Hi,
The crystals may differ in frequency by perhaps one part per million which represents 1 second in about 11 days. For this case, on average the messages from two transmitters might drift into collision in about nine months. The problem is that they may then continue colliding for many days or even weeks, totally blocking the transmissions for one or both systems. The duration of interference would depend on the length of the data messages and how long it takes the receiver to recover from a "wrong" message (e.g. AGC recovery or clock resynchronisation) etc..
If the transmitted data contains some "address" information then the receiver may recognise the "wrong" data and ultimately be able to lock onto the correct transmission again. But if there is no address (i.e. the receiver just locks onto any message within the expected timing window), then the receiver could permanently lock onto the wrong signal.
When I designed a similar system some years ago, I introduced a short, pseudo-random delay/offset (implemented with a "maximal length
feedback shift register" emulated in firmware) for each message, so that never more than a few consecutive collisions could occur. If anybody discovers how (or if) the FO devices reliably accommodate multiple transmitters, then it's probably worth starting a dedicated thread.
_____
But back to the original topic, here is an internal photo of the wind vane head. I wanted to see the arrangement of the magnetic circuit with a view to improving the "equality" of the 16 compass sectors (and maybe damping the rotation as well). I will start a dedicated thread in due course, but for now it's sufficient to say that the magnet is the small vertical silver rod at the far right and the two empty recesses nearer the centre are perhaps for additional magnets if the relays fail to operate adequately on the production line. The bearing appears identical to that in the anemometer, i.e. with external and internal diameters of 10 and 5 mm, and the hole in cental "shaft" is 2.5mm.
The method I used to dismantle the head was to insert a thin, wide strip of steel (I used a 6" engineers' ruler, but a table knife or even a thin coin might work) into the gap between the rotating and stationary parts, twisting it gently and moving continuously around the circumference. After a minute or two, the gap was noticeably larger, so I continued until the head popped off. I cannot say if I was "lucky", therefore you attempt this method entirely at your own risk.
Cheers, Alan.
Login 
Register
Enhancement requests
Search
Thank you for finding the info 
