Principles of tubular free abrasive drilling

The bottom of these drill holes were covered with a dried-out greenish pulp – a mixture of particles of abrasive, ground stone, and oxidized copper from a tube. The abrasive component consisted of corundum grains (Mohs hardness 9) a large deposit of which has been revealed at Wadi Hafafit.

Drive 

A copper tube was fixed to the bottom end of a wooden axle which was rotated as part of a bow drill. This method was used by ancient Egyptians only for drilling very small holes with a diameter of perhaps 1–2 cm (Figure 3).

Drilling with abrasive suspension is a different process altogether, as the rotation of the brace and the tubular drill bit is precessional. While wobbling, the tube shakes the pulp in the drill hole from top to bottom both inside and outside itself, and friction of the pulp against the surfaces of the drill hole and the core not onlyslows down the rotation of the tube but also makes the drill hole broader at the top. The core, consequently, widens to the bottom.

The cutting edge of the tube, originally blunt, rubs abrasive particles together and compacts them to the inside and outside, where they start beveling its end (Figure 7). When all grains are squeezed from under the tube end, they continue to sharpen it from both sides, while it touches the bottom of the cut and grinds it.

The profile of the cut becomes V-shaped. Experiments show that a sharpened tube rotates easier, but if the cutting edge becomes too thin and sharp, it gradually flares outwards under the axial load (Figure 8).

Grains of loose abrasive intrude into the soft surface of the copper tube for a short time and act as fixed file teeth of great hardness. It is in fact with these “teeth” that the tube drills the stone. Little caverns left on the surface of the tube by abrasive particles show (Figure 8) how firmly these particles intrude into copper.

Grooves

The mechanics of groove formation on the inner surface of a drill hole is related neither to the drilling process itself (when abrasive grains embed themselves into the copper near the cutting edge of the tube) nor to the rhythmic character of feeding the abrasive into the working area. The key observation is that after replacing the abrasive suspension with dry or wet abrasive, or even with the suspension with a high concentration of abrasive, these grooves disappear within 2–3 minutes, and the surfaces of the drill hole and the core become smooth. If, however, we again add to the suspension with the proper concentration of abrasive, grooves will return in several minutes.

When not agitated, the suspension separates and abrasive particles settle to the bottom of the drill hole under their own weight. The grooves start to appear on the surface above the tube cutting edge as the suspension is shaken and stirred by the rotating tube, which presses and rubs them against the drill hole wall and the core surface (Figure 9).After a grain has scratched a groove, the nearest grains start slipping into it and fill it, therefore, the nearest groove can then only be scratched to a certain point.

Holes in heavy objects were drilled for using as lifting lugs. For instance, four holes were drilled in the lid of the sarcophagus of Akhethotep (a high dignitary and probably a son-in-law of King Khufu, the 4^th dynasty) to lift, transport, and install into its place. A good example of a slanted lifting hole is in the lid of the granite sarcophagus from the mastaba of prince Kawab (the eldest son of King Khufu). The bent shape of the hole is a mere deception caused by the shape of the chip.

Tubular abrasive drilling could also be an initial step in making stone vessels. Then, if required, the drill hole could be lathed or widened in some other way.

Often, non-cylindrical hollows were also made by tubular drilling. For this purpose, holes were drilled close to each other—either with or without overlaps. (Figures 21 and 22). But the obtained cores were not knocked out immediately. Until the work was complete, they were kept in their initial positions in order to fix the copper tube from sides during each drilling cycle. Also, cores left in place took up a certain volume thus making it possible to use only a moderate amount of abrasive suspension. The level of suspension was kept constant.

The same method was used for making sarcophagi. Traces of abrasive drilling are still preserved on some of their inner surfaces because it was difficult to remove them from hard stone. Sometimes, hubs for hinge pins of large gate doors were also made in this way—by drilling several smaller holes instead of a single large hole.

2010. Drilling with a quasi-split tube

Materials: granite, corundum (a cutting disc crushed to sand), and a split tube. A low-speed drill was used as a drive. The universal joint (Figure 24) between the drill chuck and the tube was inserted in order to imitate free wobbling of a drill bit of a hand brace.

The split copper tube used in this experiment could hardly be called “split” and even “a tube”. It was made of a rolled up copper sheet and the distance between the edges of the side opening was about 18 mm (Figure 25). This opening ensured free circulation of abrasive suspension in a drill hole between the inner and the outer volumes. Appearance of grooves on the core and the drill hole wall ( Figures 28–30) also caused by this opening. Abrasive suspension was fed into the space between the hole wall and the tube without removing the latter.

The split copper tube used in this experiment could hardly be called “split” and even “a tube”. It was made of a rolled up copper sheet and the distance between the edges of the side opening was about 18 mm (Figure 25). This opening ensured free circulation of abrasive suspension in a drill hole between the inner and the outer volumes. Appearance of grooves on the core and the drill hole wall ( Figures 28–30) also caused by this opening. Abrasive suspension was fed into the space between the hole wall and the tube without removing the latter.

