• State TI Thinks Resolved
  • Locked Locked
  • Replies 11 replies
  • Answers 3 answers
  • Subscribers 62 subscribers
  • Views 390 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.

TPS40170: LM5116 vs. TPS40170 for industrial robotics environment

Jim G
Intellectual 706 points
Part Number: TPS40170
Other Parts Discussed in Thread: LM5116, LM5146

Tool/software:

I have two off-the-shelf buck modules I am testing in a noisy robotics environment to select what will eventually be the wide input voltage buck IC for this application.
16-52V input, ~12V output, ~120W.
One of these modules is based on LM5116, the other based on TPS40170.The LM5116 based module has a TVS diode on the input and only small ceramic caps, the TPS40170 module has no TVS diode and also only ceramic caps.

I am also considering LM5146 based on Webench suggestions but do not have a test platform for this IC.

In my application with various motors and switching loads, I'm seeing the TPS40170 based solution drop out and reset abruptly. I have soldered additional electrolytic caps and TVS diodes at the input and outputs, but it continues to fail during motor activity. The issue is not thermal related as the module is only running for under 1 minute prior to resetting. This is not enable pin related as the enable pin is directly tied on at the module.
I probe and can confirm this is not an OCP event at the output, we are nowhere near the max power rating of the module, and there are no long duration transients on the input or output. I can capture very brief (<1us) transients on the input and output related to what we believe is some charge accumulation / ESD event in the system, and is the only clue I have as to what might be causing a protection mode on the buck module. Adding TVS diodes and additional external capacitors has not allowed this module to work.

However, the LM5116 based module works completely fine in this environment with absolutely no added external capacitors and only a single TVS diode at the input.

My question is this: What are the IC differences between LM5116 and TPS40170 which may make the controllers more or less resilient to external transients?
Is current mode control in LM5116 more resilient than voltage mode control in the TPS40170 for external high frequency noise of this nature? What other characteristics do these switching controllers have or not have which makes one better or worse in industrial/noisy environments?

  • 0
    TI__Guru 61891 points

    Hello Jim 

    If the output of the buck converter returns back automatically after the failure during the motor activity, then TPS40170 might be current limited. The current limit threshold level of TPS40170's Rdson current sensing is wider than the current limit threshold level of LM5116's Rsense current sensing. 

    If the switching stops and never returns back without power recycle, then the device might be latched due to an excessive negative voltage due to the noisy motor activity. 

    -EL 

  • 0 in reply to EricLee
    Intellectual 706 points

    Thanks for your reply EL.

    The buck does immediately re-start and recover every time.

    I have probed the buck output using a 1MHz current probe during the failure condition, and the load current is well below the rated buck output. It is not an overcurrent condition as far as I am able to observe.

  • 0 in reply to Jim G
    TI__Guru 61891 points

    Hello Jim 

    Thermal shutdown, Overcurrent protection, and Input voltage drop, these three items can make the restart. Please check temperature too, 1min is enough to make the device enter TSD. 

    -EL

  • 0 in reply to EricLee
    Intellectual 706 points

    It is not thermal shutdown, the load current on the design is well below the rated current of the overall implementation and we can very comfortably touch the surface of the package during operation.

    It is not overcurrent protection, we have directly measured this.

    It is not input voltage drop, we have directly measured this and observe no sustained large transients on the input voltage.

  • 0 in reply to Jim G
    TI__Guru 61891 points

    Hello Jim 

    And check the switching frequency and VBP voltage. The device might stop switching because VBP voltage is too low. 

    -EL  

  • 0 in reply to EricLee
    Intellectual 706 points

    Thanks Eric. The buck implementations I'm discussing here are on fully off-the-shelf integrated modules, so my ability to adjust the actual components is limited. I'm mainly hoping to confirm the hypothesis here of why the LM5116 based module performs better than the TPS40170 based module in the same environment - my suspicion is that one of the most important differences here is that LM5116 is based on current mode control, while TPS40170 is voltage mode control and less resilient to external noise sources. I'm sure it's not as black and white as that, but I suspect it's a significant factor. Getting some confidence in that assumption will help me to proceed in selecting a controller which may be successful in a new fully custom buck implementation based around some TI switching controller.

  • 0 in reply to Jim G
    TI__Guru 61891 points

    Hello Jim 

    For PWM operation, voltage mode controller does not use the sensed current information, but current mode controller does. Because of this difference, current mode controller's PWM operation is less stable (more duty-cycle jitter) than the voltage mode controller's PWM operation. 

    For cycle-by-cycle current limit protection, both the current mode and the voltage mode controllers need current sensing, and the current limiting operation is similar. 

    Rdson current sense is cheaper but the its protection level variation is wider than the resistor sensing, causes more head ache than the resistor sensing. 

    -EL

  • 0 in reply to EricLee
    Intellectual 706 points

    What I'm actually observing is worse operation from the TPS40170 (voltage mode), which is regularly resetting in a noisy environment, compared to LM5116 (current mode).

  • 0 in reply to Jim G
    TI__Guru 61891 points

    Hello Jim 

    It means the unexpected behavior you are observing is not caused by the 'Voltage mode' and 'Current mode' differences, but caused by the device itself.

    -EL 

  • 0 in reply to EricLee
    Intellectual 706 points

    If that is the case, what in the device itself might be causing it? To get back to my original question, why might an LM5116 based implementation function without issue in an application, but a TPS40170 based implementation cannot? Both power modules are designed for >8A output, we are not tripping OCP, we are not hitting undervoltage or overvoltage on the inputs as far as I am able to scope. It is not thermal.

  • 0 in reply to Jim G
    TI__Guru 61891 points

    Hello Jim 

    And check the switching frequency and VBP voltage. The device might stop switching because VBP voltage is too low.

    There is a possibility that the MOSFET driving current is too much and it brings down the VBP below its UVLO. 

    -EL