Kenya: Fibre-concrete roofing technology: Adaptation and progress*

* This report is produced on the basis of information and data given in an unpublished draft study prepared for the UNCHS (Habitat) by Martin Fisher and Mary McVay.

The vies and opinions expressed in this report are those of the authors and do not necessarily reflect those of UNCHS (Habitat). Mention of firm names and commercial products do not imply the endorsement of the UNCHS (Habitat).

Introduction

In the previous issue of the Journal an article entitled “Stabilized soil-block technology adaptation and progress in Kenya” was included which highlighted the technological developments and adaptation processes of stabilized-soil blocks in Kenya and outlined the various block-making press manufacturing capabilities in the country. The role of Action Aid-Kenya (a non-governmental organization involved, among others, in rural development and promoting appropriate technologies for constructing schools and houses) was also elaborated.

The present report is the continuation of the report included in the previous issue but deals exclusively with fibre-concrete roofing-tile production-technology adaptation and progress in Kenya, again through the efforts of Action Aid-Kenya. Section A describes the history of fibre-concrete roofing-tile technology development and various promotional activities undertaken in Kenya, while Section B provides more technical information on different types of machines in use or developed and manufactured in Kenya highlighting their features and manufacturing processes.

A. History

The first known research on fibre-concrete roofing (FCR) was carried out by the Swedish Council for Building Research in Sweden in 1975. Later, in 1976, two researchers began their work at Kenyatta University Appropriate Technology Centre and at the University of Nairobi, Department of Engineering, with financial support from the Food and Agriculture Organization of the United Nations (FAO), the Swedish International Development Agency (SIDA), the Portland Cement Company and several local sisal plantation owners. They developed and field-tested a technique for making FCR sheets, but with rather disappointing results. Although, the laboratory test results were satisfactory, field production did not meet an acceptable standard, mainly, because the manual production process and the large size of the sheets made them weak and expensive. The sheets were manually compacted and in order to reach an acceptable strength they needed a cement to sand ratio of 2:1 and they still needed to be quite thick and carefully made. This led to a slow and expensive production process. Moreover, the large size and heavy weight of the sheets required a heavy and accurate roofing structure to prevent twisting and cracking. Installation was, thus, time-consuming and precarious. Nevertheless, the researchers and the organizations, involved in developing and testing FCR sheets pioneered the technology in Kenya and stimulated the initial interest in this technology.

In 1983, John Parry Associates, or Intermediate Technology Workshops (ITW), began to play a role in Kenya. Parry began research in 1977 in the United Kingdom with funding provided by the Overseas Development Agency channelled through the Intermediate Technology Development Group (ITDG). By 1983, he had developed a process of making FCR tiles using a battery-powered vibrating/screeding machine and a set of plastic moulds. While the concept of vibrating concrete to compact the mix is quite old, as are the general forms of the profiles Parry uses, his innovation was in the application of these practices to fibre concrete. Incidentally, a manual method of vibrating and moulding concrete tiles was developed independently in Malawi around the same time. Still Parry Associates was the first to bring such a machine to Kenya when, in conjunction with ITDG, they established the Intermediate Technology Workshop (ITW) in Nairobi to import machines, conduct training and manufacture tiles. By 1984, they had sold seven tile kits and assisted five businesses to get started.

In 1983, Action Aid-Kenya (AAK) purchased a Parry machine and sent one of its staff to United Kingdom for training. The plan was to test and disseminate FCR technology through AAK’s School Construction and Maintenance Scheme.

By September 1984, AAK had established four FCR tile-production units in Kiboswa, Kibwezi, Webuye and Kariobangi. These rather experimental production units were run by youth groups, trained and financed by AAK with the intention that, if the technology took off, they would become self-sustaining businesses. Meanwhile, they supplied AAK with roofing tiles for the schools, which served as a field test and promotional demonstration of FCR tiles. AAK also trained local artisans on how to roof with FCR tiles. This required drafting a manual with roofing specifications, a modification of the one produced by Parry, because of the generally low quality of timber available in Kenya compared with the timber in United Kingdom. The performance of the roofs was quite satisfactory and the scheme worked well in promoting the technology. There were, however, certain limitations, namely: the youth groups were not very cohesive, and had not as yet become mature enough to break away from AAK and become fully independent businesses; luck of financial-management and quality-control skills have been constant problems; in the early stages, technology being labour-intensive, required considerable training and supervision by AAK staff; and shortage of capable staff has limited the number of production units and the number of buildings roofed with FCR tiles.

Energy-efficient building materials

Nonetheless, the scheme, significantly, increased awareness of FCR tiles in Kenya and many entrepreneurs showed interest in buying machines and starting businesses manufacturing the tiles. However, in view of the rather high cost of Parry machines to be imported from United Kingdom, AAK began considering manufacturing FCR tile-producing machines locally to make them affordable for small-scale businesses.

