Working With Principles for Creating More Efficient Cooking Stoves (Ethiopia)

In rural Africa it’s pretty hard to come by fuel sources which are a good alternative to wood. Gas is too expensive. Kerosene is also pretty expensive and quite a dirty fuel anyway. Wood also has the advantage that it burns quite slowly so that you can slow cook over a fire for several hours just feeding it occasionally. If you did this with gas it would cost a fortune. In a place like Ethiopia you can also have a local worker leave a valve open overnight and come down to find an empty canister in the morning. So like most people around here, we still cook with wood. One day we may get a biogas set up going and be able to produce our own methane gas on a continual basis (see ‘Biogas’ section at bottom of this article for an example). That would really be true sustainability but we have not yet reached a capacity to be able to develop that scale of infrastructure at this stage.

The problem we are faced with is that this area is getting badly deforested. Women trudge past our site on the road day in, day out, carrying piles of firewood back to their villages or to peddle at the market — literally carrying away the forest on their backs. It’s not healthy and it’s not sustainable. So as a permaculture demonstration site in this area, what have we got to showcase as an alternative? Well, there are two strategies we employ: 1) get as much utility as possible out of any unit of wood fuel – so reducing the overall consumption, 2) use the parts of the plant/tree which can re-grow quickly (coppice) and hence can be harvested at a sustainable rate.

Since we are an eco-lodge we have a kitchen and a restaurant which caters to up to 10 – 15 guests at a sitting and we also have around 10 resident staff eating three meals per day. So we are cooking approx. 30 to 100 meals per day.

Our early kitchen facilities, due to budget constraints, were severely limited. We basically had the local norm which was a small tin-roofed shack with the cooking done in what I call “cave man style” — i.e. three stones, fire in the middle and a pot on top. There was a hole in the roof to let smoke out. This was a terrible situation, but, basically, that’s what people do here — sit on the floor with a face full of smoke for half the day. The other half they carry wood around.

During the first half of 2010 we were able to scrape together enough pennies to start building a new kitchen, which by July had a roof and a concrete floor of 5.5m by 4.5m. At this point we had a plan to install a range and were putting in a concrete base for that purpose. At that point Steve Cran came here to do a training and he took over and built a stove he said would save us 90% of our firewood. Later I observed that the women were not moving out of their blackened hovel to make use of their new stove in their spacious and airy new kitchen. Why were they were still doing it caveman style in the smoky little shack? I asked my wife. “That stove uses too much wood!” she snapped at me.

Instead of saving us wood, it transpired that the new stove was actually consuming more wood than we were already using. So I began to analyse the structure of the stove to see what was going on in there, and to read up on cook-stoves to figure out how to re-design and then rebuild it using the same construction techniques (cob-adobe) but paying more attention to the mechanics of the system.

Efficiency =
Total chemical energy stored in fuel
Total energy transferred into cooked food

The efficiency of a stove can be defined as:

In order to maximise the efficiency of a cook stove, there are two points of conversion that determine the efficiency – a) the efficiency of combustion of the fuel, i.e. converting the stored chemical energy into heat energy, and b) the transfer of that heat energy into the food itself.

To get efficient combustion which totally burns the fuel into CO2 + H20 and not CO (carbon monoxide) with unburned solid particles (smoke) we need: a) plentiful oxygen, and b) high temperature. To insure this situation we should ideally have 1) a fast and continuous air flow through the combusting fuel, and 2) an insulated combustion chamber.

The Aprovecho Design Principles for Wood Burning Cook Stoves Manual specifies, based on a wide body of research done over the years, that to insure a sustained and continuous air flow through the burn chamber we should have a continuous cross sectional area (XSA) for the air passage though the entire length of the stove. The aforementioned stove had a very large combustion chamber with a tapering floor, so that the XSA of the of the air passage narrowed down considerably along the air passage through the stove to the chimney. This situation creates a bottleneck, so that a lot of the air that has passed though the fire already ends up circulating around in the burn chamber, lowering the oxygen level, thus reducing the efficiency of the fire’s burn. Hence the stove was smoky and inefficient. When we rebuilt the stove we lowered the roof of the burn chamber to be parallel with the floor (i.e. rising from a low level at the front to a high level at the back). This gave a more constant XSA and hence a continuous air flow for a better burn. So our second stove was an improvement. But we can still do better.

