• State TI Thinks Resolved
  • Locked Locked
  • Replies 7 replies
  • Answers 2 answers
  • Subscribers 64 subscribers
  • Views 146 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.

TPS25983: Protecting TPS25983 from Transients on 24V rail

Owen Hooker
Prodigy 20 points
Part Number: TPS25983
Other Parts Discussed in Thread: TVS2700, TVS2200, TPS25948, TPS2663

Tool/software:

Hi,

I'm working on a compact board that uses both 14V and 24V rails. It has three voltage rails, one for 24V and two for 14V and needs protection against over-current, over-voltage, under-voltage, and possibly reverse current.

I'm considering using three TPS25983 devices as fuses (one for each connector) with reverse-blocking MOSFETs. I'm also planning to add TVS diodes on the outputs for transient and ESD protection.  The 14v rails have a max current of 6 amps and the 24v rail has a max current of 4 amps.

Here's my issue:

  • 24V Rail Protection: The TPS25983 has a max voltage rating of 30V, which doesn't leave much room for TVS diode selection. I considered the TVS2700 flat clamp, but its max clamping voltage is 32V—above the TPS25983's limit.  However I think The 14V rails are fine with the TVS2200.

  • Space Constraints: The board is very tight on space, so adding ideal diode controllers and MOSFETs instead of the TPS25983 might not be feasible because of size.

  • It seams to me the issue mainly arises when the reverse-blocking MOSFET is on. If it's off, the TVS2700 would be acceptable due to the MOSFET's higher voltage rating.

My Questions:

  1. Any suggestions for protecting the TPS25983 from transients at 24V, considering its 30V max rating of TPS25983?

  2. Is there a way to tweak the circuit to get proper transient protection without exceeding voltage limits?  Maybe adding a comparator to the reverse blocking diode to detect transients and quickly shutdown the reverse blocking mosfet?

  3. Are there compact alternatives to the TPS25983 that could fit within my space constraints?

I chose the TPS25983 mainly for its small size and integration with the reverse-blocking MOSFET. But given the space limitations, I'm open to other ideas.

Thanks in advance for any advice!

  • 0
    TI__Mastermind 29030 points

    Hi Owen,

    You can check the following.

    For 14V rail, you can also consider TPS25948 which is a 23V/8A device with back-to-back FET integrated and smaller package (but it is wcsp). For 24V rail, you can also use TPS2663 with external FET (solution size should be same as TPS25983 with this device is rated for 60V/6A).

    Now coming back to TPS25983.

    Owen Hooker said:
    Any suggestions for protecting the TPS25983 from transients at 24V, considering its 30V max rating of TPS25983?

    Are you expecting transients in your application? Adding SMCJ26A or TVS2700 won't guarantee clamping below the abs max but will still help against the transients. Adding higher decoupling cap will also help in absorbing some of the charge.

    Owen Hooker said:
    Is there a way to tweak the circuit to get proper transient protection without exceeding voltage limits?  Maybe adding a comparator to the reverse blocking diode to detect transients and quickly shutdown the reverse blocking mosfet?

    I am not sure how this will help. I suspect that comparator response time will be higher than the transients time.

    Thank you for your patience.

    Best Regards,
    Arush

  • 0 in reply to Arush Gupta
    Prodigy 20 points

    Hi Arush,

    Thanks again for the suggestions—I really appreciate your time.

    I did take a look at the TPS2663 and TPS25948 early on, but ended up moving away from them because of the higher on-resistance. The circuit I’m working on is an auxiliary power output to power various accessories from a battery interface. it’s possible that we could see transients (like ESD spikes) or even reverse current if a cable is unplugged under power. That’s why I initially leaned toward the TPS25983, especially with its built-in reverse-blocking capability.

    I’ve noticed the TPS25948 doesn’t inherently detect a reverse-current condition but only blocks it in its off state.  I am considering the possiblility of adding an LM74704 downstream of the TPS25983. The idea is to tie the LM74704’s enable pin to the TPS25983’s B-GATE, so both would enable and disable at the same time while allowing the LM74704 to be able to actively protect from reverse current.  Is there any issue with combining these two parts in this way?

    Another point I wanted to clarify is the TVS diode selection. The TVS2700’s clamp voltage is slightly above the TPS25983’s absolute maximum rating. In your experience, how critical is that margin? Are these ratings so tight that I should consider lowering my 24 V rail to 22 V (not ideal but might be acceptable) and switching to the TVS2200, or is there usually a bit of leeway? I’m trying to ensure a robust design without pushing components too close to their limits.  

    Thanks again for your input.

