• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Subscribers 38 subscribers
  • Views 1448 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.

Low Level V-toI Converter Re-design Oscillating (TI Precision Design)

Jordan Leiker
Prodigy 20 points
Other Parts Discussed in Thread: OPA277, INA118

I'm modifying the TI Precision Design for a Low Level V-to-I Converter to extend the bandwidth. Currently the design as outlined in the app note has a 1kHz bandwidth and my goal is to push that out to 100kHz (or even more ideally 1MHz). In addition to this I would like to change the output current from 0 - 5 uA to 0 - 50 uA. Modifying the output current was accomplished by dropping the Rset resistor by a factor of 10 down to 10kOhm (with output current being Iout = Vin/(G * Rset).

To extend the bandwidth I've picked a new op amp, the OPA277 with GBWP = 1MHz, and a new instrumentation amp, the INA118 with BW = 500kHz @ G = 10. While picking these amplifiers I prioritized bandwidth > input offset voltage > offset drift. To accommodate the new amplifiers I switched from the single ended power design at 5V to +/- 15V.

I've done a simple layout of the design on a quick turn PCB and have found that my amplifiers oscillate at about 25kHz regardless of my input voltage. I'm positive that I'm not compensating the amplifiers correctly. In fact I never changed any values in the compensation network (I still use the 100kOhm resistors and 10nF cap for feedback).  I've included a schematic of my design as well as slau507, the TI Precision Design.

Could anyone offer some guidance on how to overcome my oscillation/compensation issues?

Co slau507.pdf

  • 0
    TI__Expert 6275 points

    Circuit is not stable the way it is designed.  I will follow-up with a detailed analysis to post here later today. In the mean time there are some key stability techniques I recommend you study to understand the forthcoming analysis.  See links below:

    www.ti.com/precisionlabs

    Look at the Stability videos

    e2e.ti.com/.../2645.solving-op-amp-stability-issues

    Download all 4 parts in PowerPoint

  • 0 in reply to Tim Green1
    TI__Expert 6275 points
    Where you show the load between the bottom of the 10kohm resistor R2 in your schematic and ground, I will need the exact range of loads and their impedances you are trying to use this circuit for. Resistance, capacitance, inductance. The compensation will directly depend upon the load characteristics.
  • 0 in reply to Tim Green1
    Prodigy 20 points

    Thanks for looking in to this! It's still an open problem that I'd like to understand and I will study the PowerPoints that you linked in your previous posts.

    The goal with this circuit is to drive the micro-amp current waveform in to a 1uF capacitor (which happens to be in front of a charge amp). The 1uF cap will be cabled to through approximately 8ft of RG316 coax. My best (and possibly naïve) estimates for resistance, capacitance, and inductance would be:

     

    C = ~1uF  (I would think capacitance of rg316, approx. 240pF @ 8ft, and input capacitance of charge amp would be negligible compared to 1uF cap)

    L = ~1 - 1.5uH (based off of inductance calculators for coax and rg316 parameters)

    R = ~1 Ohm (conductor resistance for 8ft of coax)

     

    The intended current waveform is a haversine peaking at 50uA and a pulse-width of 20us, though if I could achieve a semi-step response that would be great too. Even if it's not possible to drive the micro-amp current in to this type of load reliably I'd like to understand the limitations and reasoning so that I can apply them to future designs.

     

    Thanks again!

  • 0 in reply to Jordan Leiker
    TI__Expert 6275 points

    See attached PowerPoint for an idea that would work if you can  connect to both ends of your 1uF load.  If the load MUST BE ground referenced it will require a more complex circuit with a wider BW instrumentation op amp which may end up being two buffer input op amps feeding into a wide BW difference amplifier to create a wide BW instrumentation amplifier with high input impedance (i.e. low input bias current). 

    50uA Pulse Current Source.pptx