Testing for Overheated Edges

[Mark Reich]

Tempilaq arrived today. I'm armed with 300, 500 and 700°F, and dangerous with this stuff... cuz I'm geared fer it.  

[Bud]

Yeah Mr. Mark Reich! This is going to be interesting. I've been snooping around the shop but all we have on hand now is real high temp stuff and it looks older than the hills.  We got rid of the machine that was giving us the most trouble so I guess they don't use it much anymore. We do have some thinner that I can use. From what I read the thinner makes the stuff go off quicker. I think that I want to try something lower temp than EOU did but now I'll wait to see what you come up with and then adjust accordingly. I'll be interested to know what you think the reaction time is.

[EOU]

We second Bud's enthusiasm Mark. Looking forward to your discoveries!

[Mark Reich]

My plan is simple. I'd just like to SHOW how much grinding it takes to get a blade hot. My plan involves a camera man for the entire duration of testing so everyone can see normal grinding vs over grinding and high speed vs low speed grinding and buffing.

I was going to get some 200°F Tempilaq, but I decided water is better than Tempilaq. You can see, hear and feel a 212°F blade dunked in water.

[EOU]

Sounds pretty comprehensive Mark. We think that we'll go lower in order to compliment your tests. We'll buy some 200° Tempilaq (or whatever brand Mcmaster Carr sells) and see if we can find the lower limit. Perhaps half the test area thinned and half straight. 

Applying different approaches to the same problem is often a good way to answer questions. We'll perform the experiments, collect the data, and let folks apply it as they wish. Quite frankly, we may be reinventing the wheel here in consideration of the results provided by the two studies we provided a couple of pages ago in this thread. One study considered the range of material removal from 6 - 20 microns and the other 0 - 40 microns and the grinding temperatures attained thereof. One would like to think that the researchers were studying material removal amounts that were germane to common industry practice. If we could decide approximately how much material is removed in a single light pass on the Kally then our answers may be already before us. Is it reasonable to think that only a very few microns of material might be removed by an attentive knife sharpener using a belt grinder? Here's are excerpts from a very perfunctory search of the internet; 

 SOURCE   https://en.wikipedia.org/wiki/Grinding_(...e_cutting)

"Grinding practice is a large and diverse area of manufacturing and toolmaking. It can produce very fine finishes and very accurate dimensions; yet in mass production contexts it can also rough out large volumes of metal quite rapidly. It is usually better suited to the machining of very hard materials than is "regular" machining (that is, cutting larger chips with cutting tools such as tool bits or milling cutters), and until recent decades it was the only practical way to machine such materials as hardened steels. Compared to "regular" machining, it is usually better suited to taking very shallow cuts, such as reducing a shaft’s diameter by half a thousandth of an inch or 12.7 μm."

"Surface grinding uses a rotating abrasive wheel to remove material, creating a flat surface. The tolerances that are normally achieved with grinding are ±2×10−4 inches (5.1 μm) for grinding a flat material and ±3×10−4 inches (7.6 μm) for a parallel surface.[1]"

"Tolerances for cylindrical grinding are held within ±0.0005 inches (13 μm) for diameter and ±0.0001 inches (2.5 μm) for roundness. Precision work can reach tolerances as high as ±0.00005 inches (1.3 μm) for diameter and ±0.00001 inches (0.25 μm) for roundness. Surface finishes can range from 2 microinches (51 nm) to 125 microinches (3.2 μm), with typical finishes ranging from 8 to 32 microinches (0.20 to 0.81 μm)."

It is worth noting that the article distinguishes between grinding and surface finishing. Polishing would seem to be yet still  even further down the metal removal scale. We found one reference to polishing that described it as metal removal at the "molecular level". Additionally, grind depth seems to be primarily controlled not by the machines ability to index accurately very minute distances (as in milling operations) but rather by rotational or linear speed of the abrasive, material feed rate and by applied grinding force. These same factors seem to be the ones discussed at length by skilled belt sharpeners ad nauseum.

Assuming the accuracy of the article, then yes, metal removal depths of only a few microns are not only feasible but typical during machine grinding operations. The question then seems to be "are we humans, in a hand-held manner, capable of limiting the depth of our grinding passes to sub 10 micron levels?" If we are trying to create an edge apex of sub-micron width, 10 microns does seem a-plenty. If we are able, then our overheating fears might seem completely unfounded.  If not, and for those interested in sharpening a knife inside a few minutes, perhaps we should consider investing in automated grinding/knife sharpening machines. 

In the meantime, we'll get our 200° Tempilaq on order.

[EOU]

Got ours today Mark. Tomorrow, we try to fry it.

                      

[EOU]

Well, this didn't very long at all. Same knife, but different blade section, as our original Templaq experiment and we sharpened it the same way as well, one side only. This allows us to paint and then observe the Tempilaq reaction on the undisturbed side of the blade . Brand spanking new 150 grit cubitron belt mounted on a Kally. We didn't thin the Tempilaq as planned because it was quite thin already. This batch was much different than the 325° we received from a different supplier (we think perhaps old stock). The thinner directions specify "For replacing evaporated solvent". We think now that our 325° product had suffered to a degree from "evaporated solvent". It would appear to us that this issue of thinning Tempilaq might be much more related to a balance between flowability and coverage than to reaction time. Makes sense to us now. The applied Tempilaq is allowed to dry before test so the solvents are going to dissipate anyway. Right? 

