Punch and Die (and Fun)

I do not have the genius of Leonard da Quirm, but I do share one trait with him – I get easily distracted and sometimes spend several days trying to shave off a few minutes of some task or save a few bucks. Sometimes the effort definitively pays off – as in the case of my belt grinder or my forge burner, sometimes it is a success but with a question mark whether it was worth it – like the unbender (now I know it was worth it, btw, I have used it several times already and it is time-saver), and sometimes it is a bit of a flop, as when building a vacuum pump. If I had a definitive fail, I do not remember it, and so I allowed myself to get distracted again these last two days.

I have a problem with making metal bolsters, handguards, end-caps, and pommels. As in, it is difficult to get material thick enough to make them pretty, and even if it were not difficult, the result would be overtly heavy and thus would put the knife balance totally out of whack. The proper way to make bolsters and end caps is to make them hollow, and there are techniques for that. One of them is forging – as I did in the rondel dagger project. But that is labor-intensive, has poor reproducibility, and requires special tools anyway. Or I could buy prefabricates and adjust my design(s) to fit what is already on the market. Screw that!

So I have decided to make some new tools, and test them. The inspiration was a technique of minting coins before the invention of fly screw-press, which I have seen as a child in some black and white movie which has shown the making of Prague groschen at Kutná Hora. I remember nothing else about the movie except the part where they strike a punch on a silver blank with a hammer and thus make a coin. I think there was some drama and history in there too…

First I have made a die out of 5 mm high-carbon tooling steel. It consists simply of two holes – one for the bolster and one for the end-cap  (I have chosen my small hunting knife as a pilot project because I think the design will be improved a lot by it and because I do plan to make more of these knives in the future). Second I have ground two punches out of square stock of high carbon tooling steel that I have scrounged at my previous job. Grinding the forms with angle grinder was not easy, but it was not insurmountably difficult either. I had actually a lot more trouble with welding onto it the 15 mm round stock for holding the punch in place and for striking – my welding sucks, bigly. And because at least the first strike needs to be real mighty, I have built a small wooden stand to hold the punch in place for that.

© Charly, all rights reserved. Click for full size.

With the assembly on the concrete floor, as you see it in the photo, I have given it a mighty whack with my puny Mjolnir. And I rejoiced because it was a success. To protect the floor from damage I have put it on a steel plate for subsequent tries and I went and punched four sets for the four blades that I have currently in making, three out of brass and one out of pakfong.

© Charly, all rights reserved. Click for full size.

The pakfong was a bit thicker than the brass so it gave me some grief, thus the surface is not so smooth on the end-cap – I had to whack it several times and it wandered off the die and I struck it without noticing it. But that should not be a problem, there is enough material in there to polish these dents out.

It took me mere minutes to punch all these, and after a long time, I was really, really happy for a bit. There are a few details to iron out – like making a better stand for the punch, making it so I can put it safely on my anvil, figuring out the ideal amount of overhang and so forth – but it functions as it is and it is a massive saving in time already. Whether the knives will really look better remains to be seen, but I am confident they will. Further, this opens a lot of new possibilities for knife designs for me.

I Probably Won’t Do This Again…

After a month of work, I am finally at a phase where I have something to show for it. The kitchen knives are in the tumbler for the second day now, tomorrow I shall check if they are ready or not. But it need not hurry, I have enough to work with – eight fully polished blades.

© Charly, all rights reserved. Click for full size.

Today I have etched the logos and numbers on these because it is easier to do on a naked blade than on a finished knife. Now I can finally work on finishing at least some of these into final products while the kitchen knives tumble.

Working in bulk does save significant time and even resources, but 18 blades in one go was a bit too much. I am going to reduce the batches to 8-12 in size. That way it should keep its savings, but the polishing hell will not be that long. Polishing is extremely onerous and unrewarding work because one keeps doing the same thing day after day, working through the row of belts with very slow progress. With one knife, it is one-two days of a boring slog. With eighteen knives, it was three weeks – and one of them got broke and nine only to 120 grit before going into the tumbler, if not for that, it would be even longer.

