Changes to the Home Monitoring System Architecture
Almost a year later, and a working prototype, the development of the monitoring system had reached a standstill, mostly due to the lack of time :
On the sensor side , the Arduino + ENC28J60 + DS18B20 combo works, although it is dependent on the availability of a network cable. But the hardware is functional.
On the server side , the development of a backend and frontend with the flexibility required (multiple sensor support, dashboard with user selected time intervals, etc) was starting to take too much time.
So I was faced with a decision regarding the server side of the system:
Fully development of a frontend and backend , using a LAMP stack and bootstrap templates. This would take time that I simply don’t have, and the project would probably still be stalled.
On the hardware side, I had some ESP8622 (ESP-12) modules to try out.
Server Side changes
The server side of the Monitoring system was the one that was the most time consuming.
One of the major changes was the replacement of the LAMP stack for elasticsearch and kibana for data storage and visualization respectively. This removed the need to write a backend and a frontend from scratch. The time that would be spent on writing the back and front ends was spent on the sensor module development (both hardware and software).
Elasticsearch, according to the authors:
” Elasticsearch is an open source distributed, RESTful search and analytics engine capable of solving a growing number of use cases.”
With elasticsearch, the sensor modules sends data in json , instead of sending the data via http get to a php script with the previous Monit System architecture.
And a dashboard with visualizations of the data can be done in minutes.
Although the arduino + enc28J60 sensor module was not entirely abandoned, hardware development focus was oriented on the ESP8622 based modules.
The ESP8622 have some interesting advantages over the arduino + ENC28J60 combination :
The ESP8622 can be sourced from around 1,7 € a piece (ESP-12) (and it replaces the arduino board and the ENC28J60 in one package)
It can be used with the Arduino IDE , and use most of its libraries (no need to learn a new SDK and new tools).
No additional network hardware required, since the ESP module is a wifi module first and foremost (DHCP, WPA2 supported out of the box), and thus I could place the sensor module anywhere as long as there is wifi coverage, including the exterior of the house.
Less parts per module, since the ESP8622 has an ARM CPU besides the WiFi capabilities : besides the module itself, only the sensor, a RTC (if needed) and some passive components are needed.
However the usage of the bare modules are not as “plug and play” as with arduino boards – additional hardware and wiring required . This is not an issue since the final goal is to have a custom made PCB for the sensor module.
For a long , long time I wanted to have a system to monitor, and perform data collection, on my home.
A modular system where I could gather data from indoor and outdoor temperatures to local power consumption.
The system is meant to be designed in such manner that the server side is fully abstracted from the sensor module side, more sensors modules may be implemented on different hardware platforms. There can be as many sensor modules as required, connected on a single Ethernet network.
For quite a while, since I began to feel more comfortable with using my DSLR on manual mode, I had the desire to get into film photography of a while, being pinhole photography the first milestone.
However, for me to deal with film photography, the only practical choices were 35mm film or 120 film, in regards to availability. My main choice was 120mm due to the fact that the negatives are larger, making it more practical to handle and to make contact prints in the future. The unfortunate side effect is that it is harder to load into the film spiral during film development preparation. And, of course, way less exposures per roll than 35mm film. Another point to consider is camera price. A 120 film camera is, in general, more expensive that a 35mm camera.
My budget for the camera was set to 120€. In one hand I could get a Holga 120N pretty cheap and call it a day. Lets just say I quickly changed my mind.
The other choice was to get a Twin Lens Reflex (TLR) camera. But due to most if the cameras on my price range being used cameras (and 100% mechanical), I was wary of getting then on ebay, therefore I’ve decided to source the camera locally, at a store that could sell me a clean and tested camera. Unfortunately such stores are 50Km away, in a neighboring town.
The camera that fulfilled my criteria without looking like a dug up, rusty fossil, or a plastic toy was the flexaret VII automat.
On paper, 1/500 sec maximum shutter speed sounds quite nice, but it also means that the camera uses a more complex leaf shutter design.
After being in a workshop on pinhole photography (held at offo, in Aveiro), and getting a grasp of the development process, it was time to actually start to build a simple camera, as well as to start developing its pictures.
But the initial results were not quite what I expected in terms of sharpness.