ABOUT THE HARDWARE IN BRIEF
I have designed a fully universal 8 channel differential Analog Input Module (Thermocouple,RTD,mA,0-10VDC). All the channels are non-isolated. At the input of each channel there is a common-mode(cut-off = 1kHz)and differential filter (cut-off = 50Hz). Module is designed for low frequency measurement (10Hz max). I am using ADG1207 low capacitance mux for channel switching. The mux connects channel to IN-AMP (AD8226). AD8226 inputs are provided bias current path through two 10MegaOhms resistor to ground. An SPST switch is applied in series with Rg resistor for gain switching (for few mVs in case of RTD and thermocouple gain is 150 and in case of 0-10VDC gain is 1. Reference terminal is grounded.The output goes to a divider which is referenced to 2.5VDC). Finally the output is given to ADuC834 that has 24bit Sigma Delta ADC along with Sinc Filter for 50-60Hz.
ABOUT THE FIELD PROBLEM
In an end user's application a total of four number of 3-wire RTDs is connected to this module and an additional 0-10VDC input from a VFD is applied to this module. All the RTD readings are OK. The moment 0-10VDC input from a VFD is applied to any input channel a 10degree Centigrade offset in RTD reading is observed. The 0-10VDC itself reads okay but it is affecting RTD's reading. While measuring RTD 0-10VDC channel is disconnected then how is it affecting RTD reading. I have not used a filter post ADG1207 mux. But filter right at the channel input must be enough according to me. In simple terms i observed cross talk.
Could you please throw some light on this.