These comments may be pertinent to a number of people who have contacted me privately. This is just a start but may be enough to get some people going. Rather than an inductive loop per se, you can probably get by with inductive transfer using open ferrite cores. The smallest ferrite U cores readily available will suffice - typically about 10mm along the bottom of the U. A resonant coil is wound on this core (usually on the flat part of the U and another resonant coil on the other U . By driving one coil with a signal (square wave at resonant frequency will do) you will be able to induce a substantial voltage in the other coil at distances of 10's of millimetres. 11111 22222 11 22 <-- 2 cores with poles facing 11111 22222 If the coils are not resonant you can still get substantial transfer by bringing the 2 cores close together but over a much smaller distance. This may be adequate in your application and can simplify operation. With an unknown core, resonance is easiest to identify with a sweepable oscillator. Knowing the coil inductance is a help :-) Resonance is given by the standard formula f= 1/(2 x Pi x sqrt(L x C)) You will need to refer to standard tables or charts for the core you use. As a guide around 30 turns on a U10 core will resonate at 20 KHz with a 0.22 uF capacitor. _ |_| |_ --RRR-- .------ | ] Cap ---- [ L ---- ] | [ --------------------- Drive resonant tank circuit with signal near its resonant frequency. In this sort of application signal levels can be made large enough to detect with a simple diode plus capacitor detector. Signalling can be achieved by turning the drive signal off and on. Data rate should be low enough that there are a number of cycles of signal at the resonant frequency for each data bit. Much faster signalling schemes with a data rate at about the resonant frequency are possible but not worth pursuing in this context. How to get started: Get 2 small U cores Wind 30 to 50 turn coil on each core. Apply say 10 KHz signal to one coil via a series resistor - a PIC can generate this signal. Preferably using an oscilloscope, look at result of moving cores into close proximity and away. This is basic transformer action but is probably all you are going to need for this application. An inductively coupled loop circuit could be added but is not really appropriate here. Resonate both coils (apply capacitor across coil) Compare results as above. If you are uncertain of the inductance values the resonant frequency can be found by sweeping across a range of frequencies and watching the voltage across the inductor. A simple PIC program will allow you to do this. In some applications a simple "diode detector" will be enough. As signal levels fall amplification will be required. I'll comment on this in due course if appropriate. Feel free to ask further relevant questions. Russell McMahon _____________________________ >From other worlds - www.easttimor.com www.sudan.com What can one man* do? Help the hungry at no cost to yourself! at http://www.thehungersite.com/ (* - or woman, child or internet enabled intelligent entity :-)) -----Original Message----- From: dan.lloyd@gb.abb.com To: apptech@clear.net.nz Date: Tuesday, 25 January 2000 21:38 Subject: Inductive signalling > > >Hi Russell, > >First of all, thanks for your offer of help. I am trying to knock together a >little project using inductive loops mainly because it is something I know >nothing too much about and I see quite a few problems with it, which makes it a >bit of a challenge (like high frequency noise near to the micro etc). > >I basically want to make a home brew inductive signalling scheme where I have >something along the lines of a 12C509 embedded in a typical domestic light >switch that is capable of being talked to by (and can reply to) an external >"master" device (probably a 16C74A - cos I have a few kicking around that arent >in use), using the same 'home brew' scheme. The now "smart" lightswitches will >be able to turn the lights in an empty house on and off in some kind of >semi-random manner, depending on how they were programmed. Yes, the project is >cheesy but it is a learning exercise! : ) > >Comms speed isnt a priority and half duplex will be fine, over a range of about >4cm at maximum. Obviously, this range is not through free air but through some >kind of polycarbonate (whatever they make the light switch facia out of, these >days) and in very close proximity to the 50Hz, 240VAC UK mains supply. Space >would be limited, I think, but power would not as the 12C509 would be powered by >some rough and nasty method from the mains supply. I see the 16C74 being powered >from its own power source - there is no requirement to power up the 12C509 via >the inductive loop, not when there is a viable supply source available. When the >mains is not present, there would be no need to talk to the slave device, which >would be maintaining its RTC from probably a supercap or similar. (This >lightswitch casing is getting very crowded!) > >I wanted to try to steer clear of "off the shelf" RFID and similar products and >do something that I could *learn from*, basically; I've had a look around on the >Net for theoretical and example applications but it seems that they are a bit >sparse ( I have a Microchip app note that explans the general principle behind >the RFID stuff - but it assumes an RFID transmitter being used).....anything to >get me going would be very greatly appreciated. > >I hope you can help or point me in the direction of some useful information, > >TIA, > >Dan > > >