These are not perfect, some of them have serious problems regarding symmetry, although only in one case it is visible with the naked eye. On all of them is it visible with calipers. I am starting to doubt that I will ever do a good job, but there are some signs of progress. One of those signs has, unfortunately, a bit of a negative consequence on these blades, all 17 of them – they are a bit thicker than I expected (a few tenths of a mm). That is because I have gotten a bit better at working on the belt grinder and thus I did not grind away as much material as I used to by having to correct mistakes

Mind you, they all will cut perfectly fine even so, and some of them already do despite not being sharpened yet. But a thinner blade will always cut better. On the other hand, these should be extremely sturdy and should be able to withstand even some serious abuse, and that is a plus for a hunting knife. We will see if there will be people willing to pay for these without bashing me over the head afterward.

Now to think about how to dress-up these blades and the accompanying sheaths. I think I have quite a few more weeks of work ahead of me, but now it should be creative work, and therefore much more fun.

Well, It Sucks, But…

My goal in mcgyvering a vacuum pump was to remain under 100,-€ – which I did – and get better results than I have achieved with my shop-vac setup – which I did too. Still, I do not know whether to be disappointed or satisfied.

I wanted to utilize things that I already have, which includes several water pumps that are used to water bonsai trees and vegetable beds in summer and pumping water out of the cellar in the winter and some spare piping from house renovations. So I had to buy only the things for making the vacuum pump itself – in combination with a water pump, the best option seemed to be something based on the venturi principle.

So I went and bought these parts:

© Charly, all rights reserved. Click for full size.

The parts were connected to each other more or less in the order as they are laid out on the picture. The black plastic hose connector was fitted into the brass one to lower its inner diameter. The brass hose connector right next to the right side of the chrome T junction is the inlet nozzle – I have glued an old tip from a silicone sealant tube (not depicted) in it to get the position and size of the nozzle correct. Into the upper brass hose connector was glued the white plastic 6 mm hose connector for the air suction.

So water comes in the T-junction from the right, gets squeezed through a nozzle which sprays into a slightly bigger opening in the outlet left, behind which is again a big pipe. The spray drags with it the air surrounding the nozzle and that way achieves suction through the top of the T-junction.

I am not able to write-up complete how-to, but this is the final product up and running in a vat of water.

© Charly, all rights reserved. Click for full size.

The bubbles show it is working. When I connected the suction tube directly to a new vacuum manometer, I got a suction of whopping 0,6 bar, which did really impress me. Unfortunately, I do not get anywhere near that when I connect everything to the jar. After a few minutes, it stabilizes at this.

© Charly, all rights reserved. Click for full size.

0.22-0.25 bar is still a bit more than what the shop-vac could achieve (which was 0.2). So it is usable and it is a definitive improvement because unlike the shop-vac it can run non-stop with zero risks of overheating anything, and it also makes nearly no noise, so a win there too. But when it achieves this, it still bubbles, so it still draws air. And when I close both ball-valves on the lid (one ball-valve is for pressure release, one for the suction), I start losing pressure in the jar really quickly. That tells me that the jar is not properly sealed and this here is not the maximum this setup can achieve, but an equilibrium between the pump and the improper sealing.

I had to make a new lid from five layers of plywood for this, with two ball-valves and the manometer, so there was a lot of potential for failure. But I did use water/airtight plastic sealant for everything and I went over all connections once more, yet I still cannot identify the leak(s). If it was a pressurized container, I might find the leaks with help of soapy water looking for bubbles, but I do not know how to check vacuum tightness.

So this is where I am now and this is where I leave this be for a few days at least. It has occupied me for three days already, time to go back to making knives.

This Was Not Supposed to Happen – Quality Control and Salvage

The last batch of knives did come out of quench straight-ish, but the tangs were a bit crooked on some blades. No biggie, he said. Those are not hardened, he said. Will be corrected in a jiffy, he said. And then he broke a blade in half because he was an idiot.

