Update: Michael Drinkwine sent me another report from Global where they reported factory sharpened and waterstone sharpened knives. I added it to the “CATRA Testing” section of the article.
Update 2: Roman Landes gave me some microhardness measurements of an edge after being ground with 220 grit belts. I added them to the “Effect on Knife Edges” section.
Thanks to Ken Kneringer and Nevin Ng for becoming Knife Steel Nerds Patreon supporters!
Bad Edges on Factory Knives
It is commonly reported that a new knife needs to be sharpened a few times before its performance can be evaluated, reportedly because the grinders that put the final edge on the knife will overheat the edge. Is this a real effect?
Grinding of edges to sharpen them is relatively common for knife makers and companies. So if that technique does lead to lesser performance it would be a concern. Among knife users it is somewhat less common though the Work Sharp gets some use as well as Harbor Freight 1×30 grinders:
Abrasion of Steel
When grinding with a belt or a wheel hard media abrades away steel as fine particles:
This image from [1]
The steel often sparks as it is removed from the steel:
You might think that the steel has to be reaching extremely high temperatures to be “burning up” like that as it leaves the steel. However, it is really only the particles that leave the steel that are reaching high temperatures. The formation of iron oxide (rust) is an exothermic reaction (it generates heat). Normally that reaction is slow but with small particles they have a very high surface area and the reaction occurs immediately in air. If using very fine iron particles they will react in a similar manner when exposed to air even without grinding:
Heating of Steel
Instead it is the heat generated by friction in the grinding process that can lead to overheated edges. The amount of heat generated depends on the geometry of the knife/part, grinding media, force against the sharpening media, the speed of operation, etc. At relatively low temperatures the first thing that happens to a knife edge is that it is overtempered. Typically a knife is heated to high temperature, quenched, and then tempered to the desired hardness. A typical tempering temperature for a knife steel is 400°F or so, resulting in the desired hardness such as in the range of 58-62 Rc. Here is a tempering curve for a simple high carbon steel, 1095:
This image from [2]
You can see that as the temperature goes up the hardness goes down. After tempering, if the steel is heated to some temperature below the tempering temperature, the hardness will likely not be affected unless it is held there for very long periods of time. However, if the tempering temperature is exceeded then softening will continue and it will travel down the hardness curve above. Here is an image of a high speed steel tap which is darker at the edge indicating overtempering:
This image from [3]
If the edge is greatly overheated the steel can become so hot that it can actually reach temperatures sufficient to reharden the steel, usually 1400°F or above, depending on the steel in question. In this case the steel will be untempered and brittle. Below is the same high speed steel tap which was overheated to the degree that it was rehardened at the surface, which can be seen by the bright color at the edge:
[3]
Microhardness measurements of that tap edge showed the effect of overheating:
[3]
Away from the overheating in the dark region the hardness is at the original value of 65 Rc, it then transitions to the 58-50 Rc region where it was overtempered. Then at the surface where it was rehardened it is 62 Rc, this value is lower than the original 65 Rc because it was improperly hardened.
Effect on Knife Edges
Grinding of edges can indeed heat up the edge. Knife edges have an extremely small volume of material so they are quite easy to overheat. A very sharp edge can have a diameter under a micron. You can read more about how small edges are and what controls sharpness in this article. Because of the very small volume at the apex, it heats up much more quickly than the rest of the knife. Holding the knife with bare hands or dunking in water between passes is not likely to be sufficient since the heating occurs so quickly.
This image from [4]
Roman Landes gave me some very interesting measurements of 8660 steel (German 1.2795) which is a 0.6% carbon steel with about 0.8Mn-0.5Cr-0.2Mo-0.6Ni. There is also a result from “1065 Cr mod” which has 0.65% carbon along with 0.7Mn-0.25Cr. The knives or scissors were sharpened by hand on a grinder with 3M 220 grit Trizact belts. Interestingly, there is also a comparison between bainite and martensite microstructures. Typically bainite is less affected by tempering than martensite. However, in both cases a hardness drop was measured at the edge. Up to a 5 Rc difference was measured, and of course even microhardness can only get so close to the apex of the edge.