1. Local technology adaptation

(a) Machines

In May 1986, AAK began to develop a manual FCR tile-vibrating machine to be produced in Kenya. The initial research and design work was conducted at the Housing and Building Research Institute (HABRI) of the University of Nairobi and the first product was a treadle-operated vibrating machine. After a field test with a work group of the National Council of Churches of Kenya (NCCK), the machine proved to be not efficient for operation by one worker. Therefore, AAK began to develop a two-person bicycle-powered machine (see Section B). A joint effort by AAK staff and an American missionary, who had begun work on an electric machine and on making FCR tile-moulds locally at HABRI, resulted in finalizing the designs of the manual and electric FCR tile-vibrating machines and a concrete FCR tile-mould production process. In early 1988, the first locally-produced tile-vibrating machines, which had been fabricated in Undugu Society’s Metal Production Unit, went on sale in Kenya. While Undugu Society is a charity, the Metal Production Unit operates as an income-generating unit. Its management has successfully demonstrated that good-quality tile machines can be made profitably in Kenya and sold at a reasonable price.

Meanwhile, HABRI also succeeded in designing a hand-cranked machine being sold by Hartz and Bell. Unfortunately, its performance has been quite unsatisfactory (see Section B). Moreover, Parry Associates responding to the competition introduced their own hand-cranked machine, but, at a higher price than those manufactured locally.

(b) FCR tile moulds, made in Kenya

Just as the imported machines were considered expensive, so were the imported moulds. In recognition of this, AKK and the American missionary developed a way of making tile moulds locally, from concrete. The technique, involving a series of moulds made from one original wooden “great-grand-mother” mould, was introduced to a local entrepreneur which manufactures the moulds.

(c) FCR technology: progress

FCR technology adaptation in Kenya, is still in an early stage, therefore, it is difficult to quantify its progress in terms of the number of buildings roofed with FCR tiles - or even in terms of production units established. However, all indications show that FCR tiles are taking off at a rapid pace, and that the potential market is extensive. This rapid expansion appears to be a direct result of the locally-available low-cost equipment and training.

ITW equipment was purchased in a package including: vibrating machine, 200 moulds, batch boxes and miscellaneous equipment, for Ksh. 130,000. However, an AAK machine plus 200 concrete moulds and the same miscellaneous equipment was about Ksh. 15,000 (both July 1989 prices) This means that the former requires a high-income entrepreneur possibly selling his assets or trying to get a bank loan; whereas the latter can be easily gathered from savings and family loans. In addition, since, the local machines, equipment and moulds may be purchased separately, the investment can be made gradually, which is most appropriate for a small-scale entrepreneur.

People are already taking advantage of the new opportunity. Up to November 1988, after several years in Kenya, ITW had sold 43 kits. In 1988 alone Undugu Society had sold 68 machines, over 50 per cent more, and by March 1989 Undugu had sold another 24. The majority of these machines have been sold to the private sector. Purchasers are given a brief manual on tile-making and are invited to attend AAK training when they are ready to begin production. However, some begin production without training and the actual number of businesses started is difficult to determine.

In March 1989, AAK mailed out a brief survey to ITW and customers in an attempt to try to assess the progress and needs of FCR tile producers. Although the initial respondents were few (20) some conclusions were as follows:

(a) People generally became aware of the technology through magazines, newspapers and technology shows. Others heard about it via the institutes promoting it (AAK, ITDG, ITW, HABRI), or visiting producers and buildings;

(b) Businesses with ITW equipment required around Ksh. 150,000 vis-is AAK businesses which needed Ksh. 50,000;

(c) People promote their tiles via:

(i) Demonstration buildings;

(ii) Advertisements (signs, leaflets, samples in shops);

(iii) Approaching builders on sites, or trying to get contracts;

(d) Customers inclose institutions, wealthy individuals, and lower- to middle-income people. This illustrates FCR’s competitiveness with both galvanized corrugated iron sheets and other tiles. This means that the potential market for FCR tiles in Kenya is substantial;

(e) Producers are, generally, not producing at full capacity and expressed concern about:

(i) Lack of awareness on FCR tiles, particularly by institutions, large contractors and government officials;

(ii) Problems of marketing in general: the strategy that should be used and how to finance it;

(iii) Lack of working capital - most produce tiles on orders, but customers prefer manufacturers to have tiles in stock, ready to buy;

(iv) People are not satisfied with the red-oxide colouring - it gets a white film and also fades quickly;

(f) There are wide discrepancies in cost analysis, indicating that financial issues need more attention during training, as do marketing strategies;

(g) Most people conduct some quality-control procedures, including physical examination of tiles, making recommendations for the roof structure and often inspecting the roof as it goes up. However, without site visits, the general quality of production is difficult to assess;

(h) Many producers received second-hand or no training at all, except written guidelines. This is clearly a major problem, as all promoters of the technology feel that inadequate training will result in poor-quality tiles which, at this early stage, could ruin the reputation of the technology;

(i) So far, producers have few complaints about the performance of the machines, nor have they reported receiving complaints on the quality of their tiles or roofs.

Thus, it seems that the machines and tiles are performing well. However, as increased awareness and the availability of a cheaper machine has sparked rapid expansion in the number of machines in the market, measures are being taken to ensure proper dissemination of the technology. In particular, this means increasing and improving training programmes, establishing better coordination between trainers and producers, and addressing the issue associated with quality control. Currently, official standards are being written and the establishment of an FCR Tile Association is being considered to address these and other issues.

