• State TI Thinks Resolved
  • Locked Locked
  • Replies 22 replies
  • Answers 4 answers
  • Subscribers 46 subscribers
  • Views 859 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

ADS8556: ADS8556 BUSY/INT PIN interferes with analog INPUT PIN

Song Youwei
Prodigy 50 points
Part Number: ADS8556
Other Parts Discussed in Thread: SN74LVC16T245, ADS8555, ADS8555EVM-PDK, ADS9815, ADS9811

I disconnected the peripheral circuits of the six analog input pins of the ADS8556, but measured interference noise with a 75% duty cycle on these six pins (oscilloscope red waveform), BUSY/INT pin show in green wave. When the BUSY/INT pin was disconnected, the noise on the six input pins disappeared. Please help me determine if there is any issue with the schematic diagram—no problems were found in the power supply check. 

image.pngimage.pngimage.pngimg_v3_02s1_a15b1c7a-9723-4be9-afee-fc4270ab0e1g.jpg

  • 0
    TI__Genius 9846 points

    Hi Song,

    Thanks for your question. Could you connect the inputs to ground and let me know what the behavior is? Also, could you share what the output code is when connected to GND. This will help me understand if the ADC is functioning correctly.

    I am wondering if the noise you are seeing is perhaps a mix of the switch-capacitor input structure charge kick-back, some ground bounce, and maybe the oscilloscope probe shared grounds. This type of ADC is meant to be used with a front-end RC to drive the ADC. I would recommend reading through the datasheet and watching Precision Labs - ADC training series to get an idea on how to select these components. Alternatively, if you let me know the speed at which you intend to sample and what your input source impedance is, I can suggest RC values. But I understand you may already have these components selected in the peripheral circuits you mentioned you disconnected. Do you see the same noise when probing before an RC/driver circuit (instead of probing directly at the input pin with these components removed)?

    Best regards,

    Samiha

  • 0 in reply to Samiha Sharif
    Prodigy 50 points

    hello, Samiha

    please see this link in E2E I sent.

    TL084A: The output of the voltage follower oscillates. - Amplifiers forum - Amplifiers - TI E2E support forums

    This is the op-amp TLV084 connected to the front end of my ADS8556. Initially, the abnormality was detected at the input terminal of the op-amp TLV084 — the output of the op-amp is only connected in series with a 0-ohm resistor to the analog input pin of the ADS8556, with no capacitor connected to GND. Later, I removed this 0-ohm connecting resistor and detected the issue you mentioned at the input of the ADS8556.

    ADS8556 read frequency is 200K,  The HW/SW pin and RANGE/XCLK pin are connected to GND and 3.3V, as shown in the figure below.

    data[0:15] is connected to 74ALVC164245XTS48G/TR with 33ohm series resistor then to CPLD

  • 0 in reply to Samiha Sharif
    Prodigy 50 points

    I'm not sure if there will be an issue with this power supply configuration. The AVDD is 5V, HVDD is 15V, BVDD is 3.3V, AGND & BGND are 0V, and HVSS is -15V. The data[0:15] outputs are connected to the 74ALVC164245 chip, which is powered by 3.3V.

  • 0 in reply to Samiha Sharif
    Prodigy 50 points

    hello, Samiha

    REFIO pin is connected with 0.1uF&10uF, but the datasheet recommend 470nF. Could it be possible that the problem is caused by this capacitor being too large?

  • 0 in reply to Samiha Sharif
    Prodigy 50 points

    I measured the voltage of the REFIO pin with an oscilloscope and it is 2.5V, showing no abnormalities. However, when measuring the level of data[0:15] (from the ADC to the 74ALVC164245 chip) with an oscilloscope, as shown in the figure below, there is a fluctuation of approximately 132mV in the high level. Please help me analyze the cause of this issue.