Edge Retention of Belt Sharpened Knives
A study was conducted in 2016 [5] where they compared a knife sharpened on the Work Sharp belt sharpener with traditional sharpening on stones. To test the edges they used an Anago sharpness tester which I have not previously discussed. In that test, the entire knife is drawn through a grid-like material so that sharpness values are found over the entire length of the edge:
The Anago measures sharpness on a 10 point scale where 10 would be an infinitely sharp edge (unattainable). They wore the edges of the knives by making cuts into bamboo and measured the sharpness after 30, 60, and 90 cuts. The Work Sharp led to better sharpness (9) than the stone sharpening that they performed (8.5), but dulled much more rapidly:
The belt sharpened knife was only 6.9 after only 30 cuts on the bamboo; the stone sharpened knife was still a 7 even after 90 cuts. After 60 cuts the belt sharpened knife was only a 6.5 as measured with the Anago while the stone sharpened knife was still a 7.4.
CATRA Testing
Roger Hamby of CATRA also tells me that in their testing of edge retention that over 75% of the knives they test suffer to some degree of edge softening due to the sharpening process. He reports that nearly all manufactured knives have the problem, while hand sharpened knives generally do not. I have written extensively about the CATRA test in the past, such as in these articles: Part 1 and Part 2.
The knife company GLOBAL compared their factory sharpened knives to “whetstone” sharpened knives and found a big difference [6], the information has since been removed from the website but I have hosted it here: GLOBAL CATRA testing
Image from [6]
How to Avoid Overheating of Edges
The easiest way to avoid overheated edges is to sharpen by hand, of course. However, that is not always possible. There are water cooled methods, such as the Tormek sharpener or adding a water misting system to a belt grinder:
Using water cooling prevents steel from heating up during the grinding and polishing process and therefore may prevent overheating of edges and loss of edge stability and edge retention.
Different Steels
To some extent steel choice can help. Certain steels are designed to be tempered at high temperatures. High speed steels, in particular, are designed for operating at high temperatures without softening. They do this through additions of elements like tungsten and molybdenum that form very tiny carbides at temperatures in the 900-1100F range that increase hardness:
This image from [2]
Non-high speed steels with sufficient molybdenum or tungsten additions can also be tempered in that high temperature range. Particularly tool steels like Vanadis 4 Extra, 4V, 3V, and others. Even some stainless steels like 154CM can be tempered in that range. However, high temperature tempering also leads to a reduction in corrosion resistance so it is usually not recommended if corrosion resistance is a priority. Using steels with a high temperature temper can help prevent softening if the edge remains below the tempering temperature of 1000°F or so. If that temperature is still exceeded then it won’t matter, of course. Measuring whether or not the edge has actually exceeded any given temperature is very difficult given that it is the very fine edge of the knife that we are interested in.
Summary and Conclusions
Knife edges are easily overheated during powered grinding due to the small volume of the edge and the friction buildup from grinding. This softens the edge and worsens edge retention. The use of steels capable of high temperature tempering treatments help to some extent since they can withstand more heat. Hand sharpening or water cooled sharpening are the best methods for preventing overheating of edges.
[1] Klocke, Fritz, Olaf Dambon, and Barbara Behrens. “Analysis of defect mechanisms in polishing of tool steels.” Production Engineering 5, no. 5 (2011): 475-483.
[2] Roberts, George Adam, Richard Kennedy, and George Krauss. Tool steels. ASM international, 1998.
[3] https://customer.cartech.com/assets/documents/datasheets/Bulletin_104.pdf
[4] http://www.hroarr.com/wp-content/uploads/2014/04/sharpness.pdf
[5] Mulder, Joshua, and Jonathan B. Scott. “The measurement of knife sharpness and the impact of sharpening technique on edge durability.” (2016): 1-7.
