Is it easy to amplify a crystal oscillator's output to 1/2Watts at 50 Ohms?

[Plasmateur]

Wondering is this is a fairly easy project to do.

I need a clean and pure sinusoidal signal into 50Ohms that can be amplified up to 1/2 watts through some kind of DC bias.

I can build the low pass filter that would follow the after the amplified signal. Can I just use some kind of simple transistor to amplify the signal? Would a Class A amplifier be the best approach in the instance, since the wattage is so low? A small heatsink should be sufficient. Was looking to amplify a 13.56MHz crystal oscillator.

[David Hess]

1/2 watt into 50 ohms is 100 milliamps and 5 volts so there are a few options.  Simple transistor amplifiers can do it but several fast operational amplifiers in parallel are also a possibility.  Note that if source termination is desired, then either the output voltage or current has to be doubled and half of the power is lost in the source termination.

How much distortion is acceptable?  What power supply voltages are available?

[Plasmateur]

"clean and pure" can mean like < 0.1% THD or can mean 10%+ THD.

Given your username I'm assuming you might be interested in something more like materials science equipment than QRP radio communications.

For industrial "power" semi-clean sinusoidal frequencies it is common to use switched mode oscillators / generators, or amplifiers, e.g. class e, class f, class d, or more efficient semi-analog amplifiers like class c.

If you're willing to build / add some degree of filtering for the fundamental and/or harmonics like the 3rd, 5th, and a general low pass filter that passes the fundamental you can get a "reasonably clean, pure" powerful waveform
coming out of any of those kinds of non-linear class c/d/e/f types of stages because the harmonics are sufficiently but not perfectly blocked.  And you end up with less active circuitry for the amount of gain you need.

On the other hand the only really truly linear harmonic clean options are these:
A: make a "power oscillator" which might not even be crystal controlled, probably not, but can generate the power you need in a single analog stage.

B: Make a XO that produces something like milliwatts or less of power.  Then follow that up with something like
30db of linear class A gain to get into the 1W range output and 50 ohm source / driving capability.  Usually you'd do that in 2-3 stages if you want good I/O isolation / matching, clean signals, etc.  The first stage is easily made with a MMIC gain block to get at least to the XX mW level from which you can drive a 1W PA directly or after one more gain stage.

For amateur radio purposes in the 30 MHz range or such I seem to recall people making few watt output capabilities with the BD139 (from a certain manufacturer, others' clones  aren't so good) there are 4-5 others that are commonly used with online commendations for < 30 MHz < 5W ranges that are cheap final PA candidates.
BD139 (selected mfg., I can't remember if ST was the good one or the bad one)
FQP13N10L
IRF510
...
q.v.
https://forums.qrz.com/index.php?threads/generally-good-rf-power-transistors.437708/

Even just some strong digital push/pull MOSFET gate driver might put out enough square wave power at that frequency (barely) or of course so if you add some small pass switch mode MOSFETs, but you'll have full 3rd, 5th, 7th, ... harmonics from a square wave so better to do class E/F at that point for cleaner sine output.

Wondering is this is a fairly easy project to do.

I need a clean and pure sinusoidal signal into 50Ohms that can be amplified up to 1/2 watts through some kind of DC bias.

I can build the low pass filter that would follow the after the amplified signal. Can I just use some kind of simple transistor to amplify the signal? Would a Class A amplifier be the best approach in the instance, since the wattage is so low? A small heatsink should be sufficient. Was looking to amplify a 13.56MHz crystal oscillator.

Thanks for the reply! So I already have a class AB amplifier that can go up to 500 Watts, but right now I'm using my Siglent AWG to power it. It's not for amateur radio though, just making plasma.

Right now I'm using a 0.5 Watt to 5 Watt amplifier, which is the LPF'd.
Then 5 Watt to 500 Watt Class AB which again, is filtered.

It's impossible to find an already made solution that will give me a simple 13.56MHz into 50 Ohms on the internet so I thought I would just have to build it myself. I would also like the ability to pulse it if possible, but that's for another day.  I just want everything compact so I don't have to plug in my signal generator to make it work. Thanks for the suggestions!

[Plasmateur]

1/2 watt into 50 ohms is 100 milliamps and 5 volts so there are a few options.  Simple transistor amplifiers can do it but several fast operational amplifiers in parallel are also a possibility.  Note that if source termination is desired, then either the output voltage or current has to be doubled and half of the power is lost in the source termination.

How much distortion is acceptable?  What power supply voltages are available?

Ideally I would like it as clean as what I'm able to produce with my Siglent ARB. I figured I could use a transistor amp, but wasn't sure about other components that might be needed, such as capacitors or inductors. I can throw whatever DC source on it to make it work, but would prefer 12Volts as that is also powering a fan.

[langwadt]

PS I looked at some of the TRF7960 and 6x / 7x family parts from TI.  The ones I looked at ran on either 3V or 5V and could
output 100mW or 200mW maximum depending on the chosen supply voltage.  They made references to externally controlled
modulation by a digital pin or internally controlled programmable modulation via OOK / ASK with programmable modulation depth etc. etc.
I have worked with this family of chips before (truly horrible software / documentation for their intended purposes BTW and buggy silicon).
But it seems like maybe (TBD) a dev-board for them or similar might be an easy possible signal generator that could have enough power output
somehow possibly to drive the final stage 500mW / 1W / 2W PA.

Maybe someone else has amplifier modules or more powerful ICs.

OPA2670 ?

[David Hess]

...

How much distortion is acceptable?  What power supply voltages are available?

Ideally I would like it as clean as what I'm able to produce with my Siglent ARB. I figured I could use a transistor amp, but wasn't sure about other components that might be needed, such as capacitors or inductors. I can throw whatever DC source on it to make it work, but would prefer 12Volts as that is also powering a fan.

If only a 12 volt supply is used, then an impedance matching circuit at the output is required because the peak-to-peak voltage for a 5 volt sine wave is 14.1 volts.

Obviously RF amplifiers can put out a lot more than 1/2 watt with a 12 volt supply.  Besides impedance matching, the tank coil on the output increase the output voltage swing.  So the simplest design will be a traditional single transistor RF amplifier and RF power amplifier design procedures can be followed.

A linear design like I suggested as a possibility is also feasible but will need a higher supply voltage, or more likely +/- 12 volts, or even higher if series source termination is used instead of parallel (1) unless something more exotic is done, like a 10 ohm series termination with positive feedback to make it look like 50 ohms; this is done to allow lower supply voltages.  Or no termination can be used but then proper termination at the load becomes mandatory.

There are class-AB amplifiers with the needed performance but they are awfully complicated compared to a single transistor RF power amplifier.

(1) I almost never see parallel source termination used in this sort of application.  I think it is because most designers are used to thinking in terms of voltages rather than currents.

[Plasmateur]

Here's one example of a BJT based 1mW -> 5W PA suitable (presumably) for radio broadcast.
https://www.vu2ese.com/index.php/2020/06/19/jbot-an-easy-qrp-linear-amplifier/
https://rheslip.blogspot.com/2011/12/the-jbot-linear-amplifier-on-steroids.html?_escaped_fragment_=
et. al.

Since you only need 10%-20% (depending on source/load impedance and derating, ...) of that power presumably you
could not use the four output transistors (paralleled push-pull pairs) and just use a single push-pull pair or even a single (very inefficient) device
for the output.

As for the two gain stages from crystal level to driving the PA, well, your choice, discrete BJT or some kind of opamps, modules, etc.

Here's one idea that hadn't crossed my mind before -- IIRC 13.56 MHz is an ISM frequency in many places and in particular it is also used for LF RFID, I think that exact frequency, in fact.  Therefore there are 13.56 MHz or there-about RFID reader ICs which include (presumably) a crystal controlled timebase, digital/analog/RF signal generator, modulator, demodulator, receiver circuits, et. al.

I don't recall or know specifics of the various RFID chips out there at that frequency but I wouldn't be surprised if one of more of them can output 1W into its antenna.  And from what I recall working with a couple of such ICs in years past some of them have flexible modulation control / test modes which are programmable, so you MIGHT be able to get them to put out a CW crystal frequency locked relatively clean sine wave on command.  Almost certainly you could "pulse" the waveform on and off with some level of speed / precision / control, too.

On another tangent as I mentioned recalling there are various high bandwidth power op-amps out there, some of them intended for analog video to coax cable driving, some for DSL/PLC line driving, etc.  Some of them might be able to put out enough power for your final stage either single ended or differential drive (e.g. through a transformer) or perhaps otherwise paralleled as David suggested.
Certainly a great many of such could work for your PA driver stage if not also / instead the PA final stage.
e.g.
https://www.ti.com/amplifier-circuit/op-amps/power/products.html#p1261max=24;180&p233typ=300;10000&p23typ=14;1400&sort=p233typ;desc

Of course I'd have to question whether doing those bits with discrete transistor design or using power opamps would be an easier trade-off one way or another.  At least the discrete transistors would be easier to hand build dead-bug / manhattan style over copper clad than the SMT ICs.

Hey evb149, thanks for taking the time to reply. All the information was very helpful, and I actually looked into some of your suggested methods in the past.

I went back to the QRP route as I was able to find a solution for around $50.

The source: https://www.ebay.com/itm/133686051489
The Amp: https://www.ebay.com/itm/363147258357

Now I just need to make or buy a LPF.

Edit:....wait, maybe I only need that transmitter. lol. Oh well.

[geggi1]

If you want to do this very simple just go on Ebay and search for "HF Power Amplifier" plenty of cheep kits are avalable.
In case you need some extra drive to the linear other modules can be bought.
These kits are usualy desigend with a IRF type FET or by repurposing mobile phone base transistors.
For experiementing i would chose a kit with an IRF transistor for simple replacement and low cost when you see the white smoke.
Most of these kits have a TX relay but this is simple to bypass with a short wire.
You will have to some filtering, a power supply and the heatsink.

In case you want to start form scratch look for 14Mhz / 20meter linear designs.

Most of the designs for HAM-radio is class AB.

[HB9EVI]

as long as there is no modulation, but pure sine, it doesn't have to be a linear amplification.
it's common practice to use output filters, which make from every distorted signal a sine signal
as already mentioned a good quality BD139 or IRF510 will do the job, a 7th order chebycheff filter on the output and you're done; no need for any fancy stuff like video opamps
though it can be done with a simple RFC, I'd recommend to use a transformer

[Doctorandus_P]

Would a chip like a BUF634 be useful here?
It is just a buffer, and thus has no gain on itself, but can deliver 250mA and has a bandwidth of 180MHz