This is the hardest part, but can be broken down into some simple steps. There are two things you need to design, first the geometry, second the electricity. They are very related, but element design depends largely on size of the kiln.
Geometry
I made a decision very early on to build a polygonal top-loading
kiln. The reasons I made this choice are that it does not require any
welding, and the elements hold themselves in the grooves. Ultimately,
it's just simpler. The most efficient volume for heating is a cube,
but electric kilns work by radiative heat transfer, and the efficient
volume to evenly heat that way is a sphere, so the round shape of the
polygonal design makes sense.
The little picture there is the output from a program that I wrote to illustrate the brick layout of the walls. The first step is to figure out the internal radius. This has to be determined by the availability of kiln furniture. I opted for 21 inch diameter shelves, so I decided my kiln should be about 23 inches internal diameter. That means 12 bricks each subtending 30 degrees of the circle inferring 15 degree miters.
The next step is to figure out the volume of the kiln, meaning the height. How much volume you can enclose and still fire to temperature is determined by how much power you can put into the kiln. In turn, that is largely determined by the available amperage. The other consideration is how much volume one course of bricks adds.
My studio has 60 amp service (a 60 amp sub-panel from my 100 amp home service). That means a maximum of 45 amps continuous draw. Assuming 240 volts that means I can have a maximum of 45*240 = 10800 watts powering my kiln. There are a number of rules of thumb for how many watts per cubic foot. Olsen gives a range of 1.5 - 2.07 kW/ft3.
I decided to lay my bricks flat, meaning 2.5 inch courses. So, each course adds 0.6 cubic feet. From that you can figure out how many courses you can afford.
To make a long story short, I could make a maximum of 11 courses with 10800 watts. This would be ok, the top course I do not want any elements in for structural reasons, so that gives 5 circuits of 2 elements each in 10 courses. It is at the low range of Olsens recommendations, but that should be ok because I will only fire to cone 6, and my kiln will be better insulated. However, I don't want to be at the max, so I decided finally on 9 courses. 8 courses with elements, 4 circuits. On the advice of a local kiln builder, and the kiln builders at Euclid I decided to power it with 9000 watts.
The picture above shows the brickwork on the walls. The circle inside the walls is the 21 inch shelf. The dashed line on the outside is the extent of the fibre blanket that the kiln will be wrapped in. The solid circle through the bricks is the brick that will be ground off to make the kiln a nice cylinder.
Electricity
I don't know much about electricity. Fortunately all the books that I linked to in the first page tell you everything you need to know. What it basically comes down to is figuring out how much resistive wire you need in order to put the desired wattage into your kiln.
I won't go into all the details about designing your elements, it's not completely simple, and there are lots of good resources out there. Online, check out Coggins instructions on element design. I'm not going to bother trying to improve on that.
What I will do though is provide the python source code to my program that I wrote to help me do all the calculations: kilndesign.py .
python kilndesign.py -t ["circle" | "cube"]
-R [radius] -C [courses]
-g [element-gauge]
-c [# of circuits]
-e [# of elements]
-w [wattage]
If you want to run it you have to download python for your operating
system, and be able to run it from the command line. I have no idea if
that's possible on current apples. Feel free to send me an email if
you need help running it.
Here's the example from using it to compute the parameters of my 9 course 9000 watt kiln.
cubic feet : 5.56 cubic inch : 9610.14 wattage : 9000.0 amperage : 37.5 (0.62) courses : 9.0 height : 22.5 kW / c.f. : 1.64 circuits 8 elements
wattage: per circuit: 2250.0 per element: 1125.0
amperage: per circuit: 9.4 per element: 4.688
element gauge 15
resistance: total: 102.40 per circuit: 25.60 per element: 12.80
wirelength: 47.8 feet per element watts/foot: 23.55 watts/in^2: 11.0
So the important things here are the kW / c.f. this should be (as Olsen points out) between 1.5 to 2.1 for stoneware temperatures. The wattage and amperage per circuit and per element are simple arithmetic, no big mystery here. The resistance and watt loading is where the numbers get really important. They are calculated by the constant properties of the Kanthal A1 wire that is used for the elements.
The resistance per element should be around 12 ohms, and the watt loading should be around 12 watts per inch2. If you read Olsen carefully, you will see that he recommends a watt loading in the range 6.4-9 watts per inch2. The other experts I have consulted suggest that this is very conservative. Euclids design theirs to achieve around 12 watts per inch2 with A1 wire.
N.B. my little program there assumes you wire all elements of a circuit in series. You would have to adjust the program if you wanted to wire in parallel.
I ordered these elements as well as connectors, insulators and other paraphenalia for $200 Canadian.