The First Stove Design

Our Second Stove Design

The Aprovecho Design Principles for Wood Burning Cook Stoves Manual gives 10 principles for efficient cook stove design. I would like to just go over these and compare with our second stove design, before considering how we can improve them in our third design:

Applied in our stove?
1 Whenever possible, insulate around the fire using lightweight, heat-resistant materials.
2 Place an insulated short chimney right above the fire.
3 Heat and burn the tips of the sticks as they enter the fire.
4 High and low heat are determined by how many sticks are pushed into the fire.
5 Maintain a good fast draft through the burning fuel.
Could be better
6 Too little draft being pulled into the fire will result in smoke and excess charcoal.
Could be better
7 The opening into the fire, the size of the spaces within the stove through which hot air flows, and the chimney should all be about the same size. This is called maintaining constant XSA.
8 Use a grate under the fire.
9 Insulate the heat flow path.
10 Maximize heat transfer to the pot with properly sized gaps.

1) Insulating around the fire

To get efficient transfer of the released heat into the food there are several considerations to be made in the stove design:

  • Heat can be transferred by three means – conduction (through solids), convection (though movement of fluids) and radiation (through electromagnetic wave energy – like from the sun). A fire gives out heat energy by all three means.
  • Heat moves from high to low temperature, so a cold object always heats up in a warm environment and vice versa.
  • Materials have a “specific heat capacity” which denotes the physical ability of a material to absorb a given amount of heat energy in order to raise its temperature by 1°C.
  • Objects have a “thermal mass”, which is their mass multiplied by their specific heat capacity – denoting the amount of energy the object can absorb or release per 1°C change in temperature.
  • Insulation prevents or greatly slows down the transfer of heat between media of different temperatures.

Having the fire on a concrete base means the fire will heat up the concrete. The greater the mass of the concrete (thermal mass) the more heat energy it can absorb and hold. Hence if a fire is on a massive object or inside a massive chamber, built of big heavy clay bricks over a concrete slab, that structure will absorb large amounts of heat energy, even just to raise in temperature by a single degree. It will continue to do so as long as it is colder than the fire itself. So, if in order to heat up 2kg of food in our pot we also have to heat up 2000kg of cement and clay, then we are going to need a lot of energy for a small result.

This is inefficient.

Two solutions to this are: 1) make the stove out of very light materials, and 2) insulate the burn chamber so the stove structure does not heat up.

If the combustion chamber is insulated from the surrounding thermal mass, far more of the heat energy released will go into heating up food rather than the structure. This is a problem we did not overcome in the second version of our kitchen stove, so something to crack in the third design.

Suitable insulation materials are light, full of air and heat resistant (i.e. they don’t burn). Good examples are pumice, perlite, vermiculite and wood ash. Yes, wood ash. So that’s nice and easy.

When we insulate the burn chamber we stop the transfer of heat into the surrounding structure. This will mean a higher temperature, giving a better burn and hence better combustion efficiency, but also hotter exhaust gases and more convective transfer of heat into the air-stream passing through the fire, as well as more radiation of heat from the fire.

2) Place an insulated short chimney right above the fire

Hot air rises because it is less dense than cold air. This is how fluids (liquids and gases) transfer heat by bulk movement of their mass due to density changes. So with a vertical chimney directly above the fire we allow the hot air to move the way it naturally tends — i.e. upwards. This creates an up-draft which draws in fresh air flow from below. This insures that principles 5 (Maintain a good fast draft through the burning fuel) and 6 (Too little draft being pulled into the fire will result in smoke and excess charcoal) are fulfilled.

3) Heat and burn the tips of the sticks as they enter the fire

The point here is that the combusting part of the sticks should be hot while the rest of the stick is cold. Having wood at mid-range temperatures leads to smouldering and smoke creation. Hence the sticks need to be burned at the ends only and well inside the burn chamber, so any smoke will be drawn out in the exhaust gases and not enter the room.

4) High and low heat are determined by how many sticks are pushed into the fire

This means that the regulator of the burn speed is the number of sticks added, not for example by restricting air-flow which will lead to lower efficiency.

5 and 6) Dealt with above

7) The opening into the fire, the size of the spaces within the stove through which hot air flows, and the chimney should all be about the same size. This is called maintaining constant XSA.

Maintaining a constant XSA prevents the formation of bottlenecks in the air flow which can allow air to recirculate. Since air is very light it has a low thermal mass and hence cools quickly. Recirculating air thus cools fast and the air stream will be cooled down as a result. Worse than this is if the recirculated air can re-enter the burn chamber then it will lower the oxygen level of the air, inhibiting the combustion of the fuel and making the fire smoke.

