ISECooker Construction

This construction manual is under continuous revision with new learning and improvements.  It was largely compiled during my 1-year trip to Africa/India/Nepal, which I document in this blog.  Please see our ISECooker Research Page.



Insulated Solar Electric Cooker (ISECooker) is an insulated electrically-heated cooking chamber, allowing the user to cook over a long period with low power from a solar panel, from the grid, or from a combination that includes any electrical sources.  Energy can be stored thermally or in a battery.  Low-power cooking saves money by reducing electrical bills (if grid connected) or reducing the number of solar panels purchased.  The simplicity of ISECooking technology allows it to be manufactured in locations where it is used.  This manual describes basic concepts behind ISECooking rather than a step-by-step process.  Construction details will differ in each location depending on availability of parts, resources, construction equipment and expertise.



Topics to cover:

1. What is 100 W?

2. Different designs for storing and delivering energy.

3. Thermal Conductivity

4. Different heating technologies

5. Drawing power from the solar panel

6. Insulation and construction materials

7. Safety and corrosion

8. Cooking

9. Developing a business collaboration

10. Instrumentation: Things you may want to buy



1) What is 100 W?

  • 100 W solar panels are presently manufactured in China for about $20, and retail for about $100 in the USA and $50 in Africa and Asia.
  • 100 W will bring 1 kg (1 liter) of water to a boil in an hour.  A 200 W solar panel will bring 1 kg of water to a boil in half the time.  A 1000 W microwave oven takes about a tenth as long: 6 minutes.  Hence, a full day of sunlight with a 100 W solar panel would boil about 5 kg of beans, rice, stew, potatoes,… enough for a reasonably sized family.
  • 100 W is not enough power to cook anything without insulation because the heat would escape, preventing the food from getting hot enough to cook.  Hence it is crucially important that the chamber be insulated.  The most important part to insulate is over the top of the pot – this is also the hardest part to insulate.
  • 100 W for 5 hours is 500 Wh or 1/2 a kWh.  Most car batteries are about 1 kWh.
  • 1/2 kWh would also raise the temperature of 10 kg of aluminum from room temperature to 250 C, providing an additional way to store a day’s worth of solar energy, for high-power cooking.



2) Different Designs for Storing and Delivering Energy



The above pictures show different designs, from left:
1) Heater connected directly to cook pot, providing the best thermal connection between the heater and the food.  It is simple to construct, but is hard to clean.  People do not want to use a pot that has wires connected to it.
2) Removable Pot with a “Heated Nest” is easier to use.
3) Thermal Storage can be added directly to the cook pot by making the base very thick (see below).
4) Even with thermal storage, food can slow cook in a separate container (green), especially if there is an insulating layer between the (green) cook pot and the thermal storage.


Direct Connect: A solar panel is directly connected to the heating element (section 4).  There is no energy storage except in the food itself.  It is the least expensive, easiest to build, and least likely to need maintenance.


Removeable Cook pot: In all ISECooker designs, the electric heater can be connected directly to the cook pot, which is simple to build, but hard to use and clean because there will be electric wires coming from the pot.  However, a removeable cook pot can be made by two methods:

  • Detachable electrodes: The heater is cemented to the bottom of the cook pot and the electrical connection is made via electrodes that make contact when the pot is placed in the insulated receptacle.  One sees this in popular teapots everywhere.  We built a detachable electrode into our prototype on Saturday, Dec. 24, as recorded on my blog.
  • Heated Nest (much easier to build): An ordinary cook pot can be heated by placing it into a stationary thermally conducting receptacle that is heated by a permanent electric heater.  Pictures below right are from an ISECooker made with Salma in early December.  What’s nice about it is the thick aluminium foil is on both sides of the heaters, effectively taking heat from both sides of the heaters to the cook pot.  We also made a very nice heated nest in Zambia with Clement end of February from a thin-walled (spun) aluminium pot.  We cut the sides so the aluminium cook pot would fit inside it.  At 100 W, there was only a 5 C difference in temperature between the nest and the pot, indicating good thermal connection between the heated nest and the cookpot.

The advantage of the heated nest is that the pot is lighter and simpler, and many different pots can be heated with the same heated nest.  Having the heater connected directly to the cook pot with detachable electrodes provides better thermal connection between the heater and the food, but is more difficult to build and maintain.


One very good way to make a heated nest is to buy two identical aluminium pots, as we did in Ethiopia and Nepal; below from left: cut vertical slits in the one we designate as the nest, smooth the cut edges, cut an annulus from a thin sheet of stainless steel (or make one out of think concrete), inserting the nest into the cooking surface, and the exposed electrodes with the nest and surface upside down.


An additional benefit of this design is that the nest makes a good barrier between the insulation and the food, so you don’t need to mold the cooking surface around the nest.  Rather, you can make a flat annulus (ring) to fit around the rim of the nest.  If there is concern that the slits in the nest will prevent a hermetic seal between the food and insulation, aluminium foil can be glued over the slits with high temperature RTV glue (gasket maker).


Thermal Storage Thermal energy can inexpensively be stored by heating metal (Solid Thermal Storage, STS) or melting a phase change material (PCM), but requires good insulation in order to prevent thermal loss.  We have experimented with several different kinds of thermal storage.


Heavy Aluminium Pot.  Aluminium is a good thermal conductor.  Preheating a heavy pot itself provides very high heat fluxes directly to the food.  We demonstrated this in Togo on Nov. 28 (see blog) when we heated a 5 kg aluminium pot to 250 C and boiled 1 kg of water in less than 3 seconds, corresponding to nearly 100 kW of power.  We subsequently heated the pot again and stir-fried 1 kg of vegetables about 2 minutes.  Both events are documented in this video.  Making an aluminium pot with a 5-kg, two-inch thick base is a major task in the USA, but in unindustrialized countries, every town I visited has aluminium foundries that cast pots from recycled aluminium in sand molds.  Thus, I’ve visited  foundries in Uganda (Sept. 14), Togo (Nov. 21 and again Dec. 1), Ghana (Dec. 16, and Dec 19), and Malawi: Lilongwe (Jan. 21) and Salima (Jan 27).


Detachable Aluminium Mass.  Having a heavy aluminium pot prevents one from quickly heating food in the morning (for instance), and also makes it difficult to heat something slowly.  One solution is to have two pots: one with thermal mass, and one regular pot.  Another solution is to have a thermal mass that can be moved into and out of thermal contact with the heater and the cook pot.  One device we built worked very well and is well documented in this final report.  The downside of this design is a much lower power flux caused by the heat needing to cross two metal-metal interfaces.  This design was able to achieve power fluxes of 600 – 700 W, less than 1/100 that of the “heavy aluminium pot” thermal storage design (above).  Thermal conductivity across the interface can be improved by placing a drop of oil in the interface.  However, this would require a substance that would remain liquid and also not burn over a temperature range of 25 C – 300 C.  If you know of a nonpoisonous substance that can do that, please inform us.

Phase Change Thermal Storage (PCTS).  More heat can be stored by melting a PCM than heating a solid.  Additionally, with PCTS, most of the heat is returned at the melting point, so the cooking can be more controlled.  The downside of PCTS is that you have a pot or container with a hot liquid in it.  We’ve mostly worked with two PCMs: Nitrate Salt mixtures (melting point 220 C – 300 C) and erythritol (melting point 118 C).  Sunbuckets has promoted their nitrate salt PCTS in moveable aluminium housing.  However, I’ve heard informally, that the housings can leak, and nitrates accelerate burning, so that’s a problem.  Matthew Alonso wrote his PhD thesis designing and evaluating Sunbuckets, comparing the nitrate salts to a solid piece of aluminium.  In the conclusion, if you read between the lines (which he personally pointed out to me) is that solid aluminium is better than nitrate salts.  We published our experiments with a stationary pot housing erythritol PCM.  It worked great, and I used it for 3 months in my house nearly every day.  However after 3 months, the erythritol degraded, with a reduced melting point and reduced latent heat.


We will pursue STS over PCTS for domestic use because of the safety and simplicity.  However, for industrial uses such as large restaurants and public facilities, nitrate salt PCTS would prove more effective and less expensive.  The larger scale would facilitate longer thermal retention times.  Nitrate salts are used to store thermal energy from concentrated solar fields for night-time electricity generation, providing significant precedent and guidance.


Grid Backup: The solar panels can be connected in parallel with any external DC voltage source of lower voltage than V_mppt of the solar panel.  If the voltage source has a higher voltage than V_mppt, it will prevent the solar panel from providing power, and can even drive current backwards through the solar panels.  One can buy a grid-connected DC power supply ($10-$20) of the same or lower voltage as V_mppt of the solar panel.  Alternatively, the DC could come from a charge controller connected to a battery.  Alternatively, an AC electric heater could also be used, requiring two different heaters for a solar / grid hybrid power supply.


Battery storage:  Batteries and the necessary charge controllers add considerably to the cost.  This cost, like that of solar panels continues to decrease.  In another 10 years, the battery storage systems may be very common.  If a solar panel is connected to an ISECooker and a PWM charge controller charging a battery, the ISECooker should use the electricity available in between the power pulses of the charge controller, using otherwise wasted electricity.  However, a full home solar electricity system could also operate where the ISECooker draws power through the charge controller from the battery, which is charged by solar panels.



3) Thermal Conductivity

In order to cook with low power, it is essential to insulate the cook pot.  Insulating a cooker is a foreign idea to every culture that I know… except India, where one can buy a “thermal rice cooker” that is an insulated pot into which you pour water and rice and allow to cook.  Most of us are accustomed to cooking with powers greater than a kW.  In particular, it is necessary to insulate above the cook pot, as heat rises.  It is also important to keep the heaters cool for two reasons:

  1. To reduce damage to the surrounding materials.
  2. To reduce heat lost to the outside, because the flow of heat is proportional to the temperature difference.

So, while we want to heat the food, it is helpful to instead think about cooling the heaters.  Thus, it is very important that there be good thermal conductivity between the heater and the food, and have good insulation between the cooking unit and the outside.  If the heaters are not thermally connected to the food, the heaters will rise to a high temperature and lose heat to the outside, through the insulation.
Below, please find a table of thermal conductivities of some common substances that we use.

