An optimized smoke generator by Geoff Dixon

Click on any image to expand

My latest model of HMS Lord Nelson has twin funnels and therefore needs two smoke generators. I have one commercial unit that works very well and draws 3 amps albeit the smoke oil is not cheap to purchase. With twin units the combined current consumption on 12 volts would be in the region of 6 amps which in turn dictates a large battery for a reasonable duration.

I decided to do some research on smoke generators on the web and was surprised to learn that commercial items are readily available. They are used in film and stage productions and by the fire brigade for training purposes. Further investigation revealed these units do not use smoke oil at all, and in fact technically do not produce smoke. They do however produce large volumes of fog and use a special non-flammable fog fluid that is made up of de-mineralised water, glycol/glycerin in various proportions.

I believe this fluid was invented about 10/12 years ago, and prior to that all smoke was produced using kerosene and oil based concoctions, which were both unpleasant and flammable. I think it may be from this that the common perception that oil needs to be used which has lead to all the home made options I have seen various attempts to make smoke generators using three in one oil/baby oil/diesel etc as a fuel source which has made amateur smoke generators a poor source of smoke.

Oil based devices also all leave a residue on the model and the element and can produce noxious by-products and most importantly catch fire if the oil is overheated.

All of my previous efforts have relied upon some kind of wick with one end immersed in oil and the other end with a spiral of nichrome wire wound around the wick, however the problems with such a system is control with the element either running too hot or too cold and the whole unit suffering from hydrocarbon deposits from the oil building up on the element which reduces its heat output, and causes eventual failure, plus the other disadvantages mentioned above.

Nichrome wire (sometimes called resistance wire) is an alloy of Nickel and Chrome, and has a uniform resistance, and is found in many household products such as toasters, fan heaters, bathroom radiant heaters, electric fires, hair-dryers etc. Nichrome wire has two key virtues, firstly it is almost 100% efficient in converting electrical energy into heat, and secondly provided the temperature is not too high it will not oxidize (burn out) in air. If it gets hotter than the bright red associated with an electric fire then you are getting the temperature too high and the element will begin to oxidize.

Because it is used in so many appliances it is cheap and readily available, albeit there are many different specifications of wire. A good source is broken hair dryers from boot sales as usually the element has broken in one place only. Dependent on the make of hair dryer the wire is wound in a tight coil (like a light bulb) which in turn is wound around a central ceramic core, and there are therefore many feet of wire to play with and it costs just a few pence.

The only disadvantage of nichrome wire is that it cannot be soft soldered due to the operating temperature and must therefore be crimped or silver soldered. All my connections use very small diameter brass tube that can be crimped easily.

Fog fluid is readily available from any theatrical or disco shop countrywide. I purchased 5 liters of fog fluid from Maplin Electronics in Southend for about 15/20 which seems to be the going rate. Ed note: In North America, one source would be Corona Integrated Technologies, for others do a Google search.  This was also the smallest amount I could purchase. The fluid is categorized as non-hazardous and safe for public use with glycol being one of the most commonly used chemicals in industry. You can even purchase scent to add to the fluid to improve the smell if you wish!

Now that we have a non-flammable fluid and a hot wire to play with, the basics of an effective and reliable smoke generator should be available. My first attempt used a bare " U" shaped section of copper wire (old electrical cable) with about 5 inches of nichrome wire crimped to the top end. One stem was then wrapped in two layers of fire proof cloth and the wire wound down from the top in a spiral and tied off at the bottom, again being crimped to a small length of brass tube. A small screw cable joiner (stripped of the plastic cover) was used on the free end, and in turn joined to a further length of copper wire.- (loft insulation is too absorbent to use as a wick, but a plug of insulation in the bottom end promotes a good flow to the wick proper). Experiments showed that various thicknesses of "wick" would produce more or less smoke, depending on how absorbent, and how hot the nichrome wire. The shorter the wire, the hotter but the higher the current consumption.

There should now be two parallel "U" shaped sections of copper wire which are in turn joined to a double insulated cable block (see picture attached). It is important to ensure the main cables do not touch, as otherwise there would be a direct short with the battery. It is also important to have both copper wires bare and immersed in the fluid as heat travels from the element down the wire and the heat which would otherwise be wasted is used to pre heat the fluid.

