Originally Posted By: wfaulk
Okay, why does it need a separate input for the one channel? ... Does he need an antenna pointing in a different direction than for the other channels?

Bingo!

We refuse to pay monthly fees for television signals when they are free over the air here (including some HD channels). So we have a small tower with three antennas on it:
  • one home-made dipole, aligned to receive nearly all local analog channels, as well as the CBC-HD channels.
  • one very large/fancy array pointed at the nearest distant PBS station, soon to be rotated towards a PBS-HD signal instead.
  • one UHF antenna pointed specifically at the nearest transmitter for TV-Ontario (TVO).

Our Myth PVR has two HD tuners, and two NTSC (analog) tuners. The latter are normally wired to a combo feed from the first two antennas. The third antenna is not normally connected, because it degrades the other signals when combined with them.

So, we've been using a manual switch box to select antennas for the second NTSC tuner in the Myth box. But we don't always remember to flip the switch in advance of the next recordings.

I've been wanting to automate this for some time now, and so began the effort with a simple relay kit that I found in Toronto for about $15.

At first, I tried wiring the RF coax directly into the relay, but this resulted in more noise and ghosting, and was deemed unacceptable (our reception is normally near crystal clear on all channels, better than some digital cable feeds..).

To keep the signal quality high, we had to stick with a high quality (low loss) sealed-unit RF switch. But how to actuate it? A pair of electromagnets came to mind, but the switch can be somewhat sticky at times.

On Saturday, I stumbled across this geared 5V motor in a local shop, for $7. They also had a nicely mating rubber tyred wheel to fit, for another $5.

And thus the contraption pictured at the top of this thread.

The center contacts of the relays are wired to the two motor poles. The two signal contacts of each relay are wired to GND (normally closed) and +5VDC (normally open). So if both relays are off, the motor is grounded, and nothing happens. If both relays are on, the motor gets +5VDC on both poles, and nothing happens there either. But if a single relay is "on", the motor spins one way or the other, depending upon which relay is "on".

The two relay control inputs are wired to a couple of data lines from the Myth box's parallel port, and the same box's PSU gives +12VDC (for the relay coils), and +5VDC (for the motor) in series with a diode to prevent backflow.

To complete the package, I added an "external channel change script" to Myth for that NTSC tuner. The script just pokes values to the parallel port to turn on one relay, wait 200msec, and turn it off again.

Whirrs like a charm!