Frequency Control - Quartz
July 1938 Radio News

July 1938 Radio News

[Table of Contents] Wax nostalgic about and learn from the history of early electronics. See articles from Radio & Television News, published 1919-1959. All copyrights hereby acknowledged.

Before a suitable manufacturing process was developed for synthetic quartz, the mineral (silica, or silicon dioxide - SiO₂) was mined from the ground in various parts of the world. The primary source for electronics grade quartz came from Brazil. It was not until 1921 that Walter Cady developed the first quartz crystal oscillator, and George Pierce came up with his eponymously named Pierce oscillator circuit in 1923. It was not until the1950s that a commercial-scale process was developed for hydrothermal processing of synthetic quartz. This 1938 Radio News magazine article describes the manufacturing steps required for making radio-quality natural quartz crystals.

Frequency Control - Quartz

By Frank Kirby

Crystals are used by all broadcasters and many hams to control the frequency of their respective transmitters. The author tells how these vibrating quartzes are made.

Perhaps the best place to start an article on crystals is with the raw material itself. Quartz crystals are found in the United States and often in large size, but they are usually discolored and contain flaws which make them poor oscillators, hence most of the raw quartz is imported. The larger part of the imported quartz comes from South America. The raw or uncut crystal is sold by the pound; and the larger ones weigh two or three pounds and are seven to eight inches in length. The larger crystals cost more per pound, but they are more economical to use insofar as there is less waste in cutting.

The crystals are hexagon shaped and are tapered, the taper and slope varying considerably among different crystals. It is easier to work with a near perfect crystal because the location of each axis is simplified. For, the person inexperienced in crystal cutting it is advisable to coat the various surfaces of the crystal with colored paints in order that they may be readily identified after the crystal has been cut.

The axes in the crystal are as follows: the "Z" axis runs lengthwise through the crystal, the "X" axis bisects any of the corners, and the "Y" axis is perpendicular to any face or side. The sketch shows the various axes and their relation in the crystal. Oscillating crystals are cut primarily on either of two axes, the "X" or the "Y." The blanks for the crystals are cut perpendicular to the axis for the cut of crystal desired, i.e., an X cut crystal would be cut off perpendicular to the X axis.

A few words now about the advantages and disadvantages of the various cuts should be in place. The Y cut isn't used much at present unless means is provided to control its temperature. The frequency of the Y cut crystal changes greatly with temperature (usually in jumps) and sometimes it stops oscillating completely. Unless especially ground the Y cut crystal has two main frequencies of oscillation, which in most cases is a disadvantage, for the frequencies are occasionally close together, thus making the crystal erratic in operation. The X cut has several advantages over the Y. The frequency change with temperature change is less than it is with the Y and the thickness of an X is greater than a Y cut for the same frequency.

Two variations of the Y cut are the "AC" and "AT" cuts. These cuts are obtained by rotation of the crystal about a line parallel to the X axis either 31 degrees or 35 degrees, depending on whether one desires maximum output or low drift. In this instance thin slabs are obtained from which the round crystals will later be cut. Most of the material to follow will pertain especially to the cutting of round crystals. [So cut because of cheapness of manufacture. - Ed.] However, the principles can easily be applied to square ones as well.

In setting up the crystal to cut AT or 35 degree cuts it is possible to cut either plus or minus 35 degrees from the Z axis, hence it is necessary to test the trial cut before proceeding further. In the absence of a polarized light source to examine the crystal the next best bet is to grind one crystal by hand to determine if the trial cut is the correct one.

Some AT cut crystals are harmonic oscillators, especially on their third harmonic, that is; with the plate circuit tuned to three times the fundamental of the crystal it will oscillate on its third harmonic. The output with this mode of oscillation is fair but the temperature-frequency coefficient is greater than at the fundamental frequency. The AT cut is the thinnest and the X the thickest of the three mentioned. Hence it is possible to grind a seven megacycle X cut crystal which will stand up well under hard usage, whereas, the Y or the AT cut for the same frequency would be thin and fragile.

Square-cut crystals cost more to produce because a pipe drill cannot be used.

The first operation is cutting the blanks from the raw crystal. If the crystals are to be round the cuts are taken lengthwise on the desired axis, i.e., the cuts are parallel to the Z axis for X or Y cuts, the AT cut being an exception. These slabs are later cut into discs with a special cutter. On the other hand, if the usual square crystals are desired the crystal is cut up into blocks. These cuts are made about one inch apart across the crystal perpendicular to the Z axis.

The cutting of the crystal is done with a muck saw. This saw is a copper disc of about 16 to 20 gauge and six to ten inches in diameter. The saw runs between 100 and 300 r.p.m. The last mentioned speed is perhaps a little too fast, as this will cause the cutting abrasive to be thrown from the cut. The cutting abrasive is a mixture of No. 100 grain carborundum and water, about the consistency of a thin paste. The paste is thin enough to flow easily into the cut. The soft copper cuts the quartz because the carborundum which becomes embedded in the muck saw rim. The copper saw wears quite well and one disc is enough to cut up several crystals. The saw can also be used for edge grinding crystals. If the saw does not cut well carborundum is fed to the edge of the saw with a paddle or spoon. Making the bottom of the cutting mixture box slope toward the wheel also facilitates feeding the carborundum. Normally when the saw is running it extends down into the cutting mixture about 1/2 inch. The crystal has to be guided a little until the saw starts to cut.

