Removing offset in IA AD8253

AD8253 is a very good device, large bandwidth, low noise, but with an incredibly large offset VOS, specified as 150 + 900/G [uV].

My application is a very low distortion, moderately low noise amplifier inside a multi-channel acquisition system, This device was the right compromise.

Do you have suggestions for the correction of this offset? My considerations are:

1) the usual circuit for VOS correction (consisting of a potentiometer between a positive and negative reference, either analog or digital, and a series high value resistor to decouple it) cannot be used not to worsen noise performances and CMRR

2) can the REF input be used for this? of course driven by a low impedance source of signal, like an additional OpAmp

3) can I correct the offset at a later stage, such as a buffer or a filter? this VOS is large and the IA gain can be high [ (150 uV + 9 uV)*100 = 15.9 mV at output for G=100 ]; this means to operate the IA with a large offset: is it detrimental for its electrical characteristics, such as the THD, linearity and so on?

thank you

andrea

Parents
  • Thanks Scott. ... and don't mind for any delay, you have been quick anyway!

    The acquisition system is equipped with an ARM9 that sets offset correction for each channel by charging a large capacitor that drives a pA high-Z OpAmp. And this can be done in real-time for any change of gain and other settings, together with a general calibration by means of calibrating DAC.

    As you suggest I can add another offset trimming circuit, but after along the channel chain before the ADC, so it's an additional degree of freedom: I have a low-pass filter that uses OAs with large GBP (50 MHz or so), so not as good for offset (low drift however, that's more important).

    Going to the influence on THD and linearity, since I'm using your AD7983 and slower versions, like 7980 ..., and they ensure -105/-110 dB THD, I am quite "susceptible" to this matter:

    so, I learn that THD vs. Vcm graphs are the right place to look for when in doubt of the effect of offset; I mean external Vcm and internal offset can be assimilated equal for the effects on THD?

    cheers

    andrea

Reply
  • Thanks Scott. ... and don't mind for any delay, you have been quick anyway!

    The acquisition system is equipped with an ARM9 that sets offset correction for each channel by charging a large capacitor that drives a pA high-Z OpAmp. And this can be done in real-time for any change of gain and other settings, together with a general calibration by means of calibrating DAC.

    As you suggest I can add another offset trimming circuit, but after along the channel chain before the ADC, so it's an additional degree of freedom: I have a low-pass filter that uses OAs with large GBP (50 MHz or so), so not as good for offset (low drift however, that's more important).

    Going to the influence on THD and linearity, since I'm using your AD7983 and slower versions, like 7980 ..., and they ensure -105/-110 dB THD, I am quite "susceptible" to this matter:

    so, I learn that THD vs. Vcm graphs are the right place to look for when in doubt of the effect of offset; I mean external Vcm and internal offset can be assimilated equal for the effects on THD?

    cheers

    andrea

Children
No Data