Problems with Mold Release

Ceramic water filters are typically cast-formed utilizing aluminum, steel or sometimes concrete molds. The wet mix or charge is placed in or over the molds and the male and female are forced together under pressure to form the shape. Two plastic bags on the inside and outside of the form act as separators that allow the new filter to be easily and cleanly released from the molds. Since developing the process, and in collaboration with filter producers, we have sought an alternative to the use of plastic bags. Attempts have been made to replace bags by using:

  • palm oil
  • teflon
  • waste oil
  • silicone spray
  • silicone plastic

A Ghanaian filter producer (Tamakloe Ceramics) reportedly has been able to eliminate one of the bags by carefully applying palm oil to the inside of the female mold before each pressing.

An Italian flower pot manufacturer reportedly using a thin liquid as a release agent utilizing a similar twin mold technology.

The Ram Company utilizes plaster press molding techniques in which forms are ejected from porous hydraulic plaster by employing compressed air. While this technology would work, it would require an on going supply of hydraulic plaster. 

Material input variables

Fine Vs. Coarse sawdust
What is the relationship between fine and coarse sawdust on filter strength?


1. Compare filters of gradually increasing amounts of fine, and then coarse, sawdust. Compare strength of filters of similar flow rate.
2. Refire low flow rate filters to attempt to equalize the flow rates, then Compare strength of filters of similar flow rate. (Hypothesis is that firing for longer, or higher, will increase both strength and flow rate).
3. If filters using fine sawdust and longer/higher firing results in a stronger filter of higher flow rate, compare fine sawdust filters with coarse sawdust filters of the same flow rate for bacterial effectivity.


Adding Coarse Sand
What are the effects of coarse sand on a constant mix of clay and sawdust in terms of:
a) Strength
b) Flow rate
c) Bacterial effectivity
d) Shrinkage

Adding Crushed Filter Medium (Grog)
What are the effects of crushed filter grog  in a constant mix of clay and sawdust in terms of:
a) Strength
b) Flow rate
c) Bacterial effectivity
d) Shrinkage

Pugged Vs. Unpugged Clay
What is the effect of pugging clay  in addition to mechanical mixing on the  strength of ceramic water filters.
To evaluate this it is necessary to first agree on the method of evaluation. It is suggested that the filter strength be evaluated using a modified modulus of rupture test(MOR)

1. Compare constant mixes of filter medium with variable pugging times for strength and flow rate.

Development of appropriate mixers

Developing appropriate processing mixers, pug mills , or hybrid machines for ceramic water filter factories.

Ceramic pot filters that produce potable water are made from a mixture of clay and combustible materials, usually sawdust or rice husk. Both the clay and the combustible are crushed and screened, dry blended, and then wet mixed before forming. Factories employ simple paddle mixers and in some cases also, pug mills to prepare the mix. There are also factories which do the entire mixing process by hand. While it is possible to produce filters using a variety of methods it is useful to identify methods which produce better and consistent filters so that improvements can be made.

  • Paddle mixers are commonly used in filter factories but work best when the size of charge which they are mixing  is carefully monitored, neither too small or large. Too small a charge and the mixing action  will not work, too large and the mixer will also perform poorly. This is one reason why employing a pug mill after mixing can greatly increase the quality of the blending.
  • Pug mills do a fairly consistent job at wet blending materials. Pug mills are horizontal (most commonly) augers, that differ from paddle mixers in that they must be attended by an operator. The operator puts clay materials in the hopper where it is shredded and then compacted as it makes its way out the narrowing nozzle.  Performance and speed differ depending on size and whether the machine has  single or double screws. Pug mills come in either non-de airing or de-airing. De-airing is not necessary for filter mixes.
  • Hybrid pugger/ mixers Commonly called Peter puggers, these are another potentially useful machines for filter blends. The machine integrates

the work of two machines, a mixer and a standard pug mill.
A hopper is loaded with the mix  and the machine  mixes and pugs,  depending in which direction the mixer is rotated. These machines do an excellent job integrating materials. The challenge with these machines is that they are expensive and tend to be beyond the reach of factories. Ideally an open-source version of this machine  might  be developed which would allow it to be produced locally at a reduced cost.

Use of concrete for mixer / puggers

Concrete Mixers were  first developed by Paul Soldner.

This is a paddle type mixer where the body of the mixer is made with a rotating concrete hopper  and a fixed paddle.

In his book “A Potters Alternative”  Harry Davis outlines instructions and plans for making pugmills from both steel and concrete.  The concrete version would  be a potential model for filter workshops. By following the plans in the book perhaps a set of templates could be produced and made available to factories that wish to have machines produced locally. This robust design has great application for developing world factories because of its simplicity of design and because the machine is made with off the shelf parts (Dayton). Aside from the cast formed ferro-concrete drum there are no specialized parts on the machine  which make it an ideal candidate for open-source cloning. (

Affordable presence absence testing

Fired filters are currently being assessed by factories using three tests. Filters are
dunk tested to identify cracks that allow water to leak through the filter more quickly. If
they pass that test they are then soaked and flow tested. After flow testing, water from
some filters may be challenge tested to see if they actually remove bacteria. The cost of the
presence/absence (P/A) tests is prohibitive enough that not all filters are challenge
tested in this way.

Ideally every filter would be tested so that they are demonstrably capable of delivering
safe water. A simple P/A  test on the order of 10-40 cents each would make it cost
effective for producers to test every sound filter. This test would be conducted
at the same time that flow testing is conducted, allowing the producer to save
time and guarantee that each filter is demonstrably effective.

Kiln Setting

Here we are in the workshop doing more work on the computer than on the workbench, but its nice to have the specs to share.

There has been some work done on designing a special kiln setting to fit more filters into a kiln-load. The ones in the photos below have been made with the same mix as the filter-body.

They are functional, but we wanted to see if we could make them less massive without loosing too much of their structural strength.

This is the kiln post that is being experimented with now:

Original CWF kiln post

The filter posts in use

And here is a photo of the improved version:

Sketchup screenshot of CWF Kiln Posts

or download the sketchup file: