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.

ADS1198 Defib protection

Other Parts Discussed in Thread: ADS1198

HI,


I am using ADS1198 (analog front end) in my project. when i apply a 5KV Defib pulse(in differential mode) on the input lines, AFE stops/freezes comunication with the controller. Resumes back only after system power off and power on.

Please suggest, what could be the possible reason for this and how to mitigate it

Attached is AFE input lines with defib protection circuit in my current design

Thanks,

Santosh

  • Santosh,

    I am moving your post to the correct forum. The precision data converter group will be able to help you.
  • Hi Santosh,

    Can you please show me how the ADS1198 is connected? Also it is possible that if a large current is flowing through the ground connection that it causes unexpected voltage levels elsewhere on the board that could affect the ADS1198. Try probing the digital ground for the ADS1198 during the defib pulse and see if the voltage moves significantly away from its normal value.

    Regards,
    Brian
  • Hi Brian,


    Thanks for the reply


    Please find attached ADS1198 schematics

    yes, as you said there are some unexpected voltage levels on the ground plane (shorted analog gnd and digital gnd with zero ohm on my board).

    Can you please suggest any possible solution to mitigate this ?

    Thanks,

    Santosh

  • Hey Santosh,

    In general, it is best to decouple the respective grounds as much as possible. If they are to be connected, try to do so as far from the ADC as possible. For debug purposes, is there any way to remove the 0 ohm resistor that connects the grounds for now and use a separate supply for the digital portion of the device? If you can do that, re-run the experiment in this configuration to make sure that the voltage change on the digital ground was the culprit.

    Brian

  • Hey Brian,

    Thank you fro your quick response

    Can please elaborate what do you mean by "decouple as much as possible" because i have already decoupled(with >100uF capacitance) respective grounds and power supplies.

    And as you said i am trying to isolate analog and digital grounds and doing the defib test again. I will reply you back my observation once i complete the test.

    But, even if i assume that its a voltage variation in digital ground that is making AFE loose communication with controller, and by isolating analog and digital ground i shall avoid those voltage variation in digital ground, but these voltage variation will still be there on analog ground right? will that not make analog section of AFE malfunction.

    Please provide your suggestions

    Thanks,
    Santosh
  • Santosh,

    When I said to decouple, I meant to make the grounds as electrically independent as possible so that an event like a defib pulse that draws the voltage high momentarily on the analog ground does not affect the digital part of the device. The reason I suggested that the digital ground coupling might be the problem is because after a discussion with my colleagues, we agreed that for the "stopping/freezing" in communication you described is probably related to the digital communication functionality of the device and that momentary variation in the analog supply should not shut down communication with the device. Isolating the analog and digital power supplies will either confirm this is the case or force us to explore other possible issues.

    Brian
  • Hi Brian,


    I have repeated the defib test by isolating analog ground and digital ground but no success, AFE still hangs/losses communication.

    Thanks,

    Santosh

  • Hello Santosh,

    I will need to discuss this with a few of my colleagues before suggesting another possible fix. I appreciate your patience.

    Brian
  • Hi Brian,

    Just want you to know what i am doing and how is my setup,

    i am generating the test signal in AFE ADS1198 and the same test signal is being displayed on my device display(my device has a 2.8" LCD TFT display)

    when i apply a defib pulse in differential mode, this test signal on display gets freezed/does not get refreshed.

    Just to make sure that its not either display or MCU problem i had made some characters on display to blink at certain rate and they keep blinking even after application of defib pulse only the test signal generated in AFE gets freezed and even SPI interface between AFE and MCU hangs.


    Waiting for your valuable inputs

    Thanks,

    Santosh

  • Hello Santosh,

    When the ADS1198 analog power supply rails become too close in voltage, the part undergoes a total reset (analog and digital). Since you mentioned that the voltage on the digital ground for the device spiked when the defib pulse was provided and the two grounds were connected before you isolated them, you should see if the analog supply rails get close enough to have triggered a POR (AVDD-AVSS < 2.7 V may do it). If a reset was triggered and the host has not altered its communication in any way, then the host may place the ADS1198 SPI module in an unknown state when the device finally emerges from POR.  If it's possible, please post some scope capture images here of the supply rails during the defib pulse and the inputs during the defib pulse.

    When you say that the communication hangs, what precisely are you seeing from the ADS1198? Is the digital output low? Does the DRDY output pulse? Some scope captures of the SPI bus following the defib signal would also be helpful.

    Regards,

    Brian

  • Hi Brian,


    Please find attached Scope images for SPI bus before and after application of defib pulse

    Thanks,

    Santosh

  • HI Brian,

    Please find attached scope captures for SPI bus before and after defib pulse

    When i say SPI freez , i observe that there is no SPI CLK, DRDY line not pulsing(remains high always) and SPI data line is low always.

    Once i power off and power on every thing comes back to normal condition.

    Thanks,

    Santosh

  • Hi Brain,

    Did you get my reply mail

    Thanks,
    Santosh
  • HI Brian,

    Please acknowledge if you got my reply with scope capture images. Because i did not receive a acknowledgement mail from TI for my recent posts

    Thanks,
    Santosh
  • Hello Santosh,

    I received the images you posted. In the second image, is the host controller supposed to be sending a serial clock and is not doing so?