Substance Thermal Conductivity (W/mK) Max temp (C) Specific Heat (J/gK) Vol Specific Heat: kJ/L*K Density/ water
Copper 385 1085 0.385 3.4496 8.96
Aluminum 250 660 0.897 2.422 2.7
Iron 80 1538 0.412 3.537 7.86
Steel 45 1425 0.466 3.756 8.05
Stainless Steel 15 1450 0.5 3.75 7.5
earth, stone 2 to 3 2 to 3
concrete 2 2.4
Sand 0.3 1.6 +/- 0.1
Fiber glass 0.04 1000
Wool 0.04 570
Polyethylene fiber 0.04 260
Styrofoam 0.035 ~ 100
Air 0.026


The low thermal conductivity of air does not mean that leaving a pot in open air will insulate it.  The convective flow of bulk air around a hot pot will quickly cool the pot.  The purpose of fibers in insulation such as fiberglass and wool is to stop the bulk flow of air, so that the heat must conduct through the air rather than convect.  It is actually the trapped air, not the fibers that insulate.  Thus packing more fibers of insulation will not improve the insulation.  The insulation will increase with increased distance between the hot and cold surfaces.  So, increasing the thickness of the insulated chamber will improve insulation.

Clement and I built a nice ISECooker in Zambia using the stuffing of pillows, costing about $2.50 for one ISECooker.  See finished product in section 6.  From left, inserting the receptacle into the polyethylene fibers; the thermal switch (180 C) and thermal fuse (220 C) are located beneath the top rim of the mortar receptacle, allowing the temperature of the aluminium pot (5 kg, with 5 cm thick base and flexible handles at right) to exceed 200 C (we think, but have not measured/confirmed); the section of electric heating element is glued into the nest with RTV glue (this is a bad idea, as the heater itself easily attains temperatures exceeding that of the RTV glue; thin handles made of steel cable can be casted into aluminium pots, providing handles that:
1) don’t burn your hands because they are thin and steel (especially stainless steel) is not a great thermal conductor.
2) are flexible, so they lie flat under the top insulator and spring up when the top is removed.

Clement demonstrated the thermal storage capability with this short video.


Keep in mind:

  • Heat will travel away from the cooker through the insulation and also through the supporting receptacle.  Thus, the structure that supports the cook pot, connecting it to the outer casing must be reasonably thin and reasonably nonconducting.  For instance, the supporting structure could be made from concrete or thin steel, but not aluminum.
  • gaps between metal surfaces will greatly reduce thermal conduction.



4) Heating Technologies


We’ve used three different kinds of heaters.  Resistive heaters, Diodes, and Positive Thermal Coefficient (PTC) Thermistors.

Resistive Heaters: mostly made of Nichol Chromium (NiCr) resistive wire.  The advantage is they are inexpensive and can be made from locally-available materials.  From left to right in the image below, we have made them from:

  • Sections of a heating element from an electric range or oven, cut with a hacksaw.  The resistance can be chosen by the length of the section taken.
  • Ceramic resistors available at electronics shops.
  • Section of NiCr wire embedded in concrete or strung through ceramic beads.  This is nicely described in the this complete construction manual from collaborator Alexis Zeigler.


Preventing corrosion at the heater wire junction.  An important challenge is that the NiCr wire gets very hot leading to corrosion where the NiCr wire is connected to the copper power leads.  We have sought to remedy this challenge in three different ways that all amount to providing increased connection between the NiCr wire and a steel electrode, connecting at the other (cooler) end of the electrode to copper power leads From left to right below:

  • Connecting the NiCr wire directly to high temperature, wire with high-temperature crimps.  Inside the crimp, the NiCr wire should be wrapped many times around the other wire or electrode.
  • Using a rivet that allows the NiCr wire to wind around the steel wire many times.
  • Collaborator ASEI spot welds a nail to the NiCr wire and stabilizing the connection in concrete, as shown in my blog on September 25.
  • Creating a low resistance electrode end on the NiCr wire by folding the wire over on itself twice, and twisting the four wire lengths into a single wire.


I cut an electric heater open and was surprised to see that the ~ 2mm diameter steel electrodes are not spot welded to the NiCr wire as I thought they were.  They are just wrapped around.  Also the steel electrode extends into the heating element much further than I expected… all the way to the coiled surface.


Ethiopia has an enormous cottage industry that has perfectly solved and tested this challenge… constructing stoves, especially mitads for making the traditional Ethiopian bread, injera.  While you can buy injera stoves made in large factories, they are also made in small shops by winding an electric coil through a maze carved into a ceramic disk.


The electrode connection is made by folding the coiled wire over on itself twice and tightly twisting the four wire sections with pliers.  It makes a perfect connection because it’s the same wire, and the electrode’s resistivity is less than that of the coil because it consists of four wire sections.  The pictures below show the process in this video.




The plates that hold the heater wire are etched into low strength machinable ceramic disks with a large compass-like tool.  However, in Nepal, the indentation is pressed directly into the soft clay before firing, resulting in a better surface and indents to better hold the heater wire.  At right are two 200 W heaters: One that is 8 Ohms for the ~ 36 V output of two (100 W) solar panels in series, and one is 250 Ohms for (~ 220 V) line voltage.


