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.

THS4551: Low frequency and single ended input, differential output configuration for the FDA

Part Number: THS4551
Other Parts Discussed in Thread: PGA112, , TINA-TI, REF6025, OPA365, OPA2365

Hello T.I.

I am using the PGA112 to amplify and offset low frequencies (0-1000Hz). The PGA112 goes into the THS4551 in a single input, single supply configuration. Will the methods in the data sheet for the THS4551 satisfy the requirements for balancing the differential output signals at these low frequencies? I am using a MFB low pass filter configuration with the THS4551.

Thank you in advance,

Jeff B.

  • Jeff,
    You should have no problem with the THS4551 at these low frequencies. The performance only gets better at lower frequencies because of higher open-loop gain.
    What supplies are you using and can you please let me know:
    - the amplifier gain
    - what is the output swing expected from the PGA112
    - What is the Vocm of the THS4551 set to
    - I assume the THS4551 is driving an ADC. If so, what is the expected input common-mode and swing limitation of the ADC

    I would recommend using TINA-TI to simulate your circuit to verify proper working.
    -Samir
  • Thank you for the quick reply,

    -The PGA112 will be using binary gains 1 through 32.
    -The Vocm pin is tied to ground with a small capacitor
    -Your assumption is correct. a differential swing around a 3.1V mid supply with a 2.5V reference.

    I am having trouble with TINA-TI THS4551 simulation.

    -While I have an expert, how does input common-mode output from the pga112 affect the the THS4551

    Thank you in advance,
    Jeff B.
  • Hi Jeff,

    - In  my previous post I meant to ask, what is the THS4551 gain configuration?

    - What are the power supplies of the THS4551 set to, 5V and GND?

    Also, are you telling me that the ADC is expecting an input common-mode signal of 3.1V? Usually the output common-mode of the THS4551 should match the input common-mode level of the ADC.

    Please send me a TINA schematic of your circuit so that things are clearer.

    The THS4551 input common-mode is a function of :

    - the common-mode output level of the PGA

    - The Vocm of the THS4551 (mid-supply in your case)

    - and, the feedback and gain resistors of the THS4551.

    The above 3 factors will act as a resistive divider and will set the input common-mode of the ADC.

    -Samir

  • Hi samir

    -The LPF for the FDA was configured in TI Filter Pro for 0 dB gain in the passband region. In the AC transfer characteristics model there is -6dB in the passband probably due to my modification.

    -The power supplies are coming from an LDO  and ground which supplies both the THS4551 and the ADC with 3v in single supply configuration. The digital supplies for the ADC are supplied from a separate LDO.

    -The ADC has a 2.5V reference, REF6025 IC is used.  

    -the Output Common Mode of the THS4551 is, like you said, mid-supply (1.5V). The TINA-TI model has the THS4551 pin tied to 1.5v. When I connect it to ground or leave it floating there is an error. 

    -The input of the TINA-TA model is 1Vpp centered around zero. Does this violate the input restrictions of the THS4551?

    -TINA-TI file is attached. I could not find a model for the PGA112 so it is not included.

    -thank you for the insight on the ICM of the ADC. That was actually a very good explanation of a concept that had been confusing me.

    THS4551_LPF_share.TSC

  • Jeff,

     I made a typo when I made the statement "The above 3 factors will act as a resistive divider and will set the input common-mode of the ADC". What I meant to say is:

    "The above 3 factors will act as a resistive divider and will set the input common-mode of the FDA (THS4551 in this case)".

    The Input Common Mode presented to the ADC is simply the output common mode of the THS4551, which in this case is mid-supply.

    Now to your other questions:

    With regards to the TINA circuit, the way you have connected the single ended source is incorrect. see attached TINA file for corrected circuit.

    From the perspective of the THS4551, the input signals CM is 0V and the output CM is 1.5V. With Rf = Rg = 1kOhm, the input CM at the input pins of the THS4551 will thus be the midway point, i.e 0.75. You can verify this by running a dc simulation and probing the input nodes. The 0.75V input CM will not violate any of the THS4551 specs.

    The last thing I will mention is to check the ADCs input common-mode spec and ensure that it can take a 1.5V common-mode signal when Vref = 2.5. In many cases the ADCs input common-mode range is either Vref (2.5V in your case) or Vref/2 = 1.25. Both numbers do not match the current 1.5V CM that the THS4551 is putting out.

    -Samir

    THS4551_LPF_Samir.TSC

  • Edited for clarification.

    Samir, Thank you for your contribution. I'm sorry for my poor explanations. Thank you for your patience and please let me explain the system set-up with better detail. 

    -The ADC and the THS4551 share Vcc of 3V which is supplied externally from a high PSRR LDO.

    -The output common mode voltage supplied by the THS4551 is the ADC mid-point reference.

    -The ADC has Positive and Negative Vref pins. REFP = 2.5V and REFN = Gnd. This is designed to keep the ADC output code unsaturated when the THS4551 output is within .5V of the positive supply rail. 

    -The desirable CM output from the THS4551 is 1.5V with a maximum of +/-1V swing around 1.5V.