A board is used to hold the crystal during cutting. If the crystal is of irregular shape and is hard to mount, it is mounted in plaster of Paris and fastened to the board with pegs of wood. A feeding mechanism facilitates making the cuts all the same thickness. The threads on the feeding rod are not too fine as to become clogged with carborundum. A thread of 18 or 20 per inch used.

Cutting the crystal in this manner gives slabs about three by four inches and of a thickness determined by the finished frequency. These slabs are next fastened to a flat board with wax or paraffin. This layer of the wax is melted and spread evenly on the surface of the board with a flame, then the slab is pushed down on top of the wax while it is still warm. This gives a good firm base for the slab. The circle cutter consists of a dural cylinder one inch in diameter with about a 1/32 inch wall. Copper or brass could also be used in place of the dural. The cylinder is fastened to a steel hub with machine screws. The hub has a small diameter shaft suitable for the drill press to be used. The circle cutter revolves around 300 r.p.m. or slower. The cutter should be fed slowly and be lifted when it stops cutting to facilitate feeding in the carborundum. The carborundum if stacked up around the cylinder will run in of its own accord when the cylinder is lifted or it can be fed into the cut with a spoon.

Another method to help feed carborundum to the cutter is to slit four cuts parallel to the axis of the cylinder with a hack saw. When the carborundum is fed in, these notches fill up and thus maintain the supply of carborundum in the cut longer.

The next step is edge-grinding. This removes all the cracks and nicks. Although some crystals will oscillate with nicks in the edges it is good practice to remove them, for in strong oscillation the cracks may spread. The cracks are removed by edge-grinding by hand on a piece, of flat glass with 300 grain carborundum or, by machine grinding with a revolving disk or muck saw.

Following edge-grinding, a reference face is put on each crystal. This is done by grinding one face of the crystal. Special effort is made to apply pressure evenly over the surface of crystal so that the face being ground will grind flat. The crystal is moved about on the glass in a series of small loops to keep the wear on the plate glass, on which is being ground, even. A reference face is first put on the crystal because it will not bend as easily while it is still thick. Of course, the thinner the crystal the more danger there is of grinding hollows in the reference face. When the reference face is finished, it is identified, either by sloping one edge towards the reference face, or by marking it with India ink. If a large number of crystals are being ground together a reference face is put on each one before going to the next step.

Next, the crystals are fastened to the grinding plate. A knob in the middle is used in moving the grinding plate about on the plate glass. The plate is of cast iron 1/2 inch thick and seven inches in diameter. The plate is heated over a Bunsen burner and the beeswax allowed to flow evenly over the surface, then the crystals are pushed firmly down on the plate. This plate is moved about on a large plate glass with No. 300 grain carborundum as abrasive. The motion is the same as with the single crystal i.e., in a series of small loops. After about five minutes grinding the excess carborundum should be washed off, the plate heated to melt the beeswax, and the crystal positions interchanged. Also during this operation each crystal is rotated 180 degrees. Allowing the wax to cool again, the grinding process is repeated. After four or five of such changings examination will show which crystals are thicker, or which ones are not grinding properly, and they can then be shifted to more favorable places on the grinding plate. The arrows are placed on the crystals as a help in rotating and placing them. By careful grinding and placing, and by sufficient changes, the crystals are ground quite flat; in fact, flat enough for many of the crystals to oscillate without further grinding. However, as a general rule, the crystals are usually touched up to secure greater output.

The final touching up and grinding to frequency is done by hand with No. 500 grain carborundum on a fresh piece of plate glass. Care is exercised to see that no larger grain carborundum is carried over to the finishing plate as a large grain rolling under a thin crystal would fracture it while it was being finished with the fine abrasive.

Final grinding is done with the aid of a pair of micrometers, readings being taken in nine or ten points over the surface of the crystal. These readings are written down. It can be seen at this point that a reference face is highly desirable. With thin crystals a cloth is laid on the table under the crystal to prevent accidental breakage should it be dropped. After the readings are written down a circle is drawn around the low spots and a cross put on the high ones, thus indicating at a glance just where grinding is necessary. Overall grinding is accomplished by pressure with two fingers and the thumb. The high places are ground down by pressure on the spot with the tip of one finger.

The Y and AT cuts are ground slightly convex; that is, the edges are about 1% lower than the middle. The X cut is ground flat or for slightly better output slightly concave (about 1 to 2%). Grinding thick cuts concave is a bit difficult because the crystal will not bend sufficiently to allow grinding the middle. In the above case middle is ground down somewhat by grinding over the corner of the plate glass.

While some crystals will oscillate when "out of parallel" by as much as 10% they are usually ground within 2% of parallel.

Posted January 5, 2022