AD7329 Spi Problem

Obvious question: Is MUXOUT+ connected to ADCIN+ as shown in Figure 34 in the data sheet?

Hi, nubato .

I agree with MTF_Walker 's question. Where does the MUXOUT+ connected to from your schematic? Can you share your whole schematic here?

Warm regards,
Jo

Hi, yes it is connected to ADCIN+

I would take a look with a scope at ADCIN when it is sampling.  I suspect the 2k filter resistance is causing incomplete settling before sampling.  Page 23 of the data sheet (Driver Amplifier Choice) warns that without a buffer, the source resistance needs to be 'small'.  The AD7329 is not in the ADC driver tool but I used one with a similar input capacitance and sample rate and the kickback is dramatic:

It may be better to buffer the filtered signal before ADCIN or to use an active filter instead, where the output resistance of the op-amp presents a low impedance to the switched capacitor front-end of the ADC.  In this way, the kickback error can be reduced to below 0.5 LSB within the acquisition time.  You will also need to take into account the charge kickback from the MUX, so the VIN signals need to be buffered too.  Page 25 of the data sheet shows suitable buffering options in Figures 45 to 47.  You don't have to use the SE to Diff configurations of Figures 45 and 46 (although the differential input will give you 6dB more SNR) but you should buffer your single-ended MUXIN input as well as the ADCIN input.

Thank you for answers. We will discuss your suggestions. I fixed my problem with setted Weak/Three-State bit to 0(three-state). I dont know why I cant run with weak but it fixed with three state.