(d) FCR training

AAK’s FCR tile-production training differs from ITW’s training in that it is half theory and half practical, whereas ITW’s is almost fully practical. In view of the complex management involved in FCR tile production, and, since, building a specific type of roof structure for the tiles is rather complicated, AAK feels that it is necessary to spend more time on understanding these issues, rather than concentrating fully on tile production. In addition, since, the quality of the tiles depends a great deal on labour it was found prudent to explain the reasons, e.g., why tiles need to be vibrated for 45 seconds and why it is better to cure them in a tank for 10 days, rather than simply instruct trainees to do so.

The course consists of, first, demonstrating the raw-material processing, vibration, curing, and quality control, and secondly, roofing specifications and the calculation of the number of tiles needed for various sizes and pitches of roofs etc. The FCR tile business-management course goes over costing and pricing of tiles and roofs and also marketing strategies. For promotional purposes, participants receive colourful advertising pamphlets which illustrate the advantages of the tiles, contain technical data on roofing specifications and show an easy format for estimating the cost of an FCR-tile roof.

(e) Standards and specifications

One major barrier confronting the promotion of FCR tiles is the lack of public acceptability of this product. AAK currently chairs a committee established by the Kenya Bureau of Standards, to establish official standards and specifications for FCR tiles. Without official recognition, the municipal council cannot include FCR tiles in building codes, and structures with FCR cannot be legally built in towns. However, once standards are issued, the public will see the legitimacy of FCR tiles and will have an idea on their quality.

Women trained in FCR-tile factory

(f) FCR-tile technology transfer

The transfer of the FCR-tile technology developed by AAK to other African countries has already begun on an informal basis. But AAK is planning to establish a regional appropriate technology unit to meet the rising requests for training in these and other technologies, and to establish more systematic and effective ways of disseminating the technology regionally.

(g) FCR tiles in the United Republic of Tanzania

FCR-tile production with AAK machines and moulds has already started in Arusha and Mwanza, United Republic of Tanzania. In both cases the Roman Catholic Church purchased machines from Undugu. In the case of Mwanza, the Church sent a group for training with AAK in Webuye, but in Arusha training was conducted by a volunteer, who was trained in Kenya. In United Republic of Tanzania, the demand for FCR tiles is quite high, as galvanized corrugated-iron sheets are not readily available and are very expensive. The Roman Catholic Church in Arusha has already imported a number of Undugu bicycle-type vibrating machines and established some units to produce tiles commercially.

(h) FCR-tile technology in Uganda

An independent missionary involved in designing the electric tile-vibrator in Kenya, in 1987, travelled to Uganda to conduct a training on tile-making for Action Aid-Uganda, for the United States Agency for International Development (USAID) in various cooperatives, and for churches. The missionary took three electric vibrators with him, one, of which went to Action Aid-Mityanna. Since then, AA-Mityanna has imported quite a number of machines, but was at a standstill when they requested a visit from AAK in 1988. They were not certain on how to proceed with a technology-dissemination programme for tiles, and, more immediately, they have not been taught how to make ridge moulds. The informal style of the training they received was, thus, problematic.

Meanwhile, USAID endeavoured to fabricate AAK-designed tiles and machines locally. It imported the necessary metal for fabricating machines in Gulu Metal Co-operative, assisted by Euro Action-Accord. Two people from Gulu, in Northern Uganda, visited the Undugu Society’s Metal Production Unit for training and returned to Uganda, designs in hand. However, due to the informal training and lack of follow-up, they have, yet, to produce a satisfactory machine.

In Uganda, as in United Republic of Tanzania, while things have progressed to a certain level, the problems of lack of training and local mould and machine production have not been fully solved. Owing to the AAK’s tied commitments in focusing efforts on its own target areas, regional requests for training and information were not responded properly. However, there are now plans to expand Action Aid-Kenya’s technical training facilities not only for Kenya, but for the East and Southern African sub-region.

Tiles ready to be put on building roof

B. Fibre-Concrete Roofing Tile Technology

Introduction

This section gives some description on FCR-TILE technology. It explains the theory of vibration, the critical factors involved in designing an FCR-tile machine, provides some information on FCR-tile equipment available in Kenya and gives a detailed description of the AAK-designed FCR-tile machines and tile moulds.

1. General principles

FCR tiles are thin screed, sand and cement roof tiles with a very small percentage of natural fibre added to the mortar mix. The wet mortar is laid on a flexible plastic sheet and it is compacted into a thin flat screed by a vibrating table. The compacted flat screed is then lifted on to the plastic sheet and laid over a tile mould which gives it the desired shape. It stays on the mould to dry for a minimum of 12 hours before it is demoulded for another 12 hours of drying and then placed in a water tank for curing for a minimum of 10 days. Finally, the tile is removed from the water and allowed to dry for an additional 10 to 14 days before it is ready for use.

The function of the fibre in the tiles is to help stop the formation of cracks during the initial manufacturing and drying stages. During these stages, the mortar is still wet and weak. The fibre is stronger than the wet mortar and can hold it together to prevent bending and shrinkage cracks. After the tile is fully cured, the mortar is strong enough and the fibres no longer contribute significantly to the tile’s strength. It is only the fact that tiles with fibre may have fewer micro cracks than tiles without fibre that makes the tiles with fibre somewhat stronger.

In fact, if care is taken during the forming process and if the hardening of the mortar is carefully controlled, so as to minimize shrinkage cracks, use of fibres would not be necessary.

