Introduction of

Rocket Stove Technologies

(Institutional stoves, Household stoves
and insulative refractory bricks)
Into Mozambique
March- July 2004
Draft Report

Prepared by Peter Scott

Prepared for ProBEC South
Submitted Oct 13th, 2004

Please note: the plans for the stoves described below are available on a limited
basis. Please contact ProBEC for more information
Material Outputs
ProBEC, in cooperation with World Food Program, has developed a new
institutional stove for Mozambique. The MangiMangi stove was first produced in
December of 2003. It was given the name
‘MangiMangi’ (which means ‘fast’ or right
away in Shona) by the residents of Manica
province. Initial response has been very
favorable. Users report that the stove is :

• Efficient: uses 80% less wood than

their open fire
• Fast: boils 40L of water in 30min
with 2kg of wood
• Clean: produces almost no visible
smoke
• Inexpensive (approx US$150) and
very low maintenance.

No smoke. No chimney!
In the last 6 months, 12 MangiMangi Institutional stoves have been produced
by SAVEPLA, our stove partners in Chomoio.

• 2 MangiMangi stoves were made for

Formigas del Futuro

One was the original 60 L MangiMangi shown

above. The second was the taller 80 L double
walled MangiMangi (see photo left). These
stoves have been used consistently since
December 2003. Although the
cement/vermiculite liners have had to be
replaced, the school has been very happy with
the stoves. The 60L stove was retrofitted with
clay/ sawdust bricks and delivered on Aug 31st
by Bacaiman and Zana. The 80L stove will
have to be used as is until the new bricks are
ready at Ceramica Villa Pery.

• Three (two 60L and one 80L)

MangiMangi Stoves were produced
and sold to ADPP in Chomoio.
These stoves were modified and
replaced on Aug 31st

• Four 60L MangiMangi stoves were

built for WFP. See following notes
in WFP stove section

• Two 60L MangiMangi stoves were

produced and sold to Concern
International. These stoves have not
yet been picked up. These stoves
were made with vermiculite and
cement, a combination that we no
longer recommend. In October, when
the clay sawdust bricks are ready, the combustion chambers will be
reconstructed with the new ceramic bricks and High Temperature
Zimbabwe cement.
 • One MangiMangi stove was produced and sold to Cubitsirana
• They have used the stove for 1-2 months and are very happy with it. They
also report an 80% fuel saving compared to their traditional open fire. This
stove will also need to be retrofitted with sawdust/clay bricks as soon as
possible

Since its introduction in December 2003, a number

of improvements have been made to the stove:

• Fixed wood supports -to support longer

pieces of wood - have been added
• Triangular brackets have been attached to
the 3 legs to increase stability
• The combustion chamber has been
changed from vermiculite /Portland cement
bricks to locally made sawdust/clay bricks
lined with High Temp Zimbabwe (HTZ)
Cement/ sand mortar
• The stove still uses a small amount of
vermiculite/cement to insulate the top plate.
More experimentation is needed to develop
a vermiculite free insulating top plate.

Stoves for WFP

4 MangiMangi stoves were originally made for WFP.

• One stove was sent to Sofala in Dec of 2003. According to WFP reports
they liked the stove for cooking rice but were not using it for cooking
Nsima. The explanation given was that the stove was not stable enough
for cooking Nsima. The stove should be picked up, retrofitted with stove
stabilizers and with clay/sawdust bricks that are being produced at
Ceramica Villa Pery.

• One stove was sent to Gondola School. Atanasio Augusto visited the
school two weeks after delivery and reported that the stove was working
perfectly; that it was able to cook nsima as well as rice; and that the cook
was happy with the stove. This is stove #3

• One stove still remains -as of August 31s t - at WFP Beira

headquarters. This stove was damaged in transit and/or was not deemed
suitable for cooking and so was never placed in a school. This stove
 should also be picked up, retrofitted with stove stabilizers and
clay/sawdust bricks

• One stove still remains at Savepla. This stove is awaiting new

clay/sawdust bricks. Because there has been some confusion due to
stoves being moved around so much, Savepla might need to be
convinced that one of the stoves at their workshop belongs to WFP. I feel
that there is consensus between WFP and myself that WFP still has one
stove at Savepla.