[6] https://web.archive.org/web/20160118100853/http://global-knife.com/catra/
I sharpen on my 2×72 but I run the belt at a very slow speed. On the order of 500-800 SFPM. Do you know to what degree belt speed plays a roll?
Faster is worse of course. No one has done a study to compare different belt speeds.
I use a slow speed and fresh belts, I only sharpen 7 knives per belt. I would love to test one of my machine sharpened blades to see how it holds up. But I do know I can achieve a sharper edge with stones than I can with a machine and have suspected this was the reasoning. Thanks for the great information! I will be joining your Patreon group!
The topic of burning edges on belt sanders comes up often on discussion boards, but most of the “research” is suspect or flawed. The 2016 study you site used a worksharp belt sander? I assume this is the pull thru where 1 side is sharpened edge leading & the other edge trailing? In any case I have sharpened many knives with my guided stone sharpener & many more using my variable speed belt sander with an angle guide. My experience is that the deburring process/skill has more impact than which system I use. Would love more information
I just haven’t seen any studies that show won’t overheat the very apex when using dry powered sharpening methods.
There is a really interesting one from roman landes (the guy that coined the term edge stability) that someone translated a passage from into English. That showed just how hot steel can get about a micron of so thick for fractions of a second a lot easier than most people would suspect.
If you want to do sharpening on a belt I think watercooling is the only way to go. Also roman from kase knives Switzerland does some microhardness testing if you want to look at some stuff he posts about.
Excellent article. I’m fortunate to have inherited a used Tormek sharpener and absolutely love it.
Lucky! I wish I had one.
When the iron dust ignites, is it touching the knife?
It shouldn’t be, since then the surface area would be smaller and the bulk piece would be pulling away heat.
Makes sense
DOE‐HDBK‐1081‐2014 Primer on Spontaneous Heating and Pyrophoricity
Finely divided powders of iron and steel can be dangerous if they are allowed to form a dust cloud or
disturbed due to poor housekeeping practices. Iron/steel powder that is allowed to build up on flat
surfaces has the potential to cause massive fire balls if the powder is disturbed and allowed to free fall to
the floor below
Allowing metal dust to accumulate in the dust collector of a surface grinder can make for a VERY interesting fire. I have done it once.
Great article Larrin, thanks for putting this together. I am curious if any study has been done to show yet, that water cooled grinding does indeed fully prevent temper damage.
My other question is: will a millisecond of time over the blades temper actually cause any damage? Doesn’t traditional tempering have to be held at the temp (e.g. 400f) for a long time (few hours)? Would a split second at 600f actually do damage, and if so, how much?
Great question. Tempering is a time-temperature process, so short periods of time at high temperatures are equivalent to long holds at low temperatures. And the tempering times are longer than necessary generally, ie similar hardness would be reached with a 30 minute temper vs 2 tempers at 2 hours, because it reaches a “steady state” where much longer times are required for a significant reduction in hardness. The long tempering times are recommended for larger parts and for decomposing/destabilizing retained austenite. I wrote briefly about the time-temperature relationship in tempering here: https://knifesteelnerds.com/2018/04/23/what-happens-during-tempering-of-steel/
This is interesting
Any thought on a hybrid method? I’ve been using a 120 grit belt to put an initial edge on my blunt knives, before sharpening on ceramic whetstones. I’m thinking this might remove the over tempered steel from the grinder?
(that’s no why i do it, i’m just trying to establish the initial edge quickly 😛 )
If the over-tempered portion of the edge is removed by the stones then it would work.
While this site is the only place to learn something on these topics, this article has left me with questions and no usable answers. As I am trying to asses and understand the same thing – how to sharpen best using multiple methods, here are some comments.
It is clear that one can overheat a knife on a grinder. There is no point on discussing it. But, that may be the same as saying: one can hit the fingeer (instead of a nail) with the hammer. Yes, one can.
