ADS114S08: Bad measurements returned from ADC after running for period of time

Edward Hunckler

Part Number: ADS114S08

Hi,

We are running into an issue with this ADC where the device returns bad measurements after running for a period of time. The ADC is set up to sequentially read 4 RTD channels. The schematic is linked below.

I have also linked the device register configuration. Through the operation of the device we periodically change INPMUX, IDACMAG, and IDACMUX to configure different channels for a single shot conversion. This is done by writing the register and then subsequently reading the register to confirm its state has been properly changed. This works well for a large number of the systems that have been produced. A number of the systems have experienced a failure point where the ADC no longer returns measurements that reflect the RTD values. It could be 0 or MAX returned. I now have one of the units with this behavior and am running tests trying to isolated the cause.

I was surprised to see that the registers seemed to be correct even while the device is failing. We have suspected that some communication noise might be causing this issue, but have been unable to isolate any.

The one register that is interesting is the STATUS register. This POR bit is never cleared in the firmware and we never write to that register. On the failed chip, this bit was cleared. It is worth noting that a software reset command plus re-initializing the device registers fixes this issue.

Have you seen any issues like this before or do you have suggestions related to this?

Thank you for any help!

Best,

Eddie

REG	NAME	Default	Failed State	Expected Initialization (by inspection of code)
00	ID	0C	0C
01	STATUS	80	00
02	INPMUX	01	--
03	PGA	00	00	00
04	DATARATE	14	37	37
05	REF	10	02	02
06	IDACMAG	00	--
07	IDACMUX	FF	--
08	VBIAS	00	00
09	SYS	10	11	11
0A	RESERVED	00	00
0B	OFCAL0	00	00
0C	OFCAL1	00	00
0D	RESERVED	00	00
0E	FSCAL0	00	00
0F	FSCAL1	40	40
10	GPIODAT	00	00
11	GPIOCON	00	00

Continuously updated and checked after each channel is sampled.

over 4 years ago

0 Bob Benjamin over 4 years ago

TI__Guru** 112725 points

Hi Edward,

This sounds like a possible transient. Can you show me what you have schematically in one of your input blocks? From the description it sounds like a transient may be turning off the IDACs. Adding some firmware diagnostics may help in determining when an issue takes place like when reference goes below the desired value be enabling the reference monitor and viewing the information in the Status byte.

I also find it odd that the Status byte is showing the POR bit cleared without you clearing it. This could be from a communication transient. Finding these events can be very difficult.

Schematically I see that the bypass cap is not the required 330nF or greater, but rather 100nF. I highly recommend that this be corrected for analog stability. I also see different ground symbols (some show GND under the symbol). Are the analog and digital ground the same point? And are you using a solid ground plane?

Best regards,

Bob B

0 Edward Hunckler over 4 years ago in reply to Bob Benjamin

Prodigy 220 points

Hi Bob,

Thanks for the reply. Here is the input block for the RTDs. Eventhough the GND symbols are different, they are connected to the same net. The analog and digital grounds are the same. Also, we are using a solid ground plane.

I will try and swap out the analog bypass capacitor with a 330nF. I am planning on adding some more diagnostics to log any changes of state while the device is running.

Thanks for the feedback!

Eddie

Data converters

Data converters forum

ADS114S08: Bad measurements returned from ADC after running for period of time