. Heat flow in general is inversely proportional to material thickness so a 0.025" edge will come to temp much faster than a 0.25" spine. I would wonder for example if this wasn't one of the reasons for problems with S30V if the edges were ground too thin before heat treating so as to minimize the grinding necessary after hardening.

-Cliff
 
What is the reluctance to "talk to" a certified aircraft materials commercial heat treater? They have AMS standards that they must follow and must certifify their furnaces to the tolerances of the aircraft alloy specifications. These are the same furnaces used to heat treat quite a bit of commercial blades. Stack for example has done and continues to do many of the makers in the Pacific Northwest. Step away from the internet and learn something in the "real" world. Cliff, make a trip to the Oregon show next year. Stop by Stack and get a tour of a commercial heat treater, do a presentation on knife design and evaluation and network with many of the best makers in North America. What a learning vacation.
 
Satrang, as I noted I have requested and read over actual published data on thermocouples recently. If you want to provide any additional facts, actual *published* data on the precision tolerances both as spec and in use I would be interested as well as of course any actual data on QC testing of knife blades such as I have provided in the above. As for traveling to locate the apparently endangered metallurgists/makers, I thank you for the invite, and you may not be aware of this, but there are metallurgists outside of those who work for Crucible. In fact some are local as are knife makers and I have talked to both. There are also places locally which even do heat treating.

In general private conversations are highly problematic to actual information echange because there is no responsibility/confirmation. I have heard horror stories about heat treating in private conversations with makers. If you want actual data and not gossip then you look for the actual relevant published papers and the facts which will be openly discussed and defended in *public*, not private exchanges. If you actually want to see how discussions on the internet can be informative about metallurgy then go to the Swordforums and check out the materials sub-forum. Note the focus on actual published data and the public information exchange and almost complete lack of gossip and character focus as this is completely irrelevant to information exchange.

If you have any questions as to knife steels and design then I would be happy to share any knowledge I have with you, or anyone else, and I will readily do so in a *public* forum. I will also explain the reasoning giving examples of actual use of knives and the published materials data it is based on. There are of course lots of makers who will do the same with *public* commentary such as Kevin Cashen, Phil Wilson, Ray Kirk, R. J. Martin, and too many others to name.

-Cliff
 
Trying to compare what two different people say about variance tends to be very misleading. One guy says he went target shooting and for the whole afternoon he was within 1/2 inch of his aiming point. Another guy says he was within 10 inches for the afternoon. They both scored the same, but the first guy thinks "a few flyers" don't count....
 
It is not a trivial field and in general there are not strict guidelines about procedures even in one area let along from one dicipline to another as everything changes massively then. Take something which is very trivial or appears to be for example such as judging which data that you can ignore and which isn't used to effect the final outcome. Ask a bunch of experimentalists how they determine if they can reject data or not. There are various statistical tests you can use such as probability of occourance but they depend on you "knowing" that the data really is false and not something real and your model is actually wrong. Often times people just judge by eye as in that is too far to be real, with no defined objective criteria.

Even the really well known and stock methods like least squares are highly contended. Median is great for outliers, it ignores them completely which has its own host of problems. Most people never consider the implications of many actions which can be significant. Robert le Roy for example (chemist) has developed methods to deal with the problems of rounding numbers for presentation. How can you tell for example how much you can trim off of your results before the calcuations you make from them will be signficantly different? He found that it wasn't as most people expected (round to the uncertainties), and developed a method of rounding and refitting to give maximum information with the least amount of numbers by taking into account the correlation between all the variables.

As anyone actually familiar with experiementation knows, what is chosen to publish/present is in general hand selected with the minimum variation. There are exceptions to this, some focus on the stability of the data and discuss it in detail, but usually you are seeing the "best" results obtained. The maximum tolerances are usually much wider before data will be rejected.

-Cliff
 
As posted by Cliff Stamp

"In general private conversations are highly problematic to actual information echange because there is no responsibility/confirmation. I have heard horror stories about heat treating in private conversations with makers. If you want actual data and not gossip then you look for the actual relevant published papers and the facts which will be openly discussed and defended in *public*, not private exchanges. If you actually want to see how discussions on the internet can be informative about metallurgy then go to the Swordforums and check out the materials sub-forum. Note the focus on actual published data and the public information exchange and almost complete lack of gossip and character focus as this is completely irrelevant to information exchange."

It is obvious and sad that you will never get outisde the forums to experience the benefits of personal interaction with the knife community.
 
You must log in or register to reply here.
Back
Top