The unit is immersed in a bath of fog fluid to about half its depth and the battery connected up via a switch. On Lord Nelson I used an automotive relay to handle the current and a small push-on push-off micro switch operated by a servo. I also use a 5 amp in line fuse for protection. The fireproof cloth soaks up the fog fluid which is then vaporised by the nichrome element producing a good column of thick white smoke which is easily enough for a model boat. As an experiment a further layer of fireproof cloth was wrapped around the element and the whole enclosed in a thin brass tube about 6mm diameter. This set up will draw between 2 and 3 amps on a 12 volt battery dependent on the length of the nichrome wire and this modification produced a better smoke stream for the same current consumption. Fireproof cloth can be obtained from a model engineering company as boiler insulation, or possible from a hardware store as fiberglass rope, for the seal round an enclosed household fire.

Further research into fog machines led me to understand a little more about the actual process. The heat atomises the water and glycol with the glycol condensing in the air forming billions of tiny molecular reflectors, which is why the smoke appears white. The finer the atomisation the more smoke is produced, and the longer it stays in the air. By enclosing the unit within a brass tube a little pressure builds up which shoots the "mist" upwards. The first inch should show a clear plume. Only when it mixes with the colder air does it condense and produce the telltale smoke stream. Fine atomisation also keeps any residue to a minimum.

I constructed a unit for use in Lord Nelson's second funnel and found it worked every bit as well as the commercial unit, but after about 20 mins smoke production reduced significantly until it stopped working. The problem was found to be linked to using two units because as the battery capacity dropped the unit which drew the most amps sucked the life out of the battery causing the other unit to stop. Two separate batteries would solve this but there wasn't room in my model. This is like trying to run two different sized motors on the same battery, sooner or later the larger one will dominate. I therefore decided to build a bigger but single unit along the same principals.

Further research showed the more the smoke was mixed with air the greater the volume of smoke. Commercial fog units advise for heavy smoke effect to inject a stream of compressed air directly into the heat exchanger. I experimented further and found if I blew into a thin tube parallel with the smoke stream the smoke volume was massively increased. I can now fill the spare bedroom in five minutes! I am pleased with this, my Wife and the cat less so!

A further enhancement was to introduce air into the smoke stream and this was done using a small computer cooling fan (50mm) which is ample.  The cooling fan should be wired in a way to ensure full voltage to the fan even as the heating element battery drops down in voltage, so perhaps a separate battery for the heating coil would worthwhile.  You may have to bleed off some of the excess cooling fan air to get the proper flow. In order to make this work the original concept was reverted to with an exposed element but with a double spiral of nichrome wire calculated to use about 4 to 5 amps wound round a thicker but shorter "wick". A series of square ducts was constructed using plastic card and the effects are as shown in the photographs. It proved necessary to insert some guide vanes between the fan and the heater unit as the spiral effect of the forced draft caused most of the smoke to exit one funnel only. Some further work is needed to get an exact balance, however the volume of smoke produced speaks for itself. There is also a small alloy heat shield above and to the side of the element to protect the plastic card from direct heat.  Future refinements would be replacement of the plastic ducts with allow as they showed some signs of melting from the heat.

As a bonus there is very little residue, with the funnel tops only being wet with water and no sign of any deposits on the masts or superstructure.

The only outstanding item appears to be longevity, which only time will tell, but at the cost of a few pence for the element this would seem to be the way to go. I also found it is best if the unit is tailored to a given battery, okay if its brand new, no problems, but if set up for an older battery, then a new one is used the curent consumption can increase causing overheating.

I have used 12 volts for the generator as its simply more efficient than six volts when drawing the same current. Power in watts = volts times amps. On 6 volts and 3 amps its 6x3 = 18 watts but on 12 volts its 36 watts, or double the available heat energy for the same current consumption.

All of my research has lead me to understand you cannot produce coloured fog as the dye residue coats the electric element/heat exchanger, however curiously Maplins do sell red or green fog fluid, and I wonder if mixed together brown smoke would result? Just a thought.

I have now used the generator on the pond and I have found the volume of smoke quite adequate even for fairly windy conditions where the smoke plume is still visible at a distance.

There does however seem to be a curios light effect depending on how the light hits the smoke. This can be seen from the attached photographs. In certain light conditions and angles the model can hardly be seen for the smoke!

So far longevity seems okay and I have used this generator for about three hours in total with no sign of failure. The battery seems to be lasting longer than the fluid which suggests I need to increase the supply in some way.

I think that none of the dimensions are critical, other than the length of nichrome wire which governs the amperage used, and there are numerous modifications that could be made to make the unit fit any model.

Back to Construction Articles


Version 1.1  04/04