Class D Power amplifier using ADS Song lin

Class D Power amplifier -----using ADS Song lin @utk June 30

Outline Why select Class D? Compare different device Simple Class D architecture Load-pull to give out Zout Matching and simulation results of ideal narrow band Class D PA Broad band matching

Why Class D? Class D PA works in the switching mode, with a square wave voltage and a half wave rectified sine wave of current. In its ideal switching mode, when Vds<>0, Ids=0; when Ids<>0, Vds=0. Class D PA can achieve very high frequency close to 100%. Although it is very nonlinear, we still can use the LINC technique to kill the IMD product.

Two kinds of Class D PA

Compare the device(I) For high frequency, ----the higher the better For high on/off switching speed, ----the shorter the better For high efficiency, the on-resistance of a switching device must be as low as possible to minimize the power dissipation in the switches during the positive half cycle. ----- Rs is the smaller the better. For high Power output, ----- BV(beake down voltage) the higher the better For high Gain, ----the bigger the better For power dissipate, ----the small the better

Compare the device(II) Conclusion: Also for the wide band application, we should chose the component whose Zout and Zin has very little variety in some frequency range. I suggest to use the MRF 282 SR 1. -----N-channel Enhancement-Mode Lateral MOSFETs

Basic Class D PA architecture

Load Pull to give out Zout

Narrow band input and output matching and simulation results

Wide band matching using coaxial A conventional design allows the coaxial transformer to transform the impedance to obtain a match the low end of the band, then additional low-pass matching sections to lower the impedance at the upper band edge.

Using the MRF 282 S to simulate narrow band VMCD @300 MHz

Push_pull structure

Narrow band matching (input, output)

Narrow band simulation results

Final schematic

Final simulation results(1)

Final simulation results(2)

The problem remain: 1. The Class D PA need a resonator tank to pull out the 2. 3. 4. fundamental signal, to filter out the third time signal, so I decide to divide the band into 3 parts, one from 30 to 88 MHz; 88 MHz to 200 MHz; 200 MHz to 500 MHz. We can separate the signals by filter bank. For the real device, the Rs isn’t very small, so the efficiency can’t be so high. Because of the , the Vds and some overlap with Ids, it also kill some efficiency. To achieve better performance at low frequency band, I have to increase the Vgg. ADS is very hard to converge when simulation.

Thank you!!!