Other factors which contribute to the strength of FCR tiles include the following:

(a) The sand used should be clean, sharp and well graded river sand;

(b) The sand and cement must be well mixed in dry condition using a proportion of three measurements of sand to one of cement by weight;

(c) The mortar mix must be kept fairly dry;

(d) The vibration/compaction must be adequate;

(e) The initial drying must not be too fast; and the time for curing and final drying must be adequate.

All these steps must be taught to the operator in a training course on the production site. However, the effectiveness of the compaction by vibration is to a large extent dependant on the vibrating machine used.

2. Compaction by vibration

Builders have used vibration as a means of compacting concrete mortars for over a century. Starting, simply, with the tamping of a concrete form, and ranging to the use of portable poker vibrators and the use of large vibrating tables for compacting precast concrete components. Vibration of a wet mortar acts to compact the mortar into a flat screed and to remove any excess air bubbles and water from the mix. It thus, ensures that after drying and curing the product will have a minimum of voids or air pockets and will, as a result, be much stronger.

FCR tiles which were not vibrated would require a much higher percentage of cement in order to be of the same final strength as those which are vibrated.

Physically, the vibration causes compaction of the mix by slightly shifting the position of the particles, allowing gravity to pull them into a tighter configuration, making the screed more dense and compact. The effectiveness of the vibration is directly related to the acceleration which it imparts to the mortar particles and to the number of oscillations (the duration of the vibrations in a given time). Up to a point, the longer the vibration time the more compact the mortar. It should be noted that too much vibration can lead to a separation of the mortar by particle size and weight and this can actually cause the tile to become weaker.

The acceleration is equal to the amplitude of the vibration times the frequency of the vibration squared. However, too large vibration amplitudes will cause the mortar particles to jump off the table instead of just shifting slightly. Generally, the amplitude should be less than 1 or 2 mm of motion. Good compaction is, thus, best achieved by vibrating at a high frequency with a low amplitude.

The very limited experimental results available indicate that it is the product of the acceleration times the duration of the vibration that determines the degree of the compaction. Approximately, equivalent compaction was obtained in one ninth the time when the frequency was tripled (the acceleration, which is proportional to the frequency squared, was increased by a factor of 9). Experiments also indicate that both a vertical and horizontal component of vibration are necessary for effective compaction. In particular, the vertical component is most effective for knocking down and spreading out the pile of mortar when it is first place on the table while the horizontal component is most effective for compacting the mortar after it is already spread into a thin screed.

Women tile producers in Kenya

3. FCR-tile vibrating machines

(a) Methods of vibration

A vibrating table, which, in most cases, is mounted on springs or rubber mountings, can be made by using a number of different mechanisms. It can be impacted by a hammer which will cause vibrations at both the impacting frequency and also at higher frequencies (the natural frequencies of vibration of the table top). It can be driven by a direct connection to a rotating motor or wheel giving vibrations at the frequency of the rotation. It can be driven electro-mechanical by fixing a magnet to the table with a stationary electrically oscillating coil around it. And, it can be driven by a rotating eccentric cam mounted on bearings fixed to the bottom of the table and driven either through a flexible linkage by a motor also mounted to the bottom of the table.

The last two options, electromechanical or a rotation eccentric cam, have the advantage of a minimum of wearing parts: and the last one has the added advantage that it can be easily adapted to either an electrically- or manually-driven mode. This is the mechanism of vibration chosen by both John Parry in his designs and by AAK. The amplitude of the vibration is determined by the product of the mass and the eccentricity of the rotating cam and also by the mass of the table top and the stiffness of the mountings.

The energy required to vibrate a table is also a function of the table’s mass and properties of the mountings.

An effective and energy-efficient eccentric-cam vibrating devise should have a light weight, yet stiff table-top mounting on either very stiff springs or stiff rubber mountings. In general, the eccentric cam should be lightweight and the eccentricity should be quite small. However these factors have to be adjusted to match the weight of the table top and the properties of the mountings.

(b) Critical factors for design and analysis

The following factors must be considered when designing or analysing any FCR-tile vibrating machine.

Frequency of vibration: As discussed above, the higher the frequency the more effective the vibration and the less time it will require to compact the mortar.

Mode of operation: Is it operated electrically or manually, and what is the most appropriate mode of operation for a given location of use?

Strength: What loads will it be subjected to during use and during transport, and is it designed so that it can withstand these loads without breaking? Are any of the moving components likely to wear out after a short time or are they all properly designed for a long life?

Ease of use and maintenance: Is it easy to use? How many operators are required and how strong do they have to be? How easy is it to maintain? Are spare parts locally available?

Ease of manufacture: How difficult is it to manufacture? Are any special skills or tools required? Are all the components locally available or do they have to be specially imported?

Cost: Is it affordable for members of the target group to buy? What is the pay-back time period?

4. FCR-tile machines in Kenya

(a) Parry electric and manual machines

The first FCR-tile machines in Kenya were the original models of the Parry Associates (ITW) electric tile vibrators. These vibrators have a rotary eccentric cam which is mounted between two ball bearings inside a small metal box which is bolted to the bottom side of the table top. The cam is driven through a flexible linkage (made out of a small-diameter plastic hose pipe) from a small 12-volt fan motor.

