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.

TPS2834: Low side of MOSFET driver not operating at all when in synchronous mode and using in full bridge configuration. External modification had to be performed in order to get this working.

Part Number: TPS2834
Other Parts Discussed in Thread: LM5104,

Hello,

When using two TPS2834 MOSFET drivers (and MOSFETS) in a full bridge configuration (as seen in the schematic below) and in synchronous mode (by holding the SYNC pin high), the low side driver signal LOWDR (going into the gate of the low side MOSFET) would remain low instead of switching inversely to the HIGHDR signal, this resulted in the system connected on the output not operating at all. Other sections of the design that use the MOSFET driver in half bridge configuration work without any issues. The same full bridge configuration was tested with the LM5104 MOSFET driver and works without any issue or need for modification. It is however not possible for me to use LM5104 MOSFET for this design.

In order to get the TPS2834 MOSFET drivers working in the configuration shown in the schematics, a resistor was connected across the drain and source of both the low side MOSFETs used. The LOWDR signal (going into the gate of low side MOSFET) was then seen operating inversely to the HIGHDR.

Whilst the MOSFET drivers were seen working with the modification, I am not keen on proceeding with the design, I would much prefer if it was working without this modification and would like your feedback in order to help me resolve this issue.

In the simulation that I found online I noticed that the dead time pin was pulled to ground instead of connected to the bridge of the two MOSFETS, I have yet to test this however I do not think it will resolve the issue.

in the schematics attached:
VCC = 12V DC
EX1-A and EX1-B are PWM signals that are the inverse of each other.

If you have any questions then please let me know.

Thanks

Adam

  • Hi, Adam,

    The TPS2834 has adaptive dead-time control. The implementation in this particular IC looks at the voltage at BOOTLO, and if it's high, it will not turn on the low-side driver. This is why hanging a resistor from drain to source made it function. This provided a path for the BOOTLO voltage to be discharged.

    The LM5104 has a different adaptive dead-time control approach which is why it works in your circuit. Maybe we can find an alternate part for you. Why can't you use the LM5104 in your circuit?

  • Hello Don,

    Thank you very much for the quick response.

    I cannot go into too much detail, however we did some environmental testing with the LM5104 IC and saw some early issues which rendered the system inoperable, in the same environment the TPS2834 IC operated without any issues. It is thought now that adaptive dead-time control circuit of the LM5104 was causing the issues.

    In the above schematic, it is now understood that the issue were occurring due to the BOOTLO (and thus the DT pin) not reaching below the 1V low threshold, which in term was preventing the low-side driver from switching on. This is thought to be because the system is driving a capacitive and inductive load that doesn't have a ground return path, unfortunately this is not something that can be changed.

    I believe that the potential solutions may be as follows:

    1. Use the existing solution of pulling down the bridge voltage via a resistor.

    2. Connecting the DT pin permanently to ground. This was observed in the simulation file found on the TPS2834 webpage and appeared to have no issues, however it is thought that in reality this may not be the case. This option would however mean that deadtime control is permantly disabled. 

    3.  Creating a potential divider from the bridge (BOOTLO) to the DT pin, this would ensure that when the high side has completed switching that the BOOTLO voltage was below 1V. The dead time control would still be active in this scenario.

    Please could you let me know your thoughts on the 3 options and any problems you would anticipate to see.

    Thanks

    Adam 

  • Hello Adam,

    I work with Don as an applications engineer, and will comment on your 3 proposed solutions.

    If you have the 1st solution working, pulling down the bridge voltage with a resistor, I would suggest that as a primary solution.

    Grounding the dead time pin does not sound like a good idea since the dead time will be disabled.

    The 3rd idea of creating a divider is worth exploring since in theory the divider should help resolve the issue of achieving the voltage low enough to be detected by the dead time circuit.

    Confirm if this addresses your questions or you can post additional questions on this thread.

    Regards,