While trying to make the low-resistance electrode ends for the thinner wire, the wire repeatedly broke.  I was finally able to make an electrode for the thin wire by making an electrode with slightly thicker wire winding through one layer of thin wire.  In the pictures below, the thinner wire is in the lower right.  From left, the first folding; the second folding with the thin wire woven in; twisting the electrode tight, finished electrode with the thicker wire cut away.


Diode Heaters: Because diodes have a constant voltage drop of about 0.7 V (although this voltage drop decreases somewhat with increased diode temperature), the power produced is the product of this voltage drop and the current.  Thus the power scales linearly with the current rather than quadratically as with resistors.  Diodes can be purchased inexpensively when bought in large numbers.  The main advantage of diode heating is it draws maximum power from the solar panel under varying light intensities, as detailed in our Hot Diodes! Paper and in section #5 below.  The main disadvantage of diodes is that their maximum operating temperature is limited.  Most are rated for 150 C, but our experiments show they are stable up to temperatures of 270 C and they are limited not by current, but by temperature.  That is, if they are well thermally connected to a heat sink (like the food), then they can manage much (MUCH) more than their rated current.  Flat diodes are rated for more current and are much easier to heat sink than cylindrical diodes.


Positive Thermal Coefficient (PTC) Thermistors: Made from a ceramic substance with a temperature-dependent resistance that rises abruptly at a determined temperature, essentially shutting off the power.  They provide high power without allowing the temperature to exceed a certain threshold, and therefore provide both a heating source as well as temperature control, so no additional thermal monitoring is needed.  For instance PTC heaters rapidly raise the temperature of car seats to ~ 50 C and reduce power, maintaining the moderate temperature.  PTC heaters are available for temperatures as high as 270 C, which is high enough for ISECooking, even with thermal storage.  Another advantage of PTC heating is that the heater comes complete with high temperature wires that can be soldered to the power leads allowing the wire junction to remain relatively cool, avoiding corrosion problems.  The disadvantage of PTCs is that without good thermal connection to the food, the PTC temperature will rise to the transition temperature and turn off before the food is hot.   Additionally PTCs are not locally available in any African country I have visited, and must be imported, although they are not too expensive.  Lastly, a considerable number of PTCs have broken, for reasons that are not clear to us, but likely because they were forced to put out more than their rated power.  For details, please see our PTC report.


More than one heating type can be used on the same ISECooker, providing multiple functions.  For instance, a series circuit of diodes and PTCs provides temperature limitation and keeps the working voltage reasonably close to V_mppt.



5) Drawing Power from the Solar Panel


The amount of power drawn from a solar panel depends on the panel, the amount of sun, and the voltage at which the load (cooker) is drawing power.  Below are data I took from a solar panel on my roof in California.  With the knowledge that Power = Current * Voltage; P = IV, one sees several important characteristics:

  • Lower sunlight means less current (for fewer light photons exciting fewer electrons).  But the maximum voltage remains about the same for all solar intensities. **
  • The panels produce no power at their maximum voltage, because the current is zero.
  • The maximum power is produced at about V_mppt =  ~17.5 V (for these panels), for all solar intensities.  Thus, we always want to be drawing current from the solar panels at V_mppt for all light intensities.  Because the power curves are pretty flat on the top, one only needs to be close to the V_mppt to draw maximum power.  Also, it is better to err on the side of lower voltage than higher, because the power drops of much faster for voltages higher than V_mppt.


** It’s more complicated that that.  Lower solar intensities also slightly reduce V_mppt, and the maximum voltage.  However, increased panel temperature decreases voltage.  Because the panel is usually hotter when the sun is strong,  these two effects largely cancel; meaning that yes, V_mppt doesn’t change with solar intensities.


A resistance heater will draw the most power if the resistance (R = V/I; from V = IR) is chosen to draw power at V_mppt.  However, because V = IR, the voltage on the heater will drop with decreasing current (decreasing solar intensity).  The “working point” is the current and voltage that the load and solar panel are experiencing.  The working point is where the solar panel output curve intersects the (straight line, V = IR) resistance curve of the load.  Below, left, see three working points where the red resistance line crosses the solar panel output curve for three different solar intensities.  The maximum power point for 13:00, full sunlight is about 5.7 A, 17 V, corresponding to resistance of 3.0 Ohms, and power output of 97 W.  Thus, the red line represents a 3 Ohm resistor.  At 5:00 PM, not only does the current drop to about 2.2 A, but the voltage correspondingly falls to about 7 V, corresponding to a power of 16 W, although the power output would be close to 40 W if the working point was further to the right, at the same 17 V, requiring a resistance of 8 Ohms.


Choosing the R_mppt for one intensity, compromises power at the other solar intensities.  The resulting powers are listed for the different resistances and time of day, in the table below.

R(Ohm) 9:30 13:00 17:00
3 33 97 14
5.6 59 65 33
8 40 44 47


The panel can be made to optimally deliver power at all solar intensities with the following strategies.

