What is the best variable
capacitor for my crystal radio?
By Dave Schmarder
This is a question that has been asked of me many times. It would be easy for me to just say, "Just buy some holy grail capacitors.". Unfortunately, they are not so common anymore, and when you see how much they cost at the large auction venue, you'll forget about the high cost of the litz wire that goes with them.
The purpose of this paper is to discuss various types of variable capacitors, showing their good and bad points. I have come up with some answers that I have not seen discussed elsewhere. Perhaps this work will help you find the right variable capacitor for your radio.
This paper deals nearly exclusively with a crystal radio detector unit in a two section radio. Discussion of an antenna tuner section can be found on my antenna tuner page.
Before we get down to the nitty-gritty brass tacks of the actual capacitors that you should get, let's set the basics for this page. First, anyone that wants to build a dx set should visit the crystal radio DX'ers best friend, Ben Tongue's website. Specifically his paper dealing with variable capacitors as used in crystal sets. In his article 24, section B1, his charts plainly show the difference between a holy grail and an everyday capacitor.
One even wonders how a crystal set works with such a lousy capacitor. Well, it really isn't that bad. But with losses highest at the high end of the band, plus the very squeezed tuning (10% of the capacitor rotation covers 30 to 40% of the frequency range), there is a lot of bad karma at the high end of the band.
Below are some examples of capacitor traits. They are not in any particular order of importance or the goodness of one capacitor over another.
A variable capacitor is made up of two major sections, a stator or stationary part, and a rotor or moveable section. These sections are made with plates, also called vanes or blades. The frame is usually connected to the rotor and is one connection of the capacitor. There are some instances where the rotor is isolated from the frame, but not too often.
This is the material that separates the stator plates from the rotor plates. An air dielectric is most common, and is very low loss. Air has a standard dielectric constant of 1.0. Sometimes other materials are placed between the rotor and stator plates, mainly so the plates can be closer together. This allows for a smaller capacitor size, but it is at the expense of higher losses. I like using air capacitors the best.
Insulators are the non-conducting material that holds the stator plates to the capacitor frame. The insulator material is considered by most builders the prime criteria in selecting a good capacitor. Two major types of insulating materials used are ceramic and phenolic. Ceramic is consider the very best and the only thing to use in a dx crystal set.
However, new ceramic insulated variable capacitors are expensive. For less aggressive radios, and very aggressive budgets, phenolic is the choice.
Some of the variable capacitors that you see will have a silver plating. Ben warned us in his article that what appears to be silver plating might not be. The new capacitors sold have bare aluminum plates. They are higher resistance than the silvered plates. Of all the capacitor spec's, this is the least important. But most of the older capacitors that have ceramic insulators will also have a better plating.
The wiper arms are those little springy thin pieces of metal that attach to the frame and contact the rotor. Variable capacitors are rated by the number of wipers and how much surface area is in contact. The tightness of the contact should also be considered. Tight is good.
Inexpensive capacitors sometimes have no wiper arms. While they do work, their use in a good crystal set is discouraged. Other cheaper capacitors have just one small wiper at the back frame. The contact area is small, and the contact pressure is fairly light. These are ok for TPT sets. (Radios that are low performance because the builder uses coil forms made with cardboard Toilet Paper Tubes.)
The maximum number of wipers is determined by how many frame surfaces are reachable. In a single gang capacitor, two, one at the front and one at the back is the maximum. In a two gang, it's four (front, back and two in the middle), and so on.
Most hobbyists don't consider the wiper arm quality when they are looking for the capacitor for their dream radio. I consider wipers to be almost as important as the insulator material. I have tested phenolic insulator capacitors that had very nice wipers and while they didn't test as well as ceramic types, they were still very decent.
Capacitance Minimum to Maximum Ratio
Here is something that almost nobody considers when buying their capacitors, until the builder tries to tune the entire band with a low maximum to minimum capacitance ratio and finds out that it doesn't work. The US-Canadian broadcast band is from 530 to 1710 kHz (while the US is 540 - 1710 and Europe only goes to 1606 kHz). This means that if you are going to tune the entire band, without switching in additional capacitance or with coil taps (yeech), close attention has to be paid to the maximum to minimum capacitance ratio.