    -I did not attempt to connect a single ended source in the tina-ti file. I was trying to model the differential circuit. I was looking for a single supply FDA with the proper low-pass filter characteristics as a starting point. (ie -105 or more dB at 128kHz)

    -The Vocm pin is supplied with 1.5 volts in the simulation. Should it not be floating or tied to ground for a split supply Vocm? 

    -Will I need a termination resistor to balance the differential outputs? The simulation looks fine, but I am about to breadboard the circuit.

    -The discrete RC filter is required with the MFB filter in the FDA to improve its roll-off to -60dB/decade. 

    -Is there is Tina-Ti model available for the PGA112? I want to use it to offset a bipolar input prior to the FDA. Also I would like to use it to scale the input. The transducer I am using is very sensitive.

    Thank you for the clarification, and the advice for probing the inputs.
    Jeff B.

  • Jeff, see answers below

    -The Vocm pin is supplied with 1.5 volts in the simulation. Should it not be floating or tied to ground for a split supply Vocm?

    Samir: The Vocm defaults to mid-supply, however TINA doesn't like floating nodes so the easiest thing to do is to connect the Vocm pin to GND through a 10-100nF capacitor. In a real world circuit you should do the same thing. The capacitor provides noise reduction at high frequencies.

    -Will I need a termination resistor to balance the differential outputs? The simulation looks fine, but I am about to breadboard the circuit.

    Samir: Is this at the output of the THS4551? Why do you think you need a termination resistor. Is that mentioned anywhere in the datasheet? Send me a picture of the circuit you are thinking of and I will confirm.

    -The discrete RC filter is required with the MFB filter in the FDA to improve its roll-off to -60dB/decade.

    -Is there is Tina-Ti model available for the PGA112? I want to use it to offset a bipolar input prior to the FDA. Also I would like to use it to scale the input. The transducer I am using is very sensitive.

    Samir: Doesn't look like there is a model. If you require a model, you can post to the Precision Amplifier E2E forum requesting this. Might take some time.

    Thank you for the clarification, and the advice for probing the inputs.

    Jeff B.

  • Samir

    -I will use a standard operational amplifier for now. Thank you for the pointer to the resource, I will make the request for future users.

    -In your TINA-TI file how does your voltage source, 'VG2', model the single supply PGA112?  I know the PGA can only swing rail-to-rail at the outputs in the single supply setting. Even though the THS4551 can accept more negative voltage input relative to the amplifiers ground, the PGA cannot provide this as far as I know.

    -your simulation performed great, thank you for the reference. I created a new circuit based off of it using OPA365 in voltage follower configuration with a DC offset. The problem is that AINN and AINP are no longer centered around 1.5v. From what I've read in T.I. technical literature, the DC offset and the impedance from the OPA365 causes the unbalanced differential outputs. Is this correct? What is a good starting point for balancing the outputs?

    -By termination resistor I am referring to a resistor referenced in documentation describing methodology to balance differential signals in a single to differential configuration.

    -My ADC has two Nyquist frequencies. Caused by the ADC internal filtering and the delta-sigmas  modulator frequency which peaks at 0dB at multiples of 128KHz. The MFB and discrete RC filter are an attempt to attenuate this modulator noise at -100dB or better. I noticed the filter response gradually retreats upwards after 128KHz in the circuit analysis. Will the harmonics at multiples of the 128Khz ADC modulator frequency null my initial attempt to attenuate their amplitudes if the positive slope of the filter continues its upward trend? In figure one there is a sharp change in phase for AINP. Is this the reason for the continuation of the negative gain slope for 'VM1'?

    Thank you in advance,

    Jeff B.

     figure 1: AC characteristics extended to 10MEG generated from 'THS4551_LPF_Samir'

    THS4551_opa365_single_ended_Jeff.TSC

  • Hello Jeff,

    -  The voltage source doesn't substitute the PGA112, however I need to drive the THS4551 in some manner, so I use an ideal voltage source. We 1st need to make sure each component in the signal chain works well with ideal sources before we move on to testing the entire signal chain; that's how I typically do my designs in order to not compound errors from multiple sources.

    - If you drive one side of the THS4551 with a 1.5V CM, you will need to do the same on the other side as well. Remember the fully differential amplifier rejects a common mode signal and amplifiers a differential signal. In your case with 1.5V on one side and GND on the other side, you are feeding in a differential common-mode which is amplified by the THS4551. In the case of your simulation, the left side of R3 (1kOhm) needs to be driven by  a low impedance source with an output of 1.5V. You can use a dual channel, OPA2365 to do this. Use the second channel of the OPA2365 to buffer a 1.5V output. See attached simulation.

    - For noise analysis, please run a noise analysis in the simulation rather than an AC analysis. I ran a noise analysis looking at output noise and do not see any issues.

    -Samir

    THS4551_opa365_single_ended_Samir.TSC

  • Samir,

    Thank you for the time you took to answer and give helpful advice. They are all very helpful.

    you are TIe2e#1

    best regards,

    Jeff, B

  • No worries Jeff. I think you will be pleased with the THS4551 performance. Let me know if you need any help during the debug phase.

    -Samir