Diy rf power amplifier for experimental use

[Mm]

Im looking for most affordable and efficient way to amplify single square wave frequency. The frequency is between 1mhz to 10mhz. Best would be that the amplifier have potentiometer/trimmer to adjust the amplification. Input signal is weak, less than 1W and the amplifier should be able to amplify the signal at least to 100W or even more. If possible, input power supply could be flexible between 9-60V (which then affects to the output amplitude of the signal).

I hope this makes any sense. What kinda options do I have? I have friends who can help me out with making PCBs and soldering all parts together, but they have less experience from radio tech, so I have to try figure out with help of internet how this should be done.

This amplifier is used in very experimental device and it is not used at all in conventional way. Signal generator creates the signal, the signal gets amplified and then it is transmitted to resonator circuit. No radio is used at all.

I have searched in aliexpress but havent found any good solutions yet.
Something like this:
https://a.aliexpress.com/_m0mNNRY

Any help is appreciated. Thx

[T3sl4co1l]

Is 1-10MHz the squarewave's full bandwidth, or its fundamental in which case what rise/fall time do you need?

"Less than 1W" isn't exactly "weak", but if you need 40dB gain, fair enough.

Since it's square, surely you would need a (fast) switch, more so than an amplifier (implying linearity, proportionality, and generally, poor efficiency)?  That is, an inverter?

Is this like a piezo driver or something?

Tim

[Mm]

Thx for replying and expressing interest.
The frequency is going to be one frequency between 1-10mhz at given time. The resonator where the signal, after being amplified, is transmitted will be changed between experiments and that determines the frequency (which is always just one specific frequency with the resonator/experiment).

Im experimenting one specific phenomena with zpe device and for that I need to be able to test how signal amplitide affects to the outcome after resonant frequency is found. So, basically I have experiment where I have certain type of resonator and I transmit signal to it. If the signal is correct (resonant frequency) I get very specific and wanted outcome, but that is with very narrow amplitude scale of the signal. Now I would like to experiment way wider amplitude scale.

I do not have budget to buy expensive signal generator amplifiers, but Im hope Im able to build one. The amplifier can be very crude (no fancy stuff needed).

[mikerj]

Thx for replying and expressing interest.
The frequency is going to be one frequency between 1-10mhz at given time.

A square wave has a lot of energy in the odd harmonics (e.g. 3x, 5x, 7x, 9x the fundamental).  If you try to pass a 10MHz square wave through an amplifier with a -3dB cut off around 10MHz then you'll get something like a distorted sine from the output as all the harmonics will be significantly attenuated.  Hence the question about whether 10MHz is the fundamental frequency of the square wave, or it's bandwidth.

If you want to amplify a 10MHz fundamental square wave and keep it looking like a square wave then you will need considerably more bandwidth from the amplifier, the exact amount depends on how fast you need the rising and falling edges to be.

[Gyro]

.. The frequency is going to be one frequency between 1-10mhz at given time...

This is probably a good time to note the importance of correct use of case. 1-10mHz is milliHertz. T3sl4co1l and mikerj interpreted and corrected this to 1-10MHz (MegaHertz) but you repeated the error. Such things are important to avoid confusion, particularly if RF isn't mentioned anywhere in the text or title as a clue.

[Mm]

Im not sure if this helps at all, but my estimation is that that frequency Im going to be using is 1-5mhz in each experiment. The bandwidth that the amplifier can amplify can be much more broad if it helps with this.

About how fast rise/fall can and should be is still a mystery. Im sure slower rise/fall is acceptable at this point.

[gbaddeley]

Why do you need a square wave?
Is an approximate square wave OK? (Limited harmonics, ringing, limited slewing)
Sine wave is not suitable?

[Mm]

So far square wave has given the best results in my experiments, but approximate square wave is ok too at this state

[T3sl4co1l]

Ah.....

.. The frequency is going to be one frequency between 1-10mhz at given time...

This is probably a good time to note the importance of correct use of case. 1-10mHz is milliHertz. T3sl4co1l and mikerj interpreted and corrected this to 1-10MHz (MegaHertz) but you repeated the error. Such things are important to avoid confusion, particularly if RF isn't mentioned anywhere in the text or title as a clue.

...If they're "working" with "ZPE" (Zero Point Energy?) devices, somehow I don't think it will matter.

This also explains the budget limitations.

Tim

[Mm]

.. The frequency is going to be one frequency between 1-10mhz at given time...