© Charly, all rights reserved. Click for full size

Still, I do not think the blade should have broken like this, it is tempered and I did not exert too much force. But maybe I did. Unfortunately, the tangs did partially harden on some blades, which is a major headache because I have to heat them up significantly in order to straighten them now. As my hardening experiment has shown, with this steel all that is supposed to be soft must not be heated above the 1050°C at any point in the process. Which, admittedly, I did not know for sure at the time of quenching these blades.

So what to do with this? I have spent nearly eighteen years being involved in quality control in one way or another, and I must say it is a great way to make enemies at all levels. Customers hate you for the crap your colleagues have produced, your colleagues hate you because you point out to them that they have produced crap and the management hates you because you cost them money (both when things go well as well when things go awry) without generating any profit whatsoever. And everybody expects you to tell them how they should solve the problems you found – despite that part being actually their job, with yours only being the pointing out of the problems.

Now I am my own quality control so at least now it really is my job to solve the occurring problems. For example, apart from this broken blade, one blade seemed to be softer near the tang. I have decided to continue work on it anyway and I established later on that it was only slight decarburization and the steel is sufficiently hard deeper under the surface. So far so good. But is there anything that can be done with this broken blade? Well, I surmised that yes, and I have decided to try some new things whilst doing it.

© Charly, all rights reserved. Click for full size

The broken-off tip was long enough for me to grind small (9 cm) blade out of it with enough steel left for a short hidden tang. I have simply cut it with an angle grinder to the desired shape and then thinned, formed, and polished the blade on the belt grinder. The notches on the tang are there to allow for the epoxy to grab onto.

© Charly, all rights reserved. Click for full size

I have made the handle from a piece of apple-root from a dead tree given to me by my neighbor. The bolster is from a coconut shell. The wood was partially rotten, so I have decided to try my hand at a new thing – stabilizing the wood with epoxy resin. And to utilize my resources meaningfully on this experiment, I went and cut some more wood, even more rotten, from a willow trunk that lay in the garden for years. It is full of holes, wood dust (AKA grub poop), fungal discolorations and it is extremely porous, it even crumbles in parts.

© Charly, all rights reserved. Click for full size

I do not have a compressor or a proper vacuum pump, and I was unwilling to spend any money on them at this stage, the resin itself was pricey enough. So I tried my best with things I already have at hand and I came up with this.

© Charly, all rights reserved. Click for full size

These are two huge thick-walled glass jars for canned fruit, with removable snap-on covers. They should withstand vacuum reasonably well. I have replaced the glass-covers with ones made from particleboard, with the joint and snap-on latch made from fencing wire around some screws.

The jar on the left has a ball-valve in the middle, with 6 mm tubing attached to it. The jar on the right has two 35 mm holes and one 6 mm hole in the middle. To the middle hole is attached the thin tube from the left jar, shop-vac is connected to one of the big holes and a swiveling cover is over the second hole.

I have given the knife with handle suspended in resin in the left jar. When the shop-vac was running with the ball-valve open and the big hole covered, it sucked air reasonably well out of both jars. But shop-vac cannot run with blocked intake for too long, thus the ball valve and the covered hole – after half a minute I could close the valve (keeping the low pressure for a few moments more), opened the big hole and left the shop-vac cool with flowing air for a minute or two. Then I have repeated this process as many times as it took until no bubbles were rising from the resin.

To get an estimate, I weighed the knife before and after, and it has gained about 20 g of weight (65 to 85 g), which seems reasonable, the wood was not extremely rotten, only slightly so. I have also weighed one of the willow pieces and the weight gain there was much more extreme – from ca 20g to 120 g. I have not cut the wood to see how deep the resin has penetrated, but it definitively penetrated deep enough to work on pre-formed handles, even if not on squared wood.

So this has confirmed that I can do this with the equipment I already have and that investing in some better equipment is thus warranted. Still, I am hesitant about spending 100,-€ or more on a vacuum pump, I hope I can mc-gyver something a lot cheaper. On Wednesday I am going shopping for parts. If I succeed, I will let you know.