  • Making the heater from several resistors or a variable resistor that can be changed in order to roughly maintain the solar panel output voltage near V_mppt.  The resistors could be changed manually or by a logic circuit.
  • Using a chain of diodes to heat the pot, as described in #4 above and extensively in our Hot Diodes! Paper.
  • Using a voltage converter to draw power at mppt and convert the voltage to meet the resistive load.
  • Providing a back up power supply for when sunlight is insufficient, maintaining V_mppt on the solar panel.



Changing Heater Resistance:

The current drops with lower sunlight, but V_mppt remains the same, increasing the resistance for maximum power.  For example, above left, the best working point for 5:00 PM would be (18 V, 2.2 A), requiring a resistance of about 8 Ohms.  Thus in order to provide optimal power between full sunlight, the resistance should be able to change from 3 Ohms to  8 Ohms.  This could be done with multiple resistors, but would require the user to change the connections.  Certainly this kind of changing could be done with an electronic circuit.



Chain of Diodes:

Unlike a resistor, the voltage on a diode depends very little on the current through the diode.  We can start with the simplification that if there is current running through the diode, the voltage on the diode is about 0.7 V.  To a very small extent, the voltage will increase with increasing current and decrease with increasing temperature, and is different for different diodes.  The voltage on a chain of diodes equals the number of diodes * 0.7 V.  For instance, the blue curve below left traces the IV data from a real diode chain heating an ISECooker.  Because of the nearly constant voltage, the working points for the diode chain are close to V_mppt for all solar intensities, resulting in about 100 W and 40 W for full sun and 5:00 PM.  There is a slightly reduced voltage and power in the afternoon because of the elevated temperature, because the morning started out with cool water and finished with boiling water.  Details in our Hot Diodes! Paper.


Buck Converter:

We can use a buck convertor to take current from the solar panel at V_mppt and deliver increased current and reduced voltage to the load at constant power.  By drawing less current from the solar panel, the solar panel output voltage increases to V_mppt.  For instance, below, the dotted blue line represents possible voltage conversions at constant power.  The black arrow illustrates the conversion that draws 40 W from the solar panel at V_mppt and converts it to the resistive load line, increasing the current and decreasing the voltage.


The voltage reduction ratio can be controlled with two possible feedback mechanisms:

  • Feedback power: A microprocessor can take a measurement of the power and continuously change the voltage reduction to maximum power.  This would be a MPPT controller.
  • Feedback solar panel output voltage.  A simple circuit can  decrease the buck convertor output voltage until the the solar panel output is at the target voltage.  This is less expensive, simpler and good enough, because we see above that as long as the solar panel output is close to V_mppt, the power will be very close to P_max.

We are attempting to use the cheaper, simpler solar panel voltage feedback.  Below is a picture of the circuit we tried (that kind of worked!), and Eric’s message to me indicating changes he made to the circuit that improved it.


Battery Backup or grid connection will improve ISEC in two ways:

  • It will provide back up power allowing cooking to be done in periods of weak sunlight.
  • The voltage of the solar panel and load are at least as high as the backup voltage, allowing the solar panel to deliver more power to the ISEC.

As an example, a solar panel with an MPPT of 18 V and 6 A with an ISECooker resistive heater of 3 Ohms, delivers 108 W under full sunlight.  A small 12.7 V battery can be charged with a PWM charge controller in parallel with the ISECooker, so that the ISECooker uses all the power that the PWM doesn’t use to charge the battery.  Under half sunlight, the panel COULD generate 54 W at V_mppt.  However, with a current of only 3A, the voltage across the 3 Ohm heater would drop to only 9 V, delivering 27 W of power from the solar panel.  If the battery (through the charge controller) is connected in parallel to the ISECooker (with a diode so the panel is not directly connected to the battery), the voltage on the ISECooker would be 12 V  (12.7 V – 0.7 V), delivering 4 A of current for a total of 48 W.  But the solar panel would be providing 3 A, and thus 36 W of the power (12 W more than with no battery backup).  The improvement in efficiency from the solar panel would be greater if using higher backup voltage (as long as the back up voltage is lower than the solar panel V_mppt).


BACK UP VOLTAGE SHOULD NEVER BE GREATER THAN V_mppt.  If the backup voltage is greater than V_mppt, the backup power supply will stop the current from the solar panel and possibly drive current backwards through the solar panel itself, heat the solar panels, possibly destroy the solar panels, and waste power from the backup power supply.



6) Insulation and Construction Materials


The heated cook pot needs to be suspended in insulation in a way that is secure and doesn’t conduct heat to the outer housing.  Additionally, the insulating chamber should be sealed to prevent insulation from getting in the food, and to prevent food, water and condensate from getting in the insulation.  We have experimented with two classes of suspension:  Supporting the cook pot from below and supporting it from the top cooking surface.

Supporting the cook pot from below, and sealing the top surface with a thin, insulating surface.  At Cal Poly, we have had success  using fiberglass fabric for the top cooking surface (and the bottom surface of the insulating top).  We have suspended the heater/cook pot inside of a cylinder of thin welded-wire mesh, which provides excellent insulation, but is likely not strong enough to hold a pot through robust stirring as with making nsima.  We’ve also supported the pot with fire brick, which is thinly-sliced (improving insulation) and packed between fiberglass wool under the pot to provide stability.  These are excellent methods except that fiberglass wool and fiberglass fabric is hard to clean and hard to find in the tropics.  In Togo, we were able to buy Teflon fabric, which is good to temperatures of about 200 C, but ultimately degrades.  Also, the fabrication of Teflon has considerable environmental impact.