    Brian
  • Hi Brian,

    Yes host should be sending the SPI clock but in our code it will send the clock only when it host gets a DRDY# pulse from AFE.

    When Difib is applied you can see in the picture that DRDY# always stay high it does not pulse so no SPI clk from host untill DRDY# goes low.

    Thanks,

    Samtosh

  • Santosh,

    I noticed in your schematic that the START pin was connected to a pulldown resistor and then off the page presumably to the host. Is that pin held high by the host always? If not, bring that pin high or issue the SPI START command. If the device is undergoing POR after the defib pulse and the START pin is held low, the device will not convert. In addition, try talking to the part with a command where you expect the device to talk back (i.e. RREG, RDATA) after the defib pulse and see if it talks back. To see if the device enters a POR state on the defib pulse, measure the analog supply voltages when it happens and see if AVDD-AVSS ever dips below 2.7 V.

    Brian
  • HI Brian,

    Yes start pin is pulled low, but during power up this start pin is pulled high by the host and then released back to low after some time.

    Thanks,
    Santosh
  • Santosh,

    I recommend holding the START pin high indefinitely if you wish to continuously convert. Try holding that pin high and see if the DRDY pin falls low some time after the defib pulse.

    Brian
  • Hi Brian,

    I just tried, pulled start pin high all the time and applied the defib pulse but no success.

    Thanks,
    Santosh
  • Santosh,

    Have you tried speaking with the device at all after the defib pulse? Try reading a device register or reading data on command. Perhaps try issuing the SPI RESET command and see if the device responds. I would wait a minimum of 2^16 clock cycles after the defib pulse before interrogating the device. In the meantime, would you mind sending me your layout? If you do not feel comfortable attaching the files here you can email them to b-pisani@ti.com.

    Brian
  • Hi Brian,

    Yes we have tried speaking with device after defib pulse(freeze condition), but for doing this i need do go to debug mode of controller which will reset the entire system and once reset every thing will work normal. so i could not actually read anything from AFE when it is freezed.

    And i have tried SPI reset also with a delay in some mili seconds(delay way more than 2^16 clock cycles) it works properly only when we add a huge delay.

    I think this is not something related to software, i want to resolve it in hardware because in any other case (other than defib) if AFE hangs or malfunctions i need to indicate that event to the user.

    I have sent my PCB files to your mail ID


    Thanks,

    Santosh

  • HI Brian,

    I just re-tested,
    Pulled START pin always high (no control on START pin by HOST) and applied the defib, its working! when a defib pulse is applied system goes blank and comes back for normal operation with in 3sec.
    Yesterday when you suggested to do the same i forgot to isolate START pin from HOST which was pulling it low even though i tied it to high, hence it failed yesterday.

    But once concern is what are the consequences of keeping START pin always high ?
    Is it ok to keep START pin always high.

    Please provide your suggestions


    Thanks,
    Santosh
  • Santosh,

    If you're going to be constantly collecting data from the device, I would recommend holding the START pin high at all times. The one step you must take though, is if you're going to perform any device configuration at startup you should issue the SDATAC (stop read data continuous) command after startup but before configuration. This ensures that register writes are not interrupted by new conversion data. Once all registers are configured to your desire, issue the RDATAC (read data continuous) to place the device back into continuous conversion mode. Descriptions of the functionality of both the RDATAC and SDATAC device commands are located on page 36 of the ADS1198 datasheet. A helpful flow chart illustrating a proper startup routine is located on page 66.

    Brian

  • Hi Brian,

    One observation is, there is a change in data(SPI DOUT) from AFE to HOST after defib pulse.

    I have attached the scope captures of SPI DOUT line before and after defib.

    As i said earlier i am generating a test signal in the AFE and displaying the same on my device display but after the application of difib pulse this test signal is no more displaed instead a straight line is displayed. Why this change happens ?

    Requirement is device should come back to previous operation mode even after defib hick up.

    Thanks,

    Santosh

  • Santosh,

    From your scope capture images, it looks like the test signal that is being converted after the pulse is producing nearly the same result as before the defib pulse. From what I can tell, the most significant bits are nearly identical. I do not know your SCLK rate so I'm not sure what the exact data codes are but they definitely look to be in a close range. If the codes coming from the device are correct, then the problem with the display is likely not related to the ADC.

    Brian
  • Hi Brian,

    Scope capture images are not same.

    Is there any configuration change happening because of the AFE getting reset. I dont think AFE is generating the same test signal after difib test. see the attached images test signal displayed before and after defib test ?

    does the AFE loosing its configuration after defib(or POR) ?

    Do i need to reconfigure AFE back to generate a test signal after every defib pulse ?

    Thanks,

    Santosh

  • Santosh,

    All configurations will be lost on reset. Registers will take default values when the device is powered on again. I'm in the process of reviewing your layout to see if I can identify what is causing the device to reset.

    Brian