• To participate in the 911Metallurgist Forums, be sure to JOINLOGIN
  • Use Add New Topic to ask a New Question/Discussion about Thickening, Filtering or Tailings and Water.
  • OR Select a Topic that Interests you.
  • Use Add Reply = to Reply/Participate in a Topic/Discussion (most frequent).
    Using Add Reply allows you to Attach Images or PDF files and provide a more complete input.
  • Use Add Comment = to comment on someone else’s Reply in an already active Topic/Discussion.

Silica Sand Dewatering (14 replies and 2 comments)

Tony Verdeschi
12 months ago
Tony Verdeschi 12 months ago

Silica Sand Dewatering: Cyclostack VS Fish Tail Cyclone

I am looking for experiences in sand dewatering. Any experiences with the cyclo-stack or similar fish tail cyclones? The should be able to dewater to +80% , increasing solids of the underflow.

Even though I am running on 30% solids, I haven't been able to increase more than 66% solids on the underflow.

The cyclone is D6 running at 6 psi. It the pressure relevant?

The product is silica sand and kaolin. We want to dewater the sand for an attrition stage. The mineral is 50% sand with a distribution of 80% over 150 mesh, which is the cutpoint of the cyclone.

I was planning to change the Vortex for a smaller one to increase pressure. Might that work?

Please let me know if someone has used this system. My intention is for feeding the attrition stage.

David Kano
12 months ago
David Kano 12 months ago

Dewatering sand to a particular dryness fraction by fishtailed cyclone (or separator cyclone) is a function of the feed properties, average particle size and solids SG contained in the underflow stream as well as the setup of the pumping and return system.

  • dryness can be achieved given favorable conditions.

Can you give me the solids feed rate and concentration and the feed size distribution so we can discuss further. Also given that you're feeding an attrition cell, I'd like to sight a sketch of the (intended) flow sheet. 

Victor Bergman
12 months ago
Victor Bergman 12 months ago

In hydrocyclone operation, the apex/spigot controls the density and size of the underflow product (the vortex finder diameter and length control the particle size reporting to the overflow, feed density and pressure control the efficiency of separation), so by choking down on the apex/spigot a denser (less fluids) underflow will result. The ability to achieve 80% solids is dependent on your feed and other parameters.

This concept haas been in use for many years, and does work well, but it also needs a fair amount of monitoring as performance degrades quickly as the choke wears. By adding the choke you will impact other operating conditions, in particularly the efficiency of separation (you will get a coarser overflow). If you are looking for a low % solids, high efficiency, simple to operate simple system, look at spiral classifiers.

Helena Russell
12 months ago
Helena Russell 12 months ago

It depends on how dry you want to get the sand. The "duckbill" cyclones are essentially a roping cyclone. For sand the underflow will typically be about 80% solids, almost dry enough to convey. Most operators mount these on a tower to make a pile and then allow the sand to drain further.

Screw classifiers will produce sand at about the same % solids and are probably a lot more forgiving in operation, especially with coarse sand (+150 microns). Dewatering screens will get a little drier, about 15% moisture, depending on grain size.

As with all roping cyclones, the risk is that if they are overloaded with underflow some may be forced out the top. They have a tonnage capacity, and it's relatively small for the size of the cyclone.

Most attrition scrubbers working on silica sand are fed at 70 to 80% solids, so I suppose a duckbill cyclone could be used to feed one.

If you need drier sand and don't have room for a drain pile or drain bin, there are options like belt filters that will get it down to about 5 to 10% moisture. Beyond that you will need some kind of thermal drying.

Helena Russell
12 months ago

Also, dewatering cyclones work best when there is a bi-modal size distribution of the incoming sand, with most of it well to the coarse side of the D50 split and some or none well to the fine side of the D50 split, and little incoming sand near the D50 size.

Bill Fraser
12 months ago
Bill Fraser 12 months ago

In a fishtail cyclone the siphon vacuum determines the underflow moisture content. The apex size is not a factor. Typically, with silica sand, an underflow density of 72 to 75% solids is achievable

A couple of points:

  1. Fishtail cyclone feed slurry needs to be relatively dilute - some manufacturers state a maximum of 25% solids by weight (for silica)
  2. The fishtail cyclone must be mounted high enough for the siphon to develop in the overflow pipe - typically overflow must discharge at an elevation 6 ft below the underflow discharge
  3. Fishtail cyclones generally operate a lower pressures than "standard" cyclones. Some USA manufacturers will stupulate 1 bar /15psi as the reccomended operating pressure
John Koenig
12 months ago
John Koenig 12 months ago

By "duckbill", I assume you mean fishtail cyclones (a fishtail is an underflow regulator). Cyclones with fishtails are not are roping. If a cyclone is roping it is working outside its boundaries and is behaving no better than a T-piece. 

