The term „tank“ seems to be one of the central topics of control line flying. Probably quite rightly so, since our way of flying airplanes is absolutely unique and doesn’t exist anywhere else in aviation. The tank problem has accompanied our life since the very beginning. Maybe the problems have not been realized as such. We lived with what we had, and we understood the shortcomings as a fact which just had to be accepted.
Times have changed however, and clever minds have explored new technologies which have helped to overcome the basic problems. Actually I will never understand how we could live with those antiquated systems we used yesterday. Fig 1 shows the typical control line tank design which was used by manufacturers some 50 years ago.
Perhaps the problems didn’t show up so distinctly because at that time we used smaller engines, hence smaller tanks. Maybe we ran our engines full bore all the time, and we didn’t notice the difference between slightly lean and totally lean. This situation suddenly changed with the advent of slowly flying aerobatic airplanes and the requirement for constant engine performance. The “Uniflo”-system conquered the stunt scene. Fig A tries to explain the principle.
In A1 we have a completely closed container filled with a liquid almost to the top. Gravity works on surface height H , and as a result there is a strong flow out of liquid at the bottom of the container. In A2 we have much less surface height, and the strength of flow out is much less. In A3 we enter the Uniflo system. The big difference is that the vent tube doesn’t end at the top of the container, but instead goes down to the required height H ; in our case close to the bottom of the container. Height H is now the dimension ‘bottom end of vent tube’ to ‘outflow tube’. As liquid flows out, a vacuum is forming at the top of the container. This will allow further outflow only when air has flown through the vent tube. Thus the outflow speed is held constant . When the liquid surface uncovers the vent tube bottom end, the uniflo system ends and the tank continues to work in traditional manner. This principle is called the “Mariotte bottle” and is known since the late mid age. In our case the Mariotte bottle is simply rotated 90 degrees and is positioned horizontally. We simply replace gravity by Centrifugal Force.
  Fig 2 is an example of a basic uniflo tank. Please note that the overflow pipe always has to be CLOSED during the flight. Otherwise the uniflo system cannot work.  
Before we feverishly begin to bend tubes we should know about one basic principle. It doesn’t matter how the tubes bend and run inside and outside the tank. It’s only important to decide where the tubes begin and where they end. This is decisive for their respective function. Especially the outside end position largely depends on tank position, mounting system, and preferred kind of maintenance. Very often our tanks are filled some time before we actually start the engine. In the meantime the tank shouldn’t run dry of fuel. That’s why we bend the tubes in a way that this cannot
happen.Either the tubes inside the tankare bent to the tank top, or the outside end is higher than the tank top. Fig 3 shows this arrangement.
Fig 4 shows a tank built for a profile airplane. The uniflo pipe can be bent right over the fuselage , so the end is well inside of the inner tank wall ( in this context “inside” refers to the pilot’s side, whereas “outside” means away from the pilot ). That way we prevent fuel from being sucked out of the tank during flight. This can easily happen if the uniflo vent is not clearly inside the tank body.
Lugs help to mount the tank to the fuselage. Bend these lugs carefully. The indent of a sharp bend may cause the lug to break from vibrations of the model. It’s quite helpful to clamp a piece of plywood together with the lug into the vice, and only then begin bending. That way a small radius is produced. Fig 5 shows how a tank can be mounted in a full fuselage airplane with front and rear lugs bolted to the tank floor. The length of these lugs decides the vertical position of the tank and should be chosen from recent experience. Of course the sketch shows a tank which is installed from the fuselage bottom. The uniflo and overflow vents are bent that both of them leave the fuselage exactly at the separation line between fuselage and engine cowl. This method ( after having removed the the cowl ) allows quick and easy removal of the tank without any further operations.
Horizontally mounted engines tend to give us some headaches. Normally we’d expect to not find any problems, since we can easily set the tank centre at venturi centre. In reality however this may not work properly. From my experience ( and that of other stunt flyers ) it helps to mount the tank slightly higher than on centre line; sometimes even considerably.  
The reason seems to be the fuel line which has to go over or under the engine. As an example: in my INDIGO design the needle was planned to be at the fuselage bottom ( it shouldn’t stick out from the top of the fuselage nose ). In this arrangement the fuel line runs over the engine. But this method didn’t work. The fuel line had to go under the engine. A new venturi had to be made, using an OS spraybar with fuel nipple and needle on the same side ( see Fig 6 ). Now the system worked. Don’t ask me why !
Since many years I’ve noticed engine problems on profile airplanes. Here the needle usually sits at the top. Again, with the tank centre mounted at venturi level, there is a problem: a difference in engine speed at upright and inverted flight. In most cases mounting the tank higher will solve the problem. If this is not possible, a trick will help. A different way of running the fuel line can simulate a change in tank height even if the tank stays in the same place. For this purpose a special tank was made. The feed pipe leaves the tank at the rear. A short piece of tube is soldered to the front. Now these two tubes are connected, and the fuel line runs from the front tube to the engine. Changing the length of the rear fuel line will produce a shallower or wider bend. With the wider = higher bend the engine will behave as if the tank was mounted higher. Crazy but true ( Fig 7 ). If fuel filters are used they shouldn’t be installed at the highest ( lowest ) point of a fuel line bend. The additional fuel volume can make our fuel feed go crazy.

