Home | SW Projects | HW Projects | People | Web Links | Contact |
Place the both cables close together to prevent ground loops.
Do not make a gap in no cable's shielding, otherwise you make the whole device radiate awful EM interferences. I tried to solve EMC by cutting one shield of the pair, Ronja worked but the packets were audible all over AM bands. When the gap has been shorted, the AM interference absolutely clearly ceased.
The holders must be attached to something as rigid as possible. Building wall is fine, including a wooden building. Take care of twisting wooden boards or something similar. Mounting plugs into a masonry wall and adding screws and the holder is probably the best mounting possibility.
Do not mount the optical heads above a surface rain can fall on. The drops spatter and get onto the lenses from the bottom. If you placed a hood from the bottom, the drops would spatter from the hood. So there must not be a hood on the bottom of the device and enough free space for the drops to fall through. I thinks about 40 cm should suffice. I realized this danger because I mounted the prototype just abouve my balcony railing. When rain comes, it ususlly hits our railing and the packet loss makes the link unusable. When rain falls in such a direction that it doesn't hit the railing, the transmission is completely OK.
Try not to run the beam over a surface that gets hot in summer days because the atmospheric flicker would decrease reliability.
When feeding the voltage from the wall cube to the optical heads, try to use as thick and as short wires as possible because long and thin wires cause a voltage drop so you get for example only 8V instead of 12V at the end when the two hundred milliamps flow through the wire which may not suffice.
Do not remove the anti-sun protection. Otherwise you are under a risk of fire.
Be careful that sun doesn't get just into the signal path. In such condition the optoelectronics will burn out and the link will be destroyed. Try to skew the path in this case (choosing the other end of building or similar). If you can't avoid this, compute which days are critical (from the aastronomical tables -- I don't know how to do it) and obscure the optics on these days. It sould be about two days a year I guess.