LTC2494 Pinout Spec Error

It seems, based on my measurements, that the MUXOUTP / ADCINP and MUXOUTN / ADCINN pins are switched in polarity in the datasheet when compared to the actual IC.

So, the MUX routes the positive input to MUXOUTN (pin 27) and the negative input to MUXOUTP (pin 24).

And the positive input of the ADC is ADCINN (pin 26) and the negative is ADCINP (pin 25).

Can someone confirm that this is in fact the case?

Top Replies

Parents
  • +1
    •  Analog Employees 
    on Apr 5, 2021 2:16 PM

    Hi,

    That is not the case. The pinout is correct.

  • Hi,

    Thank you for the quick reply!

    I double checked it with a different measurement set up this time, and you are right, the pin out is correct.

    I also discovered the source of the confusion.

    I connected 2.25 V to CH0 and 1.2 V to CH1, and then I made the ADC do a differential measurement on CH0+ CH1- continuously.

    I measured the MUX outputs (P: green N: yellow) and the SDO line (red) while CS was continuously pulled low.

    In double speed mode this is what I get (difference: blue):

    In single (normal) speed mode however, it looks like this:

    So it seems that in normal speed mode, it switches up the inputs for the first conversion (used for offset compensation) and then in the second conversion it switches the inputs the right way. I guess, the result of the first conversion gets mathematically negated and added to the second conversion. This way the offsets cancel out and the result gets doubled but that is solved by just discarding the LSB.

    The curious thing is though, and this is what caused the confusion the first time, is that after the second conversion is done and before the next conversion is started, it switches the outputs back to the inverted state.

    I was first only measuring these MUX outputs with a multimeter after doing a conversion with the ADC (in normal speed mode) to see if I managed to set the MUX up properly. And since, apparently, it switches to the inverted state before it goes into the sleep state instead of leaving the MUX as is, I was measuring this negated voltage with the DMM.

    It would be nice though, if this behavior was described in the datasheet, since like this users could reasonably expect that the outputs of the MUX will have the voltage that is consistent with the last measurement and try to use that voltage by connecting to it with circuitry in addition to the integrated ADC. + It also causes problems if test points are placed on these signals for QA and the production test system is commissioned parallel to the product.

    The datasheet still needs to be updated with the correct equations anyhow...

Reply
  • Hi,

    Thank you for the quick reply!

    I double checked it with a different measurement set up this time, and you are right, the pin out is correct.

    I also discovered the source of the confusion.

    I connected 2.25 V to CH0 and 1.2 V to CH1, and then I made the ADC do a differential measurement on CH0+ CH1- continuously.

    I measured the MUX outputs (P: green N: yellow) and the SDO line (red) while CS was continuously pulled low.

    In double speed mode this is what I get (difference: blue):

    In single (normal) speed mode however, it looks like this:

    So it seems that in normal speed mode, it switches up the inputs for the first conversion (used for offset compensation) and then in the second conversion it switches the inputs the right way. I guess, the result of the first conversion gets mathematically negated and added to the second conversion. This way the offsets cancel out and the result gets doubled but that is solved by just discarding the LSB.

    The curious thing is though, and this is what caused the confusion the first time, is that after the second conversion is done and before the next conversion is started, it switches the outputs back to the inverted state.

    I was first only measuring these MUX outputs with a multimeter after doing a conversion with the ADC (in normal speed mode) to see if I managed to set the MUX up properly. And since, apparently, it switches to the inverted state before it goes into the sleep state instead of leaving the MUX as is, I was measuring this negated voltage with the DMM.

    It would be nice though, if this behavior was described in the datasheet, since like this users could reasonably expect that the outputs of the MUX will have the voltage that is consistent with the last measurement and try to use that voltage by connecting to it with circuitry in addition to the integrated ADC. + It also causes problems if test points are placed on these signals for QA and the production test system is commissioned parallel to the product.

    The datasheet still needs to be updated with the correct equations anyhow...

Children
No Data