Suggetions for RF power stage amp

Thread Starter

Zeno53

Joined Sep 23, 2020
46
I need to select a power amp to match as the last stage to a previous preamp circuit.
Let's leave out for a moment the laws, regulations, usable frequencies and powers, as currently it is a simulation, this pre-amplification circuit in simulation allows to supply in sinusoidal it will be able to supply from 20mV (+/- 10mVolt) to 2Volt (+/- 1Volt) up to 100MHz then the power amp can be driven with this voltage depending on the stages used. The most important factor is the phase linearity up to 100Mhz and I thought it could be done as there are devices on the market with a bandwidth of several GHz. Without this factor I don't need it. I would prefer to use FETs but all the amps I've tested from J-FETs (very sensitive and ideal) and Mosfets don't have phase linearity. The use of Mosfets is not excluded because even if they use higher input voltages and for this reason I have put as a limit of voltage supplied by the preamplifier up to 1Volt. Furthermore, it could drive low ohmmic loads without any particular problems. Advice on which component to use?
 

Papabravo

Joined Feb 24, 2006
15,800
I need to select a power amp to match as the last stage to a previous preamp circuit.
Let's leave out for a moment the laws, regulations, usable frequencies and powers, as currently it is a simulation, this pre-amplification circuit in simulation allows to supply in sinusoidal it will be able to supply from 20mV (+/- 10mVolt) to 2Volt (+/- 1Volt) up to 100MHz then the power amp can be driven with this voltage depending on the stages used. The most important factor is the phase linearity up to 100Mhz and I thought it could be done as there are devices on the market with a bandwidth of several GHz. Without this factor I don't need it. I would prefer to use FETs but all the amps I've tested from J-FETs (very sensitive and ideal) and Mosfets don't have phase linearity. The use of Mosfets is not excluded because even if they use higher input voltages and for this reason I have put as a limit of voltage supplied by the preamplifier up to 1Volt. Furthermore, it could drive low ohmmic loads without any particular problems. Advice on which component to use?
Most RF power amplifiers operate class C (conduction for less than half a cycle) for efficiency. This compromise pretty much throws linearity and low distortion out the window. I fail to see how you can reasonably achieve your objective. Tell me something I don't know -- perhaps I've missed something. I have a pair of RF Power MOSFETS in my KPA-500 which is good enough for voice communications and FSK-441 meteor scatter. I don't think it would work very well in your application.
 

Thread Starter

Zeno53

Joined Sep 23, 2020
46
What value should I look at on the datasheets indicating the potential linear phase loss? Is it the capacitance? The only one is as mentioned in the other post to use an all-pass filter. If there was a calculator it would be really nice. If it exists. the alternative is to use more power stages divided by band only that the circuit becomes too complicated for my abilities. What do you think about J-fet?
 

DickCappels

Joined Aug 21, 2008
7,294
There are N- and P-channel JFETs. Surely somebody could put together a complimentary pair, but JFETs are not very useful for power, that if you can even find them anymore.

A series of class-B MOSFET stages (or class-A if need be) might get you the X100 voltage gain you crave.
 

Thread Starter

Zeno53

Joined Sep 23, 2020
46
Hi @DickCappels , do you have any specific models in mind that I do a couple of searches? I explain what I want to do because i need some advice if possible, I have to create an amplification stage whose input will be managed by an amplitude of maximum 100mV (but if necessary I can reduce this output value as it is a multistage) but the important thing that it must have the output in phase from 1Hz up to about 100Mhz. I did some tests with some J-Fet and I focused on the low values Ciss aka Input Capacitance and Crss aka Reverse Transfer Capacitance but more or less all of them have a loss of linearity on the phase. These values should be significant as to what is the loss of linearity on the phase. It is fine other like an OpAmp, BJT or transistor which maintains the above characteristics. And I confirm that you don't need a lot of power and class A would be preferable for the kind of sine response it provides. Many thanks!
 