Oh, and here is the finished knife.

© Charly, all rights reserved. Click for full size

Overall length 20 cm, blade length 9 cm at the spine, 3,5 cm wide, satin finish (trizact A100). The handle was polished to 1200 grit and buffed with beeswax.

I have given the knife to my neighbor because she has refused payment for the dead tree and I think she should get something for such beautiful wood. She did not object this time and I hope it will serve her well.

Different Hardening Methods for N690 – Experiment

When I was hardening the blades two days ago, I have tossed in there six cut-offs as well and I have used different methods to quench them. After that I had still one piece left so I have heated that up to the 1050 °C and let it cool in the forge. Then I performed some tests and the results are very interesting.

© Charly, all rights reserved. Click for full size

These are the samples and methods used (marked by the number of notches on the edge):

0 – left to cool in the forge
1 – untreated
2 – left to cool in the air
3 – quenched between Al plates only
4 – oil
5 – water
6 – AL plates + water + Freezer fro 2 hours

None of the samples were subsequently tempered, so they should be at peak hardness.

With the sample size just 1 piece per method I cannot of course perform too many tests. The idea was to polish the samples – which I did. And already during the polishing, I have noticed that all samples seem to be hardened, except #1.

The next step was supposed to be to etch the surface and look at the structure under a microscope. Well, that did not work at all, and the reasons are a mystery to me. Just as it happened with the knives last time, it happened again here – the electrochemical method worked on some samples perfectly, but completely failed on others. I was unable to solve the problem. Another thing was that my microscope apparently does not have big enough magnification to see a difference between the original steel and the quenched one. I could re-build it and improve it about ten times, but I am still not sure if that would be enough and I do not want to get sidetracked to that now, it would be probably more than one day of work and I have already spent two days having fun instead of working.

So I did what I could with samples of this geometry. First trying to scratch them with my hardness measuring gages .

The sample # 1 could be scratched by the lowest 38 HRC gage, which was to be expected.

All samples except # 1 could not be scratched by the 62 HRC gage, so they are at hardness 62 HRC or more. Which is something I did not expect, especially not of # 0, which was left to cool in the forge – and that took definitively several hours. I was expecting this sample to be harder than the new steel, but not hard enough for a knife – but it evidently is more than enough hard for a knife, hardness 62 is in fact quite excessive.

Secondly, I have tried scratching the samples against each other, and the results became even more interesting. All samples could scratch # 1, as expected. But none could scratch #4 and #4 did scratch all, whilst the remaining five could not scratch each other, so they are all of the same hardness.

Thirdly I have put the sharpest angle of the triangle approx 10 mm into a vise and break it off. #1 has bent easily, as expected, all others have snapped off.

What can I deduce from this? Several things.

  1. Sample #4 was hardened with the method recommended by the manufacturer and did come out as the hardest of them all, possibly somewhere around 63-65 HRC, which is as hard as steel can get. It could be a fluke (remember – sample size 1), but it could be the reason why this method of hardening is recommended. It is not surprising.
  2. From a practical standpoint, the method of quenching seems to be quite inconsequential nevertheless. The oil quenched sample would be brought down a few points in heat treatment anyway and for practical purposes, anything above 51 HRC will be usable with just a bit more edge maintenance, anything above 55 HRC will have reasonable edge retention and above 57 HRC we are in the realm of no reason to complain whatsoever. In this light, the difference between the recommended oil quench and all the other methods seems to be so small as to be trivial and only interesting from a nerd standpoint.
  3. The freezer step does not seem to have done anything for this one piece, but this does not rule it out from use on larger pieces that could not be so thoroughly and consistently heated in my setup. Did not do any harm either.
  4. Although tempering was not tested, this experiment does indicate that it is just not possible to really destroy the edge on this steel by overheating it during grinding/polishing since even cooling it from the 1050°C to room temperature over the course of several hours hardened it very nearly as well as the recommended oil quenching. I will not test tempering temperatures with regard to this specifically since there are graphs to be found on the internet that show already that the hardness of N690 does not get below 56 HRC up to 900°C.
  5. If I want to peen the end of the tang, or do any other work with it, I must be careful to not heat it above the critical temperature at any point in the process. Because once heated above certain threshold, this steel hardens, I cannot prevent it and I probably cannot anneal it again.

