Exploring the sampling rate requirements for behavioural amplifier modelling
2006 (English)In: 18th IMEKO World Congress 2006: Metrology for a Sustainable Development, 2006Conference paper (Refereed)
In this paper it is shown that for the purpose ofnonlinear power amplifier behavioural modelling, thesampling rate can be set to the Nyquist rate of the inputsignal, rather than to the Nyquist rate of the output signal bymaking use of Zhu’s generalized sampling theorem. Thisclaim is supported by measurements on a basestation poweramplifier. The findings are that the model error obtainedwhen the output signal is sampled at the Nyquist rate of theinput signal is approximately 1.5 dB higher than when thesampling rate is set to the Nyquist rate of the output signal.However, if a sampling rate of twice the Nyquist rate of theinput signal is used, which is still typically, much lower thanthe Nyquist rate of the output signal, the degradation is only0.2 dB. These are important findings that will substantiallyease the requirements on ADCs used in measurement setupsused for amplifier modelling.
Place, publisher, year, edition, pages
Amplifier Modelling, Generalized Sampling Theorem, Undersampling
IdentifiersURN: urn:nbn:se:kth:diva-7762ScopusID: 2-s2.0-84877731295ISBN: 978-162276646-8OAI: oai:DiVA.org:kth-7762DiVA: diva2:12884
18th IMEKO World Congress 2006: Metrology for a Sustainable Development; Rio de Janeiro; Brazil
QC 20100824. Tidigare titel: Exploring the Sample Rate Limitation for Modeling of Power Amplifiers. Titel ändrad 20100824.2007-12-072007-12-072014-11-05Bibliographically approved