The 6-mm thick aluminium table top is mounted on four stiff rubber bearings which are, in turn, mounted on the top of a sheet-metal box which houses the 12-volt motor. The 6-mm-thick metal tile frame hinges at the back of the table and locks down in the front. The lock handles hang below the table top and are pivoted from it on 6-mm-diameter machine screws.

These vibrators are manufactured by Parry Associates in United Kingdom and quite a number have been imported by Kenyan FCR-tile manufactures.

Parry’s newest version of the vibrator is the plastic encased “multivibe” unit which clamps onto the bottom of a 4.5-mm steelplate table-top, mounted on four rubber mountings.

By 1987, Parry had developed a hand-operated vibrator which is operated by two people. The crank turns a set of reduction pulleys which are connected by two V-belts and which drive the eccentric cam through a flexible linkage at close to 40 revolutions per second (the same frequency of vibration as the electric machines). Although the hand-operated machine seems to work just as well as the electric machine, reports from the field indicate that it is quite tiring to use.

(b) HABRI hand cranked machine

The Housing and Building Research Institute at the University of Nairobi designed a low-cost manually-operated tile vibrator and arranged for it to be manufactured by Hartz and Bell in Nairobi. HABRI has used these machines in some of their training and has arranged for some field-testing by various organizations including Canadian Save the Children in Meru.

As with the other FCR-tile machines available in Kenya, the vibration is caused by a rotating eccentric cam, mounted on the bottom side of the vibrating table top. The driving mechanism uses the hand crank, gears and gear housing from an Indian-made hand-powered grinding wheel. This mechanism turns a 6-inch diameter pulley which in turn uses a V-belt to drive a 3-inch diameter pulley mounted on the same shaft as the eccentric cam. Thus, the eccentric cam rotates about 25 times for each rotation of the hand crank. The hand crank, therefore, has to be turned at high speeds to get effective compaction.

The machine is designed to be operated by a single operator who concurrently turns the crank and screeds the mortar to make tiles. This requires a very awkward motion and makes it rather difficult to operate.

Reinforcement of FCR tiles

(c) AAK FCR-tile machines

This section describes the main design features of the vibrating table-top which is used in all three of the AAK-designed machines: the treadle, bicycle, and electric machines. Although some of these features are similar to those on the Parry machines, most of the design decisions were made based on the locally-available resources or in order to make machines compatible with local requirements.

As with the Parry machine, the vibration of these tables is caused by a rotating eccentric cam mounted between ballbearings on the underside of the table top. This cam is connected to the power source by a flexible linkage.

The size of the tile screed was chosen to be 250 mm by 500 mm and either 6- or 8-mm thick. This is the same size as tiles from the Parry machine. The reason for this was that by early 1986, when work on these machines began, there were already a number of tile producers in the country using Parry equipment. Thus, rather they introducing a competing tile size, it was decided to design machines to make tiles compatible with those already on the market. Later that same year, the Kenya Bureau of Standards (KBS) Technical Committee on FCR tiles, decided to standardize the size of FCR tiles in Kenya to the same nominal size of 250 x 500 mm.

The table top is made from 4.5-mm thick mild-steel plate. (Parry has recently also switched from 6-mm-thick aluminium to this material on his latest model.) The rubber mountings are the mountings used for the air cleaner of a Peugeot 504 which are readily available in Kenya. These rubber mountings are attached by nuts with spring washers to small stall mountings made from pieces of 50 x 25 mm hollow sections. The steel mountings are, in turn, bolted by two 8-mm diameter machine screws and nuts to the underside of the table. This arrangement avoids any problems of stripped screw threads and is, generally, a robust design.

The tile frame is hinged at the back of the table as on the Parry machine and is locked down by two locking handles on the front. These handles are welded to large-diameter machined bushings which pivot on 20-mm diameter shafts. These short shafts are bolted by two 8-mm diameter machine screws to the underside of the table. This is a much stronger design than other machines where the locking handles are pivoted on single 6-mm diameter machine screws.

The eccentric cam is mounted inside a piece of 2” x 3” rectangular hollow section which is bolted by four 8-mm diameter machine screws to the underside of the table. Machined bushings are used to mount the two sealed ballbearings. The shaft for the cam is 16 mm in diameter. The flexible linkage is made from a length of fibre-reinforced plastic tubing and is held in place by two standard screwed hose clamps.

On all the AAK machines, including the electric one, the vibrating table is mounted on a framework made from rectangular hollow-steel sections. These are also much stronger than the sheet metal and flat-bar frameworks used by other machines.

The weakest part of the table-top design is probably the mounting of the eccentric-cam bearings and the eccentric cam itself. Although the bearings are large enough and will last a long time, there is still too much machining and too many nuts and bolts involved in the way they are mounted. Also the eccentric cam is only held in place by a small set screw which is liable to come loose in time. If it is not kept tightened, it can come loose and cause a reduction in the vibration frequency.

(d) AAK manually-operated FCR-tile machines

AAK works in many remote parts of the country. In these areas it is difficult to recharge a battery every week as is required for electric machine. Thus, it was decided that AAK’s first priority should be to design a manually-operated machine. This section explains the various methods of producing vibration manually and describes in detail the method chosen for use on the AAK-designed machines.