An example of a capacitor that will not tune the entire MW band in one swoop is a capacitor that is available on the surplus market. This capacitor has everything that one would want in a capacitor, except a good maximum to minimum capacitance ratio! A picture of it is shown below. Each section measures from about 300 to 35 pF. This is ratio of about 8.6:1. The medium wave band at my location needs a total ratio of at least 20:1 if it is expected to tune without band switching. This includes the distributed capacitance of the coil, and the diode input. At 20:1 there is very little leeway and the coil has to be constructed with extra care. The cheap capacitors fall well within this range, as they have a ratio of about 24:1.
The way you can tell about a capacitors ratio at a glance is to look at the plate structure. If you see a many plates and they are small, this means that the ratio is low. This is because when completely open, the capacitor still has a fair amount of surface between the plates.
Capacitor Frequency Linearity
Those of you that have tuned a radio using a capacitance linear capacitor such as the inexpensive ones mentioned earlier, noticed the top of the band dial squeeze. This is because the capacitance is linear to the rotation. If the capacitor shaft is rotated half way, the capacitance is approximately at half the total value. This means that the center of the dial is at around 800 kHz, about 370 kHz out of the total of nearly 1200 kHz that comprises the medium wave broadcast band. The other rotation half has to cover 3/4 of the band. It gets a lot worse the closer you get to the top of the band.
But what if you could move the capacitor 50% and have half of the band covered? This is what is called a straight line frequency capacitor. In practice having a completely straight line frequency capacitor is nearly impossible, but we can come close.
There are three basic types of capacitor construction. One is where the shaft is centered on the capacitor frame, right to left. The capacitor plates are semicircular. This operates as a straight line frequency capacitor (because if you plot the capacitance vs. the rotation, there is a straight line result).
Another type which was very popular in the old superhet tube radio days was with the shaft and plates offset from the center. These types can also have a modified plate shape, which produces a slight "S" curve when the capacitance vs. rotation is plotted. These, mostly dual gang capacitors are widely sold, but not as popular as the inexpensive single gang linear variable capacitors.
You can expect to pay twice as much for these as the inexpensive capacitors, but most have the advantage of a dual wiper, which lowers the losses. The tuning of this capacitor does help the high end of the band frequency crunch. It is at least tolerable.
The third type is a closer to a straight line frequency capacitor. This has highly shaped plates plus the offset shaft position. There is a much better frequency spread across the dial, especially at the top end of the band.
Insulator Loss vs. Capacitor Plate Loss Ratio
As Ben so clearly shows in his charts, as the capacitance is increased on a particular variable capacitor, the losses go down. In his example of the cheap capacitor, the losses go down by over a factor of 10 when the shaft is rotated to about 1/3 of the way up. The more expensive capacitor also trends up, but at less steep rate. This is because the losses of the dielectric and wiper arms are summed with the losses between the capacitor plates.
This is the so called "waysalot" capacitor. I used a pair of these in my #75 dx crystal set. It is the best quality capacitor I have. This one has a vernier attached and is used as a test capacitor for determining unloaded Q of the coil under test.
This is known as Holy Grail (everyone should bow and say "I'm not worthy, I'm not worthy"). This capacitor is about 475 pF can was removed from a high quality test instrument. If you can find one for less than seventy-five bucks, grab it.
Here are examples of three different variable capacitors frequency vs. capacitance plots. The one on the left is the inexpensive ones available everywhere. The shaft is centered left to right and the plates have a semicircular shape.
The middle one also has the semicircular shape, but the shaft is offset from the center of the frame and the plates. This offers better frequency linearity than the capacitor on the left.
Finally the capacitor on the right is a great example of a straight line frequency plot. The shaft is offset, and the plates are very long. It is only near the bottom of the tuning range is where a lot of capacitance is added with little turning of the shaft. This allows for a more even frequency spread across the tuning range.
Ready to on the big capacitor hunt?
Ok, that was quite a bit on one page. Make sure to read Ben's information again.
As I write this, I am designing a crystal radio that will make use of the information on this page in a new way. This will be a dx grade radio that will use medium priced parts. I am expecting the performance vs. cost to be very good.
Keep watching my pages for the latest information. Happy building. Dave