Simulating and Tuning the MRFX1K80H 175 MHz reference circuit

[Antoine]

Hello,

I've purchased the MRFX1K80H 175 MHz reference circuit, and I'd like to tune it in order to make it functional on the band 154 MHz to 175 MHz, with a gain flatness of 3 dB max approximately. I use a 60V drain supply voltage.

I was hopping this can be achievable by just tuning some capacitors in the input and output matching network... can you confirm that it seems possible ?

In order to help choosing the rights capacitors values, I'd like to run a linear simulation of the amplifier, based on the S-parameters file of the MRFX1K80H transistors. The BOM of the 175 MHz reference circuit gives the values of all components, but concerning the planar impedance transformers at input and output, and the vertical hairpin at the output of the transistor, I have some troubles:

1. I assume that I can simulate the planar impedance transformers with a classic transformer (balun) in my linear simulator software, is it right ? And so what are the impedance ratios I have to set in my simulation, corresponding to these transformers ?

2. The vertical hairpin should be simulated by a coil, I suppose. But the value of this coil (nH) is not given in the BOM, could you precise it ?

If you can help me on these 2 questions it would be really appreciated !

Thanks in advance for your help, and have a nice day.

Best regards,

Antoine

[Antoine]

Hi Pavel,

Thanks for your reply, I will have a look to the interestings links you send !

Have a nice day,

Best regards,

Antoine

[LPP]

1.
There is no a simple way to simulate this design. You will need to perform simulation using non-linear model of the transistor and S-parameters of the planar transformers.

S-parameters of the transformer can be calculated using 3D Electromagnetic simulation. Alternatively, you can manufacture these PCB transformers alone and measure their S-parameters.

To understand the basics of this design, read this application note from NXP (now Ampleon).
https://www.ampleon.com/documents/application-note/AN10858.pdf

2. MRFX1K80H 175 MHz reference circuit could be manually optimized for 154 MHz to 175 MHz frequency band. For this purpose, you can tune C2, C7, C14, C22 capacitors. To extend lower frequency bound, the values should be increased.

Attach trimmer capacitors in parallel to existing ones. Take care about voltage rating of the capacitors (500V min for C22). Apply input frequency sweep at reasonable backoff power level (-3dBm) to avoid reliability issues.

A1. The transformer can be modeled roughly like a classical transformer with loose coupling. You can read it in AN10858. The problem is that you don't know exact parameters of the transformer used in NXP reference circuit. The model is useful for understanding the effect of capacitor values on the transformer performance.

A2. Inductance of the loop is about 23nH.

Best Regards
Pavel
NXP CAS

[Antoine]

Hello Pavel,

Thank you very much for your reply and your help.

Finally I succeed in roughly matching it to my 154-175 MHz band, by tuning the values of the capacitors (I didn't used trimmer capacitors, although it is a good idea, I just replace step by step the capacitors with new values). I reach a 4dB output power ripple in the 154-175MHz band, wich is acceptable for my application (but I'm sure it should be possible to achieve much lower ripple !).

The final capacitors values I obtained are:

• C2 = 15 pF (it was 24 pF initially on the reference design, and to my surprise it was better to decrease this value, instead of increasing it).
• C7 = 82 pF (it was 68 pF initially on the reference design)
• C14 = 68 pF (it was 47 pF initially on the reference design)
• C22 = 20 pF (it was 12 pF initially on the reference design)
• And I had to tune C13 too, to 300 pF (it was 240 pF intially on the reference design)

Concerning the simulation, I didn't have to use it finally, since I don't succeed to modelize the planar transformer correctly. But in the future it could be a good idea to measure it on a testboard, and then import the S-parameters measured into a simulation, you're totally right !

Anyway, I think it should possible to do the matching by using the S-parameters file of the MRFX1K80H transistor which is downloadable on the website, and no the non-linear model. The non-linear model is only useful to simulate the compression, harmonic, etc of the amplifier, but is not needed to match the impedances, I am right ?

Thanks one more time for your appreciated help, and have a nice day,

Best regards,

Antoine

[LPP · Accepted solution]

1.
> The non-linear model is only useful to simulate the compression, harmonic, etc of the amplifier, but is not needed to match the impedances, I am right ?

On the contrary, RF power transistor provides maximum power when it operates in non-linear mode close to saturation. If using linear parameters an amplifier would provide good gain and efficiency but lose about half of possible power (-3dB).

You can get some answers in this FAQ from Polyfet:

http://www.polyfet.com/FAQ.htm#d8

More information about matching LDMOS RF power transistors on page 3 of NXP application note:
https://www.nxp.com/docs/en/application-note/AN1526.pdf

2.
The best practice is to use NXP non-linear model for designing an amplifier.

There is an option that I didn't try but it may be useful for you. NXP provides application note "Using a Linear Transistor Model for RF Amplifier Design"
https://www.nxp.com/docs/en/application-note/AN12070.pdf

Best Regards
Pavel
NXP CAS