1

I have been trying to design an LNA in the 24 GHz ISM band using CE3520K3.

The FET is inherently unstable and one cannot use resistor loading at the input, resistive loading at the output, source inductor or drain gate feedback (RLC) circuit to stabilize the FET.

The only way to stabilize it is probably to use a bandpass filter at 24 GHz, with a very low bandwidth.

This filter might be used at its out- or input, but the best way might be to use it at the FET's output to preserve the noise figure.

Another limitation at 24 GHz is that there are no capacitors that behave like capacitors (i.e. the series resonant frequency of the capacitors is much less than 24 GHz). So has anyone designed an LNA using CE3520K3?

I have designed filters at 24 GHz using AWR MWO and Ansys HFSS. I can make the amplifiers stable using a bandpass microstrip filter. Also, I am using interdigital microstrip capacitors for blocking the DC. For dc bias at the gate and the drain, I am using quarterwave microstrip lines to get an RF open.

I have asked the guys at CEL for an application note or any circuit that they have used to design the LNA. But they do not seem to have any application note or a circuit.

This is the plot of the input and output stability circles. You can see that the pink color plots are the input stability circles and the blue color plots are the output stability circles. These circles are plotted in AWR MWO.

Dave Tweed
  • 168,369
  • 17
  • 228
  • 393
SMALL
  • 56
  • 3
  • 3
    *Another limitation at 24 GHz is that there are no capacitors that behave like capacitors (i.e. the series resonant frequency of the capacitors is much less than 24 GHz)* So as any experienced RF designer would know you would not make the resonators using discrete components but you'd "make your own" on a PCB substrate. You can only use discrete caps where the behavior at 24 GHz is less relevant like AC coupling and supply decoupling. If this information is new to you then you have some studying and learning to do before you can design a circuit for 24 GHz. – Bimpelrekkie May 23 '19 at 14:31
  • 1
    (edit: Bimpelrekkie said what I meant to say, but better) So, get one of the classics (e.g. Pozar's *Microwave Engineering*) and start building filters :) – Marcus Müller May 23 '19 at 14:31
  • 1
    *So has anyone designed an LNA using CE3520K3* Yes, the designers of that CE3520K3 should have because they want to be able to measure its behavior in a circuit. If you sell a transistor with a low NF then someone is going to build an LNA with it so you better be the first to do that because if your customer cannot build an LNA meeting the specs. you list, you will be in trouble. That customer will shop elsewhere from now on. – Bimpelrekkie May 23 '19 at 14:34
  • Are there any stability circles? What are they? – analogsystemsrf May 23 '19 at 14:45
  • 1
    I have designed filters at 24 GHz using AWR MWO and Ansys HFSS. I can make the amplifiers stable using a bandpass microstrip filter. Also, I am using interdigital microstrip capacitors for blocking the DC. For dc bias at the gate and the drain, I am using quarterwave microstrip lines to get an RF open. – SMALL May 23 '19 at 15:10
  • I have asked the guys at CEL for an application note or any circuit that they have used to design the LNA. But they do not seem to have any application note or a circuit. – SMALL May 23 '19 at 15:13
  • @Bimpelrekkie and analogsystemsrf This is the plot of the input and output stability circles. You can see that the pink color plots are the input stability circles and the blue color plots are the output stability circles. These circles are plotted in AWR MWO. The link to the picture is https://i.stack.imgur.com/bpdgL.png. – SMALL May 23 '19 at 15:25
  • *But they do not seem to have any application note or a circuit* That begs the question how they can then can guarantee the specifications and usability of their transistor. So basically with this transistor **you're on your own**. That might be a risk, if this was for a commercial application,I would shop elsewhere (I mean, use a transistor with a vendor that does have an application note etc.). Also, you do seem to know your RF stuff, then why mention basic things like capacitors not being usable at 24 GHz? It sets us on the wrong foot about your experience level. – Bimpelrekkie May 23 '19 at 15:57
  • So there ARE regions of stability. How to bias, and provide Zin and Zout, to exploit these regions? – analogsystemsrf May 24 '19 at 05:02

0 Answers0