Why DC offset required?

[LNA]

This may be a basic question: Why do we need DC offset for a modulator?

I'm working on a base band filter interfacing DAC and modulator. The modulator datasheet requires 0.5V DC offset at its input. Why? Why not the modulator generate such a offset inside?

Another question I have is : some design use DC coupling between DAC and modulator, while others use AC coupling. Why?

 

[FvM]

The first question is completely useless without telling about the exact component type.

Regarding your second question. If the baseband signal includes low frequency respectively DC components, you won't be able to process it by AC coupled circuits. Isn't it?

 

[vfone]

DC offset voltage at the I or Q outputs of the I/Q demodulator (non-zero DC output voltage when there is no RF signal) can be a problem for DC-coupled receiver designs due to subsequent amplification by the baseband amplifiers. If the DC offset voltage is too large, it can limit the baseband amplifier signal swing and degrade the dynamic range of the baseband A/D converters. DC offset can arise from mismatches in the I/Q demodulator itself.

The DC offsets are caused by a collection of effects. Imbalance in the mixers will lead to a DC output. This is generally a constant quantity and could be cancelled with suitable circuitry. Any leakage of the local oscillator to the input of the receiver will result in a DC signal being generated. If this leakage is via radiation coupling into the antenna, then this may vary with the local environment.
Non-linearities in the mixer may cause signals to be generated at DC from other interferers. The latter mechanisms can be time varying, and any offset cancellation needs to be able to respond to time variation. It is only when special precautions are taken to cancel DC offsets that direct-conversion architectures can be used.

 

[LNA]

The modulator is ADL5385. The datasheet says in page 13 "Baseband Inputs: The baseband inputs QBBP, QBBN, IBBP, and IBBN must be driven from a differential source. The nominal drive level of 1.4 V p-p differential (700 mV p-p on each pin) is biased to a common-mode level of 500 mV dc."

 

[FvM]

Isn't it obvious? The input ADL5385 input voltage range doesn't include the negative supply rail (GND). Practically, the said input offset is achieved automatically by connecting an industry standard DAC with differential output and respective load resistors. Of course, if you want to drive the modulator by a different source, e.g. a test generator, you have to provide the offset. In a lab, using a bias-T at each input would be the most simple method.