DickCappels

Joined Aug 21, 2008
7,294
The phase shift requirement over such a bandwidth will be a real challenge.

I can't recommend any specific transistors because I haven't done anything like what you want to do.
 

Thread Starter

Zeno53

Joined Sep 23, 2020
46
Up to this point it seems possible. The problem actually seems to go further. The simulation predict a gain of about 1600x. Let's not get too excited because the output, as I said, can reach a few tens of mV. But interesting isn't it? Probably maintains a fairly linear phase even over 100Mhz.



peng.JPG
 

Tesla23

Joined May 10, 2009
441
Have you looked at using a couple of MMIC gain blocks? 2V p-p into 50Ω is only 10dBm, so if you choose a MMIC amplifier with a 1dB compression point closer to 20dBm, and a few GHz of bandwidth, you should be fine.

For example, the Mini-Circuits ERA_50SM https://www.minicircuits.com/WebStore/dashboard.html?model=ERA-50SM+ has a response DC-2GHz, and about 21dB of gain, a 1dB compression point of 17.6dBm at 100MHz when run from 4.4V at 60mA.
The S-parameters show a linear phase to past 100MHz:

1619120897497.png

Cascade two of these would seem to be close to your requirements.

Your issue may be arranging the coupling/level shifting to maintain response down to 1Hz.
 

Thread Starter

Zeno53

Joined Sep 23, 2020
46
@Tesla23
Hi, yes I have looked at them and there are some interesting ones. I would need to figure out which of the S-parameters I need to look at. The simulator I use to watch them allows me to select are s11 s21 s12 and s22 in addition to the phase, they are 10Log (P), 20 Log (V), Linear V, GrpDelay. The program is called RFSim . I don't have much experience with this kind of parameters I'm working on it now. Thanks.
 

Tesla23

Joined May 10, 2009
441
The one that matters here is S21 which is the gain when fed from a 50 ohm source and terminated in a 50 ohm load. On the data it is:
1619231328225.png

This is an inverting amp - so the phase starts at -180, and changes about 13 deg over 100MHz.

Another option would be the newer ADC driver chips. Have a look on the ADI site. Here's one https://www.analog.com/media/en/technical-documentation/data-sheets/640026fa.pdf with 26dB gain, 1.9GHz BW and able to drive 2Vp-p into 200 ohms at over 100MHz. It's flat to within 0.1dB to 280MHz, phase is pretty linear:
1619231893620.png
It's a pretty impressively linear:
1619231957297.png

One of the advantages of the ADC drivers is that they are probably easier to use down to lower frequencies.

You'll need good eyesight though, it's in a 16 pin package measuring 3mm x 3mm.
 

Thread Starter

Zeno53

Joined Sep 23, 2020
46
Thanks for your kind reply @Tesla23 . I have had a look at the 6400 series and they look like really great components. The only problem is that they are drivers and they provide very low output voltages. The 6400s are powered with + - 1.5Volts and obviously cannot offer higher output voltages. The ideal stage may be something like a little Mosfet that provide a linear phase up to the frequencies but i seem to understood not exist a component like that.
 
Last edited:

Tesla23

Joined May 10, 2009
441
Sorry - I thought you were after an amplifier to provide the 2V p-p, these ADI drivers would appear to be suitable for this. This is the only amplifier you have roughly specified. Are you asking for suggestions for the power amplifier for which you have provided no other details other than it takes 2V p-p and has a linear phase?
 

Thread Starter

Zeno53

Joined Sep 23, 2020
46
Excuse me because it is probably me who am unclear.I have already prepared the circuit with the intermediate stages that allow you to drive this final stage. And further then I need a small power amp that can be driven by these stages. I attach a screenshot of the output that I have available to be able to pilot the final stage bearing in mind that it still needs some adjustments.A transistor can also be fine, even if the preference is on Mosfet circuit. But I "marry" the component that would allow me to do so

out.JPG
 
Top