March, 26 2010:
Progress report.
Hardware |
||||||||||||||||||||||||||||||||||||
|
Software
(embedded) |
|||||||||||||||||||||||||||||||||||||||
|
Software (PC) | |||||||||
|
March, 8 2010:
Modem and debug.
The porting of the TCP/IP library on the ATMega644 is completed. With a lot of optimization, we can put all the FAT32 code and the TCP/IP stack in Flash memory and in SRAM too, with all the rest of course. I reduced some oversized buffers, and I put all constant strings in Flash (without copy in SRAM). There is a little bit less than 1 Kb of SRAM for the normal functioning of the system. By cons, there is a limitation to a single socket, thereby eliminating the possibility of an FTP client. So I started doing the code to send mails with attachments, since the protocol is extremely simple.
The TCP/IP library must be tested now. I had a new problem: we need 2 serial ports (one for the modem, one to debug). So I used my old prototype card to use the SPI port and receive debugging data, and forward it in the 2nd serial port. Maybe I lost a night to do that, but I know from experience that I'll win a lot of time later to understand what is happening.
Febuary, 7 2010:
Some news.
Some information on the project progress:
- I finished the design of the main board schematic with Eagle. It has been validated with a prototype, it should consume about 4.5mA maximum including Geiger counter, magnetometer, SD card and the circuitry of the RS232 port.
- I have almost finished porting a library "TCP/IP stack over PPP protocol". I found it on http://www.avrfreaks.net/, it will provide a connection on an old GSM/GPRS modem found on eBay. For now, I try to test it with Windows to be sure it works well. This modem will give a remote access, for example to a FTP server to send daily data, or to send alerts by SMS or e-mail in case of anomalies. But there is an unknown parameter: will the microcontroller have enough RAM and Flash ?
- On a side project, I made the schematic of an analog barometer, although I received a digital version (very low consumption), which has the disadvantage of being a bit expensive for an optional sensor.
August, 20 2009:
Still alive.
No post since 3 months, but I'm still working on the project !
I found 9 crop circles in England during my holidays
!!! (and they all seem man made ...)
Crop circle near East Kennett (near Avebury, Wiltshire) - June 2009
June, 1 2009:
Portable digital magnetometer/Geiger counter.
Before doing my final «standalone box», I'll try some parts of the project in a portable box. I was quite disappointed (and happy in the same time) when I found this board on Sparkfun. For $17, you can have an AVR board with a 3.3v/5v power supply, RS232 serial communication, a reset button, and a generic button. My prototype board cost me more in time and money... So, I modified this board to have a SD card socket, a MicroMag3 socket, a buzzer, a LCD display, a keypad and an input to count radiation rays detected by the Geiger module. All works in 3.3v except the counter. Here is the result :
I'll post the final schematic/code later. I prefer to fully test it before .
May, 12 2009:
First shematic with KiCad.
«KiCad is an open source (GPL) software for the creation of electronic schematic diagrams and printed circuit board artwork». I tried it, it's not very intuitive to work with, but it's quite fast to learn, it works well, and it's free . So, here is the 1st module, the Geiger counter:
You can see 3 parts: an oscillator with a feedback, the high voltage generator with the coil and the cascade (voltage multiplier), and the tube with CMOS gates to get a clear signal, large enough to trigger an interrupt. Be careful for 2 reasons : you have 500v in this circuit, and it's not fully tested. Currently, I tested it only with replacement components, because I didn't find Maxim chips. So the power used is 2.5 times higher. Ok, it needs 0.2 mA on 5v instead of 80 µA ! All components may be found on Mouser except the Geiger-Muller tube (for example, here: SBM-20).
May, 7 2009:
Short update about the magnetometer.
Here is a small graphic to see the reading of the magnetometer:
March, 9 2009:
Magnetometer and Geiger counter.
I finally achieve to build the Geiger counter with a very low consumption high voltage power supply (450v). The prototype is ugly, but works perfectly. I have ~11.3 cpm (counts per minute) in my flat which is normal (fortunately ).
Here is a short demonstration:
As for the magnetometer, it also works. I have no photos, but it can «see» a small speaker magnet at about 1m.
Febuary, 14 2009:
Components.
I received all my electronic components: low power LCD display module, barometer, Geiger power supply parts, digital temperature sensor, SD card socket, MicroMag 3 magnetometer module, new AtMega chips, serial RAM, etc. Tomorrow, I'll test SPI devices, because it's easy. I love digital electronic .
January, 27 2009:
ADC & magnetometer.
I tested the ADC of the AtMega. It's quite bad... In reality, on 10 bits, 8 are significant. Bad news for the magnetometer which need the best precision. The good news is we don't care . I found a better magnetometer. Instead of 1 or 2 Honeywell HMC1001/HMC1002 (at least 20$ each), I'll use a 3-axis module : PNI MicroMag 3. Results are on 16 bits, on 3 axis, fully digital thanks to its SPI port. And it uses 10 times less power. A drawback is the sensitivity: 150 µGauss (Honeywell: 27 µGauss for HMC100x).
January, 24 2009:
FAT32 support + prototype.
Here is the prototype. There's nothing new compared to the breadboard, except the fact that now, wires won't disconnect randomly . I added one more 7-segments LED display to debug with decimal numbers, and switches to stop external interrupts. The MCU woks in 3.3v (compatible with SD flash card), but the RS-232 part and the display use 5v.
|
January, 13 2009:
DCF77 signal.
I still don't release technical information, because I wait until I'll have something to show. The good news are, now, I can receive the DCF77 signal on my Atmega644. It's important to have a time associated to data. The code comes from www.captain.at, I just changed a few lines to adapt registers names. By the way, I also put a complete library to support FAT32 on SD-card. It compiles, but I didn't test it yet, it's the next step .
September, 29 2008:
Solar cell and potential available power.
Before building sensors circuits, we need to know how many average power we will have. I found that we can have at least on 100-130 W/m² on *major* parts of the world (major = where people live). It's an average on 24 h (over a year). My solar cell can deliver 7.5V at 3 Watts at 1000 W/m² (it's a monocrystalline cell, efficiency: >15%). If we receive an average of 100 W/m², a simple calculation gives 3 x 0.100 = 300 mW (40 mA, 7.5V). If we have a power regulation loss of 50% (battery charge/regulator), we may have 5v at 20 mA (day and night).
Of course, it's a very simple calculation. It doesn't take in account battery efficiency (which varies with time and temperature), solar cell orientation, reduced power consumption with sleep mode or functions deactivation, etc.
August, 23 2008:
It’s alive ! The brain is working !
I successfully programmed my ATMega644 with PonyProg2000 from www.lancos.com. PonyProg is a free tool to program serial devices. I also used WinAVR to build a minimal program (a blinking led). It’s available here: http://winavr.sourceforge.net/. All you need is to configure a “Makefile” for your ATMega644 and then, to create a C/C++ project. I will put more details later.
I built the parallel port programmer with a prototyping board and the main circuit on a breadboard. You also need a standard parallel cable (prefer a short one).
Go to www.lancos.com to
have original information/schematics.