Global results can present a difference between bad (too coarse) sharpening and much better sharpening. Or they mean that only one hand sharpening on whetsones fixes all the detemepring made by the factory power grind!
We can not comment on what are the factories doing. So, lets limit onto WS as it is one of the sharpening methods I have available. If a knife is being ground by quickly pulling it throught the grinder (say 4-5 cm per second) it never (entire blade) gets warmed significantly. Very mild warming can be felt only on coarse grits (220). That is on high speed. So, move to medium or low speed. If a knife is cutting cardboard, it gets quickly hotter than can be held in hand. So, cutting a cardboard must completely destroy the edge, much worse than grinding it (as the edge will be much hotter than the blade).
Without knowing how long (how fast) someone was grinding particular spot and the grits and the steels involved, results are useless. For example, referenced paper [5] looks like a high school assignment. They were using the original Work Sharp with no adjustable speed so a full grinding speed had to be used, but did not mention that. Paper does not mention neither speed of pulling nor grits used. They say for WS: “…was found to indeed create a high amount of heat energy in the blade whilst in use.” I have never had even warmed knife at all when sharpening it (1000 and higher grits) on the Work Sharp Ken Onion Edition. only reprofiling with 220 grit will warm it a bit. They do not mention knifes or steels used, judging from the picture, it is a stamped blade <55 HRc if not worse, seems not German, so likely a Chinese one (they are from .nz).
Grinding on a 220 grit is done rarely. Likely only once if one is setting the new sharper angle on the knife. Afterwards, all sharpening is done with 1000 grit and finer. 1000 grit does not warm the knife at all (by the touch). Even on high speed. And normally for WSKO we use medium speed or lower. There are even numbers attached to these speeds.
It is funny how we can use High speed tool steels to make a hole in a concrete with the red hot drill bit. And it did not dissapear and can make another hole afterwards. But, a knife seems to be warmed to 40C for 10 seconds and will detemper and soften.
Unfortunatelly, this article did not provide any knowledge for me. Sadly, many will use it to claim that grinding is bad. I do not know that for now, and certanly not from this article. I know only that grinding can overheat a blade. Yes, it certanly can. Same thing with the hammer and the finger: yes, we can certanly hit a finger with the hammer.
220 grit belts are used regularly for setting the bevel on knives by knifemakers, which is exactly the issue under discussion. Whether an entire blade gets warm is not the question, it is whether the very edge does. And all evidence points to that being the case. Including the tests from Roman where microhardness confirmed that the very edge is changed while the bulk steel is not. Just because more information is not available does not mean that the information we do have should be dismissed. It is indisputable that edges are overheated in grinding. What is unknown is at what speed, grit, etc. overheating no longer occurs. It could be very slow or even unattainable without water cooling because of how small knife edges are (microns). The knifemakers shouting that their process is “fine” is not useful information, it’s just a defense of their products.
I did a test with cardboard on a maxamet blade, and like the other comments said it got VERY hot, too hot to Touch for more than a second. After this i feel Im not really getting the same lifetime out of my edges on this blade. Just got Another blade in maxamet and Will try to see if theres a difference. What do you think, is it possible to overheat the edge from repeated cutting if abrasive materials? Or is it just cardboard that makes the blade that hot? Cant really say I’ve ever noticed the same heat from cutting anything else. Then again one rarely cuts anything at all as fast and for as long as in a cardboard cut-test.
Not sure I’ve never looked into that effect before.
I think the study is limited by the choice of the Work sharp as the grinder. We know that the more surface area on a belt, the less it heats up the steel. Why test then, in the smallest possible grinder? At the very least a 1×30, if not a 2×42 or 2×72 should have been tested as well. Tempering temperature varies by steel, but the knives I make are tempered between 350 and 400 and to get to that temperature on the edge, I find it hard to believe that I wouldn’t be able to feel that in the steel.
I recommend not relying on belief and intuition.
With cardboard there is a good chance the heat is being spread throughout the whole blade and not just the edge (I imagine a large portion of the friction coming from the primary grind).