  • 0
    Prodigy 50 points

    I tested the ADS8556 used in another product. The read speed is still 200K in parallel mode, and the data output uses the SN74LVC16T245 chip instead of the 74ALVC164245 — all other peripheral circuits remain unchanged. Similarly, when the analog input pin is floating, the waveform below is measured: the analog input pin voltage ranges from -2.967V to -4.187V, with a frequency of 200K as well. Could you please explain why the floating analog input pin has a negative voltage and why the frequency is 200K?

  • 0 in reply to Song Youwei
    TI__Genius 9846 points

    Hi Song,

    Thanks for sharing additional details. Let me take a look at this and get back to you.

    Best regards,

    Samiha

  • 0 in reply to Samiha Sharif
    Prodigy 50 points

    hello, Samiha, Has there been any progress regarding the 200K noise issue with the analog input of the ADS8556?

  • 0 in reply to Song Youwei
    TI__Genius 9846 points

    Hi Song,

    Thanks for your patience. I am wondering if this is likely ground bounce from the digital pins (like BUSY in this case) when it goes from high to low and vice versa. This ground bounce may be reduced by placing a small series resistor close to the device on the digital pins. I see you have R200 on BUSY. Is it close to the ADC? How long is the trace between the MCU and ADC for BUSY? Could you see if placing this resistor closer to the ADC helps?

    Best regards,

    Samiha

  • 0 in reply to Samiha Sharif
    Prodigy 50 points

    Have R201 on BUSY, I change R201 from 49.9ohm to 200ohm, still have problem. and I check our other products, all with this problem. R201 is close to ADS8556.

    image.png

  • 0 in reply to Song Youwei
    TI__Genius 9846 points

    Hi Song,

    Thanks for the detail. Some thoughts:

    1. Could you show me an oscilloscope screenshot of the noise on input disappearing when BUSY is disconnected?
    2. When measuring the noise on inputs, which ground are you connecting probe ground to?
    3. Is there a ferrite bead somewhere between digital and analog ground? What is your ground plane like? One solid plane?
    4. Could you try replacing R201 with a 1k ohm resistor and measuring the current drop across it with BUSY connected?

    Best regards,

    Samiha

  • 0 in reply to Samiha Sharif
    Prodigy 50 points

    hello, Samiha, 

    We use the 16-bit serial read mode. When the BUSY/INT pin is disconnected, the analog input remains a stable 0V.

    I connect probe ground to ADS8556 AGND&BGND.

    I have removed all the ferrite beads, but the problem persists.

    still have problem when replacing R201 with a 1k ohm resistor.

     I noticed there is an ADS8555EVM. Could you please test whether the same issue exists with the 16-bit parallel read mode?

  • 0 in reply to Song Youwei
    TI__Genius 9846 points

    Hi Song,

    Thanks! Yes, I have just received the ADS8555EVM I had ordered to test this. I am checking it on my side today and will get back to you with results.

    Could you measure the current drop across the 1k ohm resistor? Since the issue goes away when you disconnect BUSY, I am thinking, when connected, the BUSY pin has a large current draw. Could you measure this by measuring the voltage drop across the 1k ohm resistor?

    Best regards,

    Samiha

  • 0 in reply to Samiha Sharif
    Prodigy 50 points

    hello, Samiha

    Does the ADS8555EVM exhibit the same issue when performing 16-bits parallel reading at 200Ksps?

    BR，

    Youwei

  • 0 in reply to Song Youwei
    TI__Genius 9846 points

    Hi Youwei,

    I see some similar behavior, although slightly smaller. This is from the switch charge injection and is normal behavior for floating pins. However, I do not see the behavior where the noise disappears when I disconnect BUSY from my board. I do not see any difference when I disconnect my BUSY pin from the board. If you measure the current as I requested and see a large current, it may be because on your MCU, the BUSY pin is configured as an output instead of input, which is driving a large current into the BUSY ADC pin. This is one theory.

    This ADC is not designed to be used while the inputs pins are floating. Please ground unused pins, and drive used pins appropriately using a driver circuit. Thanks.