As discussed in the previous section, the effective compaction of mortar requires a high frequency of vibration. The imported machines vibrate at about 40 cycles per second, and it was decided that to be competitive, a manually-operated machine should vibrate at about the same frequency. However, a person can comfortably turn a handle, operate a lever, or otherwise move for extended periods of time at a rate of around only one cycle per second. Thus, the problem of designing a manually-operated tile vibrator is really one which converts a human motion of one cycle per second into a vibration of not less than 40 cycles per second.

A number of methods were considered for producing this 40 to 1 multiplication in frequency. These included: using a set of gears; using two sets of pulleys and V-belts, each with a 6 to 1 reduction as was also used by Parry for his manual vibrator; using a flexible belt to run a small pulley off a bicycle or other large-diameter wheel, as is often used on manually-operated sewing machines and knife sharpeners; and using a multi-stage adaption of the rope-driven slip-drive mechanism in which a small roller-wheel is driven by friction from a rotating bicycle wheel and its tyre.

In this last option the small roller-wheel is mounted between two fixed ball-bearings, it rests on top of a standard 26” rear bicycle wheel and tyre, and when the wheel rotates, the roller is driven by friction at a much higher rate. The small-roller is then connected by a flexible linkage to the rotating eccentric cam which causes the vibration. This option has a number of advantages over the other options considered. It does not require any difficult or expensive components such as gears, V-belts, pulleys, flexible drive belts or special elastic and nylon cords. In fact, only common bicycles parts and ball-bearings are required, which makes it fairly cheap and easily repairable by village bicycle repairers. It can easily reduce rotation frequencies from 15 to 20 times the frequency of the bicycle wheel, and is easily adaptable to a handle, treadle, or pedal-drive mechanism.

Tile-production factory in a rural area

Its major disadvantage is that, like any drive system, it is prone to energy losses. The main losses in this system are due to the flexing of the bicycle tyre, friction in the bearings, and occasional slippage between the tyre and the small roller-wheel. It is difficult to assess the quantity of these losses, but if the bicycle tyre is kept well inflated and if the bouncing of the small roller-wheel is kept to a minimum the losses should not be very great.

Nonetheless, early experiments showed that the frictional losses caused the vibration to stop within seconds of stopping the power input to the bicycle wheel. This meant that to keep the vibration going; the bicycle wheel had to be powered almost continuously. In order to solve this problem, two 175-mm diameter flywheels were mounted on the same shaft as the small roller-wheel with one on either side of the wheel. These act to store the energy and, thus, allow the cam to continue rotating up to 5 or 10 seconds after the wheel is no longer powered. Thus, the wheel can be powered less frequently and more easily. In order to avoid expensive machine costs, the fly wheels are made from cast-iron plough-wheels which are first statically balanced by drilling holes in them to reduce any unbalanced weight.

Another problem encountered during the machine’s development was that the bicycle tires available locally are not very round. Thus the small roller-wheel needs to be able to move up and down to compensate for the changing tyre radius. Therefore, the roller and two fly-wheels are mounted on a hinged member. It was originally hoped that their weight would be enough to stop them from bouncing on the bicycle wheel, but it was found that they need to be held down with a spring. Unfortunately, this extra force increases the energy losses due to the tyre flexing and makes the bicycle wheel somewhat more difficult to turn.

(e) AAK treadle operated FCR-tile machine

The idea behind the treadle-drive machine was that a single person could both operate the treadle, to cause the vibration, and at the same time screed the mortar to make the tile.

The vibrating table top was mounted at waist level on a framework made of 40-mm square hollow sections. A standard rear bicycle wheel, sprocket and freewheel were mounted just behind the table top with the top of the wheel some 75 mm below the level of the table top. The small roller-wheel was 35 mm in diameter and was mounted in between the two ball-bearings of a standard bicycle bottom bracket in such a way that it could roll on the bicycle tyre. The two fly-wheels were mounted on the same shaft which was connected by a fibre-reinforced plastic linkage to the eccentric cam. A bicycle chain was hung over the freewheel sprocket with one end connected to the mainframe by a flexible spring and the other end connected to the free end of a treadle. The treadle was a metre in length and was pivoted at a point just to one side and behind the place where the operator stood.

When the operator pushed down the treadle with his/her foot, the chain caused the sprocket and, thus, the bicycle wheel begun to rotate, which in turn rotated the small-wheel, fly-wheels find the eccentric cam. When the operator’s foot was lifted, the spring returned the treadle to its original position. The freewheel allowed the sprocket to reverse its rotation on the upstroke without reversing or slowing the rotation of the bicycle wheel.

Because of the small diameter roller-wheel and the long treadle lever, the eccentric cam would rotate approximately 38 times each time the treadle was pressed. Thus, if it was pressed once every second, the vibration frequency was approximately 38 cycles per second.

In laboratory tests, the treadle operated machine worked well and made good tiles. However, from the initial field test with a youth group, it was determined that it was somewhat difficult for a single operator to coordinate his of her hands and feet when making tiles, and to exert the required effort to make the tiles. As a result the tiles tended to be weak due to too little vibration.

This encouraged AAK to try and redesign the vibrator to be operated by two people, with one providing the vibration and the other screeding the tiles. Although this option requires an additional worker, labour is not very expensive in Kenya. By providing this extra job the cost of a tile only increases by 6 or 7 per cent.