This is probably a good time to note the importance of correct use of case. 1-10mHz is milliHertz. T3sl4co1l and mikerj interpreted and corrected this to 1-10MHz (MegaHertz) but you repeated the error. Such things are important to avoid confusion, particularly if RF isn't mentioned anywhere in the text or title as a clue.

Oh, sorry I missed your reply. You are right, Mhz is used not mhz. My bad.

[Mm]

Ah.....

.. The frequency is going to be one frequency between 1-10mhz at given time...

This is probably a good time to note the importance of correct use of case. 1-10mHz is milliHertz. T3sl4co1l and mikerj interpreted and corrected this to 1-10MHz (MegaHertz) but you repeated the error. Such things are important to avoid confusion, particularly if RF isn't mentioned anywhere in the text or title as a clue.

...If they're "working" with "ZPE" (Zero Point Energy?) devices, somehow I don't think it will matter.

This also explains the budget limitations.

Tim

 

[wasedadoc]

.. The frequency is going to be one frequency between 1-10mhz at given time...

This is probably a good time to note the importance of correct use of case. 1-10mHz is milliHertz. T3sl4co1l and mikerj interpreted and corrected this to 1-10MHz (MegaHertz) but you repeated the error. Such things are important to avoid confusion, particularly if RF isn't mentioned anywhere in the text or title as a clue.

Oh, sorry I missed your reply. You are right, Mhz is used not mhz. My bad.

You still have not got it quite right.  MHz, not Mhz.

[Mm]

You still have not got it quite right.  MHz, not Mhz.

That is true. My bad. MHz

[David Hess]

Im looking for most affordable and efficient way to amplify single square wave frequency. The frequency is between 1mhz to 10mhz. Best would be that the amplifier have potentiometer/trimmer to adjust the amplification. Input signal is weak, less than 1W and the amplifier should be able to amplify the signal at least to 100W or even more.

10 MHz at 100 watts is a serious design challenge.  I have achieved 600 kHz at 25 watts but that was with 2 MHz output transistors.  A discrete implementation of a current feedback amplifier should be able to achieve what you want using ring emitter transistors or MOSFETs in the output stage.

[gbaddeley]

Have you considered the broadband RFI that it will produce? Are you operating inside a Faraday cage or similar?

[Mm]

A discrete implementation of a current feedback amplifier should be able to achieve what you want using ring emitter transistors or MOSFETs in the output stage.

Do you have any examples of this? Thx for your reply

[Mm]

Have you considered the broadband RFI that it will produce? Are you operating inside a Faraday cage or similar?

I understand your consern. This is kinda thought of. The device will be eventually shaped and build in the way that possible RFI is reflected back in and hopefully used in our advance. How much RFI it will be emmitting is still a question because I do not know, at this time, how much amplification is needed for this particular signal, if needed at all. Other high frequency signals going through this device are so low in amplitude that RFI is most likely not be a problem.

[Infraviolet]

60V and 100W are pretty big, but...
For smaller signals I've had success using common emitter transistor amplifiers to amplify voltage, and an NPN can trivially serve as a buffer to increase the current capability of an analog signal (at the expense of losing about 0.6V). Maybe you could find a high enough power transistor to handle your requirements.

I had a thread recently about amplifying a sine signal at 3MHz, I needed to get an output at the end of it which was linearly proportional to the input voltage, and having abit of extra current driving capability (low impedance source) than the original input was somewhat needed too.

Do you need to amplify in this manner, where the voltage of the input signal linearly controls the voltage of the output, or are you in a more "comparator like" situation, just send the output as igh as possible when input is above a threshold and as low as possible when below a threshold?

There's a youtube channel which has some good videos about transisot amplifiers, mostly focused on audio but the discrete transistor methods should work fine up to 10s of MHz. Not sure if this will still apply at high power, but high power rated transistors do exist, some are probably similar in speed performance to typical low powered ones.   https://www.youtube.com/@TheAudioPhool/videos

[Mm]

Do you need to amplify in this manner, where the voltage of the input signal linearly controls the voltage of the output, or are you in a more "comparator like" situation, just send the output as igh as possible when input is above a threshold and as low as possible when below a threshold?

