The use of gas stored as a liquid in a robust container is the most convenient way of firing a burner to raise steam in the boiler. Such gas can be purchased in cartridges and it can be decanted into another tank using the method that is widely used for filling lighters and gas torches. In section 3.1 I pointed out that that the gas tank is really a small boiler and heat is required if the liquid in the tank is to be evaporated continuously to form a gas. If the tank were to be heavily insulated from its surroundings required to evaporate liquid continuously the heat would come from the energy stored in the liquid and the pressure in the tank would fall until the pressure was atmospheric and the evaporation would stop. This is not practical. If the tank is not lagged heat could enter the liquid from the surroundings as a result of the temperature in the tank falling below that of the surroundings. Some steady temperature and pressure would be reached but the pressure would be low. One might decide to just accept whatever pressure happens to be in the tank and that the boat will move slowly. Modellers quickly feel the need to do something about heating the gas tank and perceive that a solution to the problem of heating the gas tank must be found. Heat must be supplied from somewhere if we want to maintain a useful working pressure at the burner.
The first thing to do is to get some idea of the rate of heating that is required. We can make a few calculations. One of my boats weighs 17 pounds and the steam plant drives it at a crisp walking speed so the engine is producing a useful power output. It runs for up to 30 minutes on one fill of gas. The capacity of the tank is 70 grams and, as I have never run out, I suppose that it uses about 60 grams. It would be possible to calculate the heating rate from physical data but we do not know with any confidence the ratio of the gases in the mixtures. Using a mean value of the latent heat of vaporisation of all the gases that might be used gives a required rate of heat input about 15 watts. This is what I have provided and it works very well. It must be the sort of heating rate required in most boats. It is not excessive and must be practical.
So what options are open to us?
1 Use of a big tank.
In some boats, for example a tug, the burner can be are supplied directly from the gas cartridge. This gives a generous surface area for heat transfer. If the tank can share a space with the boiler, sufficient heat can be transferred to the tank to maintain a good pressure. If, however, the tank is exposed to the atmosphere the temperature of the gas must fall to less than the temperature of the air for there to be any heat transfer at all so steaming in the winter is problematical if the air temperature is low. Starting is not so easy when the cartridge is nearly empty because there is too little liquid to give up energy to produce gas. There is no problem in the summer. It is a method that works well in the right boat. However it is not a solution for scale models of Windermere cruisers and the like where there is simply no space for a gas cartridge.
2 Use of a small gas tank and direct heating.
Where boats do not have sufficient room to accommodate a gas cartridge, gas is decanted into smaller purpose-built tanks. These do not have large surface areas and must be heated. This starts to become troublesome. An obvious way of heating the gas tank is to put it close to the boiler. Such an arrangement means that the amount of heating is unknown and the temperature of the gas and therefore its pressure could be quite high. This could be measured using a pressure gauge and a throttling valve in the gas line can be used as a crude pressure-reducing valve to reduce the pressure at the burner to about 20 psig. It is a practical arrangement but, in my view, the tank should be no closer than that required to maintain the required burner performance. An 80 psi gauge fitted to the gas line would be a great help in setting up such a system but care is needed as the gauge could easily be overloaded. I have seen this method used with a tank in the shape of a cuboid with one face against the back of the boiler. It must surely raise the pressure far beyond the necessary value even if the tank has been tested to 350 psi.
A source of heat that is potentially capable of being controlled more effectively is the exhaust steam. One can imagine a tank with a loop of copper tubing carrying exhaust steam round it or under it. I have not seen it done because my experience with other people’s boats is limited. Clearly it is a method that will respond to cut-and-try but, however it is fitted, provision to remove the tank from the boat during filling is desirable.
These methods all become troublesome during start-up on a cold day when no heating is available and the gas pressure drops so much that the burner may not even light.
3 Electric heating.
Heating the tank by electricity is quite feasible with this low rate of heating. If rechargeable batteries like nicads are used a heating rate of 15 watts for 30 minutes is possible. If we think of using a 6 volt battery (that is 5 nicads) the current required will be in excess of two amperes so we might expect the battery to give say 5 volts on load. The current would then be about 2.6 amps. The resistance of the heating coil would be about 2 ohms. The capacity of the battery would be sufficient to give 2.6 amps for 30 minutes ie 1.3 ampere hours. The sub C sized battery is quite capable of this capacity eg 1400-2000 mAh and would be adequate to heat the tank for a 25 minute run. It may be that the boat is large enough to carry a lead acid battery, which is a better option because of its low internal resistance. It would be nice to use lithium poly but the low voltage restriction adds another complication.
