cc3200 injects noise to my thermocouple reading

[Ekrem 5]

Hi,

I use cc3200 to read temperature (using ad8495 ic). I connect to the cc3200 board over wifi with my cell phone and see the temperature reading on an internet browser. I have an issue though. My thermocouple circuit picks up noise from the cc3200 LP. Whe I turn off the cc3200 board, my thercouple circuit works fine. When I turn on the cc3200 LP tranmitting wifi signal, I see noise coupled to the AD8495 output repeating every 100ms (10Hz). I wonder the source of this noise with10 Hz. Of course, my next question is how to get rid of it.

Ekrem.

[spirilis 1,265]

Does the thermocouple lines have a 10nF (0.01uF) capacitor across them close to the AD8495? (or any other kind of filtering)

 

Does the AD8495 have proper bypass capacitors across its Vcc/GND?

Might be induced by momentary power drops in Vcc too... how is the board being powered?

[chicken 630]

The noise you see probably correlates with packets sent over WiFi. Even if your application doesn't send anything, the CC3200 will occasionally send a packet to maintain a WiFi connection.

 

I can see three possible reasons:

1. the thermocouple acts as WiFi antenna

2. the trace between the amplifier and the CC3200 acts as WiFi antenna

3. noise and/or voltage drop on the power rail when the CC3200 transmits

 

Given that the AD8495 is a differential amplifier, we probably can rule out #1

 

If #2 is the culprit, you could add a low-pass filter to that connection. As simple RC filter that filters anything above a few KHz would probably work. E.g. 100 ohm in series and a 0.1-1 uF capacitor to ground. http://www.learningaboutelectronics.com/Articles/Low-pass-filter-calculator.php#answer1

 

If it's #3 make sure that you properly decoupled the power rail of the amplifier. The application note suggests 0.1 uF and 10 uF capacitors. Make sure they are placed close to the amplifier.

[chicken 630]

PS: If you're using CC3200's ADC, it could also be a voltage drop on the CC3200's power line, which will be hard to fix in hardware.

 

You could implement a low pass filter in software, e.g. by averaging multiple measurements and/or discarding measurements that deviate too much from the previous average. Or some fancier methods: http://www.edn.com/design/systems-design/4320010/A-simple-software-lowpass-filter-suits-embedded-system-applications

[chicken 630]

PS2: A fix for the CC3200 ADC could also be to feed an external, well filtered reference voltage. I'm not familiar with the CC3200, so I don't know if that's supported and whether that pin is broken out on the CC3200 LaunchPad.

[Ekrem 5]

Thanks for the replies.

I have a filter across the input terminals of AD8495 as well as at its output. The input terminals have a 1.4uF across them and two 100nF caps from inputs to gnd. Filter corner frequency set to 15Hz.

I used the connection diagram as suggested in the datasheet of AD8495 (figure 29).

I have a 10uF and a 0.33uH in parallel from Vcc to gnd. I have single supply powering AD8495. There is an RC filter (1kohm, 0.33uF) between the output of AD8495 and the ADC input of cc3200.

CC3200 is powered from my laptop's usb port and I power AD8495 from the 3.3V rail on CC3200.

The output terminal of AD8495 (pin 6) is at 0.13V at room temp but jumps above 0.5V suddenly and comes back to 0.13V after some milli seconds. This repeats every 100ms. The output RC filter doesn't do good job in filtering such a horrible spike.

I inserted a metal plate holding in my hand in between the CC3200 LP and the AD8495. Depending on the position on the metal plate, the noise completely disappears.

In figure 29 in the datasheet of AD8495, there is 1Meg resistor from a thercouple terminal to gnd. If I disconnect that resistor, the noise also disappears. But AD8495 doesn't function without the 1Meg resistor.

I want to get rid of this noise if I can. If I cannot, I want to synch my ADC readings to this noise spike so that I read the ADC inputs in between the successive spikes where there is no nnoise.

Ekrem.

[chicken 630]

Hmm, where is the AD8495 positioned relative to the LP and its WiFi antenna? If you designed a BoosterPack, do you already have a ground plane under the AD8459 and the analog circuitry on its frontend?

[Ekrem 5]

I made a metal housing surrounding my thermocouple IC and its components. This solved my issue.

 

Ekrem.

[yosh 121]

I made a metal housing surrounding my thermocouple IC and its components. This solved my issue.

 

Ekrem.

Would be nice if you could post a picture of your setup ... [emoji4]

[Ekrem 5]

  I attached two pictures. One shows the board with a red rectangle indicating the temperature measurement circuitry. On the other picture I placed a U-shaped copper piece on top of the temperature measurement circuitry. The copper piece floats, not connected to any potential because I haven't seen any change with noise level connecting or floating it.

 

  Ekrem.

[Riot 0]

Hi,

 

For info there is another thread on the same issue where a few of us have been struggling to find a solution for quite some while:

'CC3200 Wifi and analogRead'

 

I would advise to read the above thread first before proceeding further below.

 

In short its looking like an Energia WiFi Library function bug which is occasionally clobbering the multiplexer (choice of pin type) for the ADC pin.

I have a very tight design and have properly decoupled throughout. Also I've undertaken shielding but really none of this I don't believe is the real problem.

 

Basically for any low frequency sampling (such as batt level monitoring) you can get away with just slugging the ADC output with a 'hefty' smoothing cap.

For high speed sampling of course you cannot do that and your stuck with the corruption during WiFi transmit.

 

Any further comments or advice gratefully received.

 

Stuart

UK.