• TI Thinks Resolved

TCA6416A: Turn on DC motor 12V 50W nearly, erases the configuration registers

Part Number: TCA6416A


I have a problem with this I2C expander TCA6416ARTWR. everything is fine, but when I turn on a DC 12V 50W motor near to the PCB the registers of configuration are deleted. (without I2C traffic).

The sequence is this:

- turn on the electronic

- I set the I2C: reg 4, 5, 6 and 7 (00, 3F, 00, 3F)

- I read the configuration from this registers and they are fine (00, 3F, 00, 3F)
- no more traficc on the I2C bus

- turn on motor

- turn off motor

- I read the configuration register and they have been reiniciated (00, 00, FF, FF)

I don`t see Vcc noise and reset signal is 10K pull-up near to chip

I guess it is a noise problem, but this one seems too susceptive. what reason can to reiniciate the configuration? How can I improve the desing? Can you help my?

  • Hello Serio,

    There are one of two things that can reset the TCA6416A, either Vcc is lowered past it's reset voltage or the RESET pin is seeing a voltage that toggle the register reset.   I would add a capacitor at your reset pin first.  See if that fixes your problem. 

    Make sure your scope is not bandwidth limited and look at both Vcc and RESET pins during the motor startup and shutdown.  Try and make low noise measurements, hopefully you are using short ground probes and make sure the probes have a high enough bandwidth to capture the event. 

    My guess is that adding a decoupling cap on RESET will fix it.  You might have to play with the values.  It would be nice to capture the event that way we could get more information.  Please send waveforms if this doesn't fix it.

    -Francis Houde

  • In reply to fhoude:

    Hello Francis,

    Thanks for you quick answer!

    I have been doing probes as you say me.

    My scope is bad (it is a USB scope) 20MHz 12Bit and my probes are worse ... (6MHz).. What features do you thik that I need?

    With this devices, I see:

    reset pin:

    I think that this es little variation

    With a 10uF in reset pin (the problem is smaller but if I move a lot of the motor the registers are deleted. the signal with 10uF:

    Finally, I disconect the reset pin and conect it to 1,8V near directly. The problem is similar to that of 10uF in reset pin. Then, I also connect 10uF in paralel and the problem is similar. signals:

    What do you think? I need a new scope...I guess....?

  • In reply to fhoude:

    Hello Francis,

    The Exposed Center Pad isn't connected to GND. It is connected to a small electrically open area . Is it correct?


  • In reply to Sergio Valiente:

    Hey Sergio,

    The exposed center pad is used for heat dissipation purposes so I don't believe this will cause any issues.

  • In reply to Sergio Valiente:

    Hello Sergio,

    Your scope probably isn't fast enough (wide enough bandwidth) to capture you noise event.  Also, what is your ground reference for the probe and also the ground reference for the device. 

    It looks like you are using Altium.  Can you send me the packaged project so I can review the layout.  If the grounding and the location of the decoupling caps isn't done correctly then you might have issues like this. 

    Send files to francis.houde@ti.com

    Let me know where you placed the 10uF decoupling cap on RESET pin.  Also, did you try the decoupling capacitor on Vcc.  Take a look at that during transients too.

    Sorry for the late response.  I will be better about it.  I say Bobby responded but I didn't realize that your other questions hadn't be answered as well.

    -Francis Houde

  • In reply to fhoude:

    Hello Francis,

    OK, I send you the Email

    Thank you very much

  • In reply to fhoude:

    Hello Francis,
    Have you received the Email?
  • In reply to Sergio Valiente:

    Hello Sergio,
    Did you get my email and did that capacitor on Vccp fix the problem? Let us know.
    -Francis Houde