Thx for taking your time and answer to me.
At the monent I need amplifier so I can play and tweak the signal more freely. Im building a prototype of zpe device, but atm I need dev amplifier to have understanding do I even profit of using amplification and if I do profit from it, how much do I need to amplify the signal. Purpose of this dev amp is to find "sweet spot" and after finding it, develop the actual amp for the protype device itself (if it is needed). Most likely the actual amp is easier to dev knowing more parameters than I know at this point (which is kinda zero).
There is naturally other things I need to take in consideration too, like impedance, SWR etc..

The point is to have somekinda understanding, would amplifying the signal effect the output. There is no information out there that I can use to know this before testing it myself. How the signal power can effect the output? Will the output increase at all after amplifying the signal? Is the possible output increase linear to the amplification? Is there samekinda "sweet spot" like there is with frequency (resonant frequency)? And many other questions is in my mind. Crude/simple amp would give alot of answers.

[David Hess]

A discrete implementation of a current feedback amplifier should be able to achieve what you want using ring emitter transistors or MOSFETs in the output stage.

Do you have any examples of this? Thx for your reply

The closest available examples I know of are older 40+ MHz function generators from Tektronix and especially HP which use a suitable output amplifier but not to that power level.

I would design such a thing from scratch as something similar to a traditional class-AB audio amplifier, but with the output stage configured to use local current feedback which is typically not done in audio designs.

[TimFox]

When dealing with square waves, the rise time is an important parameter along with the fundamental frequency.

[tggzzz]

Ah.....

.. The frequency is going to be one frequency between 1-10mhz at given time...

This is probably a good time to note the importance of correct use of case. 1-10mHz is milliHertz. T3sl4co1l and mikerj interpreted and corrected this to 1-10MHz (MegaHertz) but you repeated the error. Such things are important to avoid confusion, particularly if RF isn't mentioned anywhere in the text or title as a clue.

...If they're "working" with "ZPE" (Zero Point Energy?) devices, somehow I don't think it will matter.

This also explains the budget limitations.

... and the continued ignorance of SI units
... and the continued lack of information about the rise/fall time in favour of "square wave".

The OP might learn why rise/fall time is relevant by looking at the practical measurements (plus a little theory) at https://entertaininghacks.wordpress.com/2018/05/08/digital-signal-integrity-and-bandwidth-signals-risetime-is-important-period-is-irrelevant/ Everybody else here already knows that!

[magic]

Why even think about amplifying square waves?

Get a variable switching PSU to set output amplitude (or spend your 100W heatsinks on making a linear regulator), then switch the load between supply and ground with a pair of complementary MOSFETs. It's still some work to make fast gate drivers, and getting the dead times right is highly advised, but sounds an order of magnitude easier than a linear amp with such capabilities

[Mm]

... and the continued ignorance of SI units
... and the continued lack of information about the rise/fall time in favour of "square wave".

The OP might learn why rise/fall time is relevant by looking at the practical measurements (plus a little theory) at https://entertaininghacks.wordpress.com/2018/05/08/digital-signal-integrity-and-bandwidth-signals-risetime-is-important-period-is-irrelevant/ Everybody else here already knows that!

Jeez, no need to yell. What triggered you so badly? Im happy for you that you know your theories.
Little quote from "NEWBIES PLEASE READ BEFORE POSTING" post:
"Regulars, please be extra nice to beginners asking questions. Do not deride them for anything, if you can't be nice, don't post."

I have not said at any point that Im expert in this field, hence Im writing at the "Beginners" section, have you noticed it? Actually Im very very novice with electronics and specially with ham radio stuff, I have no problems admitting these facts.

I already apologized and admitted my unit error, what else you want? I dont see any point to start give reasons why my SI unit mishaps happened, you can think what ever you want of it.
..and I said also this "Im sure slower rise/fall is acceptable at this point.". Im not trying to achieve 90 degree angles on the square wave, I said few times already that the amp can be crude and simple, no fancy things is needed. I did not add ns value for rise/fall but I did not ask complete solution for the amp either.

And yes, 100W is most likely overkill and too hard to achive. Maybe 50W is better and still it might be too hard to achieve.

[langwadt]

Why even think about amplifying square waves?

Get a variable switching PSU to set output amplitude (or spend your 100W heatsinks on making a linear regulator), then switch the load between supply and ground with a pair of complementary MOSFETs. It's still some work to make fast gate drivers, and getting the dead times right is highly advised, but sounds an order of magnitude easier than a linear amp with such capabilities

might try two of these in a bridge, https://www.ixysic.com/home/pdfs.nsf/0/6AC7CC624179BDA185257873005D6BE4/$file/IXD_630.pdf