I built up this FM microphone from belza.cz. The circuit is super simple, uses only a single transistor. The output of the electret condenser microphone is coupled directly to the base of the RF oscillator, and causes the frequency modulation. However since the microphone has no amplification, it is not sensitive and the user must speak close to the microphone.
Earlier FM Mics I’ve Built
I’ve built several of the two transistor FM microphones that are often seen on the ‘Net. They typically use a RF stage that is different than the one seen at Belza.cz. These typically have an NPN transistor, often a 2N3904. The base is at RF ground, typically by connecting a 1000 pF capacitor from the base to ground. There are two (or sometimes one) base bias resistors, typically 10k and 6.8k. The emitter is connected to ground through a 470 ohm resistor. The collector has the parallel resonant tank connecting it to the positive supply. This parallel resonant tank is typically a 120 nanohenry (same as 0.12 uH) coil and a 20 pF capacitor connected in parallel. Along with the transistor’s capacitance, it resonates at the low end of the FM broadcast band [see Note]. The feedback capacitor, typically 4.7 pF, is connected between collector and emitter, and causes the circuit to oscillate. This stage is preceded by a single transistor to amplify the microphone. This amplified audio is applied to the base where it changes the transistor’s capacitance and that causes the frequency modulation. This type of RF stage forms a fairly reliable transmitter.
But the FM mic at Belza.cz is quite different. It uses a PNP 2N3906. Again, the base is at RF ground due to C1. The collector has the coil, but the capacitor(s) for the coil are C2 and C3, connected in series and the center connection is also connected to the emitter. I’ve built similar RF stages before and been disappointed because they sometimes did not oscillate or they weren’t stable. I decided I would stick with the more common type without the series capacitors. But I’ve built several other circuits from Belza, and they all lived up to expectations. So I decided to try this one for a change.
I made a few changes to Belza’s circuit. I used a 470 pF and 330 pF in parallel for C1 – it’s an RF bypass cap and is not critical.
To get the right voltage, I used 2.7k for R1 (the 4.7k was too high). R2 actually measured 360 ohms. I used a 20 pF NPO (NPO means zero temp coefficient) for C2, less than shown because the other circuits I mentioned use only about 5 pF. For C3 I used a 36 pF, but when I found that the frequency was too high, I paralleled a 5 pF to it for a total of 43 pF.
The All Important COIL
The coil of the FM microphone is important because it along with the capacitor(s) determines the frequency and stability. Also it is usually small, in order to keep the FM mic small. In order to tune the FM mic, the capacitor can be a variable capacitor, but they are usually much bigger than the fixed ceramic disk and are more expensive. So many inexpensive FM mics use a fixed capacitor and a coil that can be squeezed or compressed to tune the frequency. Even if the coil is wound on a coil form with a small tuning slug, it is smaller and may be cheaper than the variable capacitor. But whatever the coil is, it should have good mechanical stability; the copper wire should be thick enough to support the coil firmly, and the coil should be small and light so it doesn’t pick up vibrations. Vibrations, also called microphonics, can be heard in the FM signal when it’s quiet. One way to muffle this is to stuff the coil with a tiny piece of foam rubber and when the coil has been tuned, drop a few drops of paraffin AKA candle wax on it. The experimenter can use a short piece of soda straw as a coil form, and some hot glue, silicone seal or other glue to hold the coil in place.
One other method of making a coil is to use a spiral shaped trace on a PC board. This is very stable but is not variable, however it can have multiple taps to give some range of tuning. We’ll just leave it at that, because it has to be done when the board is made and few experimenters want to make their own boards.
However, I’ve made the equivalent to that by using a short length of 14 AWG or other heavy gauge copper wire, and I bent it into a square shaped single turn coil. Each side of the coil was about 35 to 40 mm or 1-3/8 to 1-1/2 inches on a side. This was held up above the circuit board by putting blobs of silicone glue on the board, letting them dry, then mounting the coil on the silicone blobs with more silicone glue.
My coil was at first about 120 nanohenrys, which is typical for these FM mics. It was 22 AWG solid enameled wire, and about 5 turns, inside diameter of 5 mm or 0.196 inch. The frequency was up around 94 MHz, too high – I wanted it around 90- MHz. So I removed the coil and put in one made of 26 AWG solid enameled wire and was 10 turns close wound on a much smaller 0.120 inch diameter air core. The turns are touching each other so it is quite stable, but then it can’t be squeezed to tune it. But the inductance is about 200 nanohenrys and the frequency is about 89.8 MHz.
