Waterjettting 30c - why cavitation isn't being used in machining metal.

Most applications of high-pressure waterjets need the jets to make relatively precise cuts into a target surface. When examining different ways of improving jet cutting performance, the narrowness of the cut which is achieved is therefor often a critical factor in deciding how to make the cut.

In this short segment of the series I have been discussing some of the findings that Dr. El-Saie made in his Doctoral Dissertation, and at the end of the last post had shown that he had found that, under the right conditions, a cavitating jet would remove more material from a surface than would an abrasive-laden waterjet of equal power within the same time frame, and at the same operating jet pressure.

This work was carried out prior to the submission of his Dissertation in 1977, and – because most of the work on refined nozzle design had still to be done – the nozzle designs that were used in the study were considerably cruder than those that have been developed, by a number of companies, in later work. For this reason it is not realistically possible to compare the results achieved in the early studies with the current state-of-the-art, particularly since, while there has been a great deal of work on improving the design of AWJ nozzles there has been almost none focused on improving the cavitation destruction of a waterjet.

The important distinction to make is that while there has been considerable work on reducing the damage that a cavitating flow can achieve when it impacts on a surface, there is almost none that has been aimed at making that damage worse. Part of the problem that this lack of work has left us with is that, in most cavitating systems, part of the cavitating cloud will collapse within the nozzle assembly. While this may only be a small fraction of the total, within a relatively short time (and this has been measured in fractions of a minute in an intense erosion design) the nozzle is destroyed. It is therefore unlikely that the most efficient nozzle design for inducing cavitation erosion has yet been developed.

Part of the reason that it has not has to do with the requirement that is listed at the top of the page. Where abrasive particles are mixed within a very narrow jet of high-pressure water, the abrasive cutting is confined, so that the slots can be controlled to a high degree. We have shown, for example, that it is possible to make cuts through titanium within a tolerance of 0.001 inches of the design requirement, and with a smooth surface over the full surface of the cut. (This requires, in thicker materials, that the nozzle be slightly tilted so that the jet taper over the depth of the cut is compensated for over the desired edge cut).

Unfortunately this precision in cutting cannot be achieved (at least with the current levels of understanding of the process and controls) with a cavitating jet system. In part this is because of the omni-directional nature of the collapse of the cavitation bubbles over a surface. An abrasive particle has the great majority of its velocity aligned with the axis of the cutting jet (though this might slightly deviate if the particle cuts into the surface more than once over the depth of the cut). Cavitation attack can occur in a more omnidirectional way.

To illustrate the point consider an experiment where we fired a waterjet along the edge of a block of dolomite, in such a way that the jet did not contact the rock, were the test to be carried out in air. The jet was carried out underwater, with back pressure of the surrounding water adjusted to intensify cavitation along the edges between the waterjet and the surrounding water. The collapse of those bubbles against the side of the rock eroded the cavities shown.

Figure 1. Samples of dolomite attacked by a cavitating jet. The jet is, in both cases aimed parallel to the cut face (along the line of the red arrow) and just off the block surface.

The samples shown in Figure 1 were exposed to the jet for a minute, with the samples held under water in a cell where the back pressure (BP) could be adjusted.

At a pressure of 6,000 psi, with a back pressure of 60 psi and a nozzle diameter of 0.02 inches the damage to the rock is small. Although it should be noted that there is a hole that eats into the rock perpendicular to the direction of jet flow.

This is more immediately obvious with the sample shown on the right, which was cut with a 7,000 psi jet, against a back pressure of 35 psi, and with a jet diameter of 0.030 inches.

It is also worth noting that the hole generated is not consistent in diameter as the hole deepens. The penetration of the jet into the wall, perpendicular to the main jet flow, is caused by the individual collapse of cavitation bubbles against the surface of the rock. And this is not a consistent phenomenon along the jet length, but for varying conditions it will occur at different distances from the nozzle. In this case the hole widens and deepens about half-an-inch below the rock top surface, eating further into the rock relatively consistently for the following few inches.

In a separate experiment the resulting hole can be seen to vary in depth over the length of the cut into the rock.

Figure 2. Hole cut into dolomite by a cavitating jet. Note that softer layers of the rock are excavated more deeply into the hole wall by the collapsing bubbles. The hole is roughly six inches deep.

