Post Go back to editing

Some channels have problems with sampled output values

Category: Hardware
Product Number: ltc2320-16

Hi, I am using LTC2320-16 chip for some research projects. According to the instructions of the chip manual, I used FPGA to send the corresponding SCK and CNV signals to the chip, and observed the SDO signals of 8 channels. All 8 pairs of differential pins are grounded during the test. The test results show that some channels have an SDO output value of around 0x0000 (e.g., channels 1, 3, 8), some channels have an output value of around 0xFFFF (e.g., channels 2, 4, 5), and some channels have a value of around 0x0500 (e.g., channels 6, 7).

Attached is the waveform diagram of the test. The waveform of the oscilloscope from top to bottom is SCK, SDO, CNV

correct channel

wrong channel

This is my schematic(Differential signal pairs are grounded in the test)

Thank you very much for giving me some help!

regards,

zhu

Thread Notes

Parents
  • Hello zhu,

    The output data of LTC2320-16 is in 2's complement format. This means that 0x0000 (decimal 0) and 0xFFFF (decimal -1) are only 1 LSB apart, so getting values closely around those with grounded inputs looks OK.

    The other value you mention, 0x0500, is decimal 1280, which multiplied by 125uV/LSB turns out to correspond to ~375mV at the input. This, you should not be getting with a grounded input. Could you provide screenshots for when you get such values? Are you consistently measuring a value around the 375mV in the channel? Or is it around 0V and sometimes the higher value shows up?

    Regards,
    Emilio

Reply
  • Hello zhu,

    The output data of LTC2320-16 is in 2's complement format. This means that 0x0000 (decimal 0) and 0xFFFF (decimal -1) are only 1 LSB apart, so getting values closely around those with grounded inputs looks OK.

    The other value you mention, 0x0500, is decimal 1280, which multiplied by 125uV/LSB turns out to correspond to ~375mV at the input. This, you should not be getting with a grounded input. Could you provide screenshots for when you get such values? Are you consistently measuring a value around the 375mV in the channel? Or is it around 0V and sometimes the higher value shows up?

    Regards,
    Emilio

Children
  • Hi,Emilio

    Thanks for your reply! I also found the relevant description of binary complement in the manual. As for the reason why the sampling value of some channels is 375mv, I found that it should be due to the poor contact of the differential input ports during my test, which resulted in one port being grounded while the other port was suspended, rather than both ports being grounded at the same time.

    Thanks again for your help!

    regards,

    zhu