• State
  • Replies 3 replies
  • Subscribers 49 subscribers
  • Views 66 views
  • Users 0 members are here
Support feedback
Options
Options
Related

THS4551: Amplifier Configuration

Nishant Gupta
Intellectual 470 points

Part Number: THS4551
Other Parts Discussed in Thread: TINA-TI

THS455_Ckt_configuration.TSC 

Hello TI Team,

I am using the THS4551 in an RGT package 

Application requirements:

  • Differential output: 2 Vpp.

  • Input signal: Square-wave/DC levels varying from 0.2 V to 2.8 V. (Supply is single supply 5V)

  • Intended gain: 0.7 V/V, but I am unable to achieve the expected gain in hardware.

  • I would like to understand whether I should first validate the circuit at unity gain or directly test it at a gain of 0.7 V/V.

Could you please guide me on the following?

All measurement tell in hardware and TINA TI

  1. DC performance: How can I measure gain accuracy, gain linearity, input offset voltage, and offset variation in hardware and TINA-TI?

  2. Gain configuration: For a gain of 0.7 V/V, how should I select the feedback resistor (Rf) and gain resistor (Rg)? Should I use equal resistor values or adjust their ratio? What could cause the measured gain to differ from the theoretical value?

  3. AC performance: How can I measure the −3 dB bandwidth, gain flatness, noise density, and distortion? 

  4. Stability analysis: How can I evaluate loop stability, gain margin, and phase margin in TINA-TI? Would adding a 1 pF feedback capacitor provide any benefit, and how should I determine its appropriate value?

  5. Transient analysis: How can I evaluate stability, overshoot/peaking, settling time, slew rate, and step response using TINA-TI? What hardware measurements would you recommend?

  6. Large-signal performance: Since my input is a square wave with levels from 0.2 V to 2.8 V, how should I evaluate large-signal bandwidth and slew-rate limitations? I understand that small-signal bandwidth alone may not adequately characterize this response.

  7. Output swing: How should I verify that the THS4551 can achieve the required 2 Vpp differential output without clipping or distortion under my supply, output common-mode voltage, and load conditions?

I would appreciate a step-by-step procedure for performing these simulations and hardware measurements, including the recommended test conditions and important parameters to consider.

Please let me know if additional circuit details are required.

Thank you.

  • 0
    TI__Expert 3620 points

    Hi Nishant,

    I'm surprised you haven't been able to verify the gain. Starting out with G=1 or changing the gain could be good to debug your setup. How far off are you from your intended gain? What results are you getting?

    1. All the DC parameters are modeled in the TINA model based on just the typical numbers in the datasheet. They will be the same every time. In hardware the DC values of each device is going to vary, but the values should stay somewhere within the range in the datasheet.

      1. In TINA you can put DC meters anywhere but in hardware you usually measure everything on the output of the amplifier and can calculate to the equivalent input offsets. 
    2. Gain is the ratio of Rf and Rg so depending on your gain you set your resistors such that G = Rf/Rg.
      1. Resistor mismatch and DC offsets can create some gain error but how much gain error are you seeing?
    3. To measure 3dB bandwidth and gain flatness in hardware you would need something like a network analyzer or something like the BODE100. Using a function generator and measuring the output on an oscilloscope can give rough estimates too. Those specifications are usually small signal. 
      1. Noise density might need something like a spectrum analyzer or a DAQ with a low noise floor.
      2. Distortion would need a function gen and that spectrum analyzer or low noise DAQ. Some filtering around your signal source to get rid of source harmonics will help your capability there a lot. 
      3. TINA only models 3dB bandwidth and noise FYI
    4. I believe I gave you example TINA files for stability analysis but here it is again.
      6114.THS4551 Stability Eval.TSC
      1. In hardware stability will have to be evaluated indirectly by looking at overshoot or gain peaking you can't exactly break the loop.
    5. In TINA just use a square wave, or use a function generator and oscilloscope in hardware.
    6. Correct that your conditions are more large signal. I would just use a function generator to create the large signal steps and see the behavior on an oscilloscope.
    7. Measuring it in the lab with a function generator and scope can help you verify your Vpp is correct. Probing around with a scope can help you see where issues in your signal chain are coming from.

    -Evan

  • 0 in reply to Evan Kirkbride
    Intellectual 470 points

    Hello Evan,

    I have sent you a friend request. Could you please accept it.

    I would also appreciate your technical guidance on the following circuit design.

     The THS4551 fully differential amplifier (FDA) to convert a single-ended input signal ranging from 0.2 V to 2.8 V into a differential output. My target differential output amplitude is approximately 2 Vpp, with a gain of 0.7 and a common-mode voltage (VOCM) of 0.9 V.

    My main concern is whether applying a DC offset to the non-driving input is necessary to achieve the required output swing and maintain stable operation. I would prefer to connect the non-driving input to ground, if possible, without compromising the output range or stability.

    I would like to evaluate the following configuration

    1. Bipolar-supply operation: Would using ±5 V or ±3.3 V supplies allow the non-driving input to be grounded while maintaining the required output swing and stability?

    2. Output voltage limits: I need both output nodes to remain within the range of −0.3 V to +2.0 V under all specified operating conditions, with a differential output amplitude close to 2 Vpp, or as close as practically achievable after the required attenuation.

    3. Common-mode voltage: The target output common-mode voltage is 0.9 V, with a nominal gain of 0.7.

    The single-ended input range is 0.2 V to 2.8 V. I would appreciate guidance on the appropriate FDA configuration, feedback resistor selection, input biasing, and supply voltage to achieve these requirements.

    My main question is: Which configuration would you recommend to obtain the required differential output without applying an external DC offset to the non-driving input? If grounding this input is not appropriate, what is the correct way to configure the circuit?

    I would also appreciate clarification on the expected output swing, common-mode behavior, and stability for each supply configuration.

    Thanks 

    Nishant

  • 0 in reply to Nishant Gupta
    TI__Prodigy 450 points

    Hi Nishant,

    Evan is currently out of office, but he should get back to your request later next week. 

    Applying a DC offset to the non-driving input is necessary to achieve your required output swing. You cannot simply ground the non-driving input. If you do so, you would create a large DC offset between the two inputs and force an unbalanced differential drive that violates your output common-mode limit. The non-driving input must be properly biased with a DC offset (or tied to a matched DC ref) so that your output common-mode voltage properly centers around 0.9V.

    The recommended total supply voltage, according to the datasheet, is 2.7V to 5.4V, so ±5 V or ±3.3 V is not permitted since they both exceed the device rating. I recommend a +5V single supply. It gives more headroom for your 2.8V maximum input while comfortably supporting the desired output swing. Since your single-ended input range is from 0.2V to 2.8V and is gained up by 0.7V/V, both outputs (Vout+ and Vout-) sit at a min of 0.445V and a max of 1.355V, which is well within your specified range of -0.3 to 2V. 

    Regarding the TINA circuit configuration you sent earlier, I would recommend changing the feedback resistors Rf to 1k, and your gain resistors Rg to 1.43k, while still maintaining the same gain of 0.7V/V. The datasheet specifically recommends using 1k feedback resistors for typical applications and characterization setups.  

    Please let us know if you have anymore questions, thanks.

    Best,

    Amber Lee