delta-sigma ADC and noise cancellation

Thread Starter

kubeek

Joined Sep 20, 2005
5,733
Let´s say I have ADS1255 ADC http://www.ti.com/lit/ds/symlink/ads1255.pdf and I have its sample rate set to 50 samples per second. It looks like that at that sample rate the first notch of the digital fillter will be at 50Hz.
This tells me that there should be some pretty good rejection of 50Hz noise coming from the sensor that is connected to it.

However, there are three sensors multiplexed externally, and the multiplexer switches between those after each sample, therefore at 50Hz. Will the 50Hz rejection still work, or will it in this case ultimately have no real effect?
 

joeyd999

Joined Jun 6, 2011
4,433
Let´s say I have ADS1255 ADC http://www.ti.com/lit/ds/symlink/ads1255.pdf and I have its sample rate set to 50 samples per second. It looks like that at that sample rate the first notch of the digital fillter will be at 50Hz.
This tells me that there should be some pretty good rejection of 50Hz noise coming from the sensor that is connected to it.

However, there are three sensors multiplexed externally, and the multiplexer switches between those after each sample, therefore at 50Hz. Will the 50Hz rejection still work, or will it in this case ultimately have no real effect?
Multiplexing will have no effect on the internal filter. On the other hand, if you are doing any further DSP in code, your sample rate for each channel will be 50/3* samples per second. You must account for this.

When multiplexing inputs, this is the pertinent information you need to understand in order to get best performance:

Selection_048.png

*EDIT! Actually, < 50/3 samples per second since you are restarting the conversion slightly after the DRDY line becomes active.
 
Last edited:

Thread Starter

kubeek

Joined Sep 20, 2005
5,733
The multiplexer is external because there needs to be galvanic isolation between the input channels, but some of that excerpt sure would apply as well.
Multiplexing will have no effect on the internal filter.
I thought each channel (long wires, going to different places in a factory) could have the 50hz component at a different phase, isn´t that going to affect the filter?
 

joeyd999

Joined Jun 6, 2011
4,433
The multiplexer is external because there needs to be galvanic isolation between the input channels, but some of that excerpt sure would apply as well.
I thought each channel (long wires, going to different places in a factory) could have the 50hz component at a different phase, isn´t that going to affect the filter?
I missed the "external" MUX of your original post. Sorry.

You still need to time the conversions to your MUX action. But internally to the A/D, 50Hz noise will be rejected as per the datasheet.
 

joeyd999

Joined Jun 6, 2011
4,433
More importantly: You may run into a settling time issue now as the speed of the external MUX, it's output impedance, and the bulk capacitance of the node between the MUX output and the A/D input (including the pin capacitance of the A/D) affect your signal rise/fall time.

You will need to adjust your timing accordingly.
 

Thread Starter

kubeek

Joined Sep 20, 2005
5,733
Thank you, this will have to suffice for now, I think I understand it a little more after a bit of discussion with my collague. I am doing a new design based on an older module and was trying to understand this relationship between the mux and filtering a bit better. May have more questions in the future though ;)
 
Top