• State TI Thinks Resolved
  • Locked Locked
  • Replies 7 replies
  • Answers 2 answers
  • Subscribers 46 subscribers
  • Views 908 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.

ADS1299EEGFE-PDK: Re: ADS1299 EVK testing.

sushant rajbhar
Prodigy 130 points
Part Number: ADS1299EEGFE-PDK
Other Parts Discussed in Thread: ADS1299, ADS1298

Hello team,

 We have started testing the ADS1299 EVK and are stuck onto an issue.

The input configuration used here is single ended input

Jumper settings of the EVK are as follows:

1. JP7: no jumper

2. JP8: 2-3

3. No jumpers at JP25 (1-2, 3-4, 5-6)

The other jumpers are as per the factory default settings wherein the board is running on an unipolar supply.

I have provided positive input from the signal generator to pin 36 of J6 (channel one positive)

And the Ground of the signal generator is connected to pin 6 of JP25.

The input from the signal generator is a sine wave of frequency 1Hz, amplitude being 100mVp-p.

The waveform acquired is as below:

Scope analysis :

Register settings:

The issue here are:

1. The amplitude set is 100mV. However I am getting 103.9mV. How can I attain the accuracy required?

I have tried using the external reference buffer and also tried to increase the gain, the amplitude however was unchanged.

2. If at all you see the waveform, it is clipped for the negative edge. Even though the amplitude measured is more than the input signal, the representation is a clipped wave. Is it normal or an actual issue?

3. I can see some voltages on other channels too. Currently there is no input on other channels except channel 1. Why are the voltages appearing on the other channels?

Kindly confirm if the register settings are correct for single ended input.

Also, if I increase the frequency of the input signal, the waveform gets more distorted. Please help me knowing why is it so.

The waveforms for a 20Hz signal is attached below. Rest of the signal parameters and the EVK settings are unchanged:

My aim here is to use do the measurement as referential montage with the highest accuracy possible. Any inputs to achieve the same would be appreciated.

Thanks,

Sushant

  • 0
    TI__Guru 66816 points

    Hello Sushant,

    Thank you for your post.

    How are you driving the negative input of Channel 1? Based on your setup description, the IN1N pin is floating. The register settings show that you have selected to use SRB1. However, you do not have anything driving SRB1 because JP8 is in the [2-3] position and the U4 amplifier does not have a driving source (JP7 is uninstalled).

    For a single-ended input, the negative channel input should be set to mid-supply. You can jumper this mid-supply voltage from the output of U11 (JP6[2]). I would disable SRB1 for now and connect JP6[2] to J6[34].

    The positive signal can stay connected to J6[36], but you must increase the common-mode voltage to match the same mid-supply common-mode as the negative input. Your function generator's DC offset function can be used for this.

    The ground of your function generator is also not connected to anything in your current setup. Please move the function generator ground to one of the AGND test points (i.e. TP1, TP2, TP11, TP12).

    The sine wave becomes less smooth as you increase the signal frequency because you are only capturing a finite number of samples per period. This may not be enough to show the full detail of the signal at the peaks. Increasing the sampling rate or decreasing the signal frequency is the only way to improve this.

    Regards,

  • 0 in reply to Ryan Andrews
    Prodigy 130 points

    Hello Ryan,

     Thankyou for your prompt reply.

    I have followed the steps provided by you wherein I provide a mid supply to the negative input of the ADC and the GND of the signal generator is connected to the AGND.

    However for the offset part that you have mentioned (provide common mode voltage to the positive input of ADC), I believe it has to be the same voltage that is being given to the negative input.

    In this case, at negative input I am getting a voltage closed to 2.5V on the negative input of the ADC.

    The maximum offset that can be provided by our signal generator is 2V.

    I am getting a clipped waveform for some of the signals. Is the mismatch of voltages on the negatiive and positive pin the reason?

    Also, as I mentioned I want to measure the signals in referential montage.

    To do the same, I have done the following settings

    JP6-2 connected to JP25-6. In this way there will be a voltage that is driving the SRB1 pin. Because from the schematic, I have not used external reference buffer and hence J25-6 is connected to SRB1 pin..

    Positive of signal generator is connected to J6-36.

    J6-34 is left floating. (As SRB1 is internally connected to all the negative inputs)

    GND from signal generator is given to AGND.

    SRB1 is closed in the MISC1 register.

    The rest of the settings remain same as per the first post.

    The problem I am facing here is I get the required signal only when the gain is kept to 1.

    Please help me understand why is it so.

    For example, the input signal is 30mV. I get the signal when gain is 1. When the gain is increased to any number, I can only find noise and no signal.

    I am not able to correlate this.

    Also please let me know your thoughts on the second method I have performed.