Five stoves have been numbered (with an arc welder) to ensure better
tracking of the stoves. When referring to bricks, the numbers and letters refer
to the recipe used to make the sawdust/clay bricks. See attached excel sheet for
specific recipes.

Stove #1
This stove is for Formigas del Futuro. It was built with
lower grade clay/sawdust bricks. The liner bricks (with
the exception of the two entrance bricks - ‘E’ on the left
and ‘C’ on the right) were made with common brick
clay last year. Although I am skeptical about the
durability of these bricks, I thought it would be helpful
to make one test stove using these lower quality bricks.
If these bricks are successful, than it will greatly
simplify production.

The bricks are mortared together with 3 sand: 1 HTZ

Cement: 1 water

This stove was to be delivered on Aug 31st.

Stove #2
Note: the chalk mark says 2*1, but the arc weld
says #2. 2*1 refers to the mortar mixture

This stove was collected from ADPP on Aug 22nd.

This was one of the 3 stoves that were sold to ADPP
for 10 million Meticais, (approx US$420) in March
2004. As can be seen from the photo, the metal body
had sustained some heat damage from the stove
being used after the cement vermiculite bricks had
worn away . The stove generally suffered from cook
and student abuse. The stove was brought back to
Savepla in August and reconstructed with the higher
quality clay/sawdust bricks. It was mortared together with a 2 sand: 1 HTZ
cement: and .75 water mixture. The top plate was filled with the
cement/vermiculite insulation. The stove used:

Floor bricks: “5” Bricks

Bottom Sidewalls: clockwise from left to right - E, B, B, 3 8.
Above stove entrance: Brick 11
Middle course: Bricks A and B.
Top course: Bricks 31,32 (clay/sawdust/cement bricks, very fragile)

Stove #3

One MangiMangi stove was originally sent to Gondola,

Manica in December 2003. The stove was used
consistently for cooking rice but not for Nsima. The
reason for the stove not being used is uncertain. Lynn
Miller and Paola, from the Chomoio office, said that the
stove could not be used to cook Nsima because it was
not stable enough. The cook, on the other hand, told
Atanasio Augusto that the only problem was that the
stove was too short.

To address both of these concerns we built a step for

the cook and added stove stabilizers. It appears that
these new changes have addressed the height and
stability issues as WFP is satisfied with the stove and
wants to order 10 stoves from SAVEPLA

Stove #3 was mortared with a 3 sand: 1 HTZ Cement

: .75- 1 water
Stove #4
Savepla originally produced this stove as part of the
three stoves purchased by ADPP. Unfortunately the
stove diameter was too small to fit ADPP’s pots. The
stove sat unused for three months until I came back.
The stove was picked up on Aug 20th . and retrofitted
with ceramic bricks.. This stove is designed for the
taller aluminum 80 L pots that ADPP and WFP both
use.

The bricks were mortared together with 5 HTZ cement:

2 water. The bricks were
lined with a mixture of 2
sand: 1 HTZ Cement: .75
water.

Stove #4 is the single

walled version of the double walled stove (see photo
right) that was originally made for Formigas del
Futuro. More testing is necessary to compare these
two stoves but for the moment I would recommend
using the single walled stove as it is cheaper and
probably only slightly less efficient.

Stove #5
This Stove was originally produced for
WFP and sent to Gondola.
Unfortunately it was used before the
mortar had dried which caused it to
crack. When stove #3 was delivered,
this stove was picked up and returned to
Savepla on my last day in Mozambique.
Instructions were left to retrofit the
stove, Mark it with a #5 and then deliver
it to ADPP on the 10th of September
during Zana’s first visit back to
Chomoio.

Instructions were left to mortar the stove

with a 1 sand: 1HTZ Cement: .75-
1water mixture.
MangiMangi Construction
General notes

• The bricks should be lightly moistened on any surface that will receive
mortar. Too much water and the stove wont dry for months and too little
water will cause cracking in the mortar. Ideally the stove should be left to
dry in the shade for two days. If the bricks are moistened properly than the
bricks should remain moist, without additional water, for the first 24 hours.
If the bricks become dry in the first 24 hours then a small amount of water
should be applied.

