A few years ago I did some features on Coherers, including a test of a real coherer. Fessenden tried his hand at a different form of coherer. The usual coherer was a tube of powdered metal which cohered into a semi-solid and conducted more current when RF was applied. Coherers served as a sort of relay for the ons and offs of Morse.
As always, Fessenden decided to take the opposite road. He developed a current-centered coherer that conducted less when RF was applied. It was similar to a light bulb, with a filament in a glass bulb.
The filament was made of very fine platinum wire. Fessenden had found that the platinum raised its resistance slightly when radio frequencies were flowing through it. This was probably due to the skin effect, which wasn’t yet understood. (Higher frequencies tend to stay in the outer skin of a wire instead of filling the whole wire.)
The change was much more subtle than the full conduction of the usual coherer, so he placed a balanced pair of bulbs in a bridge. One bulb was coupled to the antenna and received RF when it was coming in; the other was out of the antenna’s influence and remained constant.
As in the sechometer I featured yesterday, the bridge is triggered by a square-wave providing transients to both sides. A tuning fork gives a steady audio-frequency square wave. The left side of the bridge is the standard, unaffected by RF, and the right side is coupled to RF from the antenna.
= = = = =
Here’s the bridge when there’s no RF coming through. Both bulbs are changing their resistance in unison, so the bridge is balanced on average.
= = = = =
Now RF is coming through the antenna, shown by a red antenna. The left standard bulb is still changing in time with the tuning fork, but the right bulb is changing much less. The bridge is unbalanced on average, and the meter is off center.
Finally, here’s the circuit in more realistic form, with Polistra and Happystar adjusting the tuning fork.
Polistra's Mill 