This article describes the thought processes, and the construction of the Precision Marine Bullet skimmer. I have been very fortunate to be involved in this project, from its conception to the final production models. Hobbyists need to see and understand what actually goes into building a skimmer, and that is my reason for presenting this report. The construction materials, testing, and final optimization of the unit, adds up to a serious commitment of time, effort, and money. I must thank the owner of Precision Marine, Mike S. for his commitment to the project, for without his expertise in constructing these skimmers, and providing the raw material, this project would never have gone forward.

Evolution of a skimmer
Background
Back in August of '98, another avid hobbyist, James Wiseman, and I had been discussing our fish tanks. In particular, we had focused on the role of feeding large amounts of food in a closed system. One concern that kept arising was how our tanks would handle this nutrient load. The outcome of the discussion was that if we wanted to feed heavily, we would have to find a way to export the extra waste. One possible solution to this problem is to eliminate the extra nutrients via skimming (foam fractionation). Both of us had used traditional air-powered or counter-current skimmers and were disappointed by their lackluster performance. We thought a more efficient, large water-processing skimmer might be a good solution. Within a few days, James sketched a skimmer design outlining what we thought was a good compromise of compactness, utility, and performance. He had placed structural limits on the design of this skimmer that were dictated by our large tanks. As neither of us wanted an external skimmer, it had to fit under a 30" cabinet. To save under-cabinet space, we wanted the skimmer to reside in our sumps, which were less than 12" wide. A third consideration was that it had to have high turnover rates and water processing ability. Therefore, the skimmer had to have a configuration that would allow for long contact times, with fast and non-restricted flow-through.

The initial sketch outlined a step-shaped, tall rectangular box that had a divider running down the center (Fig 1). On the right side of the skimmer box was the water inlet where the air/water mixture entered the chamber. This mixture would run down the length of the right side to the bottom, where a small baffle directed the mixture to the top of the left side. On the top left side, we attached an interchangeable neck and collection cup. The total height of this skimmer was 18" on the right and 14" on the left. The skimmer was 9" long and 4" wide. The total distance that the processed water would travel was 28", which was equivalent to a 28" tall reaction chamber if the skimmer was vertical. This distance was more than sufficient for our purposes.

As an important aside, Chris Paris, a Ph.D. student working on wastewater processing, posted his ideas on how to increase the effectiveness of current skimmer technology. In his writings, he presented complex equations that described the fluctuating parameters that occur in an air/water mixture when one increases the amount of air mixed into a constant volume of water. Also, he discussed parameters to increase contact time between an air bubble and a waste molecule. The bottom line was that he presented a strong argument for the use of more efficient aeration devices (foam generators), as one way to create a skimmer with a high waste removal capability. We found Chris Paris's proposal of using the patented Beckett pond foaming head [See Explanation Box Below] as a possible aeration device, very intriguing. This was just the foam generator we needed for our new skimmer. The Beckett valve possessed many favorable features such as maximum air intake, extended water processing time and compact size, which we could exploit in our design.

Once we had decided on using a Beckett head as an aerator, the skimmer plans began to fall into place. Since neither James nor I had much experience at working with acrylic or building a skimmer, we decided to approach Mike at Precision Marine. Mike is a long time local Houston resident and an expert skimmer manufacturer. At the time, Precision Marine had many airstone driven counter-current skimmers (the PM-AP line), as well as recirculating counter-current traditional venturi skimmers (the PM-CV-line). These skimmers, and his product line, were well established in the aquarium trade. We approached Mike with the idea of a new skimmer utilizing the Beckett head, and presented him with our rough sketch and our list of limitations. Once Mike was sold on the idea (thank God for margaritas), he constructed the first working prototype based directly on James' sketch (Fig 2).

Mike's first step in building our skimmer was to create a screw-on containment chamber for the Beckett head (Fig 3). The modular design allowed for two features: 1) we could mount the Beckett head onto any of the test skimmers, and more importantly, 2) the containment chamber allowed a defined amount of air to enter the Beckett head. From the literature and from Chris Paris's writings, it was clear that a Beckett valve, which was unrestricted to air, sucked in air at rates that were detrimental to stable foam formation, since excess air resulted in large bubble formation, and violent water movement from the outlet.

As a way to reduce bubble size ("tighten the bubbles"), and provide a more consistent bubble output, we would have to restrict the amount of air entering the Beckett head. One way of restricting air is to create a surrounding (containment) chamber, which has only 1 air opening (Fig3, top panel). The top and bottom of the Beckett are sealed off to air, and only allow water to pass through (Fig3, lower panels). Air enters through a single side hole, in which a user-adjustable needle valve is attached. This needle valve controls the amount of air that enters the Beckett. This containment unit was affectionately labeled the Beckett 'bullet', due to its rounded shape (Fig 3). Once this containment chamber was built, it was attached to the original skimmer, and our first working prototype was tested.

Prototypes galore
As an initial test of the Beckett head (actually to prove to ourselves that this was the best foaming head), we decided to add the Beckett bullet on a skimmer system that we were quite familiar with. We attached the Beckett bullet to a downdraft skimmer; the downdraft design is a well-established skimmer technology and works efficiently with ETS's patented spray technology. In this first test, we sought to determine foaming (aeration) quality, and to measure water flow. Since it was unclear which pump would work the best with the Beckett bullet, we tested the bullet/downdraft combination with both a Rio 3100 powerhead (900gph) and a Mak4 pump (1200gph). Our first observation was that air intake was substantial, but with obvious differences between the two pumps. With the low-pressure pump (Rio3100), the bubbles were visibly bigger, there were fewer bubbles (the water was grayish), and the water exiting the bullet was violent. However, it did foam, but the foam was interrupted and unstable. Using the high performance pump (Mak4), we observed substantially better results. The bubbles were smaller in comparison ("tighter"), the aerated water was the color of milk (many more bubbles), and the water leaving the exit tube was fast, but calm. Interestingly, tests using both pumps resulted in foam formation within 10-30 minutes, but the foam climbed higher and was denser using the Mak4 pump. Since this was only a preliminary test, the Beckett bullet performed as good as we had anticipated.