• The recommended material for the institutional stoves in

Mozambique is now the clay/sawdust brick from Ceramica Villa
Pery. These bricks should be mortared together with sand and the HTZ
Cement. The actual proportions of these materials will become clearer in
the next few months as we collect more data from our test stoves.

• We need to make some tests using a local high temp mortar but for now,
the HTZ Cement is a suitable mortar.

• The top plate will still need to be covered with the cement vermiculite
mixture. Experiments should be made with mixing the HTZ Cement and
sawdust to see if we can use that as a suitable replacement for the
vermiculite.

Way Forward for WFP

WFP should order stoves to fit specific pots (i.e. if a school has 2 60L pots and 1
80L pot then they should order at least one 60L stove and one single walled 80L
stove. This is important. The stoves will be much more stable and efficient, if the
pots fit the stoves exactly.

WFP has expressed interest in having a couple of brick MangiMangi prototypes

built so they can compare that stove design with the portable metal version. This
can be done in January, upon my return, or earlier if Zana can be mobilized to
build the stove.

Way Forward for Savepla

• We need to link SAVEPLA with a larger partner. They suffer from a lack of
resources, for purchasing materials and equipment, as well as access to
transportation. A more established partner could overcome some of these
hurdles and allow them to expand their business.
 • In the short term ProBEC should liaise with ProCIP to assist with
transportation. At present they need assistance to collect bricks from
Ceramica Villa Pery, and to bring materials to the bread oven construction
site (e.g. 2-3 days per month for the following 4 months).

• Savepla needs to have someone (such as Zana) inspect the stoves before
they are delivered to the customer. Quality control and customer follow up
have , so far, not been the strength of Savepla

• As always Savepla is desperately in need of business and marketing

assitance

Suggested Way Forward For Zana

• make links with GPZ .

• Coordinate stove order with WFP. If possible we should try to convince

WFP to buy between 10-50 stoves in the next 6 months. We should
contact Nina and Angela as well as Atanasio and Lynn to pursue the stove
order

• Contact Albertino at Ceramica villa Pery to find out when the bricks will be
ready for pick up. Once the new bricks are ready then we can retrofit the
two stoves for concern international , the tall 80L stove at Formigas del
Futuro , the stove at Cubitsirana, and the WFP stove at SAVEPLA

• We need to assess the durability of the various mortar mixtures. At

present we have 4 different recipes (see stove s 1-5) and we need to
decide which is working the best after 3 months and then use that for the
upcoming WFP stove order.

• HTZ Cement will need to be purchased for the new stoves

• Contact Albertino of Ceramica Villa Pery for information about cast iron
producers in Beira

• Link Savepla with DED small business program

• Monitor production at Savepla to ensure quality control

• Savepla made a couple of household stoves for the headmaster of

Formigas del Futuro (Senora Doka). Follow-up on these stoves would be
helpful
 • Zana needs to spend at least 4 days a month in Chomoio working on
institutional stoves in the next few months

Brick Production
Since Nov 2003, we have made two large batches of sawdust clay bricks. Over
90 test bricks have been produced at Ceramica Villa Pery, Chomoio,
Mozambique. 60 of these bricks are now being field tested in the MangiMangi
stoves. The test bricks that we made had a density of between .5 g/cc and 1.1
g/cc. The bricks between .5 g/cc and .6 g/cc were rejected, as they were too
fragile. The bricks between .9 g/cc to 1.1 8 were also discarded as being
(suspected) poor insulators as well being too heavy for a portable stove. Three
bricks were tested by Dale Andretta of SEA Ltd to establish the thermal
conductivity of the bricks.

University of Dayton is also conducting thermal shock , 3 point bending and

compression strength testing of these bricks but the data is not yet available

Here are the results from the 3 bricks that were sent to Andretta.