  • Solids is difficult to achieve for a cyclone without an underflow regulator. Based on particle size alone, the P50 at the apex would need to be around 600µm for silica at 2.65 sg to achieve 75%w/w (and allows for a 5% contingency against choking).

I don’t doubt that in some isolated cases you could achieve slightly higher results, but my example is a result of hundreds of data sets.
A fishtail cyclone separator is an assembly which produces a controlled result. And as with all equipment, it has to be set up correctly to work consistently. A cyclone with fishtail will definitely produce a slightly less sharper and coarser separation. But once again, this is predictable and I can model the product size distribution given the inputs.
I agree with your comments on the tower and advice on dewatering screens.

The apex size is a factor as its ration with the vortex finder is critical to the operation of a fishtail.
I agree with your other remarks.

Apart from DMC, there are many providers of dewatering screens which do the job as well. 

Helena Russell
12 months ago
Helena Russell 12 months ago

Here we call them duckbill cyclones. Fishtail would be the same thing. An example, probably the first one, would be the Linatex Dewatering Cyclone. Others have tried to copy it with varying degrees of success.

I agree that the overflow vacuum is the most important control as that is what holds the duckbill at the apex closed and influences the underflow % solids.

They do have many of the properties of a roping cyclone in that the duckbill on the apex intentionally limits apex flow to increase underflow % solids. They have to be fed low % solids and correspondingly low solids feed rates so that the solids don't overwhelm the duckbill and truly act as a roping cyclone by sending oversize out the overflow.

The weight of accumulated sand in the apex is what pushes the duckbill open against overflow vacuum and mechanical resistance and allows the sand out of the apex.

They are limited in dewatering by the interstitial water in the sand. The volume of underflow that is not sand will be water. That's why they are limited to about 75 to 80% solids in sand. Any drier and some of the water has to be displaced by air, thus a drain bin, drain pile, or other supplementary dewatering method.

Also if there is fine material like clay in the feed, this material goes where the water goes. Some will wind up in the underflow with the interstitial water.

We've used them successfully in foundry sand and frac sand plants followed by drain piles or drain bins. They don't dewater well enough for coarser (-300 micron) iron ore concentrate that will be shipped any distance.

Finally, translating Minnesotan to the rest of the English speaking world, apex = spigot, duckbill = fishtail (we have lots of both around here), we spell labor, color, and harbor without a u, analyze with a z, and a hot dish = a casserole. It's not any better or worse, just different.

John Koenig
12 months ago

I agree with your comment 110%

John Koenig
12 months ago
John Koenig 12 months ago

Some attrition will occur in the cyclone, so to advise you effectively I will need to know the feed conditions, 

  • Particle size distribution, 
  • Clay / contaminant % 
  • sg 
  • solids feed / process rate 
  • cut point 
  • pump circuit 
  • return sump size 
  • lift 
  • siphon arrangement 
  • and what type of pressure gauge you have at the cyclone inlet

essentially as much info as possible and you will not regret your background work.

I can suggest changes / adjustments but without knowing the whole story, it will be hit and miss. 

Victor Bergman
12 months ago
Victor Bergman 12 months ago

You stated that your feed was 30% solids at 6 psi feeding a D6, but you did not mention your feed rate. As SC Chow said a normal feed pressure would be around 9-10 psi. Changing the vortex will give you a finer cut putting more into the underflow but other factors are also important. Even running with a "fish tail" the actual size of the apex on the cyclone is important. Also how much material you are feeding is important, and is there enough coarse to actually give you the denser underflow based on the D50 for the cyclone.

Helena Russell
12 months ago
Helena Russell 12 months ago

Here a syphon is the device that shoots pressurized water into a drink. Siphon is the thing that draws water over a high point by gravity. There are thousands of words spelled differently between America and Britain, or even Canada.

Alan Carter
12 months ago
Alan Carter 12 months ago

Why are you tied to a cyclone? Yes, increasing the inlet pressure will indeed increase the percent solids in the underflow, however it also increases the solids weight to the overflow. Does your choice of a cyclone have to do with equipment footprint? Frankly in my opinion you should be looking at a spiral classifier. Properly operated it will give you a higher density, less slimes and will take variation's in feed rate much better than a cyclone.

David Kano
12 months ago
David Kano 12 months ago

This is all that we know:

Target is 80%cw underflow but currently achieving ≤ 66% underflow (solids by mass).

Feed is 30% solids (by mass) and the cyclone is a D6 cyclone running at 6 psi.

So what we can tell from the cryptic information above is that a 150mm dia Krebs cyclone operating at 6psi would be handling a volume flow of roughly about 14m3/hr given I have no idea of vortex finder size.

