• State Resolved
  • Locked Locked
  • Replies 11 replies
  • Answers 2 answers
  • Subscribers 47 subscribers
  • Views 357 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

PGA855: Question about input voltage rating

Marcel Robitaille
Prodigy 40 points
Part Number: PGA855
Other Parts Discussed in Thread: PGA849, PGA280, INA500,

Tool/software:

Hello,

I would like to use the PGA855 with a single supply and a signal of +/- 4V with a common-mode voltage of 2.5V. My question is: does "Input voltage (V_I)" of 7.5 Electrical Characteristics mean the Common-mode input voltage or the voltage of IN+ and IN- individually? Reading "Input voltage", I would assume the latter, but this seems to contradict figures 9-1 to 9-5. Furthermore, the similar-looking PGA849 has "Common-mode input voltage" in the same place in the table.

Is this a good amplifier for this application? I want something simple (I like A0 A1 A2 over SPI), with a single channel, and below unity gain options. I think the PGA855 is intended for slightly higher voltage applications, but it's the closest thing I could find. Any recommendations would be much appreciated.

Thanks,

Marcel

  • 0
    TI__Mastermind 20782 points

    Hey Marcel, 

    Sorry for the confusion in the nomenclature in the datasheet, please let me know if it would be helpful for me to update it. 

    Yes the spec, V_I is the input common mode voltage: 

    These align with Figures 9-1 through 9-5 in the datasheet's highlighted areas: 

    Vs = ±15V, Max/Min VICM = ±12.5V accordingly; similarly, Vs = ±5V, Max/Min VICM = ±2.5V. 

    As you can tell there is a boundary established in the relationship between what the max and min expected output differential voltage is and what the max/min common mode must be. I recommend leveraging the PGA85X-INPUT-OUTPUT-RANGE-DESIGN-CALC. 

    The PGA849 is for a single-ended output and the PGA855 is for differential output - these are both great options. Since you are already using SPI you may also be interested in PGA280. If not, I recommend staying with the PGA855. 

    All the best,
    Carolina

  • 0 in reply to Caro Walter
    Prodigy 40 points

    Hi Carolina,

    Thank you very much for your answer.  The calculator is very helpful.

    Can you please help me understand how VOUT (Max) and VOUT (Min) are calculated? It doesn't seem to follow the V_OUT specification in the datasheet.

    I would like VOUT (Max) = 3.3V and VOUT (Min) = 0V to read the signals with a 3.3V ADC with as much range as possible.

    Like I said before, the input is +/-4V centered on 2.5V. I'm willing to use a negative supply (+/- 5V).

    Thanks,

    Marcel

  • 0 in reply to Marcel Robitaille
    TI__Mastermind 20782 points

    Hey Marcel, 

    Sorry for the delay in response due to the Thanksgiving holiday.

    Boundary plots show you the headroom limitations of the op amps within the programmable gain amplifier.
    Here is a block that shows the resistors with the muxes and the instrumentation amplifier configuration on the inside:

    & here is a picture from a different device that explains how each op amp within the structure contributes to the limitation: 

    As you can see the input is limited by the headroom of the two first op amps. 

    An alternate solution, is to use a difference amplifier, we have an integrated version available: INA500, specifically the C variant that has a gain of 0.25 V/V. 

    Since this is integrated this is a fairly simple solution and saves on price and space with the following limitations: 

    • Reduced accuracy, works well with 8bit or 12bit ADCs 
    • Supply range is 1.7V to 5.5V, therefore the single supply of 5V doesn't let you get to the full 0V (limit is (V-) + ~30mV)
    • Limit of G of 0.25, with the input swing of ±4V (8V) and the expected output of 3.3V, the ideal gain would be 3.3/8 = 0.41V/V. The INA500 has 0.5V/V and 0.25V/V to not rail the device, we would have to pick 0.25V/V which is only 2V of usable data, this can be referenced up to be within the output swing of the device. 