Oh yeah for sure. But still the edge should be just as hot in my mind? In my previous comment I mentioned another maxamet blade, a native. The edge retention in that one was crazy. Would strop back to hairwhittling for over two weeks. Before stropping ot would still be hair popping arm hair. So my conclusion from this is that either my first blade was underhardened or the cardboard test did in fact draw put the temper. Im certain that the Native did in fact hold its primary sharpness for much much longer than my pm2, both with maxamet. The native was also noticably resisting sharpening way more than the pm2. Lowered the angle on both, the pm2 was just like any other knife while the Native was extremely time consuming. The shoulder was just resisting my strokes forever until I finally managed to raise a small burr. Then the other side was just as resistant.
For now Im of the opinion that the pm2 was probably just underhardened, but I feel there is definetly eveidence implying that something happened in my testing that affected the steel at the edge
If you are saying that wishful thinking doesn’t do much to help with the best science that we currently have, then I totally agree with you.
I’m so torn! I am majorly OCD about maximizing edge retention BUT hate manually sharpening (esp reprofiling). I need to reprofile 10 of my knives (all in S30V and S35VN) and really wanted to get a WS Ken Onion to do so without murdering my hands/wrists doing it all freehand.
Do you think after carefully reprofiling them down to 15 dps on the WSKO at low speed, that putting 17 dps microbevels on by hand would likely remove the overheated edge and give me back max edge retention?
I wouldn’t think so, but you can try it out and tell us what you find.
Hey Larrin,
I am wondering if there is a decent way for a guy in a garage without expensive equipment to evaluate this idea. I have a Worksharp with the blade grinding attachment, and generally I use it to establish primary bevels only. Then I hone a smaller secondary bevel by hand. My blades never get hot to the touch, even at the very edge, but then I obviously cannot measure the temperature at the very edge of the blade as it is in contact with the belt. Do you think an infrared thermometer be effective in measuring this?
Thanks for all the great information, by the way. I’ve really enjoyed this blog.
PS. Hopefully I’m not committing some blog faux pas by commenting on a 2 year old thread.
I don’t know if the infrared thermometer would have the resolution required to measure a tiny edge. I vaguely remember someone using paint that changes with temperature but I don’t remember how well it worked. At the point that is being ground it would be removed, of course.
Thanks for the reply, Larrin. A quick update on this. I looked into infrared thermometers but the problem is exactly as you mentioned–the area of measurement for the infrared beam is too large to accurately measure the very edge of a knife blade.
One of these bad boys would probably do the trick … https://www.tequipment.net/FLIR/T560-DFOV/Building-and-Industrial-Thermal-Imagers/
So, if sharpening with stones leads to lower temperatures at the edge and protects the edge also, then do you think the process of sharpening an edge without a burr would be a better method of sharpening? I am assuming it would since a burr is simply using enough friction (heat) to move some metal near the edge out of the way.
Burr formation does not require heat, it results simply from removing material from one side of the edge.
Now I am confused. You said that making a burr does not require heat in order to move material from one side of the knife. However, to my way of thinking, and please correct me if I am mistaken, getting a burr requires friction, and friction is equal to heat.
Friction is not required to raise a burr. Friction MAY be created, but it isn’t required. Consider a sharp knife and a dull knife. Using each knife to make the same cut in a material, will result in more friction with the dull knife than with the sharp one. The same work done, but less friction.
It would be interesting to get close up thermal footage of an edge being set by various speeds on the grinder, then perhaps a comparison with the water misting. man I wish I had a thermal camera.
https://www.youtube.com/embed/t6SEFqNbgc0 seems to have been set to private.
regarding cooling systems, apart from the usual pressurized spray systems I’ve seen the following method recommended: “a tank above that drips water, via a copper tube, onto an arm mounted sponge that rotates down onto the belt. The sponge is key to using much less water.”
certainly an interesting option if you don’t want to rely on a compressor.