8) Use a grate under the fire

As can be seen in the photo, we did include a grate in the design, but over the year and half since it was constructed the bars have become dislodged and it has stopped to function properly. A more robust grate should be included in the new design. We should also have the grate project further out of the fire chamber, since we have been having a problem with sticks falling out of the fire as their ends burn off. The sticks we use are quite long, since people do not tend to chop fire-wood short here. Hence a feeding shelf with a 30 – 50cm projection should be included in the new design.

9) Insulate the heat flow path

This is for the same reason as no 1 — to get more heat into the food and waste less thermal energy on warming up the stove structure.

10) Maximize heat transfer to the pot with properly sized gaps

The convective heat transfer into the pot can be maximised by: 1) having fast air flow, which rushes past the surface of the pot. removing any boundary layer so that there is always fresh hot air sweeping the surface of the pot, and 2) confining the air flow into spaces which make it pass right around the sides and bottom of the pot, covering the largest possible surface area for maximum heat transfer.

Note that heat could be transferred into the pot and hence the food by convection, conduction or radiation, depending on how we design the stove. But of the three, convective heating allows a design which fulfils the other principles best, hence giving us both good combustion efficiency and good transfer efficiency. Conductive heating requires the pot to be in contact with the burning materials themselves, which may be up to 800°C if our stove is working well, and will likely burn the food. On the other hand the presence of the pot itself right next to the burning material may disrupt air flow and reduce combustion efficiency. Radiation heat transfer may be effective, but also requires the pot to be close to the heat source, hence would preclude the presence of the chimney above the combustion chamber.

In our second stove design you will note the lowest pot is very close to the fire, allowing radiation as well as convection to heat the pot. However the second and third pots are heated much less efficiently due to loss of heat into the surrounding thermal mass which is un-insulated.

New Stove Design

With a vertical short chimney above the fire, and with the burn chamber and chimney insulated, more heat should reach the first pot despite less radiation transfer due to: a) better burn efficiency, and b) less loss of heat into the surrounding thermal mass. The temperature of the air flow should also fall off much less between the three pots if the air flow is properly insulated.

In the new stove design all three pots are at the same level. The key to ensuring effective air flow here is that the XSA is constant. We also have to ensure there is a tight fit of the pots into the stove rings. This may require us to get some better pots. Our current pots are the sort of dirt-cheap flimsy aluminium pots that everybody uses here and they are full of dents and deformations. If we can ideally get hold of a set of three steel pots we can make the stove rings of three corresponding sizes to fit exactly. Note that since it is a multi-ring stove we don’t want the hot air to escape around the sides of the pots but to continue on through the system to the next pot. We will add a metal pot-skirt to the rings using sheet steel to insure a tight fit, with only a 1mm or so gap between the pot and the skirt. The pot sits down inside the flow of the hot air from the fire, getting heated from the sides as well as the bottom. Small pot holders can be put on the floor of the air passage to hold the pots up a bit, allowing the air to pass underneath them.


Given that we are in the middle of rural south Ethiopia, we have a limited range of materials available to us. The Aprovecho Manual gives some good recipes for making insulated ceramic bricks with natural materials, but in this area good pottery clays are hard to come by and we also lack experience in firing bricks or the kind of equipment necessary to do it to a specific temperature. Hence we have opted to go with a design which incorporates materials that can be purchased from local building supplies or we can source from our site. We will use a steel frame for the structure which may make the stove have a shorter lifetime – 3-5 years, but hopefully by this time we can master the art of firing our own insulated ceramic bricks or source them from somewhere in Ethiopia.

The main parts of the stove structure are:

A) The fuel feeder, burn chamber and short chimney
B) The base block
C) The ceiling of the hot air passage, cooking rings and out-flow chimney

A) The fuel feeder, burn chamber and short chimney

We have available to us 4.5” (10.5cm) steel pipe, which we can use to form the fuel magazine, burn chamber and short chimney. The Aprovecho Manual recommends setting the length of the short chimney at three times the diameter of the air flow passage. Hence the chimney should be about 31.5cm high. If we allow another 10.5cm for the width of the horizontal inflow/fuel magazine we have a total height of 42cm for the elbow section, including burn chamber and short chimney. This whole section needs to be very well insulated so it will be set into a bed of wood ash. The fuel magazine will project out from the burn chamber by only 20cm. This does not need to be insulated as the heat won’t tend to escape out of the burn chamber in this direction due to the inward air flow. The feeding tray to hold the fuel must project out of the fuel magazine by another 40cm to hold the firewood and prevent long sticks from falling out if their ends burn down.