    If these limitations do not work, then I recommend designing the difference amplifier following this reference design: Difference amplifier (subtractor) circuit (Rev. A) a precision high voltage op amp (±5V) can be used, low tolerance resistors, and precise gain of 0.41V/V. 

    Please let me know which path forward you choose, and I can reassign to my colleagues as applicable. 

    All the best,
    Carolina 

  • 0 in reply to Caro Walter
    Prodigy 40 points

    Hi Carolina,

    No problem. I completely understand.

    Is it possible there's a bug in the calculator? It doesn't reflect my measurements of PGA855EVM the at all.

    Here are my settings in the calculator:

    Here is a measurement I took with my oscilloscope with the PGA855EVM. Ch1 and Ch2 is the differential input signal. Ch3 and Ch4 is the output of the amplifier. Math1 (top right) is Ch1-Ch2. Math2 (bottom right) is Ch4-Ch3. The settings are identical to the calculator (VS+=LVDD+=5V, VS-=LVDD-=-5V, VOCM=1.65V, Gain=1, VICM=2.5V). While the calculator gives VOUTP (Max) 1.85V VOUP (Min) 1.45V, I'm measuring more than +/-3V differential on the output.

    I'd prefer to stay with the PGA855 if possible. I'm already done the design using it and we're on a tight schedule. It seems to work.

    Thanks,

    Marcel

  • 0 in reply to Marcel Robitaille
    TI__Mastermind 20782 points

    Hey Marcel, 

    Interesting! I will take it to the lab tomorrow and verify. 

    All the best,
    Carolina

  • 0 in reply to Caro Walter
    Prodigy 40 points

    Hi Carolina,

    Were you able to verify this in the lab?

    Best,

    Marcel

  • 0 in reply to Marcel Robitaille
    TI__Mastermind 20782 points

    Hey Marcel,

    Apologize for the delay - the e2e servers have been unreliable lately.

    I spoke with the engineer who designed the calculator.

    The input/output ratings are conservatively guardbanded to ensure the device works over temperature, over corner and lot variations while ensuring inside a linear operating range.

    Thank you for attaching the oscilloscope screen captures, although I am mostly counting divisions - it appears your input is ±2V centered around VICM = 2.5V. This is likely why in lab it hasn't appeared as dire of a slam to the supply voltage rail.

    The design is outside of the linear range and will likely slam to a power supply rail. I do not recommend proceeding forward with this design as it is not robust. Is there a way to reduce the 2.5V VICM on the signal? or is there another power supply rail that is higher than 5V, maybe 12V?

    All the best,
    Carolina

  • 0 in reply to Caro Walter
    Prodigy 40 points

    Hi Carolina,

    Thanks for the answer. Unfortunately, no, the sensor has a fixed centre voltage of 2.5V. I can't do anything to change it. Yes, 12V is available. I guess 10V would be ideal, right?

    All the best,

    Marcel

  • 0 in reply to Marcel Robitaille
    TI__Mastermind 20782 points

    Hey Marcel, 

    Wonderful! With the following conditions, I suggest the following design:

    • VS+ = 12V, VS- = -5V
    • LVSS+ = 5V, LVSS- = 0V 
    • VOCM = 1.7V (increase from 1.65V)
    • Input (cannot be changed) = ±4V input centered VICM = 2.5V 

    With the larger VS+, the input is no longer limited due to VICM = 2.5V. 

    With a gain of 0.5 V/V, the expected output is 0.7V to 2.7V (for the expected 0 to 3.3V). 

    Please let me know if this works for you. 

    All the best,
    Carolina

  • 0 in reply to Caro Walter
    Prodigy 40 points

    Hi Carolina,

    Thanks very much for your assistance. That sounds good.

    Is it necessary to change VOCM to 1.7V? I would prefer to keep 1.65V. Also, I was using LVSS-=-5V instead of your 0V.

    All the best,

    Marcel

  • +1 in reply to Marcel Robitaille
    TI__Mastermind 20782 points

    Hey Marcel, 

    If you keep LVSS- as -5V, there is no need to change VOCM. 

    All the best,
    Carolina