The cut is also much broader than the originating jet, as can be seen from the cavity eaten into the rock surface perpendicular to the jet at the bottom of the sample. The cut is much more ragged than the precise line that would be cut by an AWJ, so that, although more volume is removed, the removal path is not as precisely defined as is needed in most applications. The irregularity of the slot shape can be seen where a cavitating jet is traversed over a 2-inch long block of dolomite over a period of five minutes.

Figure 3. Traverse path of a cavitating jet eroding a slot into dolomite. The block is roughly 2 inches long.

Clearly, at this stage in its development, there is little precision to the slot that is being generated in the rock, and it has little application in machining.

However the operating jet pressures of these cuts are within the range of those pumps that are available at relatively low cost at hardware stores, with the simplicity in use that this implies. And yet they are capable of disrupting, into the constituent grains of mineral, even the hardest of rocks that will be encountered. Gold ore, for example, can be relatively expensive to drill, because of its strength and toughness, and yet it can be penetrated in the same way, and with the constituent minerals separated, as was the dolomite.

Figure 4. Cavitation erosion of a sample of gold ore.

The problem comes in collecting the very fine grains of the mineral from the rest of the ore sample, once it is disaggregated.

Cavitation therefore, at present, is a tool better used in applications beyond those of the machine shop. It does not, however, have to stay limited to that restriction.

Waterjettting 19b - California gold mining

While mankind has directed the flow of water against earth and rock faces for millennia, as a way of eroding and removing material, it was not until the days of the Gold Rush in California around 1850 that the idea of confining the water into a hose, and through a nozzle crystallized.

Gold was originally found in the gravel in 1848, when James Marshall was helping John Sutter build a sawmill on the banks of the South Fork of the American River, in what is now Coloma, CA.

Figure 1. Sutter’s Mill on the American River (Replica by California Parks ).

The news that gold had been found in the tailrace led into what has been referred to as the greatest mass movement of people in the Western Hemisphere, as people flocked to California over the next few years, during the period of the California Gold Rush.

Figure 2. Location of Coloma, CA (red circle) relative to Sacramento and San Francisco – Lake Tahoe is by the 395 sign in the upper right. (Google Earth)

As the prospectors panned the gold from the stream beds, so they moved north east along the valleys and rivers, seeking the sources of the gold particles that millennia had washed down from the Sierra Nevada. One such source was found at American Hill, just north east of Grass Valley. Here the gold was found in beds of a weak sandstone, lying relatively close to the surface.

Figure 3. The American Hill Diggings, with plaque. The original height of the hill can be seen in the background.

The gold settled to the bottom of the sandstone, and so the miners would tunnel into the side of the hill, seeking to find the richest layer. Unfortunately as you dig out the bottom of a hillside, the overlying rock has a habit of falling down, with mildly fatal results to those caught in its path. This made mining somewhat dangerous, given the soft nature of the rock as Edward E. Matthison found when he was nearly buried when he was working the property. So with partners, he decided that a more remote method of digging out the gold was needed. So, with the help of a local blacksmith named Miller, he fashioned a nozzle on the end of a canvas hose he ran from a water reservoir at the top of the cliff (initially a nail keg) and used the resulting stream to wash the ore (and overlying rock) into a channel that was later turned into a flume, with a series of strips to catch the gold.

Figure 4. Early hydraulic mining

The method had many advantages since, in the process of washing the rock from the solid it was broken down into individual particles. This separated the gold, sand and clay particles, so that while the gold particles would be trapped in the flume, the lighter sand and clay particles would be carried further downstream with the water. By 1853 they were paying a water bill of $153 a week (with water at $0.75 per miners inch this meant they were using 2,000 gal/min) but making the four partners a profit of $50 a day. Larger and larger monitors (the name given to the nozzle and pivoting assembly) were built, throwing water at greater distances, and mining at much faster rates.

Figure 5. Monitors at work at the North Bloomfield mine.

This, in turn, required increasing amounts of water, and this was carried in flumes down through the Sierra Nevada, with agreements being made between companies for distribution, collection and the passing on of water. The nozzle diameters of some of the larger monitors grew to more than 8 inches, and they were capable of mining tens of feet from the operator.

