• State
  • Locked Locked
  • Replies 5 replies
  • Subscribers 46 subscribers
  • Views 747 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.

LMP7300 Reference Output Impedance

Sheldon Rubin
Intellectual 540 points
Other Parts Discussed in Thread: INA122, LMP7300, OPA333

My application is a lithium battery pack for use with underwater robots, The LMP7300 is part of a circuit that detects if the battery pack is unplugged and disconnects power from the output connector to prevent damage or hazardous conditions in case this happens underwater. A Wheatstone bridge, powered by the raw 21 - 29V battery voltage, is amplified by an INA122 and used as the input to an LMP7300. The purpose of the Wheatstone bridge is to detect if the resistance of a short circuit through the connector rises above 100 ohms so that the connector can be safely de-powered.

I'm using the REF output of the LMP7300 to drive the REF input of the INA122. Based on the 0.5 mV/mA Load Regulation specification of the reference, the DC output impedance seems low enough to preserve the high CMRR of the INA122.

The problem we're seeing is that a sudden load change on the battery, and the resulting change in battery voltage, causes a significant transient on the output of the INA122 that triggers the LMP7300.

The INA122 has a gain of 20. The LMP7300 hysteresis is set to plus and minus 5 mV.

I can think of two possible causes of the problem:

  1. The INA122 is also powered directly from the battery voltage and the PSRR may be insufficient
  2. The CMRR has been compromised by the REF input drive impedance

I don't think the problem is a layout issue. The majority of the load on the battery is from BLDC motor controllers. The miniature size and dense population of the BMS board makes rework or tinkering painful. Before I start making hardware changes I was hoping to reduce the range of possibilities, it wouldn't be easy to power the INA122 from a more stable power source. I was hoping someone could comment on the suitability of using the LMP7300 REF out to drive the INA122 REF input with no buffering or filtering.

Thanks for the help!

Sheldon

  • 0
    TI__Guru 61516 points

    Hello Sheldon,

    How fast is the battery load change? What is the edge rate?

    Can you share a schematic?

    Do you have the recommended R and C on the reference line? (see LMP7300 DS Section 7.3.1)

    The LMP7300 is a micropwer device. The reference output amplifier is relatively low bandwidth. While it may be able to source current at DC, it cannot handle (react fast enough) the transients. The Ref output is not designed to drive large currents (to save power).

    The INA is a conventional 4-resistor instrumentation amp, which means the reference input is expecting a low source impedance (both DC and AC). Fast common mode transients are a high-frequency signal - and the AC CMRR of the INA will degrade over frequency - which results in a "glitch" during any fast transients. This is expected behavior.

    The 5uF reference cap should be able to supply the current during the transition. You can play with the value of the R and C (but don't go lower than 100 ohms - as it may oscillate). Try tapping of the top of the cap for lowest AC impedance.

    Monitor the VREF line during the transient. If everything is okay (with the 7300), there should only be a very small glitch. Using a large cap with low ESR (ceramic) should provide enough charge reservoir to supply the current.

    Regards,

  • 0 in reply to Paul Grohe
    Intellectual 540 points

    Hello Paul,

    Thank you for the very timely response. The load is primarily BLDC motor controller with 6-step trapezoidal commutation and a 40kHz PWM frequency. Four motors can run simultaneously (but not sunchronously) at 10A average current. The battery is a 7S6P with Panasonic NCR18650B cells. The actual vehicle will have more chassis wiring that my test bench so there will be more inductance but I see switching times of 400nS or less on the output of the motor controller. I've added the TPS7A4901DGNR to provide a more stable excitation voltage to the Wheatstone bridge and the analog circuitry. Previously, the excitation was the raw battery voltage.

    I currently only have a 47nF X7R on the REF output, I can change it to 5uF X7R with a 130 ohm resistor. How confident are you that this will work? Should I buffer REF with a micropower OP Amp and, if yes, can you recommend one that has sufficient drive capability?

    The BMS schematic is attached. Everything is working great except for the Flooded Connector Detection Circuit.

    Thanks for the help,

    Sheldon

    DiNIS-005-SCH-B.PDF

  • 0 in reply to Sheldon Rubin
    TI__Guru 61516 points

    Hi Sheldon,

    I assume the "Kill" signal is connected to the battery negative terminal on the battery side of the harness.

    My "guess" is that there is either:

    1. An inductively coupled transient from the power lines (+ or -) into the "KILL" line. It is surprising how much parallel wires talk to each other!
    2. Ground bounce between PACK- and SGND, causing a large enough imbalance to cause a trigger.

    I would put a scope probe *DIRECTLY* across C37 (with a short, direct ground wire soldered directly to C37 - no clip leads!) to monitor the input signal.

    Yes. Try adding the 5uF. You can temporarily replace the 47n with the 5uF to verify (w/o the resistor), but check for oscillations on the reference pin. If it looks good, then add the resistor.

    Also try adding some series resistance to the KILL line (with appropriate compensations) and increase C37 to create a LP filter. You will need to find a tradeoff between detection time and false triggers.

    Regards,

  • 0 in reply to Paul Grohe
    Intellectual 540 points

    Thank's Paul,

    I'll keep these points in mind as I continue to test. Can you recommend a micropower op amp that would be suitable to use to buffer the REF line. It doesn't seem possible to tell from the datasheet what op amps have good AC drive capability, preferably with low Vos and a small package size. It may take less board space and reduce risk to do this than to add the a 5uF capacitor with a low voltage coefficient.

    Sheldon

  • 0 in reply to Sheldon Rubin
    Intellectual 540 points

    Paul,

    We decided to err of the side of caution and buffer the LMP7300 reference with an OPA333. We don't have time in our schedule to spin another board if the resistor and capacitor don't work. We have a hard deadline to be ready for the UN38.3 shipping tests.

    Thanks for the help,

    Sheldon