10 mW NBFM Transmitter Schematic

Thank you Harry. I’ve ordered 100 X 16V 1W Zener diodes from eBay for only $2.5 USD. They ship from China. I’ll have to play with them to replace the varactor diodes as electronic hobbyist in India who are trying to build my circuits are running into non availability of components. Same goes for RF transistors like BF199, etc. 2n3866 is still available so my next FM TX projects will have 2n2222 in oscillator/buffer section and 2n3866s in driver and PA stages. I’ve seen several designs on the web where they have used 2n2219 in FM transmitters for 0.4W output.

I am unable to share links here (new members can’t for 7 days) so emailing you my PLL FM Transmitter circuit which was published in electronics for you magazine. The oscillator/buffer section is inspired by your designs.

Hi Joy,

In my experience, the Zener diode is about equal to a varicap/varactor diode, and works to many 100s of MHz without problems.
Your circuit drawing looks good :-)

1N4007 I found to have a lower capacitance and crap out at about 100MHz.
1N914 / 1N4148 perform well but have a very low capacitance.

BR Harry

Okay Harry and the arrangement of 1n4007 diodes are fine? Isv149 varactor diode is only available on ebay from china so I need to replace it with in4007s in parallel or a 16v 1W zener diode would be more useful here? Attached is the curve of isv149.

Hi Joy,
Yes, you are on the right lines. Looks a lot more practical.

The only difference is that I would skip the 10n and first filter tuned circuit, and place the second filter beside the tank circuit - inductively coupled. But it should work exactly as you drew it.

BR Harry

Yes Harry you explained it to me. To simplify the circuit I’ll test this one transistor circuit soon after I order the usb spectrum analyzer. It will be used in a simple FM intercom. This looks good? Let me know your thoughts on the schematic I am attaching. I plan to substitute the isv149 varactor diode with several IN4007 in parallel. Audio quality was surprisingly good when I tested this single transistor TX with isv149 back in 2015. There was just a single tuned circuit at the collector. No BPF.

Hi Joy,
The parallel tuned circuit at resonance has capacitive reactance equal to inductive reactance. If the capacitive reactance is 500Ω (3.3pf) then the inductive reactance is also 500Ω (1uH, 7 turns, 10mm Dia.). A 50Ω tapping is therefore 10% of the impedance = 30% tapping (square root of 10:1 = 3:1) = 2 turns. Choose a little less than 30% so that the load on the inductor is a little lower and the Q will be a bit higher (BW is a little lower :-).

You would need a lot of filtering if you had a higher power, like 10 Watts, but with 10mW the spurious radiation is already 1/1000th part of that, so you don't need much (passive/lossy) filtering.

Just place a simple tuned circuit beside your output tank (parallel tuned circuit) coil to get coupling and you will get 50Ω out.

Have I explained this ok?

BR Harry - (still in ES).

One more question Harry, how did you come up with 50 ohm impedance for the filter inductors? Is it the capacitive reactance?

The output BPF is probably wrong in values. It did help reduce the harmonics. Here’s the spectrum Analyzer shot. I was learning about frequency multiplication back in 2015 when I made this circuit and it’s not perfect. Inductive coupling between two coils tuned to 90 MHz and 50 ohm output taken from 2 turn tap is better idea. I’ll use your suggestions.

Yes, it looks good, but I just wonder about the values and the impedances. If the design is published then it probably worked in some application.

But the output feeding the load via a 10pf capacitor looks a bit strange. The reactance of 10pf = 170Ω at 90MHz. It could be that the original design fed an antenna that was inductive. The impedance at the top of the filter inductors is only 50Ω, and the coupling between them is 1pf (about 1800Ω).

I would be more inclined to be a bit more conventional, use two tuned circuits with a top impedance of about 500Ω (3.3pf and about 7 turns). Feed the 50Ω RF in and out to a 2-turn tapping on the inductors, and physically position the inductors to get mutual magnetic coupling. The output tuned circuit could also be used in this way at such small powers.

I once built such a filter using 4 cavities about 20mm square, made from copper/PCB board. Cut a slot in each so RF could leak out into the next cavity. Size of the slot determines the coupling. I had a VERY narrow bandwidth, but if you are using this for WBFM then two tuned stages should be enough. Too many performance filter stages and you could restrict the FM bandwidth. You need to pass about 150kHz to 200kHz almost flat. Increase coupling and stagger the tuning and you can get a nice response curve, flat at the top, with just three stages.

BR Harry - SM0VPO

This is the alternate circuit with 90 MHz BPF. Would this work?
Audio sounds pretty good. I have tested it without the BPF and range is 50 meters maximum.

Hi Harry,
I would like to share NBFM TX circuit that I designed in 2015.
Output power is 10 mW. With RFC1, R3, and C5 in place the audio sounds louder but transmit frequency is 89.97 MHz. When they are removed and the 30 MHz crystal directly connected to D1, output is 90 MHz. I was thinking if there is a better way to achieve the same results with less components. A single transistor xtal oscillator feeding a band pass filter. All unwanted emissions are below 45 dB in the above circuit.

Joy