A new upload at American Radio Library opens up a fascinating Road Almost Taken.
Carborundum is synonymous with abrasives. The brand is still active, still making a variety of papers and grinding wheels. Needless to say, their original factory in Niagara Falls was closed by EPA, like all factories outside of China. The brand is currently owned by Saint-Gobain, an EXTREMELY WOKE company ferociously compliant with all ESG buzzwords. I can’t tell if they have any real factories outside of China. One former Carborundum factory in America seems to be running under a different company.
= = = = =
Long before anyone imagined the nightmares of EPA and ESG, Carborundum tried to become the universal radio material, with some remarkably clever technology. In the depression they retreated to their main skill, which is always the smartest move for a company; but the smart move lost an interesting opportunity.
The alternate road started in 1906 with research at DeForest, which was also developing the vacuum tube. Dunwoody, formerly of Signal Corps, was trying various crystalline materials to find the best detectors. Carborundum is silicon carbide, one atom of silicon and one atom of carbon, and it showed a mixture of the characteristics of both elements. Carbon for resistance and silicon for semiconductance. Dunwoody found that a single Carborundum crystal, not the familiar mass of compacted crystals, worked as a detector.
Carborundum wasn’t the most sensitive material, but it had an interesting conductance curve.
Tanh. Exp at start, then saturation. Familiar semiconductor diodes generally have an exp curve. This tanh curve is an ideal response for an audio limiter or compressor. It’s also Nature’s favorite curve, the response pattern of neurons. Carborundum-based decision modules could have formed a more natural computer, and of course they would be SMALL.
The detector was sold as a module, with the biasing battery and the variable resistor and a bypass capacitor all in one unit. The knob was meant to be a front-panel control along with tuning and volume and bandswitch. Adjusting the bias effectively pulled the start of the curve down toward zero, for increased sensitivity. Adjusting could also move the nonlinearity up or down for different response characteristics.
Here’s the circuit of the detector module.
The whole module would be placed in a receiver in the same place where a conventional diode would normally stand, typically in series between the RF tuner and the audio amp stage.
The module appeared as unidirectional, with an effective cathode and anode, because of the bias. Internally, the crystal itself was bidirectional.
= = = = =
Carborundum also made resistors as tubes of varying thickness, using the carbon side of the material’s personality. These had a clever solder-or-screw connector, making a solderless receiver possible. Hypothesis: These might have failed because carborundum has a negative temperature coefficient. Heat makes resistance decrease, causing more conduction. Most materials have a positive coefficient, so they tend to self-regulate. When increased current warms up the material, resistance rises, counteracting the increased current.
The British branch of Carborundum carried the radio ball longer and farther than the US branch. Several articles in Wireless World were discussing and recommending the detector in the late ’20s, and the British branch made some modules that didn’t appear in the US brochure.
This module included all the resistors and capacitors needed for coupling between tube stages. Note the Big Chief medallion proudly declaring an American invention! Radio parts were openly visible in the pre-streamlining era, so it paid to advertise.