Automating Building Manufacture

At Auroville I took the work on automating their CSEB manufacturing machine – Auram 6000. Compressed Stabilized Earthen Blocks are fast becoming popular for their obvious benefits. These blocks have superior thermal insulation than concrete. With proper design, they compare well with Concrete on earthquake tolerance. These are normally made from locally available material, thus reducing transportation cost. Curing time for earthen block is low.  As these blocks have better dimensional accuracy, it is possible to use these blocks without using mortar. These help to improve productivity and at the same time is the technology is friendly to environment.

Aureka makes manual and semi-automatic CEB presses. These can make around 65 types of normal and specialized  bricks. It has now come with fully automatic machine that will be able to produce around 500 full size (1′ x 1′) blocks in a hour. An average size 2500 Sq.ft building will normally require 15000 blocks. Automatic CEB machine can produce these bricks in 30 Hrs. At this rate CEB will rival the construction speed of concrete. In order to use mortar-less construction the blocks need to be within +/- 1 mm accuracy. Molding sand and soil to this level accuracy proved to be a challenge.

In this machine I am using PIC 16f4877a microcontroller to control the hydraulics. Micro-controller helps to automate movement of different rams. These speeds up the process and relives the operator from monotonous lever shifting work.  First day when we tried, the operator pointed out that for some specialized blocks, machine needs to be stopped at a certain stage for manual intervention. So I had to modify the microcontroller program. Next we tried a small sample batch. Position variations were very high. I noticed that the sensors were placed in a manner that is responds to change in position quite slowly. Oriented sensor positions to improve the sensitivity issue. Next issue was the problem of inertia in the control circuit. As sensor detects position, it responds by closing a relay in hydraulic. How much the ram will travel after receiving stop signal depends on:

  1. Response time of solenoid
  2. Viscosity of oil
  3. Plasticity of soil

These three variables depend on ambient temperature, humidity, condition of machine, amount of water added in soil. In simple words, it is impossible to control the inputs to provide consistent output purely on sensor feedback. Solution was to introduce some physical limiters at vantage points. This CEB press operates at around 100 bar pressure. In these pressures it is quite easy to damage structures unless proper design precautions are taken. These physical limiter, in turn required  modifications in the structure. After taking care of all these, machine started producing blocks with proper dimensional control. Here you see the trial run on the prototype setup.

With 65 different types of blocks to produce, next challenge was to find a convenient process to set the block height and mold depth. Each block require different setting. Sometime setting may need to be changed to take care of type of soil being pressed. For this I found it best to consult the person who will actually do the setting. After some brainstorming we settled on a process that is easy to use in field with minimal equipment. The process was documented.

Next stage was to take the machine for a full scale trial. CEB press works with auxiliary equipment –  crusher is used to break solid lumps of soil and dry mixture is used to mix sand and cement uniformly with the soil. These are required to feed CEB press at its required speed.  Here you see the full setup for production trial.

Old Hobby in New Age

After my (mis)adventure with ordering a micro controllers in QFN pack, I decided to play it safe. Meaning take the new things in baby steps. To start in the uncharted territory ordered a development board and programmer from the “source” – Microchip. For the beginner PICKIT2 with LPC development board looks safe one to handle.

LPC+Pickit is good. One can program it without taking it out from the circuit. It has nice LEDs switch and potentiometer that suffice many applications. But soon I realized the full meaning of LPC – L is for limited. One needs more space to connect various things with the LPC. So came a bread board.

 Now, connecting LPC with bread board is not a easy task. Connecting 14 sets of wire is a chore. A bridge connector makes the 14 wire to one connection.

 Concept of modularity gets carried further into break out boards. Here is an array that I got myself end of a couple weeks. From top left is a DIP switch with extra long connector to get firm connection on breadboard, a set of eight pins shorted for getting Vcc to DIP switch, a power supply from battery eliminator, I could not find a resistor block in the local store so made one, next is a LED array to check the states of a port, next is to check 1 bit – green and red, then a piezo buzzer from a discarded clock to make some noise and a set of four LEDs to check a nibble size.

After all the LEDs it is time for some real output – here is a 4 segment LED display. I am using here a CD4511 – BCD to 7Segment decoder/latch/driver. This IC reduces the no of pin required quite a bit, then I use a program to multiplex the display.

Now, connecting the LPC with this board requires some 10 set of wire. This is made simple with another bridge connector. For protection, I encased all the wires in a good dose of glue. Now it will develop any loose connection with multiple handling.
Here is a small timer on my LPC prototype. Requires 3 connections.

 Working with 4511 requires less pin from microcontroller but it has limitation. One can only display numbers with it. Here is another display that is totally software driven and multiplexed. This is made on a stripboard. Connections from LED pins are taken out below the LED resulting in a rather neat appearance.

Next step is a LCD display. LCD is good companion to uCs. It does not need to be refreshed continously, has more number of character and gives better look and feel.

I have put header pins on the LCD display for quick connection to breadboard. But I realized, there is something wrong with this display. Checking LCD display is rather difficult. The display has a uC on its own

 and requires complex sequence of instruction to show anything. There is a nice blog by Suhas [] that describes how to check a LCD. COncept is nice, but I felt the PCB can be improved. Some more search on the net got me a nice CAD package – VCAD that supports strip board layout. A weekend on VCAD, I could get the circuit on a 16×16 piece of board. Here is the layout. that you can also use in some place.

While making the board, I found some used components and modified the layout a little bit to use those old rusty pieces.

Result is slightly different, but still needs less footprint and desk space and goes into breadboard without any other additional piece of wire.

Hobby Electronics is not same any more

Gone are the days when electronics hobby meant spending a tidy sum in building most funky amplifier in the town. In my student days it was mostly transistors, with some ICs making its appearance. All electronics had their legs that went nicely inside the holes of a vero board. Sometime a few more adventure loving specimen went ahead and actually made their own PCBs.

I had left electronics in the nineties and thought of picking up the lost thread. In the current state of economics, building amps are no more viable. So I thought of piking up the latest fad in town and ventured into the world of micro controller.

After some research on net and then after reading a bit more on the new animal in technology, I finally ventured into ordering some part. After mousing through the catalog I zeroed on 12f519. Supposed to be a neat tiny piece of silicon. It packs more computing power than what I had used for my first computer project at BHU on a 3rd generation computer.

After a week or so I get a nice package – some 6″ x 3″ cardboard box. I open it, inside there is another box of foam, open that. Now all I see is an envelop.

Inside the envelop there is a tiny pipe. Not to be intimidated, I open the sealed pipe – out come two tine pieces of plastic.

Welcome to the world of qfn package – chips with no legs.

It is nice for miniature piece of appliances, but for hobbyists I hope, there were some better alternative.