• Locked Locked
  • Replies 11 replies
  • Subscribers 50 subscribers
  • Views 3287 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.

INA240: Lightning ESD Protection

Hayden_Hast
Intellectual 1515 points
Part Number: INA240
Other Parts Discussed in Thread: OPA192, INA300, STRIKE

Hi Team,

I'm using the INA240A2 in my system and am currently trying to add protection in front of the device to protect from lightning strikes. The schematic snipit of the INA240 is attached to this thread.

The test for lightning is performed by injecting a 1500V, 60A SIN wave into the TMS1LS pin, where it would pass across the MMZ16085601A shielding bead, 30V clamping diodes, and across R630. 

We thought that the INA240 would sense the over-current and quickly turn off Q24 (not shown but INA240A2 & OPA192 control gate of Q24, but we are seeing that the INA240 is failing and blowing up Q24.

Do you have any ideas on how we could protect the INA240 and subsequently Q24 from the lightning testing?

INA240_ESD.docx

Regards,
Hayden

  • TI__Expert 4915 points

    Hi Hayden,

    Thank you for posting on the E2E Forums.

    I looked into your circuit and it is not immediately clear how the mosfet should turn off, assuming the current flows to ground. Would you please elaborate by adding some current flow arrows and circuit waveforms/expected behaviour, where possible?

    As for your setup, I would recommend using the INA300 or 301. This is the right application for those devices and they can act in as little as 1uS.

    With that aside, I would like to express my concern to the fact that switching the mosfet off under the conditions you mention is likely going to destroy the device, rather than protecting it. Once the device is off, the VGS voltage is likely to rise above the 100V breakdown threshold

    In the IGBT world, drivers often come with a feature called active clamp. The device is switched off from saturation but kept operating on the linear region so as to dissipate the energy without exceeding is breakdown voltage. This is valid when the device is suddenly turned off too: Any parasitic inductance on the bus and the sudden decline in the currents will likely cause an overshoot and damage the device.

    I am also not satisfied with the placement of the current sense resistor, this should be placed before the mosfet, tied to GND, otherwise when the mosfet opens, all the high voltage from the strike will go though the INA240, effectively destroying it. There should be nothing on that path with exception of the TVS diodes.

  • in reply to Carlos Silva85
    Prodigy 50 points

    Guys,

    Here is the complete circuit. Please comment on it now.  

    (1) We are sensing current on high side and low side of motor. Top mosfet is always on. Bottom mosfet switches at 40 Khz.  Processor is turning ON and Off at 1.25mSec ( 40Khz).

    (2) We are applying lightning at sense resistors PINS.

    (3) We are thinking that Lower sense amplifer is getting destroyed ( because of excess common voltage) when lightning pulse is applied at lower sense resistor.

    (4) Lower MOSFET is getting destroyed beacuse of all the lightning current  passing through it ( since sense amplier not able to turn-off MOSFET).

    How VGS goes up when MOSFWET turns-off.    Please comment on it.

  • in reply to kondala ruthala
    TI__Expert 4915 points

    Hi Kondala,

    There are no attachments with your last post.

  • in reply to kondala ruthala
    Prodigy 50 points
    Carlos,

    Please look at my comments below.

    • I looked into your circuit and it is not immediately clear how the mosfet should turn off, assuming the current flows to ground. Would you please elaborate by adding some current flow arrows and circuit waveforms/expected behaviour, where possible?

    ---- comments kondala : We have high side MOSFET ( Q35), which is always ON. But low side MOSFET ( Q24), which switches at 40Khz. Current sense amplifier on low side senses the voltage across sense resistor and develops a logic high to set OC_SHUTOFF latch to turns off the MOSFET driver and low side MOSFET.

    As for your setup, I would recommend using the INA300 or 301. This is the right application for those devices and they can act in as little as 1uS.
    ---- comments kondala : Common mode input voltage on INA300 also very low. INA300 or 301 may not be suitable.

    With that aside, I would like to express my concern to the fact that switching the mosfet off under the conditions you mention is likely going to destroy the device, rather than protecting it. Once the device is off, the VGS voltage is likely to rise above the 100V breakdown threshold

    ------ comments kondala : How MOSFET is getting destroyed. How VGS goes up. Do you have app note to explain it?

    In the IGBT world, drivers often come with a feature called active clamp. The device is switched off from saturation but kept operating on the linear region so as to dissipate the energy without exceeding is breakdown voltage. This is valid when the device is suddenly turned off too: Any parasitic inductance on the bus and the sudden decline in the currents will likely cause an overshoot and damage the device.
    I am also not satisfied with the placement of the current sense resistor, this should be placed before the mosfet, tied to GND, otherwise when the mosfet opens, all the high voltage from the strike will go though the INA240, effectively destroying it. There should be nothing on that path with exception of the TVS diodes.

    ----- comments Kondala : Want to discuss on it. We may not able to measure current , if place resistor some where else. Want to have a discussion on it.
  • in reply to kondala ruthala
    TI__Expert 4915 points
    Hi Kondola,

    In order to comment further I will require:

    Oscilloscope plot showing current/voltage waveforms on the moment the lightning strike is applied (Both ID, VGS, IN+ and IN-)
    Waveforms showing the reaction time under normal operating conditions when a short circuit is applied, so as to verify the reaction time of the INA240 circuit.

    I also suggest moving the low side current sense resistor to the ground path, before the mosfet.
  • in reply to Carlos Silva85
    Prodigy 50 points
    Carlos,

    Are you available to have quick conference call today? Please let me know. We will make a plan to capture all the waveforms etc on the hardware board.
  • in reply to kondala ruthala
    TI__Expert 4915 points
    Hi Kondala,

    Yes, send me the invitation. I should be available between 3-4PM UTC-7.
  • in reply to Carlos Silva85
    Prodigy 50 points
    Carlos, please send me your email address to "kondala.ruthala@woodward.com". I will schedule a meeting to plan on INA240 debugging
  • in reply to kondala ruthala
    TI__Expert 4915 points

    I will discuss this with the INA240 Expert and will get back to you. In the meanwhile here's a document that may be of use.

  • in reply to Carlos Silva85
    TI__Expert 4915 points

    Hi Kondala,

    I want to clarify one more thing.

    You have previously mentioned that the node TSM1LS- where the 1500V 60A lightning strike is being applied is protected by two TVS diodes, with a total conducting voltage of 60V. We have looked at the datasheet of the inductor in use and it seems to be rated for 500mA with a considerable impedance at high frequencies.

    Can you try to replace this bead with a short to guarantee that the TVS diodes operate as expected and verify whenever the failure persists?

    Just for our reference, how long is this 1500V, 60A pulse applied for?