Post Go back to editing

AD7785 internal full scale calibration

Hello,

 

We are use the AD7785 in a Oxygen measurement system. At this time we manufactured ca. 600 pieces. At the last 100 we have problems with the precision of the AD7785.

 

The problem is the value of the full scale register. The failed devices (ca. 75pc from 100) have all the Factory default value of exact hex 500000 at power on in this register.

We can set the full-scale register manually to an value that the full-scale voltage and the read out of the AD7785 is proper, but after an internal full-scale Calibration the value of the full-scale register is again exact hex 500000.

 

We test about 30pcs from the 75 and they have all the same properties.

Is this a failed charge ?

 

Best Reguards

Thomas Dauner

Parents
  • Hi, Thomas.

    Ideally, having 0x500000 should be acceptable as the power on/reset value for full scale register is = 0x5XXX00 wherein x is a don't care. However, it's very surprising to have the same values for many devices. May I know at what gain and output data rate did you perform internal calibration? Have you also tried to perform a system calibration and see if the same results? How about writing and reading back on the other registers, are they all working well except this?

    Thanks,

    Jellenie

  • Hi Jellenie,

    i write a little programm to read the registers

    Here is one of the 75 Devices with power on values in the Full-Scale and Offset Registers.

    At AI1 there are 1,0000 Volt

    after an internal zero-Scale Calibration the Offsetregister has a litte variation.

    when i set the Full-scale Register manually to 54FB00, the readout of the datareg. for

    the 1,0000V is ready, but after an Full-Scale Calibration the Full-Scale Register is always hex 500000

    at all bad devices. 

    If I change the AD7785, and this is a "good" device it is all OK, in the same Hard- and Software. 

    Today i will test this with the System Calibration.

    Thanks

    Thomas

  • p.s.

    this is a "good" dvice at the same state

  • Hi,

    I am still not sure what is really the problem and why it is fixed with that value so I still need some more details. I would like to clarify when you write to Full scale register, all devices response correctly? But if you performed internal calibrations only the good devices you have named changed its value? Is that correct? May I know how did you performed calibrations? Are you polling the RDY bit or RDY pin before reading back the registers? And what is that 1V AIN? Are you not getting correct conversions due to the value of the FS register?

    Thanks,

    Jellenie

Reply
  • Hi,

    I am still not sure what is really the problem and why it is fixed with that value so I still need some more details. I would like to clarify when you write to Full scale register, all devices response correctly? But if you performed internal calibrations only the good devices you have named changed its value? Is that correct? May I know how did you performed calibrations? Are you polling the RDY bit or RDY pin before reading back the registers? And what is that 1V AIN? Are you not getting correct conversions due to the value of the FS register?

    Thanks,

    Jellenie

Children
  • Hi Jellenie

    the problem is a malfunction at the internal full-Scale calibration. The Full-Scale Register is always setting to 0x500000, at all bad devices. This is also the power-on value.

    we buy this (bad) items from an unauthrized dealer, we think this is not original AnalogDevices.

     At the last days we test 30 pcs of the AD7785 from an authorisized Dealer and they are all OK

    But a little difference to the datasheet is always present: the read out of the ID-Register is 0x83,  in Datasheet the desired value is 0xXB 

    (we are polling the RDY-Bit)

    Thanks

    Thomas

  • Hi Thomas

    Just wondering would it be possible to get a picture of a part that is not working as expected? If we can get the date code and lot id of the part that would be really helpful in relation to helping us understand this a bit more.

    Thanks

    Aine