The field test also demonstrated that bicycle bottom bracket ball-bearings are not suited to the high rotational frequencies of the small roller-wheel. They required frequent greasing and continually became loose in their housing.

(f) AAK pedal-powered FCR-tile machine

This machine is very similar to the treadle-operated machine except that the bicycle wheel is driven by a standard bicycle chain, chain wheel and pedals replacing the treadle. Two operators are required, one to pedal and the other to make the tile.

Again, the vibrating table top and the bicycle wheel are mounted on a framework made of 30 mm square hollow sections. A demountable drive assembly with pedals, a chain-wheel, a bicycle seat and handlebars are bolted to one side of the framework. It is positioned so that the operator faces the table top when pedaling. The whole machine can be levelled by screwing in or out its six adjustable feet. The chain-wheel and pedals rotate in a standard bicycle bottom bracket which can be moved back and forth on the frame of the drive assembly to adjust the tightness of the drive chain.

The small roller-wheel is 50 mm in diameter and is gnarled to give a better grip on the tire. It is mounted on a 20-mm diameter shaft between two sealed pillow block ball-bearings with one fly-wheel mounted on either side. This bearing and the wheel sub-assembly is pivoted from above and pressed against the tire by a compression spring.

Since there is a reduction in diameter between the main chain-wheel and the freewheel sprocket, each time the pedals are turned the small roller-wheel rotates almost 38 times. So if the operator pedals at a rate of one revolution per second then the vibrating frequency will be 38 cycles per second.

Women carrying finished tile

The pedal-operated vibrator worked well in field tests making up to 250 tiles per day with five workers, and over 50 machines were sold in the first year and a half of production. The machine was sold for Ksh 8500 (July 1989 price). In general, the idea of a manually-operated machine is well received by customers who see the advantage of not using a battery and of providing more jobs. However, they find the price somewhat high and the machine too large to transport easily.

(g) AAK electric FCR-tile machine

The main problem with making an electric tile machine is the availability of a 12-volt motor. A convenient motor to use, and one which works at the appropriate rpm, are the motors used for the interior cabin fans in automobiles. Technically, these are the same motors used in the imported machines, but in Kenya new motors of this type are usually unavailable or prohibitively expensive. However, second-hand fan motors can be bought very cheaply from junk yards around Nairobi. Many of these motors are in very good condition because, in general, people in Nairobi do not use the interior fans of their cars very often.

The main body of the electric machine is made of 40-mm square hollow sections welded into a rectangular framework and then covered by a sheet of metal. The rubber mountings of the vibrating table top are bolted directly on to the hollow sections and the motor is mounted on a piece of sheet metal welded between the hollow sections. The whole machine is substantially stiffer and stronger than the imported machines. Four 6-mm thick metal tabs with 12-mm holes in them are welded to the bottom of the machine and can be used to bolt the machine to a workbench. An electric switch is mounted on the sheet-metal cover and two wires with cable clips are used to attach the battery. Over 100 of these machines were sold by the Undugu Society in the first one and a half years of production and so far very little trouble has been reported.

The biggest problem is the electric switch. It is very difficult to buy good-quality electric switches in Kenya and the best ones available, still, tend to break down quickly. This is also a problem with the imported machines and it just means that the user will have to be a bit careful and be willing to install a new switch every once in a while.

(h) FCR-tile moulds

The most crucial design features of any FCR-tile mould include the following:

Shape: Determines the type of tile manufactured. Most FCR tiles are of the pantile shape, but recently interlocking and “Roman”-shaped tiles have been introduced by Parry Associates and are now made in Kenya.

Alignment mark: The tile moulds must have some type of marking so that the mortar screed can be placed on the mould in the correct position and alignment. If the screeds are not properly aligned on the moulds, then the tiles will be skewed and will not lie on the roof properly.

Weight: Tile moulds should be reasonably lightweight. They have to be moved around the site during tile production and if they are too heavy this will become a problem.

Stacking mechanism: With over 200 moulds at any manufacturing unit, there should be a method of stacking the tile moulds so that they do not occupy too much space. Also the wet screeds must be kept out of the sun and wind so they do not dry out too quickly and crack. Thus, it is useful to have a stacking method which protects the tiles from these elements.

Strength: Each mould is handled several times a day, and they should be strong enough to withstand the stresses.

Cost: Since about 200 tile moulds are required to start a production unit, it is important that their unit cost should remain quite low.

Moulds for ridge tiles are usually made from marine plywood. They are “V” shaped with the angle fixed according to the pitch of the roof. Some have been made using hinges at the bottom so that the angle can be easily changed for differently pitched roofs.

Small-scale production of FCR tiles

(i) Parry FCR-tile moulds

Parry Associates manufacture and sell hot-pressed plastic tile moulds. These lightweight tile moulds are made in United Kingdom and sell for Ksh 350 each in Kenya (July 1989). The moulds are mounted on wooden frames and they stack on top of each other in such a way that the freshly formed tiles are protected by the mould above from the sun and the wind.

In order to ensure that tiles are correctly placed on the moulds, the moulds have a small protruding ridge which has to be very carefully aligned with the bottom edge of the tile. However, the ridge is rather small and is difficult to see because it gets covered up by the plastic sheet.

