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.

UCC27211: problem on the Vhs pin

Part Number: UCC27211

Hello everyone,

I use the UCC27211 in my project. In particular, I make the electrical circuit as in the datasheet in the half bridge configuration. In this moment, the Vhi and Vli pins are connected to ground and the Vhs pin is not connected to the transformer. I use 9V for UCC27211 and variable boot voltage from 6V to 24V for half bridge mosfet.

My problem is on Vhs pin because I always measure a constant voltage (close to the supply voltage). But if Vhi and Vli are at logical level voltage, in Vhs the voltage it should be zero. It's right?

If I disconnect the high side mosfet, the voltage on Vhs is always present.

Please, can anyone help me? 

Thank you

Best Regards,

Donatello

  • Hi Donatello,

    I am an applications engineer on TI's high power drivers group and will work with you to solve your questions.

    Can you elaborate on the following, "But if Vhi and Vli are at logical level voltage, in Vhs the voltage it should be zero. It's right?"

    Do you mean if HI and LI are both high?

    regards,
    Mateo

  • Hello Donatello,
    I am an applications engineer at TI who works with Mateo in the high power driver group. What I understand from you description, is that LI and HI is low (ground), VDD is 9V, and the HB to HS voltage is 6 to 24V.
    One consideration is there is a boot diode from VDD to HB so if VDD is 9V, the HB voltage will be a diode drop below 9V. Also if HS is floating, low side MOSFET and high side MOSFET is off, the current charging the HB to HS capacitor, from VDD will cause the HS to rise.
    In a typical application, a transformer winding or inductor is connected to HS so there is a path from HS to the power train. Also the HB to HS capacitor is charged when the low side MOSFET turns on and HS is close to the power train ground; at this time HB-HS is charged to VDD minus the boot diode forward drop which is 0.6 to 0.7V.
    I think your test setup is allowing HS to float high since there is a high impedance on that node.

    Regards,
    Richard Herring
  • Dear Mateo and Richard,

    thank you for reply.

    I did a step-down switching that works perfectly (150Khz), while in half bridge configuration and high frequency (1MHz) I have many problem. At present, unfortunately I burned 6 MOSFET driver. In my test circuit, the LI and HI pins are always low (ground), VDD is 8.7V. Only low side MOSFET is connected and the other pins (HB, HO, HS) are disconnected. There is a voltage of 6.6V on the HS and HO, while on HB there is 8.1V. if I connect a capacitor or a resistor between HS and ground, the driver supplies the load until it burns.

    if I connect the side high MOSFET and the bootstrap capacitor, on HS there will be the bootstrap voltage. The voltage on HO will be slightly higher than HS. After the driver burns.

    In the step-down synchronous configuration, when HI and LI are low, no voltage is present on HS and HO.

    Unfortunately I can not solve the problem.

    Regards,

    Donatello

  • Hello Donatello,
    Can you provide a schematic diagram, and component values of the test circuit you are having the failures? Also it sounds like the test circuit is different than the target application schematic with the converter. Can you provide the schematic of the target application? Please include the VDD voltage range and converter input voltage range, and output voltage and current.
    The UCC27211 is capable of high frequency operation and is quite often at high switching frequencies.

    Best Regards,
    Richard Herring
  • Hello Richard,

    this is my schematic:

    The UCC27211 is supplied with 8.5V. The Vpower is variable from 6V to 30V. The signal in HI and LI pins are TTL with frequency of about 1MHz and 50% DC.

    If I disconnect the transformer and Vpower, there is a voltage of about 6.6V on the HS pin and 8.1V on HB pin. In this case, if I connect a small capacitor on HS, it charges.

    If I connect the transformer and increase the Vpower from 0V to 30V, sometimes the driver breaks down just after I enable the input signal on LI and HI. 

    I'm currently using more MOSFET in my tests, that have the following characteristics: Vds>=60V , Rds<200mOhm and Low gate charge <= 10nC .

    Regards,

    Donatello

  • Hello Donatello,
    Thank you for the details on the schematic and the application parameters.
    In the case of the 1st condition when there is no transformer and Vpower, the HS pin is floating since the low side FET is off. In this case the HS pin can have a voltage potential. The HB pin will be a diode drop lower than VDD so the 8.1V on HB makes sense.

    In the case of when you connect the transformer and increase Vpower, I have some questions. You mention that HI and LI are 1MHz switching frequency and the duty cycle is 50%. If the LI and HI are out of phase with 50% duty cycle, there will not be any dead time from the HI and LI inputs rising and falling times. I suggest confirming there is some dead time between the high side FET drive and low side FET drive, start with 100 to 150ns to confirm operation. With 150ns dead time the duty cycle will be 35%, with 100ns the duty cycle will be 40% with 1MHz operating frequency.