What are the characteristics of this heated tank? One of my gas tanks is in the bottom of the boat under the saloon. The temperature in that space must be dependent on both the temperature of the air and that of the water in the lake. However let us suppose that my tank is at 20°C and the temperature round it is 20°C and the burner is started and 15 watts of heating is supplied and cycled on and off to maintain a steady pressure in the gas tank. The gas temperature will remain steady and the burner will perform at a steady rate. But suppose that the temperature round the tank is 10°C and the tank starts off at 20°C. If the heating is only enough to evaporate the gas and not enough to meet the cooling that will take place to the air the temperature will gradually fall to a temperature above that of the air surrounding the tank and the burner will run at a lower rate. Conversely if we start at 20°C with an air temperature of 30°C the heating will meet the evaporation but there will be heat coming from the air as well so the temperature and the pressure of the gas will initially rise. However it will quickly fall to the heater setting. So providing just enough heating to keep the temperature steady at 20°C with an air temperature of 20°C will give satisfactory performance at higher temperatures but will not give a wholly satisfactory performance in cold weather. One must think of increasing the heating rate to say 25 watts and relying on the pressure switch to keep the pressure steady and perhaps accepting a somewhat shorter run in winter or even making little muffs for the ends of the tank.
This all means that a pressure switch has to be made. It is not too challenging as the Bourdon tube from a foot pump for car tyres can easily be adapted. In figure 29 the Bourdon tube is the ring in the centre. It is a normal oval tube that tends to straighten when pressure inside it increases One end is fixed to a vertical tube that is connected to the gas supply line to the burner. When the tube straightens its free end separates from the adjustable contact on the left. This contact cannot carry 2.5 amp so it must be unloaded by using a relay. Both sets of contacts require suppression condensers. The red object in the middle is an LED that indicates when the heater is on.
Figure 30 shows the heating system in the hull. The battery is on the right. The red band round the tank contains the coil and its insulation. (The lead weight is to trim the bulbous bow.)
On a practical matter the heating coil is best made from brass wire of say 0.020² diameter. The resistance per foot of brass wire is much lower than nichrome wire and a long length is needed. This long length gives lots of turns of the gas tank and spreads out the heating area. If the tank is first covered with thin robust paper the wire can be wound on to it and spaced with a thread wound on at the same time. When the wire is made off with terminals and insulated with thin polystyrene the coil never gets hot but still provides the necessary heat.
All this means that a reliable supply of gas is quite feasible for anyone to make.
I have already described the various modes of heat transfer in section 3.1. Now I have to look at this burner when it is inserted in the end of the lower flue.
In the standard Bunsen burner the gas is formed into a jet at the bottom of the upright tube and air, called the primary air, is drawn through two large holes on a moveable ring to mix with the gas to form an inflammable mixture. When this is lit at the top of the tube it burns with a flame that burns partly in the oxygen from the primary air and partly in oxygen from the free air surrounding the flame, the secondary air. The flame can be altered by adjusting the flow of primary air.
In the burner shown in figures 28 and 29 the flow of primary air is adjusted by moving the whole jet carrier to partially obstruct the air holes. As the flame produced by the burner is to be inside the lower flue some arrangement must be made to supply secondary air. The standard design has four small holes that seem to be very small. It looks as though the designer intended the user to adjust the primary air and the flow of gas to give a good flame. This neglects the possibility of fitting a ring with many holes to permit too much secondary air to enter the flue and then rotating the ring to restrict the flow as required. With this arrangement fitted the flow of gas, the primary air and the secondary air and, if necessary, the jet size, can all be adjusted to give a combination that is the best that can be managed in this flue. As far as I can see the only guide to the combustion is the sound of the flame but a start can be made by setting up the flame outside of the boiler. Using this burner and a form of burner control and the heating of the boiler is as effective as is likely to be possible for a miniature boiler.
If the design is based on carrying water in the boiler there must be some limit to the volume of water that can be used. Presumably there is no question that the boiler cannot be completely filled at the start because some space is needed to permit expansion of the water as it is warmed and to contain the steam when it is generated. There is a limit to the length of the sight gauge set by the practical constraints. Experience suggests that if the water level is visible in the sight gauge there will be adequate steam space.
It is not so easy to set a lower limit. In the presence of any water the boiler will not get excessively hot but if the water is boiled away completely the flame will heat the boiler. Nothing will melt but the lagging and the water gauge seals may suffer. The boiler should not be allowed to run dry. There will be other levels where the lower flue is only partially covered with water. This is not an efficient way of running the boiler but is not harmful to it. So I suppose a lower limit is the top surface of the lower flue and a little below. This gives a margin for misjudgement.
When Cheddar Models were in business they marketed a system of burner control. At a preset boiler pressure the system turned down the burner to simmer. The burner was turned on fully again when the pressure fell below the set value. The system used a thermistor directly in contact with the boiler to detect indirectly pressure change and, using an electronic unit running of the radio battery, switched a servo to operate a two-position valve. In one position the flow of gas was not restricted but, in the other position, the flow could be adjusted to make the burner simmer. It was a good unit and it completed the control system for the firing of a miniature boiler.
Figure 33 shows a Cheddar unit in my paddle steamerr. Note the pressure gauge reading pressure at the burner. It is repeated along with a gauge for steam pressure in one of the windows so that I can monitor both when the boat is running.