• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Subscribers 64 subscribers
  • Views 332 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.

TPS76833-EP: Voltage Regulator design/schematic questions

Daniel Mueller
Intellectual 890 points
Part Number: TPS76833-EP
Other Parts Discussed in Thread: DRV8301, TPS54160, TPS768, TPS746, TPS746-Q1

Hello TI, I wish you a happy new year!

I have a motor controller board that has a few issues with quite a noise (80-100kHz) on the 3.3V rail. It’s the board of a friend and I think the noise comes from messing up the ground.

The TPS54160 Buck Converter (DRV8301 Buck Converter) is used to generate 5V and a the TPS76833 is used to generate 3.3V from the 5V line.

He asked me to help him on the circuit and layout and we have created this:

In the picture you can see the new board. Now there is only one ground connection point at the bottom of the DRV8301, next to that is the VDD connection point (purple lines). Both with star connection from the input connector. I think that should be right?

I placed the TPS76833 a bit away from the Gate Driver and Buck Converter.

 

 

In the picture below there is the DRV8301(Buck Converter) schematic.

I added a filter before the PVDD Buck input (yellow line) and on the output of the Buck Converter a ferrite bead (purple line). What do you think about that?

I’m also thinking about adding a ferrite bead on the input of the Gate Driver (blue line).

On the input of the 3.3V Regulator I added an RC-Filter with a resistor value of about 1-3 ohm, would that make sense?

Best regards,
Daniel

  • 0
    TI__Genius 12365 points

    Hi Daniel, 

    Happy New year to you as well!

    Thank you for reaching out. 

    It is hard to read the screenshots. Is it possible to share the schematic and layout files in a PDF format, or the actual design files? 

    Do you also happen to have a scope shot of the 3.3V rail showing the noise? What is the switching frequency for the TPS54160 and the PWM for the Gate Driver? 

    Best, 

    Edgar Acosta

  • 0 in reply to Edgar Acosta1
    Intellectual 890 points

    Hello Edgar,

    Of course, sry for the bad pictures.
    A few things to say, the board isn't finished yet, the drv/buck part is almost ok but the position of the 3.3V regulator can varry a bit. 4 layer, 2oz inner and outer layer. Stackup for digital part of the board is: Top(POW, SIG), L2 (GND), L3 (POW, SIG), BOT (GND).

    The switiching frequency of the buck is about 600kHz. The pwm switching  frequency is 20kHz to 40kHz depends on the motor we are using.

    I will drive my friend later and take a few pictures of the noise. I'll post those later.

    It would be intressting where the noise exaclty comes from, the old boards consits of two parts, the MCU board and the inverter board.
    There are several ground connection between the two boards. The schematic of the old and new boards are almost equal execpt the things i mentioned in the first post (buck input filter, ferrite on the ouput of the buck, rc filter on the voltage regulator input). I think the noise is due to the layout. I will also do a few measurements, maybe i can say more then. I could also share the old board if that would help?

    What I have observed so far, at the moment at the pwm turns on, the noise occurs (without any motor spinning or no motor connected).  
    It's really quite a noise and I didnt see that of any on my boards before.


    Here are some pictues with high resolution:

    Board:


    Buck:



    Buck:


    Schematic DRV8301/Buck:


    Schematic Voltage Regulator:

    Best regards,
    Daniel

  • +1 in reply to Daniel Mueller
    TI__Genius 12365 points

    Hi Daniel, 

    Thank you for the update. 

    No worries, I understand the design is still ongoing. 

    I agree with you, it does seem like it was some sort of issue with the GND connections, especially when you mentioned that the pwm switching frequency is 20kHz to 40kHz and the noise is 80-100kHz since 80 and 100 are harmonics of 20 and 40. 

    Having the Load on the same board, and or same side of the board is best practice and preferred. Also, having only one GND is as well best practice and preferred to prevent potential differences. 

    Also, your new stack up will show improvement by having: Top(POW, SIG), L2 (GND), L3 (POW, SIG), BOT (GND) and 2oz copper. 

    As of now, based on the information and the design, it doesn't seem that the issue is coming from the Buck -> LDO, therefore, the Pi Filter used on the output of the Buck could be avoided. It doesn't hurt having the footprint, you can always test this without the ferrite bead and just solder jump.

    Consequently, the RC filter on the LDO input might not be necessary, especially if the filter on the Bucks output will be used. Also, keep in mind that this will create an RC delay on the input and having that R will dissipate some power and there will be some IR drop to take into account.

    Similarly, having the Pi filter on the output for the LDO doesn't hurt as you can always measure before and after the filter. 

    As for the LDO itself, was there a specific reason why TPS768 is being used? I ask since this is one of our older devices and we do have newer devices that might be a better fit for your application. I also mention this since some of our older devices have a stability requirement, and this device requires a minimum of 60mOhms and max of 1.5Ohms. Attention must be paid to the capacitors placed on the output to make sure that ESR is still within the region of stability. Newer Ceramics tend to be low ESR and placing them in parallel reduces this ESR as well. 

    Lastly, be wary of any potential of Inrush Current as you have 20||10||10||.1 uF, some of our newer LDO's have some sort of soft start feature to help minimize the inrush current. 

    Best, 

    Edgar Acosta

  • 0 in reply to Edgar Acosta1
    Intellectual 890 points

    Hello Edgar, thank you for your answer, it helped a lot!

    Yes it's cleary to see that the noise is correlated to the pwm switiching on the oscilloscope. When the mcu board is removed from the inverter board and connected with jumperwires and only one jumperwire is used for the ground, the noise reduces dramatically.

    Because it's development we will add the additional footprints for filters and see how it goes.

    They are used because my friend has a lot of them, I also suggested to change the LDO. What LDO can you recommend for the application?

  • +1 in reply to Daniel Mueller
    TI__Genius 12365 points

    Hi Daniel, 

    You are welcome. Understood the reason of using this particular LDO. 

    Here are some LDOs to look into. It will all depend on the application needs like Load, load and line transients, automotive, etc:

    TLV757P

    TPS785-Q1

    TPS746-Q1/ and TPS746

    Those are rated for 1A, but if lower currents are needed these have also a lower current version, but these could be a good starting point, especially if these are being hand soldered. 

    Best, 

    Edgar Acosta

  • 0 in reply to Edgar Acosta1
    Intellectual 890 points

    Thank you Edgar, we will use the  TPS746-Q1.
    Best regards,
    Daniel