ADS8354: 20mV Noise

Hi,

Hoping to get some feedback on this.

Thanks,

Hello David, 

Apologies for the delay, I must have overlooked the initial notification. 

Thank you for all of the information shared on configurations, schematics, and data collected, it is all very useful.

Could you perhaps also look at how the Vocm looks like, both at the output of the ADC (REFIO) as well as after the OPA/input of the THS?

Also when measuring the AIN pins could you make sure that the common mode the ADC inputs are seeing are withing +/-100mV of VREF/2 or VREF? 

David Bradley said:

If I short at the AMC OUTP and OUTN I read between 0V and -1mV, but mainly around -600uV.

Is this measured with the ADC or the same DMM used to measure the 13mV previously?

How does channel B look? the ADC input does need a differential input with a common mode voltage mentioned before, the pins being shorted to ground are concerning. 

Best regards, 

Yolanda

Thats ok Yolanda, I figured the message had just been missed. Thanks for taking a look at the issue.

With DMM:

• REFIO_A to GND = 2.4997V
• Pin 3 of the OPA320 to GND = 1.2498V
• VOCMA of the THS to GND = 1.2499

When shorting at the output pins of the ACM3330, the rough -600uV (0 to -1mV) is what is seen on the microcontroller serial monitor, as read from the ADS8354 over SPI. The 1 to 2 LSB seen when shorting directly at the ADS8354 was also measured with the microcontroller over SPI.

Channel B has essentially been discarded. This was due to poor component selection on my behalf, and assuming writing this channel off would be easier than starting again with a new ADC, which for the specs I was looking for would also have required implementing an external VREF. I only require a single channel ADC.

Would this issue simply be explained by poor routing, or likely something else? I can connect and send you the Gerber or ODB++ if that will help?

Thanks again for your time,

Dave

Hello Dave, 

Thank you for understanding. I appreciate it.

Given the speed and the tests where you shorted inputs before and after the THS, it seems like the layout from the AMC3330 to the ADC seems sound. But if you could share at least an image of how the layout looks like, it could help confirm that.

I do have some additional questions/asks if possible to continue understanding the situation: 

Could you probe the ADC inputs with an oscilloscope? does it look as noisy as the ADC reads or as stable as the DMM used? Maybe also thee AMC3330 outputs? 

Where is the 5V DC supply being measured coming from? Does this go through the complete signal chain that leads to the ADC?

Could you also look at the data lines (SCLK, CS, SDO)?

Does REFIO_B match REFIO_A? Have you read what the ChB data looks like? The ADS8354 has a pretty good channel to channel isolation so it shouldn't be an issue even with ChB shorted to ground, but it would be good to know if there is any noise there too when making this measurement. 

Besides the 2-channel to 1-channel, was there any other ADC related change from the previous system the ADS8354 was used in? 

Also, if interested and it's still an option, I can make suggestions on single channel ADCs that could match your requirements that include a internal reference?

Best regards, 

Yolanda

Hi Yolanda,

I have PM'd you some images.

I also did forget to mention sorry. I Have shorted the input INP & INN on the AMC3330, and the noise did not resolve as it does when shorting the OUTP and OUTN of the AMC or AINP_A and AINM_A of the ADS.

Unfortunately I do not have access to an oscilloscope. I assume the signal is actually very noisy as per the ADS8354 readings, and the DMM is simply averaging as it is a true RMS DMM, so it would mask any noise.

The 5V is coming from a breadboard PSU.

This is the first use of the ADS8354.

Yes please, happy for single channel ADC recommendation.

Hello Dave, 

Apologies again for the delay, we had a holiday here that extended the weekend. 

Thank for sharing the images of the layout and the additional information. 

From what you mention about seeing the same noise when the input of the AMC3330 are shorted, but not with the THS nor the ADC I would think maybe there could be an issue there. 

Let me add a colleague that works with the AMC3330 to see if they can help with this issue. 

As for single channel ADCs there is not a large list especially when there is a need for an internal reference and differential input, most of those devices are 2+ channels. I do have some options though:

• ADS8681W : has similar specs to the ADS8354 and has an integrated analog front end (PGA)
• ADS9326 : This is a recently released device, and yes it is a 2-channel, but the 2nd channel can be set in Power Down Mode (Another like this is the ADS9224R)
• ADS8372 : is one of our not so recent devices, but it fits the requirements mentioned above. 

Best regards, 

Yolanda