This is when things got interesting. Over the next weeks, Mike built a few working models of the skimmer which we had outlined for him. These prototypes (Fig 4) were tested using the screw-on Beckett bullet powered by either the Rio 3100 pump or the Mak4 pump. Our first observations were, once again, that pump pressure differences were significant. It had become quite apparent that weak pumps would not make consistent bubble size, and this resulted in "burping" of the air/water mix. After we resolved the pump situation, the next issue appeared: the Beckett head was so efficient at aerating water, that air leaving the bullet rapidly expanded in the entry chamber. This resulted in pressurizing the reaction chamber and forcing all the water out. This problem was not observed during the downdraft test, but to our dismay was an inherent problem with our step-shaped reaction chamber. Fortunately, there was an easy fix.

We focused on efficiently moving air through the skimmer, and displaced the water baffle at the bottom center of the reaction box. Displacing the center baffle forced the air/water mix away from the downside, diverting it upward towards the collection cup. This simple fix allowed for the rapid passage of expanded air, and solved our pressurizing problem. This new design resulted in the angular step-shaped "Bigfoot" skimmer (Fig 4, 4th and 5th unit from left). The Bigfoot skimmer was an excellent test model. It was compact, contained over 30" of mixing distance, and allowed air and water to rapidly move through it. However, the curvature was not conducive to mass production of the unit. {Author's note: I personally like the look}. Using the information gleamed from the Bigfoot shape; we felt we were one step closer to a final design. We learned that a displaced baffle was required when using a direct injection scheme [see explanation box below]. Furthermore, water diverted from the baffle would foam directly adjacent to the water inlet. Mike made 4 additional models (Fig 4, farthest right unit) with differently shaped diverter baffles, such as circular diverters. This circular diverter allowed for uniform spread of the water mix (Fig 5). However, none of these uniquely shaped diverters prevented "burping" or moving the air/water mixture any more efficiently than the original displaced baffle design.

The Final Shape Arrives
The box-shaped mixing chamber was a natural evolution of the Bigfoot's curves, but without the requirement for elaborate shapes, which created problems for the mass production of the unit. Next, we required an area where the foam would accumulate and stabilize. Mike incorporated a clear cast acrylic reaction column, 6" tall by 6" wide, to replace the collection side of the Bigfoot. However, incorporating this column required the reaction box to be wider than we had hoped.

The final box measured 6" wide and 10" long. The inlet side where the Beckett bullet was attached was removed, and replaced by a single piece of PVC (Fig 6). The reaction column was fitted with a threaded flange, which permitted multiple sizes and shapes of collection cups to attach onto it. This final skimmer fulfilled our initial design criteria. It was compact, it would fit sideways into a 12" wide tank, and the total height was 24", including a 10" tall collection cup with the restrictor neck. Finally, after 12 prototypes, we had a final design. However, the skimmer was not complete without finalizing the perfect neck size, collection cup setup, and exhaust port. The exhaust port is critical in determining the volume of escaping air, in so much as an inadequately sized exhaust port may exert backpressure in the skimmer, reducing its efficiency. Unfortunately, the determination of the collection cup size and the exhaust port diameter was done empirically.

Mike constructed a series of necks and cups with varying heights, widths, and exhaust port sizes (Fig 7). Then, each neck was added to a test skimmer, and monitored to determine airflow, the neck's ability to support and produce dry foam,and the ability to project this foam over the neck. The final design was a compromise between sufficient airflow through the skimmer, and our height limitations. The first working models were built in ABS plastic and were tested on both of our tanks (Fig 8). As a test of each configuration's water processing ability, we measured and compared the volume of water going into the skimmer, versus the volume of water exiting the skimmer. Utilizing a Mak4 pump to power the Beckett bullet, we measured approximately 1200gph in and 500gph out. As a point of reference for comparison, we used a Mak4 pump to drive an MTC HSA1000 skimmer (another Beckett head skimmer), and noted it had an output of 450gph.

Image Legend

Fig 1. Cartoon of James Wiseman's sketch. 1)-water inlet; 2)-"downside" area of initial air/water mixing; 3)- mixing area "reaction area"; 4)- water diverter baffle; 5)-drain outlet; 6)-"upside" collection side where foam coalesces and stabilizes; 7)-interchangeable collection cup and neck.

Fig 2. A working prototype of the skimmer taken from James' sketch.

Fig 3. The "Bullet" containment chamber. Top panel: this is a completed containment chamber. Shown is the user adjustable air valve (white/blue) the Bullet chamber with single air intake hole. The bottom of the chamber faces left, and has a 1' screw thread. The water inlet (top of the Bullet) faces right. Lower panels: Left side, looking down into a Beckett Bullet (note the spike divider and its four diverters); Right side, looking up into the bottom of the Bullet (note the rounded surface of the ball).

Fig 4. A few of the skimmer prototypes. Note that all skimmers have a 1" screw-on adapter on the inlet (to attach the Beckett Bullet) and a universal flange with screw threads on the collection side which allows a quick interchange of any sized collection cup and neck.

Fig 5. Circular diverter.

Fig 6. A line up of different sized and shaped collection cups with restriction necks.

Fig 7. The down tube. Note that the Beckett Bullet threads onto the top.

Fig 8. The finished product, a working Bullet2 prototype.