    Are you using a function generator to generate the driver inputs? If that is the case, it is OK but you need to confirm the sequence of powering up the power train. Many function generators have some cycles of unexpected timing when enabled.
    If using a function generator: 1st power up VDD supply to the driver, enable both HI and LI inputs, confirm LO and HO is as expected and there is some dead time, then slowly increase Vpower.

    With 10nC gate charge, the total gate drive power dissipation at 1MHz will be: 2 x Qg x Vdd x Fsw= 170mW for both driver channels. There are additional losses with IC operating currents, but the driver power dissipation is usually dominated by the losses from driving the MOSFET's. Gate drive dissipation should not be the issue with 10nC Mosfets.

    Please confirm the control drive timing and if there is adequate dead time from the high side and low side FET gate drivers.

    Regards,
    Richard Herring
  • Hi Donatello,

    We haven't heard from you in a while. Did you resolve your issue? Can you share with the community how?

    I will be closing this thread out. Let us know if you need any further help.
  • Hi Don,

    these days, I've been designing a PCB following the advice of the forum. Now, I will have to solder the components and then I can do some good tests.

    Therefore, I need more days before I can share results.

    Thank you

    Regards,

    Donatello

  • Hi Donatello,

    Thanks for the update. Let us know how you progress.
  • Hi,
    I made a PCB board with UCC27211. The frequency is about 994 kHz and Duty Cycle is 50%. The dead time is about 100ns.
    Below, the oscilloscope screenshot:

    The yellow waveform is the gate signal high side and the blue waveform is the gate signal low side.

    As you can see, there are peaks on the yellow waveform:

    How can I improve the waveform for the high side?
    Thank you


    Regards,
    Donatello

  • Hi Donatello,

    Sorry to hear you're still having issues. Richard will get back to you next week. Thanks for your patience!

  • Hello Donatello,

    Thank you for you continued work with the UCC27211. I have some questions about the waveforms in the scope plots, to be able to give advice on making improvements.

    I see the scale is 10V/div on the yellow gate signal high side waveform. Is this waveform with the scope connected to HO and ground of the converter? The driver high side output is HO with reference to HS, so looking at HO to ground includes the HS switching signal.

    Can you provide waveforms with LO to ground, HO to HS with differential probe, and HS to ground? If you do not have a differential probe, then can you record the LO to ground, HO to ground and HS to ground; use the same scale and scope ground reference for HO and HS so the difference in HO to HS is clear.

    It is possible the small pulses you see on HO (yellow waveform) exists on the HS to ground, and maybe not the HO-HS gate drive signal.

    Also to confirm the timing, can you provide scope plots of the LI and HI signal inputs, and the HO and LO signals on the same plot?

    Regards,

    Richard Herring

  • Hello Richard,
    I checked the two waveforms: LO to ground and HO to HS. This signal have no small pulses and are clear. At the moment, I do not have problems with UCC27211. Maybe the small pulses between HS to ground (or HO to ground) are produced in resonance by the transformer. I tried to mitigate the small pulses but they are difficult to eliminate, I hope these do not damage the high side of the UCC27211. The ON / OFF gate signals of the MOSFETs are never overlap using a good dead time.
    The half bridge currently works correctly, a heating (about 60°C) only for the high side MOSFET while the low side MOSFET is almost cold. The UCC27211 is quite hot (about 45°C) but I can lower the temperature by connecting the thermal pad to the ground plane. These days, I will try to determine the cause of the higher temperature of the high side mosfet.

    Regards,
    Donatello
  • Hello Donatello,
    I am glad to hear the half bridge is working as expected. From what I recall on the waveforms, it is likely the HS pulses are from resonance on the half bridge when the output current is at lower levels.
    Connecting the thermal pad of the UCC27211 with improve the thermal performance substantially, so we recommend connecting the thermal pad to a ground referenced power pad for reduced thermal resistance.
    For the high side MOSFET there can be a couple of reasons for the higher thermal rise. I do not see if there are resistors in series with the gate of the MOSFET to the driver outputs. Higher gate resistance will reduce ringing and HS switch node peak overshoot, but can add to switching losses in the MOSFET's. You can experiment with reducing the gate resistance if in place.
    Another cause of loss during switching during higher load operation, when the inductor current is always over 0, is the body diode recovery time of the low side MOSFET will result in a short period when the high side Mosfet is turning on but the diode on the low side FET has not turned off yet. Mosfets with faster body diode recovery times can reduce these losses.

    Please confirm if this helps with the issues, and select the green icon on the thread.