• State TI Thinks Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 2 answers
  • Subscribers 63 subscribers
  • Views 740 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.

TPS61087: SW pin short to GND at high output current / short-circuit on output

FlorinDumitrache1
Prodigy 60 points
Part Number: TPS61087
Other Parts Discussed in Thread: TPS55340

Hello,

I designed a boost circuit using TPS61087 to boost the voltage of a Li-Ion cell up to 16V. The battery is monitored on the PCB by a uController which will turn off the DC/DC when the input voltage is lower than 2.9V.

The circuit is being used to drive a small brushed motor with speed regulation via PWM at aroung 31kHz.

The schematic is the following:

The output voltage is stable, with small ripple when the motor commutates the internal windings. The motor stall current is around 0.55A.

I monitored the output voltage between MOTOR+ and GND_DCDC during motor stall and there is no overshooting. The voltage drop over L202 during operation is negligible.

If the stall is longer than a few seconds the TPS61087 fails short between SW and GND. No other components fail. This was replicated several times with different results:

- "soft" short between SW and PGND (10-100 Ohms)

- "hard" short between SW and PGND (<1 Ohm)

-  "active" short, meaning that the device measured outside the circuit measures no short but when it is inserted back in the circuitm the TPS61087 dissipates >1W even when turned off.

I would need some help to understand why the IC does not survive during high output power.

What is the current/voltage limiting mechanism when the output power is high?

What does the IC do when it's internal temperature rises to a value close to Tj_max?

What does the IC do if it detects it cannot regulate the output voltage to the proper level?

As the input power source, for the usual tests I use Li-Ion cells, but the defect also appears when using a lab bench power supply with remote sensing the voltage directly on the PCB.

  • 0
    TI__Mastermind 40080 points

    Dear Sir,

    what is the duty cycle of the 31kHz PWM signal? 0.55A load add at the output only when motor on , right?

  • 0 in reply to Helen chen
    Prodigy 60 points

    Hello Helen,

    Helen chen said:

    what is the duty cycle of the 31kHz PWM signal?

    The duty cycle varies between 60% and 100%.
    Helen chen said:

    0.55A load add at the output only when motor on , right?

    Yes. This would be the maximum load current, with the motor in stall and 100% duty cycle on the PWM signal.
  • 0 in reply to FlorinDumitrache1
    TI__Mastermind 40080 points

    Thanks for the update.

    16V/0.55A already exceed the maximum output capability of TPS61087. The peak switch current will reach as high as 3.7A.  Considering the SW spike, the peak voltage across the SW pin easy to exceed to the 20V abs max if the layout is not good.

    Please check the SW to GND waveform at the same time.

    If possible, suggest you change to higher voltage/current  rating components like TPS55340.

  • 0 in reply to Helen chen
    Prodigy 60 points

    I checked the voltage at the SW pin with short ground lead at the osciloscope probe and could not find any overshoots.

    This is during normal operation (16V / 0.15A load):

    And this is the behavior during stall condition. Now the IC did not fail, same conditions as before:

    The measurement shows 18.4V but this is higher than real (bad interpolation on the osciloscope; capturing with a smaller timebase shows no overshoot).

    Due to space limitation I cannot go with an IC with bigger package.

    The layout is done so that the hot loops are minimum (red is top, blue is bottom):

    Could it be a stability issue? I previously had issues with TPS61087 because of the compensation network. Using the equations in the datasheet I got 100kOhm and 1.6nF. If I used 100kOhm the IC failed at high output currents, total output power about 4-5W.

  • 0 in reply to FlorinDumitrache1
    TI__Mastermind 40080 points

    Dear Sir,

    The layout looks pretty good.

    For the SW spike, I guess you use 20MHz bandwidth instead of 500MHz or 1G Hz. So the ringing was filtered by the osciloscope.

    The load current is pretty big, I guess it enters into the OCP state because the peak switch current higher than 3.5A.