• State TI Thinks Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 47 subscribers
  • Views 7433 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.

OPA659: 50 Ohm coax cable driver

xyz xyz
Intellectual 1000 points
Part Number: OPA659
Other Parts Discussed in Thread: OPA656

Hello

I want to amplify a small signal with a gain of +1V/V or +5V/V with OPA659.  (alternative amplifier for me is use of OPA5657 with a gain of +2V/V or +10V/V)

as per page 17 of OPA659's datasheet (section 9.1.3.3) we should use Riso between the output and the load to drive a capacitive loads.

a coaxial cable such as RG174 or RG316 have the below spec:

R=0.28 Ohm/m L=200nH/m C=101pF/m

the standard solution to drive a coax cable is like it:

section 9.1.3.3 has mentioned we should check the Phase Margin to check the quality of the circuit.

the phase margin is 87.675' and it's larger than 60'. I think the system is a little slow like a first order system. (in aspect of Step response)

what should I do?

How much Riso should be?

OPA656 unity gain amplifier has a PM of 55'  for my application. is it a good PM for data acquisition systems?

Regards

  • 0
    TI__Genius 15975 points

    Hi,

    The OPA659 being a voltage feedback amplifier, higher the gain it is set to the lower will be the closed loop BW (which makes the system slower). This means the system become more stable with higher phase margin and less overshoot. One way to reduce the operating phase margin from 87.7' to 60' is by reducing the OPA659 gain from 5V/V to 2V/V. This effectively increases BW of the system thus reducing rise time of the pulse response (governed by the relationship trise/fall = 0.35/f-3dB for a first order system).

    While driving a transmission line or coaxial cable, it is always recommended to match source and load impedance of the system to the characteristic impedance (Zo = sqrt(L/C)) of the transmission line for maximum power transfer at the load. For a 50-ohm system, I would recommend the Riso to be 50-ohms at the OPA659 output. It is important to note that setting the Riso to 50-ohms should have already decoupled the loop gain phase shift from the coaxial cable capacitance; thus setting the op amp phase margin solely on the closed loop gain of the op-amp set by Rf and Rg.

    A PM of 55' with OPA656 unity gain amplifier should be good PM for data acquisition systems. You could see from below curves that 55' PM should result in not very high overshoot. However, I wanted to point out that the OPA656 has much lower slew rate (290V/us) as compared to the OPA659 slew rate (2550 V/us) because of the difference in large signal BW. Lower slew rate could affect the drive capability of the OPA656 for large signal swings while driving longer coaxial cables - something to consider for data acquisition system part selection.

    Best Regards,

    Rohit

    Q vs Phase Margin:

    Transient Response vs Q:

  • 0 in reply to Rohit Bhat
    Intellectual 1000 points
    Dear
    I have two stage amplifiers. I want to build a circuit to measure an known Impedance. Actually the accuracy of Module and Phase measurements should be less than %1 and 0.2 degree respectively.
    which behavior is better for the first and second stage of amplifier? Bessel or Butterworth?
    we have a little ringing in the Butterworth but it's faster than the Bessel. But I think the Bessel is better for arbitrary signals. the Chebyshev filter has the smallest transition band and I think it's good for noise suppression.

    Regards
  • 0 in reply to xyz xyz
    TI__Genius 15975 points
    Hi,

    If you are interested in sweeping the frequency to extract the phase information, then I would suggest Bessel filter as it presents constant group delay (constant phase shift) across frequency. If not, then I would suggest going with the Butterworth filter as it would give magnitude accuracy for a given frequency point.

    Best Regards,
    Rohit