It would be interesting to see what is the exact influence on toughness/strength once tempered. I could not find it, so I will have to test it myself. But for that, I will need another sample geometry. So maybe next time.

All in all, the N690 seems to be pretty remarkable steel. It does not have the label (and price tag, otherwise I could not use it) of “super steel”, but it is no wimp either and apparently is not very fussy about the heat treatment, apart from the requirement to heat it above 1050°C.

Quenching Jig

I hope you are not bored with my writing about knife making yet. It is not that I do not care about other things, but knife making is where most of my focus is right now. And this week I had barely the strength to do anything else at all.

I have started a new batch of blades, 18 pieces altogether – 4 small hunting knives, 4 badger knives, 5 chef knives (prototyping new design), and 5 universal kitchen knives. So this whole week I was drilling, cutting, and grinding steel every day. I have progressed reasonably fast, despite being also slow. Because I simply cannot handle more than 6 hours net a day at the grinder, and I have to make a substantial pause every two hours. My hands are doing reasonably OK and I have been pain-free for a few months by now (despite never having a diagnosis about what was wrong last year), but even so, the vibrations are a strain on the fingers. And after two hours not only the glasses start getting foggy – the mind does so as well and thus the risk of injury increases.

© Charly, all rights reserved. Click for full size.

Today was a big finale of that busy week and all knives went into quench and tempering. This time I have decided to use quenching foil on all blades, despite bad experience on some blades the last time. I have decided to do so because the protective coating that I have concocted did work well but also was a real pain in the ass to remove afterward. And that did cost me a lot of time and I destroyed two belts before I figured it out. But I have tried plate quench with on this steel with moderate success), and I have hoped, really, really hoped, that doing plate quench on newly ground and straight blanks will lead to straight blades without having to scrub off a hard crust.

And following the maxim of Scrooge McDuck “Work smarter, not harder”, I have done my best to make the plate quench easy and reliable – I have built myself a jig. The construction is very simple and it did only take me a few hours.  I have used the locking pliers and aluminum plates from last time, but I have connected it all into one piece that can be easily used in one hand. So instead of one hinge in the middle of the plates, I have added two hinges on the sides. Then I have drilled holes in the fixed jaw of the pliers (and re-ground it a bit) so I can screw one of the plates to it with two M4 screws. The movable jaw is not fixed to the second plate, but it does fit into a groove cut in it and it was also re-ground for a better fit at the angle where most blades will be gripped.

© Charly, all rights reserved. Click for full size

© Charly, all rights reserved. Click for full size

The idea was that I will hold this jig in my left hand, with my right hand I pull a blade out of the forge, insert it between the plates (still in the foil), lock the pliers and dunk the whole thing into a bucket of water. And it worked well! None of the knives came out of the quench with a perm or as a banana-imitation, none have cracked either, at least I did not notice it yet. All but one blade quenched properly (and that one I have re-quenched OK the second time) with hardness above 55 HRC, most even over 62.

© Charly, all rights reserved. Click for full size

© Charly, all rights reserved. Click for full size

Hardening all these blades took me only 15-16 minutes per blade, so even with all that foil wrapping, I was well within a reasonable time. There was only one small problem towards the end – my new burner worked really well and I have reached and held 1.050-1.080°C without problem with four blades in the forge, but after a few hours, it started to struggle. I thought at first that I am running out of propane, but that was not the case. As it turns out, the propane bottle got too cold (just 10°C) after that long continuous decompression, and the gas was evaporating less. I am not sure yet how to solve that problem for the future.

Now the blades are in the second tempering cycle (each cycle two hours at circa 180°C) and they did not seem to develop any bends or curls in the oven either. So far, so good, let us hope it stays that way since I am going to try some new techniques with this project again.