Fig 8 is a three view drawing of a typical stunt tank as I build them. Please note that the uniflo vent ends precisely in the outer most corner of the wedge, about 2 to 3 mm off the tip. In order to do this the feed pipe is slightly bent inside the tank to allow for this arrangement. The end of the feed pipe is still exactly in the corner of the wedge. That big tube in the centre of the tank is my way of mounting it. There’s a 6 mm blind mounting nut in the tank compartment floor. A 6 mm nylon bolt through the copper tube ( 6 mm inner diameter ) holds the tank.

If you don't believe in the necessity of a baffle, just forget it. Nobody has ever proved that we really need it. If you believe in it ( I do ), place it exactly at the point where the feed tube ends. We don't want to completely stop the fuel from flowing from one place to another ( that might keep it from entering the feed pipe ). We just want to restrict it's flow exactly where the feed pipe ends. Whatever small amount of fuel is shaked around in the tank at the end of a flight, it should be kept at the same spot where the feed tube ends. There's no need to make the baffle wider than shown in the sketch: a full tank doesn't need a baffle. The baffle cutout in the corner of the wedge should be slightly larger than the feed tube diameter. Tube bending outside the fuselage depends on your preferances and fuselage shape only.

People who don’t want to mess with soldering irons and flux need not renounce the uniflo system. The RC market offers a wide range of plastic tanks which can be built up to suit our needs. Usually some accessories are supplied in the package, and these - supplemented with some pieces from our scrap box - can easily make up a full function uniflo tank. Fig 9 shows a basic version where the uniflo vent is made from copper tubes, bent and installed to the principles mentioned above. The feed pipe is flexible silicone tube with a heavy clunk at the end. Arrangement of the vents depends on how the tank is installed, in which kind of airplane, and what system you prefer. Manufacturers recommend to run all pipes through the front cap with rubber plug ( they don’t know that control line exists and what we need ). Of course this is possible. In Fig 10 you can see this kind of plumbing. Only the overflow pipe is a metal tube, bent and fixed to the top of the tank. Uniflo and feed pipe are flexible fuel lines, tied close together, ending at the rear end of the tank. In this case silicone tubing connections to fuselage mounted metal tubes have to be arranged, depending on tank installation.
I fully understand that buying a plastic tank can save hours and hours of hard work. And if the final result can fulfil our needs there’s no reason to ignore this possibility. However what we can find on the market may not be the ultimate solution for our problems. I prefer to have a tank which completely fulfils my demands concerning size, dimensions, capacity, and system of a tank. Commercial plastic tanks don’t quite fulfil my needs. But I still use them for sport models or those less serious creations. The choice of size is limited. So you always have to use a tank bigger than necessary. This requires to redraw some fuel from the tank before each flight. I hate this. I don’t want to do any unnecessary action at the beginning of a flight when my concentration is aimed at the essential aspects of flying. Also, for reasons which concern flying characteristics, I want to have a fuselage nose as short as possible. With all those tubes coming out of the front of the tank cap ( plus some additional room for plumbing ) the tank compartment would be longer than desired. That’s why I don’t like this arrangement - and the plastic tanks in general !
If you still want to use one try to find one with an oval cross section. It’s as close as you can get to a wedge shaped tank.

  Deciding on the dimensions for a new tank is mainly based on experience from earlier airplanes and engines. Since the width of the tank compartment is already given by fuselage dimensions, we can only vary tank height and length. For tank height we have some limited freedom only ( we have to consider tank location relative to spraybar location ). What I usually do is: I make a provisional sketch of the tank cross section ( see Fig 11 , bottom sketch ).  
I calculate the front area, using dimensions H, w1 and w2. With a pocket calculator it’s easy to find out different tank capacities at different tank lengths L. If I’m not happy with the result I can change dimension H ( usually I try to avoid this ) or I change the depth of the wedge ( w2 ). When I’ve come up with the desired ( estimated !) tank volume I do an exact drawing of the tank cross section, to find out the length of the wedge sides ( dimension S ). Some people may be able to do this mathematically, but I can’t; so I have to do it via drawing. Now I have all the necessary dimensions to draw and cut out the tank mantle. The exact procedure how to build the tank is explained in the Workshop department .
If somebody tells you that Uniflo tanks are difficult to build - just don't believe him. Once you've understood the basic principle you'll be able to easily change engine running characteristics, reduce fuel milage, and confuse the competition by just adding some more scientifically approved tube bends.