Low Frequency antenna design and issues

[picclock]

Hi
This is just an idea I'm trying to fill out for a small robotics project. Its aim is to detect the angle of the transmitter from the device, the range being up to 20 feet.
The scheme is to have a local very low power transmitter in the 100-200kHz range, and use dual ferrite rod aerials angled at +-40 degrees horizontal or so to detect the signal. The relative signal strengths could be measured and thus derive the angle to the transmitter. For example if both strengths are equal the transmitter would be at 0 degrees.

What I need to know are options for the transmitter antenna. Will a ferrite rod with a coil of wire suffice? I thought to mount it vertically.
Is this the correct frequency range for this sort of thing. I want to avoid signals being reflected as this would screw it up.

I was planning on encoding the rf signals with am pulses to make sure the correct signal would be acquired, and also just transmitting a 100mS burst every 1.5 secs or so.

Any advice or comments most welcome. I have no problems with hardware or software for this but the rf bit is daunting.

Best Regards

picclock

[TheUnnamedNewbie]

This would very much be a near-field device, as your wavelength at 100-200 kHz is in the range of kilometers.
I don't think antenna design is the right way to approach this as a result.

And yes, you might not have real 'reflections', as you need a propagating wave in order to have those (which you don't) but the field will be strongly influenced by metals and dielectrics in the area, so you are just shifting the problem from a far-field one (reflections and using ray-models) to a near field one (field distributions and impedances)

[jmpowell]

What you're describing is similar to magnetic head tracking systems used in military helmet mounted cueing systems. You'll also find the approach used to track the position and orientation of medical devices stuffed inside people. The antennas are basically square chunks of ferrite with orthogonal windings on them. So, basically the approach works and is well researched and documented. (Google is you friend) Note that the transmitted field has a dipole pattern and is curved. Your measurement will give you orientation relative to the field pattern not to the walls of the room the sensing antenna is in. Also note that metallic objects in the field will distort the shape of the field. In military applications this requires a calibration map. 

[picclock]

@jmpowell, TheUnnamedNewbie
Thank you for your interest. I failed to mention that this is for outdoor use, although predictable behavior in the presence of chain link fences and brick walls is desirable. Sunlight and darkness would cripple optical systems, as would rain. Ultrasonic systems have very poor off axis results, and whilst I could rotate the sensor, response would be very slow.

I can see that a 'square chunks of ferrite with orthogonal windings' would give very good results, especially for close distances. Your references to head tracking systems was most useful, although I am only sensing in one (horizontal) plane.

From the helmet tracking info it would at least seem that the idea is viable, which is a good start.

Many thanks

Best Regards

picclock






 


 

[coppercone2]

he means a head tracker in a helicopter or airplane that has a largely aluminum cockpit and is calibrated for the pilots head position.

I don't think you will get anything near predictable behavior outdoors

[picclock]

@coppercone2
>>I don't think you will get anything near predictable behavior outdoors
Can you explain the reasoning behind your statement ?

I found the headset in a cockpit idea useful clarification. The only difference between an enclosed and open environment is that inside you have more reflections, versus outside more interference. At least AFAIK.

I ordered up some receiver chips, SI4734-B20-GM, to check it out. These seem very good devices, which operate over a wide range of frequencies and interface well with a processor. It should be possible to get the rssi values by reading it directly from the chips.

Best Regards

picclock