The Quenched Spark Gap
"...and here's the practice o' it."
The quenched gap, as popularised by Telefunken, was a great success owing to the very high rate of switching which lent increased efficiency to their transmitters. It was said that a 500 watt Telefunken set with its quenched gap could outperform a Marconi set of 2.5kW with its rotary gap, and this was mostly down to the efficiency of the Telefunken gap, an early and poor quality illustration of which was included in "Alternator, Arc and Spark."
I decided at the outset of my attempt to replicate a spark transmitter that I wanted to use a multi-plate quenched gap and hence had to set about making one. At the time, I could not obtain copper sheet in the sizes I wanted, nor copper rod at all, so the whole thing was made from brass. The electrical and thermal conductivities of a metal are related through the Wiedemann-Franz law which states that the ratio of these conductivities is a constant, independent of the metal and varying only with temperature. Regarding electrical conductivity, the difference is not enough to cause concern as the skin effect gives some degree of compensation (as will be shown later) and the periphery of each gap is large, but regarding the thermal conductivity of brass I was a little concerned that the inferior thermal properties might be the ruin of my plans. I need not have feared. For the relatively short time that I fire the thing up, the gap never gets more than tepid at the most; however, had I the choice and was starting over again I would use copper, although it is much less pleasant to machine. Aluminium would also be a good choice for the cooling flanges, though probably not for the gaps.
The cooling flanges are about 15cm/6 inches in diameter and are made from 1,6mm thick brass sheet. The eight flanges had to be cut by hand using a 52 tooth-per-inch piercing saw with occasional lubrication with cutting fluid. This took a long time as may be imagined, each flange having a circumference of 47cm/18¾ inches, all eight amounting to sawing a line 3¾ metres/12ft 6 inches long, and consumed several weeks' worth of evenings and weekends, with considerable wear-and-tear on elbow and wrist joints and many tired fingers! The blades are fortunately not expensive, which is just as well as each blade lasted little over one flange.
The next stage was to mount all eight flanges together, drill a 1/8 hole through the centre of all eight on a friend's drill press (thanks Dennis, G7OGN) pin them to prevent their moving relative to one another and attack the edges with a file until they were all about even; an electric drill with carbide roughing disk speeded this process considerably, but flying burrs necessitated good eye protection with polycarbonate safety goggles. Following the smoothing of the edges and removal of many sharp burrs with suitable care to avoid impaled fingers, not always successfully, the eight flanges were mounted together on the trusty Unimat 3 and three holes drilled at 120°, each hole starting out one eighth in diameter, and by progressive stages being enlarged to half an inch, pinning these holes in turn to prevent rotation of the flanges between or during drilling operations. Drilling these three holes necessitated keeping one hand on the Unimat motor (for those who are not aware, the Unimat 3 motor is not continuously rated) and switching off when it became too hot to hold, which was often. This was a long, slow, laborious business, even with plenty of cutting fluid. That brass sheet is tough old stuff, and there's 8 x 1,6mm = 12,8mm (a bit over half an inch) of it in total to be got through.
The next job was by way of light relief. The brass cooling flanges were now supported by nailing them very loosely to a piece of scrap wood through their centre holes and having masked their centres (where the electrical contact with the gaps would be made) spray painting them with black matt barbecue paint. There are chemical methods for blacking brass, which I would use if I had access to the chemicals (my PhD is in chemistry!) but sadly I didn't at this time so had to make do with paint. If the authenticity bug bites very deeply, I may just strip all the paint off and do the job properly, but I doubt even I could tell much difference.
