1
$\begingroup$

enter image description hereI cannot understand how to simulate Solid State Power Amplifier (SSPA). When I directly write the Rap's SSPA model in Matlab, it does not produce BER of OFDM like published papers by other researchers'. Even no change is observed between BER with and without SSPA. But theoretically there should be significant changes between these (for low nonlinear parameter r, Asat). I have seen several matlab code in Matlab fileexchange, but those could not produce BER like published papers. What is the wrong in my code given below.

Please help me out. enter image description here

$\endgroup$
  • $\begingroup$ You need to make sure your signal is actually driving the model into saturation. What is the average and peak power of your signal? $\endgroup$ – johnnymopo Nov 26 '15 at 16:25
  • $\begingroup$ Have you taken the time to view what you are exactly feeding to the nonlinear model? $\endgroup$ – johnnymopo Nov 26 '15 at 23:45
  • $\begingroup$ Thank you very much. I am in urgent need of help. How can I use the signal to drive the HPA model? If x is the output of ifft, then the peak power will be pmax=max(abs(x).^2) and average power pavg=mean(abs(x).^2); How I can use it? Please tell me. Please help me out. I have been trying it for two months but still stuck at this point. @johnnymopo. $\endgroup$ – Shakir Hossain Nov 28 '15 at 3:41
  • $\begingroup$ Hi~I have the same question just like you.I've been trying this for a long time and still can't figure out. If you have done this, can you give me your code let me try? $\endgroup$ – Austin Shen May 10 '17 at 7:34
2
$\begingroup$

The model works fine. You need to slow down and check your (or someone else's? ) steps. I just worked through one pass of this, without pskmod, and only one value was high enough voltage to push the model into a nonlinear region. Work through it step by step, you'll see that your x values are too low in magnitude to really get out of the linear region. The x values have to approach A to get distortion.

| improve this answer | |
$\endgroup$
  • $\begingroup$ Thank you. I have tried that (using decimal number) and it works well. But problem arises when I try to simulate OFDM using parameters [mentioned in Table 2 in the newly attached image]. I can not reproduce BER for 'No PAPR Reduction' curve. Even no significant difference is found between with and without SSPA model. What is going wrong? $\endgroup$ – Shakir Hossain Dec 1 '15 at 9:16
  • $\begingroup$ Can you add a link to the paper, or the title? They say HPA saturation is 3 dB but dB is a relative number (e.g. dB vs. dBm) $\endgroup$ – johnnymopo Dec 1 '15 at 16:00
  • $\begingroup$ Also, distortion is a little unrealistic without AM-PM. Why do you not add phase distortion? $\endgroup$ – johnnymopo Dec 1 '15 at 16:28
  • $\begingroup$ Note PA simulations that only account for the AM/AM and AM/PM nonlinearity will deviate from actual results, in particular the spectral regrowth, for larger PA's where thermal effects are a significant factor. I do not have a reference handy but have come across Markov based modeling approaches that account for the "memory" from these thermal effects. $\endgroup$ – Dan Boschen May 10 '17 at 21:22
  • $\begingroup$ @Dan Boschen - I would personally like to see the Markov reference, if you can find it. Thanks! $\endgroup$ – johnnymopo May 11 '17 at 15:50

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.