Ad7927 has abnormal voltage jump of channel 0

The abnormal voltage jump of channel 0 occurred in our application, which occurred in many projects. Attached are our own test results and problems

Test conditions

Test condition description

Ad value status

Filter pool status

Report voltage value

Normal state

Normal state

All channels are normal

All channels are normal

All channels are normal

The software writes 0x0800 to ADC

Simulate default state

Only 0 channel value can be adopted

It jumps periodically between 0 and channel voltage. Because of the complement problem, the acquisition value is too large

Channel 0 jumps periodically between 313 V and 313 + true value / 2, while the voltage of other channels remains unchanged

The software writes 0x0840 to ADC

Simulate the default state and change the channel to 1 channel

Only one channel value can be taken

It jumps periodically between 1 and channel voltage. Because of the complement problem, the acquisition value is too large

One channel changes periodically between 313 V and 313 + true value / 2, while the voltage of other channels remains unchanged

The software writes 0x0870 to ADC

Analog default state, 1 channel, power mode changed to 11 (00 is illegal, 11 is normal application)

Only one channel value can be taken

Stable acquisition

1 channel acquisition value = 313 + true value / 2, the other channel voltage is unchanged

Normal state, short circuit ADC power supply during operation

Analog ad power down

Only 0 channel value can be adopted

It jumps periodically between 0 and channel voltage. Because of the complement problem, the acquisition value is too large

Channel 0 jumps periodically between 313 V and 313 + true value / 2, while the voltage of other channels remains unchanged

The power down test state of analog ADC is consistent with the state of software writing in 0x0800, so it is suspected that 00 is the default state of chip register after power on. In this state, the channel is 0, and the power mode is illegal. The output ad value will jump back and forth between the real voltage and 0. However, due to the problem of complement, the software will be too large after calculation. At this time, if the voltage mode is written, the chip will work normally without voltage jump.

Question 1: is the default state of the chip 00?

Question 2: is the state of pm01 related to ad jump?

Question 3: in the state of 0x0800 or power failure recovery, why does the ad value jump between the real value and 0?

Question 4: is there any other reason or abnormal condition that can enter this fault state? Such as: power overvoltage, undervoltage, long time no communication or frequent communication, communication error, etc

Parents
  • 0
    •  Analog Employees 
    on Feb 25, 2021 1:57 PM

    Hello Jack,

    I believe the AD7927 could power up in any mode. That’s why you need to issue a dummy conversion to properly set the PM1,PM0 for the following conversions.

    The PM0, PM1 = 00 is an invalid combination. It is a not supported mode, similar to the autoshutdown mode indeed, so that would explain why you see 00 data results every second conversion. So please make sure you write the proper PM on DIN pin.

    Regards,

    Lluis.

Reply
  • 0
    •  Analog Employees 
    on Feb 25, 2021 1:57 PM

    Hello Jack,

    I believe the AD7927 could power up in any mode. That’s why you need to issue a dummy conversion to properly set the PM1,PM0 for the following conversions.

    The PM0, PM1 = 00 is an invalid combination. It is a not supported mode, similar to the autoshutdown mode indeed, so that would explain why you see 00 data results every second conversion. So please make sure you write the proper PM on DIN pin.

    Regards,

    Lluis.

Children
No Data