800mW FM RF Amplifier

More LtSpice..
Testing a transformerless complimentary amplifier.
Output matching network 15 to 50 ohm yields an Po of 0.6W with 10mW input at 96MHz and 0.4W with 5mW input. Increase the cut off frequency of output network to 200MHz for broadband operation up to 150MHz (160mW output).

Hi Ivan,

The shiny Chinese ones are fake with only 4dB gain at 96MHz. It wasn't like that 7 years ago when all were genuine pieces on eBay. In India I've found a source for genuine 2n3866.. online website and cost is around $3 per piece. The PN2222A is still cheaper and available on the same website for 10 cents a piece.

dare4444 wrote:Too many homebrewers are trying to build FM transmitters but they get frustrated when 2n3866 isn't available.

Hi,
what is the problem? Look e.g. here. I know there may be fake goods in some shops, but all is O.K. in most of them.

BR from Ivan

I met an electrical engineer and he gets 40 emails a day for FM TX kits. He wants to build kits out of these and sell. There's only $2 profit per kit but goal is to encourage home brewing and provide hobbyists with good stable designs. Most LC based kits drift a lot and output power is low. This is the kit he's selling

https://youtu.be/9MdroBUfqpk

Check this YouTube video for prototype pictures and schematic.

To bring down drift to acceptable levels, I designed a 8 transistor FM TX last year with months of trying out different configurations, buffers, and extensive temperature testing with the board laying open on a wooden table at room temperature. I used commonly available 2n3904 from oscillator to driver stages and it made it tough to achieve high reverse isolation. The final circuit showed S12 of 58dB on LtSpice. So the reverse isolation is at least 55dB on 100MHz with 2n3904 BJTs! Switching on two ceiling fans only changed the frequency by 5KHz or so (it decreased). Board was not in a box. If it were some other design the drift would have been in tens of KHz.. maybe 50KHz or more. The biggest culprit in my experiments turned out to be the power draw from oscillator. It should be loaded just "right" for exceptional stability. Drawing DC power from oscillator and buffer section also disturbed the frequency stability. I was feeding the base of next stage directly from emitter of buffer transistor. Using separate biasing network and coupling capacitor got rid of the unwanted drift.

The biggest contribution was provided by you. Remember the 3 x 5.6V zener in series trick for the varactor diode? Temperature stability improved tremendously. A single 16V zener was causing the remaining drift. Last modification was switching common emitter driver stage to a common base configuration. The two PN2222A parallel transistors in PA stage are delivering 350mW of RF output. Yes, the power output is lower but frequency stability is awesome. I have left the design as it is with two parallel BJTs in final stage.

I've scanned the internet and got thousands of schematics of various TX and RX circuits but I've never seen this particular arrangement elsewhere (using 5.6V zeners). It showed me that BJT based VHF TX an be stable too as JFETs are more preferred in high frequency oscillators.

All that R&D takes considerable time. didn't have LtSpice. Got introduced to Spice few months after I made the design final. Testing tools were a multimeter, frequency counter, and a handheld fm receiver to check for instability and that was it.

Driver is 2n3904 (more gain than 2222).
Final is plastic PN2222A . The plastic ones seem to have more gain? I tried twenty year old metal can 2n2222A and it failed. I'm not sure why.

Build a 2-3 transistor buffer based on 2n3904 to drive this Amplifier with +10dBm from your FM V5 mic. Don't use RF transistors as it would undermine the project's goal. Publish the work on your website for others to benefit along with your own investigation and test results of the amplifier.

It would be a fun project transforming the low power V5 mic to a reasonable powerful one. The amplifier is stable with both 50 and 75 ohm loads. A vertical half wave dipole antenna would be nice

Hi Harry,

The schematics are all going to be published in Indian magazines to help your average hobbyist. Too many homebrewers are trying to build FM transmitters but they get frustrated when 2n3866 isn't available. I'm not using them right now but college students build these circuits and a few college radio stations are using my designs. Now everyone has got a pocket computer or smartphone. They are too busy with it and radio is loosing its charm.

Hi Joy,
That looks like a nice project. I really love the idea of the PN2222 devices.

I also found that audio and switching transistors will work into the VHF spectrum, as in the V5 bug. Gain is low, and at higher frequencies it is not really so useable, but that is also good for VHF stability 

It would be interesting to see how that works with my V5 bug? I could replace the BC557 (0-gain) buffer with a better RF transistor and get a few more mW to drive the amplifier. I was never so interested in higher powers from the bug and other FM transmitters. In those days there were loads of pirates being caught, but doday and in some countries (esp Spain) the only local FM radio stations are not licensed    As long as they operate sensibly and responsibly then "the authorities" turn a blind eye, and I know that on some occasions even assist in the event they cause unwanted interference 

(eg. Coast FM Gold, interfering with Malaga airport approach)

I remember using a cassette tape recorder up a tree with a transmitter, and watching it being ripped down by "the (UK) authorities" after they had detected it using a van with a rotating roof-rack. I also built another for a friend outside Cambridge. He put on some music and went for a ride. After about 5km he drove home, fast, and never dared to switch it on again.

One question to you, are you actually broadcasting? or is it for covering, say, a private building? It would be really interesting to know how you are using these FM transmitters and amplifiers.

Best regards from Harry - SM0VPO

Final PA efficiency is good. LtSpice showed 200mW dissipation in PN2222A when Po = 900mW and Pin = 10mW. R4 helps improve efficiency by limiting current to the single transistor PA stage. For maximum gain, keep the input frequency between 88MHz -96MHz. 2nd and 3rd harmonics are -30dB and -50dB down. Say goodbye to obsolete and expensive RF transistors like 2n3866 and 2n4427.

Del