Unknow source of Noise on SAADC input

[Noon]

Hello,

I need to read resistive sensors values with SAADC. These sensors going throught multiplexer and amplifier.
Infortunaly, the signal going into the SAADC is noisy. It seems to be radio event, but when i use a radio_notification flag, the noise is'nt synchronized with the radio ON/OFF events. I'm sure my sensors acts as an antenna, but i can't find the source of this noise. Any idea?

[Håkon Alseth]

Hi,

 

Do you have anything happening in your firmware every ~90 ms (seems to be the period between the noise)? For instance a SPI/TWI clock signal, which is routed next to this analog signal?

Is there any load on this pin? If yes, how much?

 

Kind regards,

Håkon

[Noon]

Thanks for your reply! Nothing in my firmware is set up to 90 ms. The same noise happening even if i stop all activity on the SPI bus. The problematic pin is P0.03 of DK52, wich is connected on the output of an amplifier (MAX9939).

[Noon]

It seems that the noise comes directly from sensors, as i was thinking, my resistive pannel act as antenna. If i disconnect P0.03 from the amplifier output, it's the same problem. Vcc is also clean. 

I have already done this hardware with other dev boards (not BLE), and i never had this kind of noise before, it's why i was thinking about radio interference. 

The picture of the zoomed noise looks like radiated frequency modulation to me, more than a capacitive coupling of high speed digital tracks. But, the frequency of noise is about 1 MHz, so i don't know where this comes from (my probe is not RF, but the bandwith is supposed to be 25MHz).

[Håkon Alseth]

Have you tried connecting a small cap on the input (between input and GND; 100pF ish, and do the same on the output for that matter) to see if this removes the noise source?

 

Kind regards,

Håkon

[Noon]

After a more precise measurement, the noise frequency is around 660kHz. Do you know the use of these radio frequency bands?

I put 150pF on the input and output of the amplifier, and the noise is still there.

I will implement active low pass filter, there is an free OPamp on the MAX9939 package.

[Håkon Alseth]

Hi,

  

Noon said:

the noise frequency is around 660kHz. Do you know the use of these radio frequency bands?

No, the radio operates at a much higher frequency. Is this noise present if you power off the nRF? For instance, if you set the states needed for the opamp and the other external circuitry, then enter sleep on the nRF (system off for instance, disables all clocks, thus radio etc)

What else is running in your firmware except for the radio? Any bus-peripherals UART, SPI, TWI, PWM?

 

Kind regards,

Håkon

[Noon]

Same thing happens when i turn th nRF into sleep mode.

There is SPI and TWI on my hardware.I'm doing 8 readings of capacitive touch driver data each 20 ms (TWI), and multiplexing + analog convertion of 6 resistieves channels. Each opérations is about 1.5 ms. In the picture, low state of yellow signal represents length of readings, and the analog signal is the blue one.

One important thing is that i've embedded the same battery management than the thingy (XC6804A4E1 + XCL210C33) on my hardware, but, even without battery i have this noise still happening...

[Noon]

Well... Nothing with Nordic, i've found that the capacitive driver (CAPA1208) caused this issue. The noise corresponds to the sampling rate and scan duration. I have to synchonise it with a flag and take care when doing ADC measurements.

Thanks for the support.