PN2222A based FM TX delivering 120mW

Thank you for the info. really interesting.

I did a quick 30-second search and found these on e-bay
https://www.ebay.com/itm/MRF5S9070MR1-880-MHz-70-W-26-V-RF-POWER-MOSFET-Freescale-NXP/333331596847?hash=item4d9c1c862f:g:vOoAAOSwRtpdf5p6

£5 each, but that is cheap, especially as they are supplied from the UK. Sweden impose a £10 additional fine on all goods sent by post from China. Plus they charge a customs handling fee, and demand duty on the value of the goods (including the postage paid).

But the link also shows the device you mentioned and gives some idea as to how it is mounted.

Thank you once more and very best regards from Harry - SM0VPO
(Corona-virus quarantine in Spain)

I've used the MRFE6S9125N in a power amplifier and it's good to 100W. These go for around $3 for a salvaged transistor around here. There is also the MRF5S9070N and MRF9060M which should work too (for a few tens of watts) and these go for $0.5 - $1 (salvaged). I did a quick search and can't find them on aliexpress, but I do have a stash of these and can provide some free samples if needed. Soldering these is tricky and I don't think you can deadbug it (transistor must be soldered to a thick copper plate), but 2 layer PCBs are very cheap nowadays so it's doable. Soldering involves heating up the copper plate (either with a hot plate or a massive soldering iron).

Copy paste from some forum.

"I have driven an IRF510 at 50mhz and gotten 4.5W out of it. To do so
I had to drive it as a 5 to 10ohm very reactive input, use a 24V
supply and load the output to around 10ohms, and bias it for 100ma
minimum. It was unstable thermally(tended to runaway toward device
burnout). The output spectrum (harmonics) and IMD were poor.
At 12V power out was nearly RF power input.

It was more effort than driving a 2SC1307 (old cb final) at 12V, more
gain and the IMD was better. At 12V the 1306 gave 6W carrier power
with good stability. A pair of 2sc1971 driven pushpull easily give
10-11W carrier at 12V and are fairly clean."

He's expecting 15ohm input impedance per device without the gate resistor. (Reactance of 200pf gate capacitance at 50MHz).

15/4 (no. of FETs on each side) = 3.75ohm X 2 = 7.5ohm X 9:1 input Z ratio= 67ohm approx

Is my math correct?

Is 2N7000 suitable for VHF amplification? Its gate capacitance is only 60pf.

The 3 turn is primary input transformer winding. Here's the schematic

https://ibb.co/VghPfGh

It's step down -

"The input transformer (9:1) uses 4C65 cores, and the output transformer (1:4) uses thin air!"

" T1 is made by threading a piece of wire three times (the primary) through a piece of coax outer conductor (the secondary) around four 4C65(*See note below on core material) cores glued together using Cyanoacrylate (Superglue), thus creating a 9:1 impedance transformation and a very reasonable input match. "

Hi again,
In Peter Frennings homepage he does not state which way round the input transformer is. I believe it could be step-down.

A 9:1 impedance ration to feed eight devices means that he was expecting an input impedance of about 5.5 Ohms. Another constructor had self-oscillation problems and added another 4.7 Ohms to each transistor. That means he was expecting about 50 Ohms per device, at 50MHz.

You can realistically expect about 25 Ohms at 100MHz.

I have not had much time to look into the finer details yet, but I hope I can get a flight and go home in a week or so. Then I will do a couple of experiments.

I have a lovely RF generator that delivers up to +30dBm out, form 10kHz to 3GHz. that will make my experiments a lot easier.

BR Harry (still in quarantine in Spain)

What's the input impedance of IRF510 at 100MHz?
Is it not the reactance of 180p gate capacitance or 8ohm?

Maybe a different device re-marked? I have seen that before.

According to the 50 MHz amplifier design at http://frenning.dk/OZ1PIF_HOMEPAGE/50MHz_IRF510.htm there should be about 12dB gain at 50MHz with a 1:9 impedance ratio step-up balun at the input (turns-ratio = 1:3), but that was to feed 8 devices in parallel.

But my quarantine ends in two weeks, so I can do a few more experiments when I get back.