B) The Base Block

The base block is a massive structure which holds the cooking surface, the air, the combustion chamber and fuel magazine stable in their appropriate positions. It is set on top of a reinforced concrete platform (already constructed) which is 1.5m x 3m in dimensions with surface set 50cm above ground level in the middle of the kitchen. The platform is held aloft on two rows of cement blocks. On top of the platform the base block will be fixed to a height of 50cm, so that its surface is 1m above ground level. It will comprise a solid block of mud-brick. We have already produced the bricks ourselves. At the front end of the platform a cavity will be left in the brickwork with dimensions of 50cm high x 30cm x 30cm. This will be bounded by one layer of bricks at the front. The elbow section will be set into this so that the mouth of the fuel magazine opens out at the bottom centre of the front wall of the base block. The cavity will then be filled in around the elbow section with wood ash. The top tier of bricks will also be only set around the periphery of the block and the whole middle part of the top tier shall be comprised of a bed of wood-ash 10cm deep. This will be covered over by a layer of chicken wire and then rendered over with a cob-render mix to seal it in place.

C) The horizontal hot air passage, cooking rings and out-flow chimney

The ceiling of the hot air passage will be set above the base block to divert the outflow from the short chimney laterally along the hot air passage. To achieve this we will first set a low wall of bricks around the top edge of the base block. This wall will be two bricks high in the sections adjacent to the cooking rings, but only one brick high in the intervening sections.

Next, insulation will be set around the edges of the passage, using glass wool salvaged from an old boiler. This will be fixed in place using chicken wire nailed into the bricks. This will then be covered over with cob-render. We cannot use ash to insulate around the horizontal hot air flow passage due to its poor binding qualities. Glass wool is not an ideal material being industrially produced and also potentially hazardous to health, so we limit to minimum use. (Wear gloves and mask if handling this stuff.) Once it is set into the wall of the passage, the surrounding bricks are insulated. Now we will use a mould to set the cross section shape for the passage in between the cooking rings (Section R in the figure above). This will be a hollow metal tube shaped 5cm high by 17.3cm wide and with a length of 5cm. This is laid down into the passage and cob render, and chicken wire laid down over it. Once the render dries the mold can be removed. Now glass wool is laid down on top of the set cob before more chicken wire and then the steel reinforcement frame and finally more cob to top it off.

The hot rings (Section Q in the figure) will meanwhile comprise metal pot-skirts. These will be set into the ceiling. The pots fit snugly into them with very little room around the edges for hot air to escape into the kitchen. Before setting these pot skirts the walls of the horizontal passage are insulated and the width of the passage formed over the insulation with cob-render so that there is a 2.5cm gap around the pot. Knobs will be set into the passage floor to hold the pot 2.5cm off the ground. This assumes a pot of 30cm diameter which sits 15cm deep in the horizontal hot air-flow passage. The design should be modified according to the pot size to maintain a constant XSA.

With the floor of the passage insulated and set, the pot skirt can be fixed into the passage. Ideally the pot skirts will be welded or riveted onto the structural frame of steel re-bar. Then more cob will be added around the upper part of the pot skirt, sealing the roof for the horizontal passage and holding the pot skirt in place.

The long out-flow chimney will be fixed into the roof of the horizontal hot air-flow passage in a similar manner to the pot-skirts. It comprises a length of sheet metal rolled up to the same diameter as the short chimney. It will likely be insulated with glass wool, held in place by wire mesh and sealed over with cob.

The pot skirts will also have lids so that they can be closed when not in use. These lids will have to be large to fill the same volume as the pots themselves. These can be made using old pots, hammered into shape. They will be filled with insulation (ash) and the tops riveted and sealed shut so that they do not absorb and dissipate heat out of the system.

Finally, a finishing layer of lime plaster will be applied over the whole cob structure.

It is actually planned to build two units like the one described, which will need two out-flow chimneys (so that when both are in use there is a constant XSA. We also envision making further use of the hot air flowing out of the stove system and we will get more info on that to you in the next write up we do. We are planning on implementing this stove upgrade during our internship program which we will be running here in Konso, South Ethiopia in April-May 2013 just after our April PDC. If you want to learn how to practically implement systems like this, come along and help us out.