Clay Sawdust water Pre fire Post volume Actual Thermal

Bag weight fire density conductivity
2 weight m*K/W
700 500 600 1800 625 973 .64 9.2

Tile # 8
Clay Sawdust water Pre fire Post volume Actual Thermal
Bag weight fire density conductivity
1 weight m*K/W
800 500 600 700 978 .71 8.3

Tile # 10

Clay Sawdust water Pre fire Post volume Actual Thermal

Bag 1 weight fire density conductivity
weight m*K/W
1100 500 600 940 1009 .93 6.6

As you can see from the above charts, a small increase in the quantity of clay in
each brick results in a large increase in the density and a corresponding
decrease in the insulative value (thermal conductivity) of the brick. These 3
samples were made in November 2003.

In March 2004, we made another batch of test bricks (see attached excel sheet)
and found that we could increase the strength of the mixtures by allowing the
clay/ sawdust/water to sit overnight before being formed into the bricks. A few
bricks from this batch have also been sent for testing at University of Dayton.

In August 2004 we placed an order for enough bricks to make 50 MangiMangi

stoves. This requires:

300- 29cm*19cm*6 cm bricks

200- 18cm*18cm*6cm bricks
100- 29cm*29cm*6cm bricks

(Note: these are the mould measurements and the fired brick sizes will be
smaller)

The ratio for the mixtures is

• 12.5 kg dry clay, pounded and sifted through a 2mm screen
7.5kg of dry sawdust sifted through a 2mm screen
12.5 kg of water

This is left to sit overnight and then formed into bricks.

Then dried for at least two weeks
Then fired for 2-3 days at 950 degrees Celsius

These bricks will be ready by mid-October and will be used to fulfill the
impending WFP contract. These bricks can also be used to fulfill other stove
orders.

Albertino, the owner of Ceramica Villa Pery was not able to quote us a price on
the bricks, although he has always been very supportive of the project and in the
past he gave me a rough estimate of US$0.20 per small brick. The bricks should
be paid for by ProBEC and then sold to SAVEPLA as needed

These bricks will be mortared together with a mixture of HTZ Cement , the ratio
to be decided after the current field test to assess which is the ideal HTZ cement
/sand mixture.

Rocket Roll Oven

Since December 2003 we have made three prototypes of the Rocket stove
Portuguese roll oven.

The second prototype, built in April, was favorably

received by the baker and the owner of the bakery,
Senor Joao. At first Senor Joao was reluctant to try
the new design, but after a little prompting by Werner
Klaus, the baker used the oven for two months and
was very pleased with the design . The Rocket roll
oven used less wood than the traditional oven and
cooked the rolls in 15 minutes instead of 20 .

Werner Klaus reported daily fuel consumption of

30,000 Meticais with the new oven (each small log
costs 10,000 Meticais)

compared to 100,000 Meticais for the

traditional oven . When I inquired
about fuel consumption a month later,
the bakers reported that the fuel
consumption of the traditional oven
was only 80,000 Meticais (40,000
Meticais for each large log . It is
unclear what caused this discrepancy
between our figures but even
considering the lower figure for the
traditional oven, the new oven still offers a 62% fuel savings. It is estimated that
the new Rocket Bread oven will reduce their annual fuel costs from US$ 1327 to
US$498 thus saving the bakery approx US$830 per year. The third prototype,
built in August, incorporates another set of improvements. A fourth prototype will
probably be required before we have a final marketable design See attached
report for more information.

Household stoves
Note: The cement/vermiculite household stove has
been discontinued.

A number of household test stoves were made at

Ceramic Villa Pery in March 2004 . These were made
with sawdust and clay.

A 110 mm diameter pipe is

cut on a 45-degree angle
and then placed inside a
30cm by 30cm Bucket.

The bucket is then

filled with clay sawdust.
The mixture is tamped
firmly into place.

The pipes are then carefully removed

from the bucket.
 The bucket is turned upside and removed,
leaving the stove inverted. A wire is used
to cut the stove in half. The stove is dried
for three weeks and then fired at 950C for
two days. See above brick recipes for
possible clay sawdust recipes.

These stoves still need to be placed

inside of a metal bucket or inside of a mud
stove and then tested. They offer one
potential for building a cheap (US$1-2) and insulative rocket stove liner.