Crystals for the Central Electronics 10B

Now that the 10B is alive again after what was probably several decades I figured it was time to make some crystals for it the old fashioned way. Yes, I know what a VFO is, but that is not the point. I’m a sucker for the lost art of quartz crystals. There’s so much mystery and old wives tales around that that it has become some sort of dark art. In my limited experience its not magic, so long as you know your materials and keep in mind the laws of physics, and obey them.

My immediate goal was to make a crystal that, when mixed with the 9 MHz local oscillator in the 10B, would put me on 1945 kHz for the AWA Top Band net. This is an SSB net, so I would need to be within 100 Hz or so to be tolerable. This is more precision than I would need for AM phone. This is an antique radio net, so it ‘s not uncommon for people to drift off a little, but I don’t want to stick out. I had a 7050 kHz plastic/Bakelite holder that was a good candidate for the nominal 7055 kHz target frequency required.

Warning…what I am doing here involves dangerous chemicals. You should not attempt this or anything like it without a full understanding of the chemicals involved and the safety practices that go with them.

I should mention why I’m etching this and not getting out the abrasives to physically grind the crystal. There are two reasons for this. This crystal was presumably already etched to frequency, which means that its surface is relatively smooth. Any physical grinding at this point is going to be very hard to control at the sub-kHz level. Secondly, any grinding is going to result in a rough surface, and this is likely to result in shedding bits of quartz off the surface during use (this is why crystals typically age upward in frequency). By etching only, I avoid both these problems. With a larger move (>15 kHz) I would consider physical grinding then etching to frequency.

Because the LO in the venerable Central Electronics 10B is not exactly 9.000 MHz, I decided I would need to test the crystal in the radio itself to get on frequency. My LO is about 500 Hz shy of 9.000 MHz, which would make it easy to overshoot the target frequency if I didn’t do this. Given that the crystal had already aged up to 7051.4 kHz and the LO was on 8999.5 kHz, I only need to move about 3 kHz, not 5 kHz as all the labels would lead one to believe. Like old tubes, nothing beats testing it in an actual radio…

I much prefer the FT-243 holders with metal lids and retaining nuts rather than these Bakelite ones. These plastic holders always seem to be more flimsy, and I worry about striping the threads with all the disassembly/reassembly this process might require. Fortunately this one is is in decent shape after 70 years. Those thin metal tabs connected to the electrode contacts have a tendency to break from stress–it’s not uncommon to find a “dead” one which has just snapped from vibration or corroded through from moisture ingress. You see the rubber gasket that was meant to seal the thing from moisture…which it probably did for a while. With these plastic holders it is sometimes a little tricky to get the gasket back into the recess when reassembling, but it’s nothing anyone should be afraid of tackling.

Gently lifting the top electrode contact with a screwdriver
The “sandwich” of two electrodes and the crystal blank slide out from its holder

Most WWII-era crystals have a “sandwich” of two steel electrodes with the quartz blank between them, which is all held together with spring pressure. They come in various sizes and dimensions, but this one is about 1/2″ square which is pretty typical. The plates have raised corners where the physical “grip” on the crystal actually occurs. This leaves a very small air gap over most of the surface of the crystal to allow for vibration. When putting this back together its important to make sure the crystal touches the proper side of each electrode. This crystal happens to be quite clean, but it is not uncommon to have dirt on the surface from the disintegration of the container. If this happens it can usually be cleaned up with a toothbrush. It is literally a piece of rock (probably imported from Brazil), so scrubbing, detergent, most any solvent–it’s all allowed so long as it doesn’t leave any residue. In fact, in the pre-war days that were filled with coal dust, cleaning the crystal with carbon tetrachloride was part of good station maintenance! It’s important to hold the crystal by its edges so we don’t put any dirt, dust or oil on the the surface where it could fill up that air gap.

I use a plastic “gripper” I cut from a broken mechanical pencil to hold the blank in the etching solution. I got the idea from this guy. Because I’m using a fluoride-ion etchant, everything has to be made of PVC or other plastic that won’t react. The gripper holds the crystal by its edges so that it will be immersed under the surface of the HF solution….and the less HF I need to do the job the safer. This is a chemical gloves-and-goggles operation, even with the 15 mL of 1M HF solution I am working with.

How long to etch is always a question. The etching rate is faster the rougher the surface of the crystal, and it slows down as etching goes on. The exception to this is in the first few minutes, where the rate is very hard to control, especially if the surface is rough. I know from experience of rough-grinding 7 MHz crystals with 250-grit powder that 6 kHz in 10 minutes with this solution at room temperature is not unheard of. However, this crystal has already been etched in manufacturing, so it should be a good bit slower than that. This was a real problem for quality control during the war. If you are interested in the variables involved, check out John M. Wolfskill’s patents, particularly this one and this one. If you are grinding your own crystals there is a treasure trove of expert tips in there from the master of the art.

This is a relatively weak concentration compared to the industrial stuff (which is typically 49% by wt.) used in the semiconductor industry, but I don’t mess around: goggles, chem gloves, ventilation, and plenty of neutralizing solution in case there is a spill. I use as little of the stuff as possible at any one time–just enough to immerse the blank. The reaction with quartz is very slow at room temperature. The products of the HF and SiO2 reaction are ultimately water and hydrofluorosilic acid…i.e., the “fluoride” in most municipal water supplies. While that sounds innocuous enough, its reaction products with the human body are none too pretty. I acquire this in small quantities off the shelf at this weak solution, and I will not even dilute my own from a higher concentration for this very reason.

Long story short, this is slightly-educated guesswork, so I opted for a conservative 5 minutes for the first round to get an idea of how long this would take. I’m glad I did as it moved faster than I expected. Five minutes got me to within 1200 Hz. Another 5 minutes go me within 100 Hz, and another minute put me spot on. Usually its not that easy. I should mention that I had the 10B running and warmed up for all this so that the crystal oscillator in the radio was at temperature equilibrium. Each time I removed the crystal from the etchant I neutralized it in a baking soda solution and then rinsed in distilled water, and finally dried with a clean paper towel. I keep some isopropyl alcohol around in case I accidentally put a fingerprint on either the blank or the electrodes when trying to reassemble it. With this level of frequency precision the holder must be completely reassembled and tightened as the spring pressure will affect the frequency few 10’s of Hz.

The following day KK1K just managed to hear my 5 watts of SSB on 160 meters as the July sun was setting. Success! Now to make some for 75 meters…and probably think about a period-appropriate amplifier.

Leave a Reply

Your email address will not be published. Required fields are marked *