Just quick rules of thumb suspect to variation in different designs: Capacitively couple the input as well as the shunting branch of the negative feedback circuit. This will set the amount of DC gain to unity and isolate the input from any DC offsets of the source. Then make sure both chip inputs see about the same impedance considering DC current’s path to common: e.g. in non-inverting configuration, if non-inverting input is referenced to common with a 22K resistor then use a feedback resistor of 22K as well. If the shunting part is not capacitively coupled then use input’s reference resistor that is equal to shunting resistance.
First try matching the input and gain resistors to 0.1%.
DC offset arises from current flowing out of the input pins, searching for a path to ground. The min/max values of current are given in the data sheet. The IR product is multiplied by the gain of the amplifer.
You can be more certain of eliminating DC offset by using a servo amplifier -- basically a portion of the output is sampled, rectified, and integrated (with a capacitor). This DC value is inverted and fed to the amplifier, nullifying the offset. You can see examples of how it's done in the National Semiconductor application note 1192.
http://www.national.com/an/AN/AN-1192.pdf
DC offset arises from current flowing out of the input pins, searching for a path to ground. The min/max values of current are given in the data sheet. The IR product is multiplied by the gain of the amplifer.
You can be more certain of eliminating DC offset by using a servo amplifier -- basically a portion of the output is sampled, rectified, and integrated (with a capacitor). This DC value is inverted and fed to the amplifier, nullifying the offset. You can see examples of how it's done in the National Semiconductor application note 1192.
http://www.national.com/an/AN/AN-1192.pdf
sadly input bias currents are not equal, so perfect matched input resistance does not improve things very much.
Even with perfect compensated input currents the input offset voltage swamps most of the effort.
As audio is about AC, setting DC amplification to one (use a feedback cap!) seems to me the most economical approach.
Of course DC-servos can eliminate offset perfectly.
regards
Even with perfect compensated input currents the input offset voltage swamps most of the effort.
As audio is about AC, setting DC amplification to one (use a feedback cap!) seems to me the most economical approach.
Of course DC-servos can eliminate offset perfectly.
regards
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.