In 1984, Parry licensed a local Nairobi fibreglass manufacturer, Sai Raj, to manufacture and sell fibreglass tile moulds. These are of a very similar design to the plastic moulds, but are in general heavier and stronger than the plastic moulds. However, at about Ksh 200 each (July 1989) they are still very expensive for many small-scale producers.

In order to avoid these high costs, many of the first groups to produce tiles in Kenya (including the AAK groups on the suggestion of ITW) made their own tile moulds by filling a plastic Parry mould with mortar to form solid concrete moulds. These were then mounted on simple wooden frames so that they could be stacked. However, these solid moulds are too heavy and the surface is usually quite rough.

(j) AAK concrete FCR-tile moulds

In October 1986, an American missionary, started working with HABRI at the University of Nairobi on designing locally made FCR-tile moulds. HABRI came up with the idea of thin screed concrete moulds, cast from a single solid wooden mould which was called the “grandmother mould”. The idea is to cast a number of thin screed (15 mm thick) “mother moulds” from the grandmother mould and then to use them to cast thin screed “tile mould”. The top surface of the mother mould is a direct negative of the grandmother mould’s top surface and thus the top surface of the tile mould would be an exact replica of the grandmother moulds.

These tile moulds are made from a manually compacted mix of two measures of sand to one of cement. They are about 600 x 350 x 13 mm and they weigh around 5 kg each. Their cross-sectional shape is the same as that of a tile with an additional flat section added on either side. The junction between the place where the tile lies and the flat side section forms a sharp edge which runs along the whole length of the tile mould. It can be easily felt through the polythene paper and this allows the alignment to be accurately performed.

The moulds are designed to be stacked in simple wooded racks with the flat sections on the sides of the moulds resting on thin wooden runners. Each rack holds about 40 moulds and is covered with a polythene sheet to protect the moulds from wind and sun. These moulds are also reinforced for strength by two ribs which are formed across their bottom surface. These ribs also act as handles allowing the moulds to be easily pulled in and out of their stacking racks.

The main problem with the original manufacturing method was the expensive and time-consuming process of manufacturing the solid wooden grandmother moulds. However, later it was determined that the grandmother mould could also be made from concrete by the introduction of a wooden-framed “great-grandmother mould”. This mould consists of a rectangular wooden box which has no top or bottom and which is bolted together by two bolts at each corner. The box is made of one-inch thick hard wood, and the top edges of the two ends are cut to have the exact profile of the tile mould.

To make a grandmother mould the “great-grandmother” box is placed on a flat table and it is filled with a dry mortar mix of two measures of sand to one of cement. A very straight and stiff metal angle-iron is then held across the two ends of the box and is moved back and forth and up and down along the end profiles of the box to shape the mortar into the shape of the two profiles. It is crucial to use a very dry mortar mix (to avoid any slumping) and to keep the angle-iron parallel to the sides of the box at all times. This step, the making of the grandmother mould, only has to be done once at any given site, but it is vital that it is done perfectly. Nonetheless, the wooden great-grandmother is much easier to make than a solid wooden grandmother mould. One has simply to trace the tile profile on to the two pieces of 1 -inch by 6-inch hard wood, cut them out carefully, make the other two sides of the box and proceed as above. In this way it is easy to make tile moulds of almost any shape including tapered moulds which have a tighter curvature at one end. To date the technique has successfully been used to make moulds for pantiles, interlocking tiles and the tapered “Roman”-type tiles following the same basic patterns as those introduced by Parry Associates.

Once the grandmother mould is made and cured for five or six days, it should then be placed back inside the box, then, be oiled with old motor oil and the box (great-grandmother mould) should be lifted up by 25 mm by placing a piece of wood under each end. It now forms a 25 mm high frame around the edges of the concrete grandmother mould. This frame is filled with mortar and the same angle-iron straight edge is used to compact the mortar into a 25 mm thick screed.

The newly-formed mother mould is left to dry on the grandmother mould for 12 hours before it is removed and put into a water tank to cure for seven days. Thus, only one new mother mould can be made each day. After a mother mould has been fully cured and dried, it is oiled with old motor oil, and a 13-mm-thick metal frame which has been formed to have ends with the same shape as the moulds is placed on to the mother mould. This frame is filled with mortar which forms a tile mould by using the same angle-iron straight edge. Two strengthening ribs are added on the back side of these moulds. After 12 hours on the mother moulds, the tile moulds are demoulded and placed in a water tank for a day or two. They are then removed and the top surface are painted with a thin mix of cement and water to fill in any holes and to smooth out the surface.

Loading of finished tiles

With one grandmother mould, one mother mould can be made per day. The number of tile moulds that can be made per day is equal to the number of mother moulds available. Thus, starting with a great-grandmother mould, it takes 35 days of producing mother moulds and tile moulds before 200 tile moulds are made and ready for use. If a mother-mould kit, which consists of five mother moulds, a metal frame and angle-iron straight edge, is used, it takes 45 days before 200 tile moulds have been made and are ready for use.

These tile moulds are being manufactured and sold by a private entrepreneur in Nairobi who sells them for Ksh 23 each. He also sells the mother mould kits at Ksh 1000. The cost of materials for making moulds is about Ksh 8 per mould. (All prices are based on July 1989 quotations.)