Other tasks performed, but not photographed, were:
the cutting and drilling of end plates from Tufnol paper/phenolic composite (takes the edge off a plane in minutes, giving the blade the appearance of having been attacked by a file) this composite being chosen as it looks very like Bakelite;
turning, cutting with a die, hardening and tempering the pressure screws and their mounting nuts, which are used to compress the gaps tightly together;
turning thick fancy brass washers to place under the nuts of the tension studding;
cutting and trimming Tufnol tubing to act as insulators over the studding, the studding and insulator passing through the three half inch holes in each cooling flange;
sawing with an eighty-teeth-per-inch piercing saw blade mica washers out of two inch mica disks 1/16 thick. These had to be smoothed down to a flat, parallel profile and this was very tedious indeed, being accomplished by a combination of rubbing the washer against a piece of sandpaper stapled to a board (and losing half the fingerprints from the tips of my fingers in the process - ouch! - and the blood on the mica doesn't help its insulating properties either) and holding the washer in the three jaw chuck of the lathe, which was free to rotate, and bringing up upon it a small grindstone spinning at 4000rpm in the drill chuck (shown above) - unfortunately the soft mica rapidly plugs the grain structure of the grindstone. I started with forty disks and ended up with about fifteen washers between 1-1,1mm thick from which to select the best ones for the gaps - obtaining complete uniformity was impossible;
turning twenty brass disks, one and a half inches diameter and around a quarter of an inch thick to act as the electrodes for the ten gaps. Ten gaps was the number selected on the basis of 1kV per gap. This operation was photographed but unhappily the flash unit mistimed and the fault was not noticed until the job had been done and the film then developed. A sketch of an electrode disk is appended. The recessed groove was cut later as a result of experience and it very effectively prevents the spark from "walking" its way through the insulating washers. The height of the protruding sparking surface was matched to the washer giving an overall gap length of around 8mm for 10kV, or 0,8mm/ 1/32 inch per gap. The rim was progressively reduced in thickness in order to increase the pressure applied to the washers (pressure = force / area) and improve the sealing, but obviously the strength of the brass limits the extent of this thinning process. These gap electrodes were silver plated by a local firm on the basis that a chapter on spark transmitters in an old book I have says the electrodes were silvered - the silvering lasted mere minutes under normal conditions of use, so the electroplated layer of ten microns was obviously far too thin. The finished gaps do look nice though.
The photograph above shows one completed rack of five gaps at the top, comprising four cooling flanges and ten silvered brass electrodes, each pair of electrodes separated by a mica washer, barely visible as five thin black vertical lines between gap electrodes. The end two gap electrodes have no cooling flange and are accordingly made much thicker; contact to these is made by 4mm radial holes into which a plug can be inserted. Below the completed rack lie the components for its partner. The blue thing at the top centre is a tube of silicone high vacuum grease with which to help seal the gaps to the washers (not entirely successful.) Immediately below the tube of grease are four mica washers and to the left and right of them the two Tufnol end plates. The Tufnol plate to the left has the pressure screw resting horizontally just above the captive nut through which it normally passes, whilst the right hand plate simply has a hardened bearing for the stack of gaps to rest on. The four cooling flanges are on the outside to left and right. Below each Tufnol plate are four gap electrodes. Below the four mica washers is an assembled gap consisting of two electrodes (one invisible underneath) and a mica separating washer, the rim of which can be seen. The mica washers are translucent and in operation, dim light from the discharge can be seen to cover the whole area of the discharge surface. A small hole is visible on the back of the upper electrode. These were machined in pairs, one of each pair having a hole, the other having a pin, and these pins are pushed through similar holes in the centre of the cooling flanges and the partner electrode then mounted on the other side of the flange. The centre holes in the flanges do not really show up in this photo, but they do on the picture showing the flanges mounted for spraying as there is a nail through the black masking circle on each one. The spacing of the three half inch holes on the cooling flanges is such that when the three Tufnol insulating tubes are passed, with their tension studding, through the flanges, the tubing just clears the two inch diameter mica washers. Below the mounted gap are the three horizontal insulating Tufnol tubes, to the right and left of which are the six (total) fancy brass washers. Below the three tubes are the three lengths of studding and below them, six nuts and six plain washers. The studding, plain washers and nuts were purchased. Everything else was made. In all, this project occupied around four months of spare time. Around a kilo of scrap brass was produced, a small mountain of scrap mica, and large numbers of piercing saw blades and Unimat drive belts were broken.
These gaps are NOT for sale!
The next section looks at some questions of efficiency.