Then at that flow rate and 30%w/w, the mass flow rate of solids must only be 4.2tph based on an sg of 2.65 for silica.

I guess that this is why an existing cyclone is better than forking out for a new classifier. Also, given that we have no idea of the size distribution or target D50 - why bother with arguing about equipment. We are being told a small part of the story and the questions we are asking are being ignored.

The original questions:

  • It the pressure relevant?

Answer: Of course and a fishtail works best at between 11 and 14 PSI and don’t forget the syphon control requirements already mentioned above.

  • I was planning to change the Vortex for a smaller one to increase pressure. Might that work? Answer: Yes, that might work but the apex cannot be more than 70% of the vortex finder diameter and remember that the apex has to be large enough to handle the production rate.

12 months ago
JohnnyD 12 months ago

I have the experience of working with a spiral classifier before entering the attrition cells. The spiral fed with hi solids the attrition cell and it ( the A.cell) worked perfectly.
Our units are 4 octogonal Wemco manufacture of 800mm.

We want to avoid installing the spiral because of space restrictions, but from my past experience is that is how I know that with hi solids the attrition cell will do the work.

The fishtail "sells" 80% solids on the underflow, but up to now we have only achieved 67%.

We will try the following:

  • increase pressure to 10 psi by adding water to the pump that feeds the cyclone.
  • change the vortex, also to increase pressure.
  • try dispersants.

If we are not able to increase solids so that we have an efficient attrition, we'll have to add a dewatering screen. But, that is the last alternative. We would rather make the fishtail works. It has been sold by Krebs engineers so we must deplete alternatives before giving it back. There have been many comments of experiences that achieve high solid content in this discussion.

The distribution is 50% sand 50% kaolin clay.

The sand is 100% below 2mm. The cut point intended is 140 mesh. Below 140 mesh there is a 20% of the sand. So then, 80% of the sand we want to dewater is above 140 mesh and balance of 20% is above 325 mesh.
The distribution in between 2mm and 140 mesh is variable depending on the mineral being processed and the ore natural variations.

The solution intended has to work for with out attention from the operator and needs to be practical and to work with different feed distributions.

The vortex is 2,5" and 1,25" apex. The inlet is 2,2sq inches. The cyclone is intended to remove the clay from the sand.

Helena Russell
12 months ago
Helena Russell 12 months ago

The one control you are not talking about is vacuum at the overflow. This has a direct influence on underflow solids. You must have an imitation if this isn't a control option. The original Linatex has a duckbill on the overflow too, along with a vacuum regulator.

Weir Cavex now owns this line.


David Kano
12 months ago
David Kano 12 months ago

Actually, in the original Linatex Australia days the set up included a cyclone separator featuring an overflow box which sandwiched the vortex finder plate between it and the tangential entry feed box. The overflow box also facilitated a syphon standpipe with elbow and a gate valve to regulate the syphon effect and the oflo pipe came straight out of the periphery of the box. I know the UK version (and possibly the US) had the syphon valve arrangement coming off the o/f standpipe as shown here. 

A duck bill at the end of the oflo pipe was taboo in Australia until about 1990 although I knew of installations within Australia. Every Linatex subsidiary around the globe had their own set up based on a common theme because they started off as concessionaires (type of agencies) until eventually bought by Harrison's and Crosfield (Australia I believe was the first acquisition whereas Camberley, UK, and Malaysia made up the original company).

The new CEO of the group brought us all together to nut out a common design for pumps first, hence the conceptual birth of Linapump in 1990, and then other products followed although the variations of the cyclone were too difficult to commonise at that time. So the volute entry cyclone was eventually designed to compete with Krebs in NAm but for mineral processing first and not applied to dewatering or sand washing.

The oflo pipe was submerged back into the pump feed sump which effected the syphon and closed the loop for simple circuits. To achieve the desired syphon, the bottom of the fishtail had to be at least 6'0" above the surface of the sump water level.

A pressure gauge was always fitted to the inlet branch to assist commissioning and after market performance and wear assessments.

 I estimate an average agglomerated particle size of around 60 mesh to the coarse fraction.
Given this, it will be challenging to achieve 80%w/w without second stage dewatering by screen.
Conservatively, 70%w/w is understandable. 

Would you be able to confirm the volume feed rate or the TPH process rate through the attrition stage.
I know you said that 50% of the solids is clay, but does this mean that from my previous calcs the total dry feed rate is 4.2 TPH or 8.4 TPH? I imagine that some of the clay would be liberated and the proportion of which would report with the water fractions. 

I plugged in some of the variables and now a 6" cyclone seems over-sized for the duty at 4.2 TPH. I agree with you that more water certainly wouldn't hurt.

Please join and login to participate and leave a comment.