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.

TUSB2046I: Oscillator questions

Expert 4630 points
Part Number: TUSB2046I
Other Parts Discussed in Thread: TPS2041

Dear e2e support,

My customer uses our hub to get multiple USB2.0 ports attached on our OMAP.

3 ports are used: on one we have a WIFI module, on another one we have another board for HMI, and the 3’rd port is used for an optional flash disk.

The HUB is designed as self powered, the HMI is always connected, the WIFI module can be powered OFF or ON according to the application needs, and the 3’rd port is used occasionally.

Performing some pretest by connecting the circuit with the HUB to a Windows PC, sometimes the HUB and the onboard Wifi module is enumerated correctly, sometimes not.

Happens also the HUB goes in suspend mode, monitored on the dedicated output. (see below oscillograms 7 and 8, blue = vcc, green = suspend out, red = oscillator)

oscillogram 6 for reference:

oscillogram 7:

oscillogram 8:

Questions:

  • Is there a minimum oscillator level needed to have the HUB working correctly?
  • We can see different oscillator levels with different crystals, would be a better solution to feed with 48MHz from an external oscillator?
  • Is normal that the oscillator stops when the HUB goes is suspend mode?
  • Sometimes happens that the HUB try to enumerate, goes in suspend mode several times, and finally remains in suspend mode (oscillogram 6)

Regards,

SCH_IO4K-R5_vP4K-GW24V_HUB (002).pdf

  • Greetings,

    Please note due to the thanksgiving holidays, response will be provided by Monday afternoon November 27th.

    Regards, Nasser

  • Hello:

       Can you take waveform during power up for VCC, crystal clock, and reset? Maybe 1uf on Reset is not enough for power on reset.

       How do you supply 5v for all downstream ports?

    Regards

    Brian

  • Hello Brian, Sorry for late answer. Here is the ocillogram you asked. (ch1=vcc, ch2=reset, ch3=suspend, ch4=osc). Let me know waht you think. In your opinion is better to feed the hub with an external 48MHz oscillator instead of using a quartz?

    The downstream ports are powered form a local 5V through TPS2041. The \EN signal is generated through an OR gate in such ways that the power control comes from the HUB but also from a GPIO of the main MCU in case we need to disable one of the devices in the downstream power (e.g. Wifi module, see below image). Let me know if you need more details.

    Regards, Zoli

  • Zoli:

        It looks like crystal was not stable yet when GRSTz goes high, maybe it's good idea to use 48Mhz clk for a good power on reset.

    Regards

    Brian

  • Hello, thanks for your feedback. We will do some test by using an external oscillator. Do you have some additional comment on our design? The instability we see, can be caused only by the oscillator? We found an error in our design, the EEDATA/GANGED line is floating without PU or PD. We don't use an E2PROM on our device, and the overcurrent detection sholuld be not GANGED. Can this flaw cause other instability? The datasheet says that there is a 100uA PD, but is not clear if this is valid also in case of no external E2PROM is used. Thanks.

  • Zoli:

      Even if there is overcurrent happened, it just  turn off downstream port vbus, but I'm not sure how do you implement power control for downstream port.

    Can you send schematic for review?

    Regards

    Brian

  • Hi Brian, the only port where we need power control is the USB-A connector to connect a flash disc. The schematics is very similar of what you see above, where the WIFI module power control is. But the question is if the EEDATA/GANGED pin left floating can have some effect in the correct functionality of the HUB? Thanks, Zoli

  • Hello Zoltan:

        EEDATA/GANGED pin only affect downstream ports power switching. Once overcurrent happened, it either disable Vbus for all DS ports or individual port and should not making hub into suspend mode or disable clock.

    Regards

    Brian

  • Zoltan:

        Any update to try 48Mhz external clock?

    Regards

    Brian

  • Zoltan:

        Happy new Year!

      Any update to try 48Mhz external clock?

    Regards

    Brian