last changed 8 February 2008
Following the relocating of my SAM magnetometer sensor unit to a point some 30m away from the house, I experienced problems due to the cable length (about 50m) used. There was lots of noise on the lines from the two sensors, and there were periods of oscillations when the apparent fields varied widely over a few seconds. Adding 74HC14 Schmidt trigger inverters with RC networks on the inputs did improve matters, but I decided that a better solution was to add differential line drivers and receivers to the system. By building an external adapter, minimal changes to the SAM unit itself were needed. The results are well worth the effort involved. Since moving to my current location on the Isle of Mull, the system continues to work well with nearly 70m of cable. The sensor box (IP67) is now contained inside a wheelie bin filled with sandy soil, which is buried almost up to the lid. This provides a dry and thermally stable environment and keeps everything dry, just as well as the field it is in is basically a peat bog.
The steps taken were:
Here is the circuit of the remote sensor unit and the receiver (small .pdf files, printable in A4 landscape)
My system is interconnected by 8-way screened cable although twisted-pair would have been a better choice, I used what was to hand. In practice the nominal 100R terminations for the lines were increased to 150R, to reduce the dissipation in the drivers and their regulator, with no ill effects.
Why the 65LBC184P?? It's just about the most robust low power line transceiver on the market, with built-in transient protection, and is relatively cheap. I used 65 prefix industrial temperature range model, since it was available at no extra cost.