    Best regards,

    Samiha 

  • 0 in reply to Samiha Sharif
    Prodigy 50 points

    Hello, Samiha

    Yes, our BUSY pin is configured as an output. In practice, we have not performed any register configuration on the ADS8556 and are directly using an FPGA for data reading. The MCU determines the timing for reading ADC conversion data by monitoring the output of the BUSY pin. If the phenomenon is normal when the analog input pin is floating, is the 200k pulse noise observed after connecting an operational amplifier to the analog input also normal, and will it affect the data reading accuracy?

    BR,

    Youwei

  • 0 in reply to Song Youwei
    TI__Genius 9846 points

    Hi Song,

    As the ADC BUSY pin is an output, the MCU pin receiving BUSY should be configured as an input. Can you try this and see if it resolves the issue?

    Best regards,

    Samiha

  • 0 in reply to Samiha Sharif
    Prodigy 50 points

    hello, Samiha

    The busy detection pin of my MCU is configured in input mode, but the issue persists. Could you please send me the waveform of the floating analog input pin of the ADS8555? I want to confirm whether the noise from your ADC's analog input pin is also transmitted to the signal acquisition op-amp circuit.

    Thank you for your support recently. I have been unable to resolve the square wave signal on the floating analog input pin, and the edges of this square wave have affected the ADC sampling device. A stable level on the floating analog input pin would actually be acceptable, but the transitions of its edges are problematic. Could you confirm whether the transition edges of the floating ADS8555 analog input pin can also range from -3V to -4.2V?

    BR,

    Youwei

  • 0 in reply to Samiha Sharif
    Prodigy 50 points

    hi, Samiha

    Could you confirm whether the transition edges of the floating ADS8555 analog input pin can also range from -3V to -4.2V?

  • 0 in reply to Song Youwei
    TI__Genius 9846 points

    Hi Song,

    Thanks for your patience as we were on Thanksgiving holiday in the US. I tested the floating inputs using the ADS8555EVM-PDK. I removed the jumper on JP0, and probed the input at A0. I removed C05. I also removed R71 so the MCU would not be connected to BUSY signal. You can see the schematic here on page 23 onwards. Here is what I see (Yellow signal = A0 input signal, Blue signal = BUSY):

    The floating signal is switching between roughly -2.5V to -3V. This is typical.

    Best regards,

    Samiha

  • 0 in reply to Samiha Sharif
    Prodigy 50 points

    hello, Samiha

    Very thanks, The noise is generated by the edges of these square waves, which propagate backward through the ADC's analog input to the sampling circuit. Could you please provide any solutions? Or alternative types of ADC part numbers. Are there any references (docs/links) for floating input square wave signals(Yellow signal = A0 input signal)? I want to dive deeper into the noise generation mechanism of this ADC type.

    BR,

    Youwei

  • 0 in reply to Song Youwei
    TI__Genius 9846 points

    Hi Youwei,

    All SAR ADCs where the sampling capacitor is exposed to the input pins will experience some charge kick-back like this. The inputs must be properly driven using a drive circuit to avoid such a kickback. Some resources regarding this:

    In this case, the kickback is more drastic as the plates of the sampling capacitor are connected to HVDD and HVSS. This is larger potential difference than ADCs not using HVDD/HVSS. That is why the kickback in this device is more drastic. If you are looking for a device without such kickback, I would recommend our ADCs with an analog front-end, where an amplifier is integrated.

    You may look into the following:

    1. ADS8688: 16-bit, 500kSPS (MUX) 8-channel single-supply SAR ADC with bipolar input ranges
    2. ADS8588S: 16-Bit High-Speed 8-Channel Simultaneous-Sampling ADC With Bipolar Inputs on a Single Supply
    3. ADS9810: 8-channel, 18-bit, 1MSPS, ±16 ppm INL simultaneous sampling ADC with integrated front end
      1. other options include ADS9815 (1MSPS/CH, MUX), ADS9811 (1MSPS, simultaneous-sampling, better INL).

    Let me know if I can help further.

    Best regards,

    Samiha