Figure 6. Later design of monitor. The wooden beam usually had a box holding rock on the other end in order to balance the weight of the nozzle section.

The nozzles were made longer, as they were made larger, in order to get the jet to throw to greater distances, but this made steering and control of the jets more difficult. The gooseneck swivel was invented in 1855 to help swivel the nozzle, and a monitor operator noted that when he stuck his shovel into the stream of water it deflected the nozzle. This was Dave Stokes at the Malakoff mine and led to the invention by his Supervisor, Henry Perkins, of the rotating system for sprinklers that is still used to this day.

Figure 7. Modern rotating sprinkler showing the deflection plate. (Aliexpress )

The largest mine in the region was the Malakoff, and in the region around it there were some 425 companies operating and, between 1871 and 1880 they produced $121 worth of gold (at the price of the day).

But there was costs to this operation outside of just the mining ones. For while the gold was captured in the flumes, the sand, and more particularly the clay, was carried in the water until it became less turbulent. And that was when it reached the Yuma, American and Beam rivers flowing out of the Sierra Nevada and down towards Sacramento. As the water slowed, so the clay precipitated out, and the river beds filled with sediment. Thus, when the rains came, the water overflowed its banks, flooding the neighboring fields.

Foregoing the fact that it was the mining that had brought the farmers and many others to the region, the floods were not acceptable, and following the floods of 1880 there was an increasing effort to contain the mining sediments. This led to the court ruling by Judge Sawyer in 1886 restricting the practice of hydraulic mining, and the technology fell into abeyance. It was restarted at the time of both World Wars, but in recent times there was only one small mine that had been “grandfathered” still in production. Its role in developing California is not greatly recognized at present, and the remaining legacy is more seen in the vertical bluffs and large flat areas of mined sand that are left north of Grass Valley, together with the old wooden water flumes that still thread their way around the edges of the valleys.

Figure 8. View of the Malakoff Diggings

I’ll talk more about the spread of the technology next time.

Waterjetting 2d - Adding cracks to Nature

In the last few weeks I have focused on demonstrating, with examples, that water effectively removes material by penetrating into natural cracks in the material and causing them to grow. But what happens when there are not enough cracks to remove material at an economic rate? The modern approach has been to raise the pressure of the water so that smaller cracks grow faster, thus providing the production rates needed, but that option wasn’t available in the past.

I mentioned last time that miners in the Caucasus Mountains of what is now Georgia used the power of mountain streams to erode gold deposits over 3,000 years ago. Perhaps learning from that, when the Romans came to Las Médulas in Spain, some 2,000 years ago, they though of water again as a way of mining the gold-bearing sandstone of the local hills. And though they had to modify the initial idea, the result became the most important gold mine in the Roman Empire. It is now a World Heritage Site.

Figure 1. Location of Las Médulas in Spain. (Google Earth)

The sandstone was more resistant than soil, and so the Romans came up with two ideas to improve the rate at which the gold ore could be removed. The first idea was to run galleries into the sides of the hills, creating large chambers underground, with support for the roof from wooden supports that were left in place.

Figure 2. Tunnel driven into the bottom of the hill at Las Médulas.

Figure 3. Underground room at Las Médulas.

At the same time that the mining preparations were going on local streams were being diverted and dammed so that a large volume of water was held in reservoirs and then carried by manmade channels to a point over the mining chambers. With the water ready, the timbers were set on fire, which initially weakened the overlying rock so that it began to fail, falling into the opening, and as the support burned away more rock fell into the opening until the cavity worked its way up to the surface. At this point the reservoir gate was opened and water flooded down the channel to fall into the cavity. As the water fell it further broke the rock into grain-sized pieces, and carried these down and out through the original opening in the hillside.

Figure 4. A Collapsed cavity, not the two figures at the arrows to get a sense of scale.

The water and debris flow was directed into flumes, in much the same way as modern miners in Alaska practice today, except that where carpet is used to catch the gold particles in Alaska, in Spain the Romans used plant stems (silex) to catch the gold. After drying the plant could be burned, easing to recovery of the gold. (In more modern times Spanish miners have lined the flumes with oxen hides.)

Figure 5. Artist sketch of the troughs used to capture the gold particles at the Spanish mines.