2016. Drilling with a quasi-split tube

Materials: granite, corundum (a cutting disc crushed to sand), and a solid (quasi-split) copper tube. A hand-made brace was used as a drive. Its plaster flywheel ensured vertical load and gyroscopic effect for drill bit stabilization. The drill bit itself was a solid copper tube approximately 2 mm thick. To ensure circulation of abrasive suspension, several through-holes of 10 mm diameter were drilled in the tube body instead of a longitudinal cut in a split tube.

This experiment went wrong from the very beginning. Firstly, the selected abrasive material was too coarse and its grains started to tearthe granite surface, dislodging individual crystals. Secondly, the abrasive suspension was oversaturated for most of the experiment. It was not taken into account that, being once fed into the process area, a portion of abrasive could not be consumed completely — its volume was only reduced a little due to mutual grinding of grains. It was not the same as with granite sawing when the suspension was constantly spilling out of a cut back and forth and needed to be replenished in large amounts. So, lacking experience, we continually poured suspension into the gap between the hole wall and the tube making the pulp ever thicker. Each time a new portion of abrasive was added to the pulp (corundum, granite, copper) when the previous portion had not yet been abraded. The drilling process was accompanied by a dry scraping noise although the granite block being drilled was put into a bucket full of water. By the end of the work, when the tube sank into the cut to approximately 5 cm, it was very hard to rotate the brace: there was so much abrasive that it filled the whole space between the tube and the walls from both sides. The result is shown in Figure 31.

The inner surface of the drill hole bears some rough and almost erased circular grooves, while on the core surface they are practically invisible. This is because there was now an opening in the tube body large enough for pulp circulation: small round holes were a poor substitution for a longitudinal cut. Also, there was no vertical splashing of suspension against the core surface.

May 2017. Drilling with a quasi-split tube

Materials: granite, corundum (a cutting disc crushed to sand), and a solid (quasi-split) copper tube. This was a successful attempt to document the skill of applying corundum abrasive optimally. The brace and the tube were the same as the year before. Then the task was to shoot a video for the first time ever demonstrating techniques used for drilling hard stone with primitive tools. The experiment was a success in the sense that an ancient hand brace was reconstructed and that drilling granite with a copper tube and abrasive was proved to be possible. But still, it failed in the sense that “truly ancient Egyptian grooves” were erased soon after emerging.

Still, the main task was to figure out the working conditions that ensured this particular pattern of holes. The physics and mechanics of how grooves appeared in granite had been understood long ago—many hinge hubs in granite lintels from ancient Egypt bare such grooves—but details of the process remained unknown and unclear. We should understand how ancient workers actually operated. For this purpose Nikolay Vasyutin had conducted more than 20 experiments before the conditions of granite grooving became clear and the repeatability of such results was obtainable.

In this experiment, in 2017, the abrasive was fed under the close supervision of Nikolay, who took it on himself to rotate the brace. Making “truly Egyptian” grooves came easily to him, while precession of the brace axle in inexperienced hands could have damaged them even if the abrasive had been fed correctly.

The moment when the corundum abrasive should be replenished was identified by the processing sound — a quiet hissing. A portion of abrasive was put on the flattened tip of a knitting needle directly into the gap between the tube and the drill hole wall. At that moment the processing sound changed from hissing to scraping, which continued for around a minute or two. The noise then turned to hissing again, and this signified that it was time to add a new portion. The correct amount—a 4–5 mm lump of wet corundum—was found by trial and error: it should provide easy rotation and efficient drilling without blocking the processing area or erasing grooves.

Another method that was tried was to heap the abrasive near the cut—inside the plasticine collar plastered around it—and flick the correct portion into the processing area at the right moment (at the changing of the processing sound) or flush it with a thin trickle of water.

The tube — the same one as in the previous experiment—was fixed to the axle in a more authentic way: with wooden wedges and a rough cord. The tube edge continued self-sharpening and in total has sharpened from 1.8 to 0.3 mm.

Such a result was unacceptable. The answer was to increase the radius of the pole at the point where the string was wound around it. Figure 41 shows a pulley fit onto the pole to increase the radius of application of string tension force. Here the arm of force is elongated and the torque that overcomes resistance in the processing areas is thus increased.

The force that presses the tube to the nearest side of the hole wall and to the farside of the core decreases because with the elongation of the arm of force its point of application shifts aside from the shaft axis.

As it has been supposed, this tool is only useful for drilling small holes. If the drill bit diameter—not the shaft diameter—is small, it takes less effort to rotate the drill. Besides, in this case the shaft does not librate considerably: being jerked by the string, it is rather rotating than librating. The speed of drill rotation is lower but the precision of drilling is higher: the tube is not pressed to the drill hole surfaces in the way illustrated in Figures 39 and 40 and, thus, the ovality of the drill hole and the core decreases noticeably (Figures 42 and 43).