  • 0 in reply to sushant rajbhar
    TI__Guru 66816 points

    Hi Sushant,

    In the first scenario, you would have an input offset voltage of about 500 mV due to the difference in common-mode voltage. This could still work as long as the PGA gain is not greater than 4 V/V assuming VREF = 2.4 V.

    You could try connecting the negative side of the function generator to the same mid-supply voltage as IN1N and setting the function generator DC offset to 0 V. This should give you a sine wave output with the same common-mode as IN1N.

    Please provide screen captures illustrating the waveform clipping and specify the channel settings for the measurement. Remember that the differential input to the ADS1298 cannot exceed +/-VREF / Gain.

    In the second scenario, you will need to move the JP8 jumper. Your original post said JP8 was in the [2-3] position, which connects SRB1 to the output of the U4 reference buffer. Please move JP8 to the [1-2] position to connect JP25[6] to the SRB1 pin. If SRB1 was previously floating, that may explain why increasing the gain caused an issue.

    Best regards,

  • 0 in reply to Ryan Andrews
    Prodigy 130 points

    Hello Ryan,

     Thankyou for the reply.

    Honestly I did not test using the first scenario. As our primary requirement is using a referential montage wherein there will be separate ref and bias electrodes lets take the second scenario forward.

    Apologies for the confusion on second scenario.

    In the second scenario the JP8 was connected between 1 and  2. In this case, the external buffer for reference is bypassed and the SRB1 pin is directly connected to REF_ELEC.

    And as REF_ELEC is connected to pin 6 of JP25, I have provided 2.5V coming out of pin 2 of JP6 to pin 6 of JP25.

    This is supplying a voltage at the SRB1 pin and hence is not floating. Also in the GUI I select the SRB1 to be closed. So now, the SRB1 pin is internally connected to the negative inputs of the ADC.

    Hence at negative input, voltage is 2.5V DC.

    At positive input, voltage is 1mV with an offset of 2V.

    Still if I increase the gain from 1 to some other number, I get no signal but only noise.

    Here the difference between two inputs is almost 500mV. I have used internal Vref. Hence Vref=4.5V.

    So ideally the gain can be set upto 8. Please correct me if I am wrong.

    The GUI and register settings are as below:

    I have an observation. For the signal with frequency range 1Hz to 65Hz, I get the closest waveform if the data rate is set to 1K or 2K.
    At frequency of 1Hz, I can see noise induced in the waveform.

    Attached is the excel file in which I have captured the results. Kindly review the results and let us know if this is normal.

    I believe if  our aim is to measure the signal of amplitude upto 1uV, the results should have been better than what we have observed.

    Test report.xlsx

    Also, I tried to attenuate the signal from 100mV to 1mV and 100uV.

    For 1mV, I am able to measure the signals using above GUI settings and jumper settings.

    But for 100uV, I am unable to get the signal on the GUI. Do I need to change some more jumper, GUI settings?

  • 0 in reply to sushant rajbhar
    Prodigy 130 points

    To add up to the observation, I disconnected the 2.5V supply that was being given to JP25-6 from JP6-2 and have provided the 1.8V supply.

    Two reasons for this:

    1. I can now match the offset that my signal generator can provide (it provides max of 2V)

    2. The ripples on 1.8V supply was comparatively less than the 2.5V supply.

    I thereby tried reducing the ripples on 1.8V further by changing C25 of the controller board to 1uV.

    The accuracy as a result increased a bit.

    Your thoughts ?

  • 0 in reply to sushant rajbhar
    Prodigy 130 points

    Hello Ryan,

     Any update on this?

  • 0 in reply to sushant rajbhar
    TI__Guru 66816 points

    Hello Sushant,

    The ADS1299 alone, with an internal short, is capable of 1 uVpp input-referred noise at Gain = 24, 250 SPS. There are several factors in your setup that are adding noise to the results. First, using Gain = 1 is going to make the ADC noise look much larger compared to the incoming signal. Increasing the gain will decrease the input-referred noise of the ADC. You can see this effect in the Noise Measurements tables for any given data rate.

    Second, additional noise will couple through the inputs from your signal source and from the ambient noise coupling from mains power. This is where high common-mode rejection becomes important. Matching the impedances in the signal path with low-tolerance components will help maintain high CMR and reduce common-mode to differential noise conversion. Additional filtering in post-processing will help to remove this noise.

    If the offset of your signal generator is set to match the voltage applied to SRB1 (1.8 V), the DC offset of the ADC channels results should be close to 0 V. There might be an incorrect setting somewhere. I would suggest probing the inputs to ensure that the function generator output is centered around 1.8 V with respect to the EVM ground (not with respect to AVSS).

    Regards,