The use of heat to weaken rock before using water pressure for cutting has been tried with a couple of interesting wrinkles both by researchers at Rolla, and at the then U.S. Bureau of Mines and in Colorado, among others. But those more modern trials will be described later in the series. Using water streams to erode surface outcrops of mineral survived as “hushing” in the North of England and elsewhere until fairly recently.

Move forward some 1800 years or so from Roman Spain, and at the turn of the 19th Century miners in both Russia and New Zealand had a problem in mining coal. In both countries there were good quality coal seams, but they sloped at a steep angle that made it difficult to move men around without their slipping and falling. It was also difficult to support the roof, which was achieved at the time by sawing wooden props to length and wedging them between the roof and floor. Both nations had the idea of modifying the Roman idea of using water to remove the mined coal, but coal was thought to be somewhat stronger and more resistant than the Spanish sandstone.

In the New Zealand case the mountainous countryside makes it expensive to drive roads and as early as 1891 wooden flumes were being used to carry coal to the consumer. However it was then realized that the water could be used to also remove the mined coal, particularly that which was left in regions of the mine where it was not safe for men to go. The coal was therefore initially blasted, and then the flow from the nearby streams was directed at the debris pile. The volume of water, and the slope of the mine combined to remove all the mined coal, often overnight, so that a new area could be worked the following day. It was not until 1947 that pumps began to be used to drive the water at greater pressures. At this point, with the higher pressures that pumping brought, it was no longer necessary to pre-crack and break the coal with explosives.

While the New Zealand coal seams outcropped at the surface in very hilly ground, the situation was somewhat different in the Donets coal seams in the Soviet Union, where the seams were thinner, and production was barely economic. The seams in these mines were much deeper than in New Zealand, and so jet pressure could be provided from the drop in height from the mine surface to the location of the large nozzle or monitor that was used to aim the water flow at the coal. As with the New Zealand experience the Soviet miners (at the Tyrganskie-Uklony mine) initially blasted the coal with explosives to weaken it with a high density of cracks, before applying the water. However the miners found that not only did the water double production (to 600 tons/shift) the streams were powerful enough that it wasn’t necessary to pre-blast the coal. The nozzle diameters of the time were up to 2-inches in diameter, and could throw a jet up to 60 ft.

Figure 6. Early Soviet underground coal miner

It was from these small beginnings that hydraulic mining began, it was, in its time the most productive method of mining gold in California, and was used for many years around the world for mining coal, and other minerals. But that again is a subject for more detailed discussion at a later time.

The combination of explosives and water power remains in use in harder rocks, particularly in South Africa in the gold mines. Here again the seams of gold are very narrow and can slope or dip at a steep grade, the working area is thus kept very cramped and difficult to work. By blasting the ore with explosive, it can again be moved with water pressure, although there is an additional advantage to water here that I will further explain when I write about cleaning rust from plates.

Gold, as is shown by the way it can be collected in flumes, is very heavy, and part of the problem in the South African mines is that small pieces can get trapped in small pockets on the floor of the seam. The higher pressure water flows can flush out these pockets driving the gold particles down to a common collection point. In that particular the practices haven’t changed that much in three thousand years.

Gold Rush Alaska - a gentle cough about shaker tables

I do not normally yell at my TV, nor usually want to throw things at it. However I almost reached that point of reaction on Sunday. I had been told about the show “Gold Rush – Alaska” by a couple of folks in my old Department, including a graduate advisee, but had not thought to go and look for it. But setting up to exercise I ran through the channels and an episode had just started. So I watched it as I worked out. (And it did get me more exercised than usual). In this particular episode they were having problems with their Wave table. (Which looks as though it works in a similar way to a table of similar appearance called a Wilfey table, and which I have recently used, as I will show in a photo added to the bottom of this post).

It became very clear, early in the episode, that the folks running it had no clue as to what it was supposed to do, and after a demonstration (which I am not sure wasn’t faked for the camera) where it failed to separate gold from the material run through it, the table was shut down. So I want to explain in very simple detail how the table is supposed to work, and show a photo, albeit with lead rather than gold ore – but that was what we were looking at when I last ran it). This is not rocket science.


The table starts out as a flat surface, onto which a number of thin strips of wood are attached. (Modern ones are of molded plastic ) The strips taper a little as they move along the table. The table is mounted so that it can be tilted in two dimensions, what I will call down the table, and along the table. And for the explanation I am going to use sketches initially.

