This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

ADS7952-Q1: Part locks up at cold temperature.

Part Number: ADS7952-Q1

I am using the ADS7952-Q1 with a TI DSP (TMS320F28375S) in an inverter application.  Everything works correctly most of the time.  I am able to read the ADC values correctly in Manual mode.

When our inverter enables it does create more electrical noise in the system.

What I am experiencing is that the ADS7952 seems to lockup and not respond to the SPI port anymore.  We operate fine at room temperature and +55C.  When we go to cold temperature (<-20C) is when we see the lockup of the ADC.  Sometimes when it locks up the data from the ADC stays locked at a specific value.  Sometimes it locks, and then the value will change, but not be correct.  

An example would be we are looking at the voltage across a thermistor, and calculating temperature.  When operating at -50C the calculated temperature might become +60C and lock.  Sometime this value will just stay, even when we send a reset.  It will reset when we cycle power. 

Wondering if there were any ideas on what might cause this and if there are perhaps some critical layout issues that should be looked at?

If the ADS7952 were to get some bad data sent to it would it cause it to lock up or get into a mode that wouldn't respond?

Is there some particular sensitivity to noise?

  • Hi Doug

    Thanks for posting your query on E2E

    In your post you have mentioned that you are operating at -50C. I would like to confirm if this a thermistor temperature and not the device operating temperature. The device is designed to operate from -40C to 125C. So just want to make sure you are not violating device operating conditions.

    The ADC latch up can be because of various reasons. Can you please check below things

    1. Please probe +VA,+VBD and REFP pins when operating at low temperature. If there is glitch on any of these supplies then that could potentially put device in to latched state (due to +VA level going below POR level). Please check if this is not the case. If possible also monitor TMS320 power supply

    2. If you do not see any glitch on the power supplies then probe at input channel of the MUX, MXO (mux output) and AINP. If you are getting expected analog voltage at MXO and AINP and not getting correct code then we have to look in to the digital section

    3. Also make sure rest of the components and ADC peripherals are rated for -40C to 125C.

    Please let me know your observations

    Thanks & Regards


  • Hi Abhijeet,

    Thanks for your response.

    I checked the +VA, +VBD, and REFP pins.  They looked good.  One issue we have is that this is a noisy environment and we have a fair amount of noise pickup on the wires we use to bring the signals out of the chamber.  We twist the wires to try and reduce this, but still pretty noisy looking on a scope.  I did not see any dips in voltage or instabilities.

    We tested the system with all the analog input shorted and we still get errors.  What we count as errors is when we request a channels data, and the returning data is for a different channel.

    One thing we do note is the errors do not occur when our inverter is not operating.  When we get bad values from the ADC, we can turn our inverter output off, but the bad values keep coming from the ADC.  I can increase the temperature to room ambient, once the ADC lcoks, and the bad values are still coming from the ADC.  What I think I am seeing is that once the ADC starts sending bad values it keeps sending bad values.  The only way to stop this has been to cycle the power on the ADC.  Is there a mode that the ADC could get into that could explain this?



  • Hi Doug

    When inverter turns on there may be noise that is getting coupled through ground plane / power lines which may be latching up the device. Though I have not seen this behavior before.

    When you see device gets locked up at cold temp, can you please try putting device in power down mode and then waking up it again. The device can be powered down in the following two ways

    1) Setting DI5 = 1 in the mode control bits. (Available in QFN and TSSOP)
    2) Configuring GPIO3 as power down (Available in TSSOP)

    This will allow you to see if any of the device registers are getting over written due to noisy environment.

    Thanks & Regards
  • Hi Abhijeet,

    Thanks for the tips.  I'll see if we can get them incorporated and give them a try.

    Yesterday I tried looking at the SPI bus at room temperature.  When the inverter was not operating, but the SPI bus was communicating, the signals looked normal.  When the inverter was enabled the data would become corrupt exactly as if we were at cold temperature and lock.  I would disable the inverter and the values would not change.  One thing I did notice is that sometimes the signal coming out of the SDO pin would not pull up to the +VBD.  I did find that the SDO pin had gone high impedance, and the /CS pin was being pulled low before SCLK sent, and pulled high after the 16th cycle of SCLK.