I use the WS-KO Blade Grinder and my blades don’t even warm up from ambient temperature… I would HATE it if even that fries my edges… Im getting Great results from the Blade Grinder attachment… but all my edges would be fried by now and i don’t have the money to buy a new knife to test ^^”
If it already is frying my edges, would spraying the blade down with water from a spray bottle between each pass on the abrasive help? I Invested a lot of money in that thing 🙁 i would hate to know that it damages my knifes….
You really shouldn’t rely on a study with a sample size of 2.
I looked at a number of reviews of the water cooled Tormek systems and clones. After using the water cooled hard wheel, they all finish up with a stropping wheel with no cooling at all, which it seems could itself heat the edge enough to lower retention. I wonder if anyone has studied the effect of the finish stropping heat on Tormek type systems?
Yes, it has been studied. Tormek was found to work fine all the time. Paper wheels on a grinder were found to work well, albeit conditionally. You might want to go directly to the PDF with the study:
http://knifegrinders.com.au/SET/Effect_of_felt_and_paper_wheel_on_edge_retention.pdf
Or,
You might want to peruse a group of topics related to the Edge Stability Testing, at the URL here:
http://knifegrinders.com.au/16SET.htm
BTW,
One of the guest authors at the link above is Dr. Larrin Thomas. Recommended reading, I mean throughout the website knifegrinders.com.au
Very helpful. This tells me that my method of hand sharpening with silicon carbide sandpaper is good. I had thought that I would get some belts for my belt grinder, but I think I’ll only use the belt grinder prior to heat treat.
I was amused to see the discussion on the sparks generated during grinding where they said that sufficiently fine particles will auto-ignite. I was working with a researcher who ordered some expensive very fine aluminum nano-particles. Each bottle was about $300. He opened the bottle and $300 immediately flashed!
Ha!
Thank you for putting this together. The information contained is very valuable and has informed by decisions to some degree (i.e. avoiding power grinding for the most part). But still, the lack of more data on this point is very frustrating–I really want to be able to get more information about what degree of softening we see on different steels, whether the lower options on the variable Ken Onion Work Sharp grinder (as opposed to the single speed original Work Sharp, which it appears was used in study 5) have an impact, to what degree regular misting offsets the problem, etc.
Having that information would help people make informed decisions. For example, I’ve been contemplating a setup like the following: Ken Onion Work Sharp on low speeds with automated misting to keep cool, then brief time spent with whetstone sharpening to replace the very narrow section of the edge that may have still suffered some softening. In theory this may be sufficient to avoid most of the negative effects we see in the data above.
But the reality is that I have nowhere near enough data to determine whether or not that is a good option; I’d just have to go for it and trust my intuition or very low sample size testing. Seems the only people who could actually answer this question for us across a variety of methods and steels would be some of the major knife manufacturers. Perhaps they actually have collected the data privately for their own use, but I’m not holding out hope that it will reach us.
Buy a 1×30 and 220 silicon carbide belts and keep the belt wet with a sponge. Problem solved.
I have a wet/dry 1×42 grinder and can drip water onto it with the top guard on. I only use it wet to sharpen knives for the most part as grinding hardened blades wet makes a mess. But at least it doesn’t throw out a lot of dust.
I might want to add that the first time I saw a Work Sharp grinder I thought to myself what a fast way to ruin the edge of your knife. I sharpen with a wet 1×42 as I said, even just wetting the belt between blades and be fast will still keep the blade cooler than a Work Sharp.
I feel like anyone who has ground a knife very thin behind the edge.can back up these results. It can take less than a second, or just a touch too much pressure, and you ruin the temper on a blade.
Now imagine, if instead of it happening on a blade only . 010 thick behind the edge, it’s on the micron or less thick edge, and think about potentially how quick that would overheat on a belt, or abrasive wheel.
I feel like the people that are arguing that this isn’t occuring, either have never ground a knife. Or have never ground a knife with a very thin geometry.
Thanks Larrin for this article. I’ve just ordered your book.