    Owen

  • 0 in reply to Owen Hooker
    TI__Mastermind 29030 points

    Hi Owen,

    Owen Hooker said:
    I initially leaned toward the TPS25983, especially with its built-in reverse-blocking capability.

    TPS25983 do not have integrated back-to-back FETs.

    Owen Hooker said:
    I’ve noticed the TPS25948 doesn’t inherently detect a reverse-current condition but only blocks it in its off state.

    I didn't get this. TPS25948 RCB can be controlled using RCBCTRL pin. RCB is effective during on state also. This is valid for TPS25983.

    Owen Hooker said:
    adding an LM74704 downstream of the TPS25983. The idea is to tie the LM74704’s enable pin to the TPS25983’s B-GATE, so both would enable and disable at the same time while allowing the LM74704 to be able to actively protect from reverse current.  Is there any issue with combining these two parts in this way?

    I think there is some misunderstanding on the functioning of the dvdt/b-gate pin. 

    I don't understand what will a ideal diode driver circuit help in the power path when you already have external FET driver in this device.

    It might be possible that I am misunderstanding your requirements. Can you explain more on your exact requirements.

    Owen Hooker said:
    Another point I wanted to clarify is the TVS diode selection. The TVS2700’s clamp voltage is slightly above the TPS25983’s absolute maximum rating. In your experience, how critical is that margin? Are these ratings so tight that I should consider lowering my 24 V rail to 22 V (not ideal but might be acceptable) and switching to the TVS2200, or is there usually a bit of leeway? I’m trying to ensure a robust design without pushing components too close to their limits.  

    It is recommended to evaluate this on silicon. Ti cannot guarantee proper functioning of devices if there is abs max violation. In general the device is capable of handling slightly higher voltage for shorter duration.

    Best regards,
    Arush

  • 0 in reply to Arush Gupta
    Prodigy 20 points

    Hi Arush,

    I apologize for the confusion in my last reply; I mixed up some part numbers, which made things unclear. Let me clarify my explanation.

    For this design, I am only referring to the TPS25983 and the LM74704.

    I understand that the TPS25983 does not have built-in back-to-back MOSFETs and only provides a gate driver to control an external current-blocking MOSFET. My understanding is that the TPS25983, when paired with an external MOSFET connected to the B-GATE pin, can block reverse current, but only when the device is off.

    Since the TPS25983 does not detect reverse current in the on state, I am considering using the LM74704 to handle reverse current detection during operation. In this setup, there would be no external MOSFET connected to the TPS25983. Instead, the LM74704, along with its external MOSFET, would take on this role. My idea is to tie the ENABLE pin of the LM74704 to the B-GATE pin of the TPS25983. This way, the LM74704 would automatically be disabled when the TPS25983 is turned off.

    The goal is to achieve reverse current blocking capability in both the on and off states of the TPS25983.

  • 0 in reply to Owen Hooker
    TI__Mastermind 29030 points

    Hi Owen,

    Owen Hooker said:
    I understand that the TPS25983 does not have built-in back-to-back MOSFETs and only provides a gate driver to control an external current-blocking MOSFET. My understanding is that the TPS25983, when paired with an external MOSFET connected to the B-GATE pin, can block reverse current, but only when the device is off.

    Yes, this is correct.

    Owen Hooker said:
    Since the TPS25983 does not detect reverse current in the on state, I am considering using the LM74704 to handle reverse current detection during operation. In this setup, there would be no external MOSFET connected to the TPS25983. Instead, the LM74704, along with its external MOSFET, would take on this role. My idea is to tie the ENABLE pin of the LM74704 to the B-GATE pin of the TPS25983. This way, the LM74704 would automatically be disabled when the TPS25983 is turned off.

    On top level, this looks like it might work.

    Still I have one question for you. For 14V rails, why not just use TPS25948 which provides integrated back-to-back FET with RCB during on and off state.

    Best Regards,
    Arush

  • 0 in reply to Arush Gupta
    Prodigy 20 points

    Hi Arush,

    The only reason for not choosing the TPS25948 is the higher on resistance and my ability to handle heat with the small size of my pcb. My application could be up to the 8amp limit with this device.

    Arush Gupta said:
    On top level, this looks like it might work.

    Are there any specific concerns you have about this that I should be aware of?

    Thanks for the help!

    Owen

  • 0 in reply to Owen Hooker
    TI__Mastermind 29030 points

    Hi Owen,

    Owen Hooker said:
    Are there any specific concerns you have about this that I should be aware of?

    I haven't seen this type of implementation so cannot tell you in very specific details on this. It will be best to evaluate this with EVM or protoboards. As such taking general considerations like abs limits and general functioning of device should prevent maximum issues.

    Best Regards,
    Arush