Schematic of the basic components of a Wifley table.

The crushed ore feeds onto the table in a slurry and water sprays along the top edge of the table are set to give enough fluid to allow the vibration of the table (caused by some sort of eccentric cam resting on the underside) to provide a partial buoyancy to the particles, as well as helping with separation. The combined action of the water flow and the vibration help to move the crushed material both down and along the table, until it hits the top bar (or riffle).

The vibrating action helps to lift the lighter and smaller particles so that they float over this riffle, but the denser valuable particles are not lifted enough. (Remember Archimedes) Instead these then move along the feed edge surface of the riffle and table. If the riffles are of reducing height along the table this means that at some point intermediate weight ores can be separated from the lightest, (which run almost straight down the table) since although initially confined they can lift over a lower barrier. They are also separated from the heaviest ore (gold or lead), which remains confined by the riffles and thus runs down the far end of the table. Smaller particles of the heavier material that get over the top riffle do not have as much water on the lower riffles, and thus become trapped and fed over to the collection stream at the end of the riffles, but lower down the table.



The adjustments to the table are made so that the slope is enough, and the water flow enough, so that the mineral to be collected does not have enough buoyancy from the water and table action to get over all the riffles. It therefore collects at the far end, while the waste material is carried over the riffles then down and off the table. If the table is tilted too steeply, or the flow of water is too high, then even the heaviest particles will be swept over the riffles. (Which was one of the things they wre doing wrong). On the other hand, as is noted below, gold is not a rich ore and so there will be a lot of material swept over the table for very few ounces of recovery.

That is the basic principle, and by more careful adjustment it is possible to separate a mixture of different minerals into separate streams, as I just mentioned, as the particle move across and down the table, and these can be collected at different points along the bottom of the table (ours has holes in the table that feed to collection buckets).

There are different forms of table, based on this initial concept. An initial Google search showed this one at an on-line tutoring site

There is a video of a table working here, and one that, as with the second illustration uses groves that the heavier ore can’t escape from, here using a gold sample.

At the end of writing this rant I did find the Web site where the table makers respond to the Gold Rush Alaska video. They comment (in part)

Mike happened to be up there at John Schnabel’s . . . . so the two of them went over to the Hoffman’s site. Mike adjusted the table, ran a sample that Dorsey had, and got a gold line. It was filmed and will hopefully be shown on the next episode.

here’s our answers to what they did wrong…..

They destabilized the table by taking the slab out of the ground and loading it onto a floor jack.

All the raising and lowering of the table was wrong. Once the material is screened properly, you find the correct height adjustment and leave it there. Dorsey almost had it running, and then it was sabotaged. (Ed note that was the first thing I noticed).

No classification – large flakes should have never even been on the table (according to Dorsey’s blog, it probably was not even on the table).

The wave table does not make gold, it recovers gold. . . . .From the onset, their desperation (and script acting) caused mistake after mistake. No professional miner would work this way. There’s definitely gold on this property, but 30 buckets of concentrate and only 2½ oz of gold total! Wrong area to work……

Sadly it is often shows such as this that lead folk to believe that technology is some form of black art, whereas with just a little more accuracy and demonstration it could have been shown to be a very valuable tool.

The ore that we were processing the other month was a lead ore, and the table was set up just to show that we had liberated the galena, so that it was not tuned to give the separation right on the edge of the riffles (I had too much dip along the table) but you can clearly see, in the photo below how the heaviests parts of the ore had been carried to the edge of the riffles, and the galena (the silver stream) is clearly separated from the rest of the minerals.



It has sadly been my experience that folk often spend large amounts of money on equipment (in my field usually pumps etc) but fail to focus their investment and knowledge on the critical aspect of the entire operation that determines whether or not it works. In my case this is the small nozzle at the end of the delivery line that controls the jet that comes from the pump, (which because it wears out is usually of a poor quality, because they are cheap) - in this case the entire operation was centered around the use of the table to achieve the final separation of the gold. But without that running properly the entire investment was threatened. (But then, as a thought, if it all worked properly maybe there wouldn't have been enough drama to justify the series - tsk, tsk, what a cynic!)