Holding the cook pot in/on a strong surface that is supported on the rim of the outer housing.  This surface must also be reasonably thin and insulating to limit thermal loss.  This surface could be made from thin stainless steel, high temperature composites such as Bakelite (which is expensive), or a thin slab of concrete, plaster, or ceramic.


Concrete slabs must incorporate steel mesh, such as rebar in large concrete slabs.


Care must be taken to make the slab uniform thickness and keep the mesh near the center of the concrete.


We used plaster of Paris (Dec 16, 17, 19, 24, Ghana), and wall plaster (Jan 15, Malawi), below.


In Lusaka, Zambia, Clement and I used Concrete, and even made the bottom surface of the insulated lid from concrete.  Using two wash basins (one deeper than the other) the ISECooker closes like a clam shell.  We used polyethelene (melting temperature ~ 260 C) insulation from pillows (cost ~ $3.00 per ISECoooker), and we put a thermal switch (T = 180C) and thermal fuse (T = 220 C) on the underside of the concrete receptacle.  The maximum working temperature of the aluminium pot is determined by the distance between the aluminium pot and the thermal switch.  See section 7.


Clement demonstrated the thermal storage capability with this short video.


We’ve usually made the outer housing from plastic wash basins, although we got a large aluminium pot at a good price in Nepal.  The plastic wash basins are seen as ugly garbage cans, which is probably a deal-breaker for the consumer market.  We are also working with ceramics shops to build the entire receptacle surface and outer housing out of clay, producing a beautiful product. Clay may however be more expensive and heavy… not a problem for a stationary cooker… like in the USA, we don’t move our ranges around.  However, so far people have prioritized ability to move the cooker.  Because of the expertise required for making clay pots, and the time required for drying and firing, the other mesh-reinforced surface are better for prototyping new design ideas.  It may be sufficient to make just the top (annulus) cooking surface from clay (ceramics), but we haven’t tried that yet.


Thermal degradation of concrete: It’s widely accepted that concrete is not high-temperature stable – that it disintegrates at temperatures above 70 C.  However, we’ve not experienced this from either our concrete nor plaster receptacles in the Malawi vocational schools, where temperatures routinely exceed 300 C.  Likely, thermal disintegration is caused by differential expansion because of uneven heating.  We may be protected from thermal degradation by doing two things:
1) The steel mesh stabilizes against cracking.  Indeed Andrew at ASEI reports that without internal steel mesh, the receptacle surface disintegrates with a singe heating.
2) The receptacle surface is thin, reducing the temperature difference across the thickness.

Lastly, if thermal disintegration appears, it may be worth investigating improvements by using Firecrete Super Refractory Castable Cement, reputedly high-temperature cement.


Stainless Steel:
In Nepal, we bought a 4′ x 8′ sheet of .5 mm stainless steel (SS) for $36: enough for eight cookers (see blog, April 28) – $.50 per ISECooker.  Stainless is the most popular cooking surface and, with a thermal conductivity of 15 W/mK, a 0.5 mm SS surface will lose less heat than a surface made from 7 mm-thick concrete.  Because we had a full heated nest (pictures below), there was no need to mold the receptacle surface around a pot.  The top receptacle annulus surface was cut from the stainless steel sheet in a few minutes.  Below, notice from left (cook pot in heated nest), upside down pot in SS rim with protruding heater electrodes, directing the cutting of the SS sheet, assembling the prototype, and workings grinding down the cut edges.  Although the thin stainless is not as rigid as a 7 mm concrete slab, it is strong enough to hold the pot in place when the perimeter is screwed down to the aluminium pot rim.  Ultimately, we inserted a piece of ceramic (from the ceramic heater manufacturer) between the nest and bottom of the outer shell, to prevent the nest/STS/pot from sagging under its weight.


Final Design manufactured in Kathmandu

For details, see the blog, Friday, July 28, 2023


NYSE has finished manufacturing the ISECookers, in fine form.  Above, from left: The closed cooker, the cooker opened exposing the heated nest, the ceramic heater in the heated nest, the 7 liter cook pot, the 8 kg STS (solid thermal storage) on in the heated nest, the 4 liter cookpot in the STS in the heated nest, the 7 liter cook pot in the STS in the heated nest.  Because all the pots are made in the same shape, the pots are interchangeable.  However, the insulating lid cannot be closed when the large, 7 liter pot is heated with the thermal storage.

September of 2023, I received the data below from Ashutosh at Kathmandu University.

We are able to estimate the thermal power loss from the temperature rate of change of the 9 kg of aluminium in the ISECooker.  We see that we need 30 W to keep the ISECooker at the boiling point of water, and that given the ~ 120 W of input power, we will probably not achieve a temperature much higher than 300 C.  This could be good news if it limits the ISECooker temperature to less than the melting point of aluminium.



7) Safety and Corrosion


Don’t start a fire!  We can prevent a fire from starting by keeping all temperatures below the combustion temperatures of the surrounding materials.  Thus, there are two solutions:

  • Insulate the cook pot / heater with materials that don’t burn, such as fiberglass and rock wool.
  • Control the temperature of the cook pot.