Not shown on the schematic but is absolutely necessary at this frequency is a 0.01 uF ceramic disk bypass capacitor across the power supply lines. I usually run these FM mics with no antenna because the antenna is connected to the oscillator and any changes to the antenna change the frequency. Without the antenna the FM mic may be picked up 300 feet or 100 meters away, but the distance is heavily influenced by the battery voltage, which determines the power. This one uses only 3VDC and draws about 3 mA, which totals only 9 mW, and consequently the signal can’t be heard more than a few dozen feet. I could add the antenna but I don’t really need the greater range.
Microphone Sensitivity
I was surprised at the sensitivity of the microphone. It’s not sensitive, but it’s not insensitive. I can put my “ticking time bomb” circuit a foot away and clearly hear the ticking in the FM radio a few yards away. I have to talk about a foot or two away, whereas the circuits with the second transistor as microphone amp can pick up sounds a dozen feet away, So this single transistor circuit can be sensitive enough for use as a close microphone. One other point that I’ve run into. The electret condenser microphone is a strange beast. Its range of sensitivities can be wide, as can its current. The circuit called for a 4.7k for R1, but I had to use a 2.7k to get enough voltage across this resistor, which determines the transistor’s operating point. So each microphone will have to have its resistor selected to match it. In other words, don’t expect the circuit to work good if you put a 4.7k in there and forget to check the voltages. Like it says in the schematic, there should be about 1.3 volts across the resistor.
Another problem I’ve found with the microphone is its low frequency sensitivity. The human ear doesn’t respond to low frequency sounds, such as when a door closes and causes the pressure in the room to change. This circuit has the microphone connected directly to the transistor, so its low frequency response is very good, probably too good. When some low frequency sound is picked up, it will cause the FM signal to vary widely and this could be a problem. The circuits that use a microphone amplifier can be limited in their low frequency response by using lower value coupling capacitors, so the frequency response drops below 100 Hz, where there are little or no human voice input. This circuit doesn’t have that option – you get maximum bass response from the microphone.
Frequency – Changes
One other point that might be a concern. As more receivers today are all digital, and they have channels that are on the FM frequencies such as 88.1, 88.3, 88.5, etc., it may become harder to pick up one of these FM microphones if it is on a frequency that is not a channel, such as 88.2. And if it is close enough to a channel it may be picked up but as the battery voltage and/or temperature changes, it may drift to a frequency that can’t be received. So it might be a good idea to use a coil that can be spread or compressed to adjust the frequency somewhat.
Another way to tune an oscillator is to put a diode across the capacitor, and connect it to a potentiometer so that a variable DC bias voltage can be put across the diode. But this usually requires more than 3 volts across the diode to give enough tuning range. If the diode has from 3V minimum to 9V or so, it will give a reasonable range of tuning. One way to get 9V for the diode is to use a pair of 3V lithium coin cells. This when added to the 3V supply will give 9VDC. The diode doesn’t draw any current, but the pot does. But if the pot is 1 Meg, the current drawn is so low that the coin cells should last for years. Another cheap way to tune it slightly is to vary the base bias. But this also changes the power output.
Note: Some information on how to determine the capacitance and inductance for this type of microphone. Let’s say you want to use a coil with 120 nanohenrys or 0.12 microhenrys. You can plug this into a calculation on a scientific calculator, and you will come up with the reactance at 90 MHz, the frequency I’ve chosen.
I multiply 2 times Pi times the frequency 90E6 Hz or 90 Megahertz, times 120E-9 or 120 nanohenrys. The result givces about 68 ohms inductive reactance.
The capacitor has to have the same reactance at the resonant frequency, so we plug this into a calculation to get the capacitance.
I divide 1 by 2, by Pi, by the frequency 90E6 or 90 Megahertz, and by the reactance, 68 ohms. The answer 2.6 E-11 has to have the decimal point moved one to the right, to give 26 E-12 or 26 picofarads. This is the total capacitance across the coil. There is a small ampunt of capacitance in the transistor’s collector, about 4 or 5 pF for the 2N3904, so the actual capacitor value should be about 20 to 22 pF. If the coil can be stretched or compressed, it should give a range of tuning to make the frequency somewhere around 90 MHz at the bottom of the FM band.
Back to experimenting…