Here's what the edge looked like after paint and before any grinding. This is taken at lower magnification and wider field of view in order to present a little more global view of the paint versus unpainted surfaces.

                                                          

Now we made just a few very light passes and felt for a burr. We could feel a very slight burr so took a picture. Pieces of the burr are barely visible at the apex but no apparent reaction with the Tempilaq. We hypothesize that the bulk of the burr is hidden by the Tempilaq.

                                                          

So we made a few more light passes and could then feel very apparent burr formation with the thumb. So we took another picture. 

                                                         

A little more apparent burr in the picture but no apparent Tempilaq reaction. This surprised us to some degree. We did not remove nearly the magnitude of metal from the bevel and edge in this experiment as we did in the 325°F Tempilaq experiment but still, this is 200°F stuff. So what does it take to trigger this stuff? In the 325° test we used a heat gun just to make certain the product was working. In this experiment we decided to try reduced feed rate and increased grinding pressure. We simply held the blade/edge in one spot (2 or 3 seconds) against the belt with maybe a pound of force. Here's the final picture taken at higher magnification.

                                                        

What a difference this made! A bunch more apparent burr and the Tempilaq has all but fully triggered.  Note to self when using a belt grinder to grind (sharpen) knife edges. Keep the feed rate up and the pressure down.

[Mark Reich]

I couldn't recruit a camera man on short notice, but I spent some time on this project today too. I didn't get the Tempilaq out, but I did some background research in order to try to quantify how much material we are talking about removing.

I chose an old Old Hickory with 1095 carbon steel at about HRC 55-56, which is way too soft for 1095, but you still get a nice heavy shower of sparks which is unavoidable with carbon steel.

I used a half worn 220 grit belt, which generates much more heat than a new belt, and removes much less material than a coarser belt. I drew lines across the blade so I could measure accurately, and they also helped keep my speed across the belt consistent.

Hopefully these pictures are worth a dozen words...  I was consistently removing about .001" of blade height with a normal pass on each side. That's a whopping 25µ !!  

Trying to get the blade warm, I started chisel grinding the edge. I took one pass/side which removed 15X the material. Nothing. I took another pass which removed 40- 50X more than a normal pass. 

I'm sorry, I'm getting practically no heat. It's warm to the touch, but that's it. I was able to blue a minuscule part of the tip, but it took some effort to build the heat to accomplish it.

[EOU]

Good to see that you have entered the fray here with your experiments Mark! The more the merrier we say!

A question for you:

You say " I was consistently removing about .001" of blade height with a normal pass on each side. That's a whopping 25µ !!". 

 What then would you say the "depth" of the grind on the bevel was? I suppose that if we knew the bevel angle of the knife and the thickness of the blade where the bevel begins, we could calculate it. We also see that you made a pass on each side and then measured the reduction in blade height - right? So, assuming a perfect grinding world, the blade height was reduced .0005" per grinding pass?


Have to tell you the truth about Tempilaq experiments. If we can't get 200° to go off during normal grinding procedures we wonder now how much luck you'll have getting much higher temperatures to trigger. There is no doubt in our minds that your touch and control with the Kally is going to be far superior to our own. Perhaps we could send you our 200° and 325° and you could work with those ranges thereby either verifying or contradicting our results. Lord knows that one bottle of Tempilaq would test 5,000 knives so there is no shortage of supply here. Perhaps you may be trying a different approach to the question though and, if so, that's good as well.

By the way, here's an interesting tidbit advanced in this research paper "The chips take away most of the heat generated in the grinding process."  https://www.witpress.com/Secure/elibrary/papers/LAMDAMAP03/LAMDAMAP03001FU.pdf. Sort of explains sparks doesn't it?

[Tman]

We're finding about the same thing here at our facilities. We turned off the water on the belt grinders and smeared 275 degree and 375 degree (Fahrenheit) temp sensor fluid on the flat side of some skive blades. These blades were ones we use in-house and were in need of sharpening. Skive blades have a lot of bevel surface area so I felt it would be a good test. Most of ours are a 1/16th inch thick tool steel and single bevel ground at 30 degrees. Hardened to around 58. Maintaining the same feed velocity and belt pressure through a typical operation we couldn't melt either sensing fluid. We kicked up the belt speed (x2) and that melted the 275 degree but not completely along the edge so I think we must have just about hit it on the nose for the test parameters. Speed kills. With the belt speed increased 2X and the feed rate cut by a third the 375 still didn't melt. Cut the feed rate by 2/3 and the 375 turned transparent completely down the edge. Lack of speed kills too. All the skive blades (melted fluid or not) tested when finished to our typical 190 BESS +/- 20. I guess we'll put them back on the press trimmers and see what we see service/performance wise and then I'll have to think about all this.