In the next stage, the shaft was shortened almost up to the pulley and this made drilling even more convenient. In the next experiment, a flat river cobble was drilled through with a lock-seamed solid copper tube (Figure 37 above) in less than 5 minutes.

May 2018. Large diameter drilling

The following experiment was conducted in order to test a method of drilling large diameter holes similar to hinge hubs in temple pylons. Essentially, this drilling method should be the same as those described above. But the aim was to check whether it was at all possible to turn a 20-cm tube by hand while overcoming the resistance of the abrasive material.

A 0.6 mm thick copper sheet was wrapped around one end of a log of 19.5 cm in diameter and nailed to it with five copper nails. A gap of approximately 1 cm was left between the strip ends for free circulation of the abrasive pulp (Figure 47). Before the start, the copper sheet and the workpiece were weighed. This workpiece was a granite tile of approx. 25×25×2.5 cm (Figure 48). In addition, a wooden supporting structure was specially built for fixing the brace, i.e. the log (Figures 49 and 50).

Surprisingly enough, drilling itself didn’t take much effort (Figure 53). The abrasive was again fed based on the telltale processing sounds: when it changed from a scraping to hissing noise, a small portion of corundum was flushed by a trickle of water from the wooden stencil into the processing area (Figures 52 and 54).

Taking turns, the four of us worked for four hours and then checked the interim result (Figure 55). The brace was lifted, the workpiece was taken from underneath and flushed. The result turned out to be long known and expected: a slightly tapered profile of the drill hole with a rounded bottom of the cut.

At the top, the cut was wide enough (4 mm) because the wooden stencil hadn’t completely prevented the copper tube from wobbling from side to side (Figures 56 and 57).

The tool was then reassembled — this time without the wooden stencil as the depth of the cut was sufficient to hold the tube firmly (Figures 58 and 59).

The next day we worked for about five hours. At a certain moment, the water suddenly drained out from the cut—an arc of around one third the diameter of the circle was sawed through. This happened because the tile thickness was not uniform: 23 to 27 mm.

The workpiece was washed of pulp but in the process of knocking the core out it cracked in two halves (Figure 61). The cutting edge of the tube was predictably covered with small caverns left by abrasive particles (Figure 62).Walls of the core and the drill hole had also been covered with grooves—somewhat erased by the excess of abrasive material (Figure 63).

The profile of the bottom cut on both sides also looks quite characteristic and predictable (Figure 63).

Figure 64 below shows cone angles of V-shaped bottom cuts. On the left picture, there is a well-known find—the hinge hub drilled in a granite lintel located near to the sixth pylon in the Hall of Annals of Thutmose III at Karnak. The right picture illustrates the result of our experiment.

After the experiment, the granite and copper materials were once again weighed and measured. The obtained values are the following:

Copper tube thickness:                                       0.6 mm

Granite tile thickness:                                        23 to 27 mm

Circular cut diameter:                                        195 mm

Max. drilling depth:                                           23 mm

Max. width of the cut at the top:                       4 mm

Cone angle of the cut at the bottom:                  ~9°

Corner radius of the cut at the bottom:           ~0.5 m

Total time:                                                         9 hours

Granite wear:                                                     4128–4038=90 g / 33.3 cm³

Copper wear:                                                      366–315=51 g / 5.7 cm³

Copper to granite mass wear ratio:                    51/90 g or 1/1.8

Copper to granite volume wear ratio:                5.7/33.3 cm³ or 1/5.8

The last experiment

One more experiment, interesting for us to conduct, was drilling with a tube home-cast with true copper ore. So, we needed a thin cast and forged plate that should be then annealed and rolled into a tube with a large diameter. Unfortunately, what was ordered and delivered to us turned out to be a ready-made cast tube with thick walls and with almost the same diameter as was used in previous experiments (Figure 65).

This would reveal nothing new or interesting: the drilling process would be very inconvenient due to excessive thickness of the tube walls and no circulation of the abrasive as the tube was solid. Also, the cut width in this case would be enormous—as well as the time needed for grinding the corresponding amount of granite into pulp. For that reason, it was decided to drill the softest workpiece to hand—the same as was used in previous experiments with bow drilling.

The abrasive had been fed irregularly this time.

The results of this final series of experiments are shown in Figures 68 and 69.

Conclusion

To show the exact methods of stone processing that were really in use in the ancient Egypt was not the aim of our experiments. Rather, ours was an engineering approach: we started from the investigation and identification of traces left by ancient tools.

In the course of our experiments, we examined practical techniques of working with such tools in order to understand what they could have been in ancient times.

From the technical point of view, the tools were simple enough and the task was to clarify the details of their use. Then, if those techniques were reproduced correctly, traces on test workpieces should be exactly the same as those seen in ancient materials. Thus, our aim was to understand the principles of hard stone working.

***