The temperature of the cook pot full of wet food will never exceed 100 C, but the temperature of the heater could be considerably hotter, especially if the heater is not well thermally connected to the food.  Additionally, at some point, a user will leave an ISECooker on without food in it, so we can not rely on there to always be food in the cooker.  Apart from this, extra care should be taken to insulate around the heater as well as thermally connect the heater to the cook pot – for instance by encasing the heater with the cook pot within an aluminium sheet or an aluminium pot, we call a “heated nest”.


Fiberglass and rockwool are ideal high-temperature insulations, but are difficult to find in Africa.  Perlite is a locally available insulator, and wool is very good with its high combustion temperature.  Organic substances such as cotton, wood chips, paper, rice hulls, and straw are poor choices: they can burn, absorb water, and rot.  However, these substances could be used as part of the outer insulation shell, far away from the hot cooker, but are not recommended.


Several temperature switches and thermostats are available (see section 10) but be careful because DC current is hard on AC switches.  There is an inexpensive DC thermal switch available for different temperatures, the highest temperature switch turns off is 180 C and turns back on at 140 C.  In case the thermal switch fails in the “on” position, a thermal fuse of higher temperature can be added in series.  If you want temperatures higher than 180 C, the switch can be moved further away from the cook pot and heater.  For instance, the temperature switch can be placed on the opposite side of the supporting receptacle (see right), or even further away from the cook pot.  For instance, if you place a 180 C switch half way between the cook pot and outside shell it will turn off when the temperature at this point is 180 C, which should be about the average between the temperature of the cook pot and the ambient (T_a): We estimate the temperature of the cook pot then to be about:

T_cp ~ 2*T_s – T_a

So, if you use a 180C thermal switch in an area with an ambient temperature of 30 C, the cook pot will get to 330C before the thermal switch turns off.  The further the switch is from the cook pot, the hotter the cook pot will get before the switch turns off.  As we discuss above in “corrosion”, DC is hard on switch electrodes. You should be careful not to use AC switches and thermostats on with DC power supplies because the electrodes may burn out.  As an extra safety precaution, you can include a thermal fuse next to and in series with the switch.  The fuse should blow at a higher temperature than T_s, so that it only blows if the thermal switch breaks in the “on” position.


Be careful to confirm that any switch/fuse you use can handle the voltage and current of the circuit.


Toxicity of Recycled Aluminium:

When recycling aluminium from (for example) automobile engines, impurities may be introduced, which can leach out into food, especially if the food is acidic (such as tomatoes).  A 2017 study showed that food cooked in pots made in unindustrialized countries from recycled aluminium contained alarming amounts of aluminium, lead, cadmium, and arsenic.  A 2022 study indicates that leaching of heavy metals is greatly accelerated by both abrasive cleaning as well as higher temperatures.  An NPR story indicates 5 million deaths annually from lead- induced CVD (Cardiovascular Disease), and introduces Pure Earth, an organization dedicated to reduction of lead and mercury pollution.


Teflon makes poisonous vapors at high temperatures (> 300 C)

Many of us have used Teflon-coated fiberglass membrane to separate the metal heating cooking assembly from the insulation.  At temperatures greater than 300 C (572 F), toxic vapors are released that cause “polymer fume fever” which can be fatal.


More Difficulties with DC electricity:

  • Stopping DC current is harder on the electrodes in a DC switch.
  • DC voltage is more likely to corrode wire junctions.

If you were pushing a car to speed it up, the momentum would grow with the moderate force you provide.  If you needed to stop the car very quickly, you would need a very great force.  For instance, if you ran the car into a brick wall, the brick wall would stop the car quickly, but with a destructively large force. When you close a switch, completing a circuit, the voltage pushes the current to rise to I = V/R.  This current running through the wires has electrical momentum.  When you open the switch again, the electrical momentum ends in a short time, which you see as a high-voltage electrical arc, that degrades the electrode surfaces in the switch after many switches.  This is less of a problem for AC because AC voltage and current pass through zero 200 times per second, so the pulse and discharge ends by itself in 1-4 milliseconds.


Metals at a junction will corrode if there is a voltage difference between the metals.  This can happen if there is an applied voltage or if the metals are different, such as copper being connected to a NiCr wire.  The more positive metal will corrode.  This is less of a problem with AC circuits because the polarity of the junction constantly changes 50 times a second, so the applied voltage has no effect.  We find corrosion to be a particular problem at the junction between the NiCr heating element wire and the copper power leads… likely made worse because of the high temperature of this junction.  It is not recommended to solder NiCr to copper for two reasons:

  • Solder doesn’t stick to NiCr well.
  • The solder will certainly melt during operation.

Melting solder by itself is allowable in other situations.  For instance, if heating with a diode chain, a small amount of solder can be applied to the twisted wire junctions.  The solder will melt during operation but not run if only a small amount is used.  Care must be taken to not allow the molten solder to short the diodes, and the entire junction should be buried in high-temperature glue.


