The early Audion vacuum-tube detectors were expensive, unreliable, and short-lived. But for many experimenters, they were also irresistible. This extract of Carl Dreher's memories of his adventures as an early radio amateur comes from his monthly "As the Broadcaster Sees It" column.
Radio Broadcast, September, 1926, page 408:

Memoirs of a Radio Engineer, Part XIV (Carl Dreher)

    We left off, in the July number, with the founding of the radio club at the College of the City of New York. Among the pieces of equipment there I recollect a long-wave regenerative receiver, with coils of annunciator wire wound on cardboard tubes about three feet long. Mr. F. T. Dickey, at about the same time (1915) had a private set of the same type, on which he let me listen to Glace Bay and other sustained wave stations. These, however, were not my first experiences with vacuum tubes. I had had one, in fact two, of my own. But they were gone. They had died, so to speak.
    Nowadays janitors and poor working girls own five-tube sets, but ten or twelve years ago the possession of a single tube was a mark of distinction, among radio amateurs, equivalent to ownership of a Nubian lion by a poor Roman, say. The comparison is slightly crippled, but it has this in its favor--a Nubian lion will consume the entire income of his master, if the latter is in middle circumstances or below; and the audion bulbs did just that for the poor radio addicts of the pioneer days. First, there was the job of getting one. The early tubes were not sold outright; you were entitled to buy one only if you were the owner of an amplifier, which cost several hundred dollars for some of the models, and less, but not enough less to come within reach of the average amateur, for a single-stage outfit. However, if you brought a burned-out tube to the manufacturer, that was accepted as prima facie evidence of ownership of an amplifier. The company then graciously sold you a bulb for $5.00 or $7.50, depending on the type of filament. Naturally there was a thriving trade in burned-out tubes; it was almost as great a triumph to get a defunct audion as a new one, inasmuch as there was no possibility of attaining the consummation of desire without going through this preliminary step. Anyway, there was not so much difference between good and burned-out audions, the former turned readily into the latter. My first lasted about fifty hours; my second, about forty minutes. But it had an extra filament, Allah be praised. The second filament, unfortunately, while it would light, showed little merit as an emitter of electrons. The best I could do on it was NAH, the Brooklyn Navy Yard, about six miles away, and he was not as loud as on a crystal detector. I therefore rescinded my praise to Allah, and lamented bitterly and lengthily the loss of my five dollars--a Christmas present. I wrote to the manufacturer of the tube, but without obtaining satisfaction. No doubt he had his own troubles, and it is probable that he needed the five dollars as much as I did. I then tried to weld the ends of the broken filament together, by applying the battery voltage and slapping the bulb in such a manner that the ends might make contact and stick together. I have never concentrated more ardently on any problem, in the years that have passed since this audion betrayed its trust, than on that sixty-fourth inch air space which showed through the glass between the broken ends of that filament. As I gazed fixedly at the small but overpowering gap, it sometimes seemed as if it must close up in obedience to my will, but the laws of matter were not suspended in my behalf. Surrounded by the protecting vacuum, the little gap continued to flaunt its presence in my face. If I ever took any stock in faith healing, it disappeared then.
    More desperate measures failed likewise. One of these methods of resuscitation was to apply the voltage of a spark coil secondary to the filament terminals. Sometimes the ends would leap together and stick. In my case they did not leap together, and hence they did not stick. I used a larger spark coil, borrowed from another surgeon. I reversed the polarity. I reversed it back again. Finally I wrapped the audion in tissue paper and went back to my crystal detector. The faithful crystal took me back to its bosom, but I could not love it as before. As usual, faithfulness was unrewarded. My heart and imagination remained with the scintillating and coquettish audion. It could be unfaithful and costly, and get away with it, because of the incomparable moments it brought its admirers. For, when they were working right, the old audions were not to be despised as detectors. They were imperfectly evacuated, and their characteristic curves sometimes had kinks and loops where the rectification was first rate. It is true that the tube factory had little more control over the operating characteristics of their product than over the annual frequency of sun spots, but this element of gambling in the purchase of an audion only added to the thrills. Flung into deepest despair by the demise of a beloved tube, or the failure of a new one which never worked at all, the audion speculator would save up his pennies and plunge again.
    The early vacuum tubes were rounded into graceful shapes, like a Greek vase. They were not severely rectilinear, like the standardized, uniform, efficient products of the present day. They were not made by machine, but by hand, and sometimes, apparently, the hand was not a steady one. No two tubes looked alike, and few acted alike. The innards of the creature were wide open to view, for they were not within a cylindrical plate element, as nowadays, but in flat shape. The plate was a metal tab about half an inch square, then came the zig-zag grid, somewhat askew, and then a feeble looking flat loop of wire, the filament. Insulated flexible leads issued from the bulb, the grid lead green, the plate red, and the two filament leads usually in plain rubber. It was a romantic looking object, and no mistake.
    The amateurs took strange measures to increase the sensitivity of their audions. They would heat the glass tube over a Bunsen flame, until the wall softened and an indentation resulted, in order to regulate the vacuum. Another stunt was to suspend a good sized horseshoe magnet with the bulb between the poles, thus subjecting the electron stream to a magnetic field. This sometimes increased the sensitivity remarkably. Some latter-day genius is going to rediscover this trick one of these days, and have his afternoon in the radio supplements.
    Some, apparently, of the more sensitive specimens regenerated in the circuits in which they were used. I recollect one I had in 1913, which brought in Boston stations in New York in daylight. This tube, like many of the audions, had chromatic qualities. When at work it glowed internally with a delicate pink-violet light, like a rather refined and emasculated Geissler tube. Irritated by strong signals, it would turn a deep blue in synchronism with the dots and dashes, so that, with the room darkened, one could read the signals visually. It was a sweet and charming companion, and I hereby dedicate this article to its memory. I regret that I never gave it a name.