BR Harry

Fake IRF510 from China?

That was a surprise.
One of the links I included was a 50MHz amplifier using the IRF510.
I also have a VHF-L amplifier circuit somewhere. I will try to find it.

I am sure that you know the IRF510 is an FET with a high input impedance.

I have a stash of these devices and I will do a few tests with them myself, then we can compare results  

BR Harry

Experiments showed that IRF510 is useless at 99MHz. 50ma standing current. 180-200mW input and 120mW output.
It's less than unity gain.

But my 2n3904/PN2222A based amplifier can be used to boost the output of one transistor FM transmitters. Tons of schematics of single transistor FM transmitters are there on the web. Output should be taken from collector only. The 4.7K sets the input impedance of my amplifier.

Say no to hard to find RF transistors. The amplifier's gain is almost flat from 90-99MHz. Po = 180mW both into 50ohm and 75ohm dummy load. It will drive a quarter wave antenna or half wave dipole antenna without any problem.

I got no gain from IRF510 at 12V. I biased it for 50ma standing current and drove the gate directly with 50ohm output from the parallel PN2222s. Less than unity gain. Gate capacitance is quite high, 200pf or 5ohm reactance at 99mhz.

Should I use a broadband transformer to drive the gate? 9:1 impedance or 3:1 turn ratio on a binocular core. 3T primary 1T secondary.

Using only two PN2222A is giving me 180mW at 99MHz. All I did was increased the standing current. Each transistor now draws around 45ma when delivering 90mW. Not bad. I cam finally stop using the 2n3866. Will build the IRF510 amp on a separate copper clad board and feed it with 180mW from my FM exciter.

So no need for LC input matching for the IRF510? I can just use a 3dB pad and 50ma standing current. Let's see how much output I get from 180mW input.

Ok, thats a good plan and I am sure you will be surprised.

I have used 4 of these to make a 50Hz, 50W inverter, 2 to make an HF linear amplifier, a bench lashup (copper-clad board) to make a VHF amplifier. I dont remember too much, but the gain did fall off a bit above 30MHz, but still useful to over 100MHz.

Try these links for ideas:

http://frenning.dk/OZ1PIF_HOMEPAGE/50MHz_IRF510.htm
https://www.g0kla.com/scpa/SimpleCheapPA.php

BR Harry

Thank u I'll try.

I dont really know, but you could just try putting a few mW into one and see what you get.

You will need to look at the datasheet to see the input capacitance, then be sure the drive impedance is lower than the capacitive reactance.
I did some experiments when I forst got these and found that a 3dB resistive attenuator at the input gave a very stable and more predictable drive.

BR Harry

Oh, I completely forgot about the IRF510. If I get 5 watts at 100MHz I would be so happy. What would be the power gain of this device at 100MHz?

Hi again,
Thank you for sharing the ideas and information.

The 2N3866 is obsolete but I recently bought a few 2N4427 devices. I found that the 4427 was more stable and even gave a slightly higher output.

I used the 2N2222 / BC547 in HF linear amplifiers and four of them gave me a very healthy 250mW without any trouble whatsoever.

As a matter of interest, have you thought about the IRF510 or similar? They are cheap, common and can deliver many watts. I have seen a design for a 144MHz 50-Watt amplifier using them in parallel, but you should be able to get 10 Watts from one device.

BR Harry
Corona-quarantine in Spain.

Update: After adding one more PN2222A in final stage I'm getting 180mW of RF output at 99MHz.

So a single PN2222A is good for 90mW output.

I added a 15K resistor in parallel to the 10K biasing resistor for increased output or 90mW from a single transistor. Current increased to 43ma.

I used two PN2222A in parallel at the output and each transistor is delivering 60mW at 99MHz with 25ma of collector current. I got 120mW out at 12V. I am thinking to make a 1W FM amplifier based on just PN2222A. It would require 15 transistors in parallel. RF transistors like 2n3866 are now obsolete in many parts of the world. Someone in Bangladesh is having a hard time building my TX because the 3866 isn't available.

The experiment shows that PN2222 plastic transistors can deliver usable output on the domestic FM band. I'm curious if several of them in parallel can deliver 1W.