The corrosion problem is commonly solved in industry by spotwelding the NiCr wire to a steel wire that is long enough to be cool at the other end, to which the copper wire can be connected without corroding.  ASEI has been able to spot-weld the NiCr wire from electric stove heating elements to small finishing nails and then stabilize the junction in concrete; see blog for September 25.  Solutions to the corrosion problem are discussed in section 4 (heaters).  In particular, please note that in Ethiopia, they fold the heater over on itself at each end making a low resistance (and therefore cooler) electrode to connect to the power wires.



8) Cooking


What EVERYONE does WRONG: Your ISECooker will not function unless it is insulated.  The ISECooker’s insulated top must be on it ALL THE TIME.  Please check yourself: feel all surfaces.  If you feel warmth, this is where the ISECooker is losing heat and this is a problem.  You can open the top of your ISECooker only for a moment to put food in, take food out, or stir… just for a very short time.  People like to leave the top of the ISECooker open to watch the cooking.  If you leave the insulating top off the ISECooker, it will lose the heat and it will not work.  Most of the pictures on this website show the ISECookers open, to show the inside.  HOWEVER, ISECookers need to be closed when heating up and when cooking.  If it feels hot, then heat is being lost


Cooking without thermal storage is good for slow cooking, boil and simmer, and sautéing: You fill the ISECooker in the morning and the food is done in the evening.  If there’s no sun, you have to use a different cooking technology.  With some adaptation, you can do many different things.  If there is water in the ISEC, the temperature will not exceed 100 C.  So, if I want to brown my onions or mushrooms, then I sauté them first.  When they are caramelized, I add everything else.  At the low power of 100 W, cooking chicken and vegetables produces a lot of “soup”.  When I come home at the end of the day, I push the food to one side with a spatula and add a cup of dry rice.  It will cook in the hot ISECooker even if the sun is down.  This way, I have stew and rice.  And the rice is the best rice you’ve ever tasted.  Also, for reasons not clear to me, cooking dried beans in the direct connect ISECooker produces the best beans I’ve tasted, and at 100 W, unless they dry up, you’ll never burn them.



9) Developing a Business Collaboration


We started building ISECookers in 2015, but we still don’t have large scale, or even medium scale production.  There are minor challenges in both the technology development, the industrial production and the dissemination (adoption).  We must do everything we can to build a technology that is effective, inexpensive, convenient, beautiful, and understood.  It is NOT important that one person or shop does all the work by themselves.  And this tendency that “I can do it all myself” must be abandoned in order to bring all the benefits of a wide collaboration necessary to make the best product.  We are not in competition with each other, we are in competition with large multinational corporations outside of Africa.  We need to work together.


Business Development: There must be a factory, even a small factory as well as a team with deep academic understanding of electricity and thermal flow.  If you have had electricity/electronics training, this is NOT the same as having a PhD in electrical engineering or physics.  It is important that there be a university or engineering partner – at least until we have a clear step-by-step path to building a working ISECooker in your area.  Likewise, an academic team is not set up to produce ISECookers either in motivation or infrastructure.  Having an industrial partner is necessary.


Production: we should build a collaboration of every enterprise that is the best at what they do.  For instance if you need specialty pots, then find a collaborator who casts or spins pots, rather than developing that technology yourself.  Another example is if you want clay pots to be part of the ISECooker, you should partner with a clay studio.



10) Instrumentation: Things you may want to buy

Power Meters:
There are some specialty electronics that are helpful in building and using ISECookers.  While digital multimeters are necessary and can be used to measure power (P = I*V), an inexpensive power meter is available: $10 power meter (PZEM-031), that continuously provides voltage, current, power, and integrated energy over time.  They are placed between the solar panels (or other power source) and the ISECooker.  These are helpful in experimentally developing your ISECooker prototypes.  However, for only $10, it may be a good addition to all the ISECookers you sell.  The integrated energy (lower right) can be used to calculate the total CO2 over time displaced by the solar electric cooking.


Thermal Switches:
A thermal switch or thermostat and/or thermal fuse prevents the ISECooker from gaining excessively hot temperatures.  The cheapest functional thermal switch we’ve found (pictured at right) is for DC motors and has an “off temperature” as high as 180C.  The “on temperature” is 40C lower.  So, the one that turns off at (T_s) 180C turns back on again at 140C.  Lower temperature switches are available.


The use of thermal switches (above left) and thermal fuses (above middle) is discussed in section 7.



Measuring Temperature:
I like to use thermocouples because the temperature sensor is just a junction between two wires of different metal composition.  Thus, the probe itself is tiny and the wires run out to a reader that shows and/or records the temperature.  There are a number of hand-held thermocouple thermometers for as low as $20 on the market and data loggers for more than $100.  However, there are also several inexpensive ones I’ve seen (such as the TM-902C for about $2 through Alibaba), but never used.  It would be a good idea to try an inexpensive one to see if they can be used to display temperature on every device.


Some thermocouple readers take more than one thermocouple probe, allowing you to read/record temperatures in more than one place.  When using these, be careful to see if there is “cross talk” between the probes.  That is, if both probes are connected to the same piece of metal, does it affect the reading of either probe?  If there is “cross talk”, then you must make sure that the thermocouple probes are electrically isolated from each other.  One way to do this is to cover each probe (or all but one probe) with a thin coating of high-temperature glue.