• State Resolved
  • Locked Locked
  • Replies 10 replies
  • Subscribers 32 subscribers
  • Views 2288 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

TUSB2077 design considerations

Atif Bhatti
Prodigy 220 points
Other Parts Discussed in Thread: TUSB2077A

I am thinking about using this chip in my Fixture design to be able to talk to 14 FTDI Chips.

I have a few questions

1). Are there any layout considerations regarding distances between this chip and upstream/downstream USB ports?

2). Any recommendations for Trace width etc?

3). since I have to communicate to 14 Targets I am planning to use three of these chips in cascaded mode giving me 19 targets. Do I need to use 27 Ohm resistor with 22 pF caps (as suggested in recommended design) at the input of second and third stage of cascaded TUSB2077 chips?

My application requires atleast 5 GPIO and 1 serial port communication capability per target, due to this I have been using FTDI FT232RL chips. in order to simplify my fixture to have less cables (reduce 14 USB cables down to one) I need to go to this chip design. I am open to consider any device/chip that will further reduce the complexity of my circuit.

Thanks in advance

Atif Bhatti

  • 0
    TI__Guru* 76130 points

    Hello Atif,

     

    For your questions 1) and 2), please refer to the following white paper for  design reference. although it is aimed for HS signals, it is also applicable for FS.

    http://www.usb.org/developers/docs/hs_usb_pdg_r1_0.pdf

    For your question 3) the answer is:  Yes, you must populate the series resistor and the edge rate control capacitor at the input of the second and third stage, also you must populate the 15K pulldown on D +/- as showed in the recommended design.

     

    Regards.

  • 0 in reply to Elias Villegas M.
    Prodigy 220 points

    Hi,

    Thanks for your previous answer. I have my design completed and wanted you to take a look if possible. it will save me bunch of time and money if I can get some sort of feed back on this design.

    7534.Hub Design.pdf

     

    8053.RC Filter and Transient Sup.pdf

     

    Please take a look at the configurations especially (Pin 8 and Pin 10). all of my targets downstream will be powered by the Hub and I want to kill the power to only the downstream device that draws excessive current (not Ganged).

    I have 14 Targets devices to talk to. I will be cascading three of these chips. Feeding one output port from previous stage to the next stage, this will give me 19 USB ports, 14 for my target devices and 5 extra to be able to attach other deivces.

    Atif

  • 0 in reply to Atif Bhatti
    TI__Guru* 76130 points

    Hello Atif,

     

    We will check your schematics and reply to you soon.

     

    Regards.

  • 0 in reply to Elias Villegas M.
    Prodigy 220 points

    Please also check

    "RC filter and transient sup.pdf "

    Since all of these are on the same board, which includes

    1). All three stages of USB Hub (TUSB2077A)

    2). All of my targets

    3). Extra I/O chips

    Can I get by without having transient supperssors for Down stream devices? or atleast have them on the PCB layed out as JIC?

    THis board is part of my test station and will be in the RF Enclosure.

    Atif

  • 0 in reply to Atif Bhatti
    TI__Guru* 76130 points

    Hello Atif,

     

    I've reviewed the hub schematic and have few comments:

     

    1: For the RC circuit you are using in the reset input of the TUSB2077, I recommend  a 20K resistor instead the 2K.

    2: You must  populate a 15K pull up resistor to 3.3V on every OC# output of the TSP2044.

    3:  Question: What are the C and R values of the RC network  placed on downstream ports D0-D7 ?

     

    Everything else looks ok, I will continue with the RC filter... section.

     

    Regards.

  • 0 in reply to Elias Villegas M.
    TI__Guru* 76130 points

    Atif,

     

    Please take no attention  to " 3: Question..."  

    I haven't seen the other attachment.

     

    Regards.

  • 0 in reply to Elias Villegas M.
    TI__Guru* 76130 points

    Hi Atif,

     

    Your RC network looks also OK.

     

    If you need anything else, I would be happily in helping you.

     

    Regards.

  • 0 in reply to Elias Villegas M.
    Prodigy 220 points
    • I had 2K and 0.1uF for Reset Circuit according to 100 uS to 1ms pulse requirement described in TUSB2077A Speicifications document. It will not be a big deal for me to change from 2K to 20K. I Just need to understand why we will be doing this?
    • Since all of these are on the same board, which includes , All three stages of USB Hub (TUSB2077A), All of my targets and Extra I/O chips. Can I get by without having transient supperssors for Down stream devices? or atleast have them on the PCB layed out as JIC?

    FYI: This board is part of my test station and will be in the RF Enclosure.

    Atif

  • 0 in reply to Atif Bhatti
    TI__Guru* 76130 points

    Atif,

     

    Giving a long reset pulse(1.5 to 2 ms) is a good practice we've seen, if you can change these resistors without problems  it wouldn't hurt.

    You can remove the  transient suppressors  of the downstream ports of the hub, but the 22pF caps, 15K resistors and 27 ohm resistors are required.

     

    Regards.

  • 0 in reply to Atif Bhatti
    TI__Genius 16440 points

    Atif,

    A 2K and 0.1uF give a reset pulse of 200us.  Per the TUSB2077A data sheet, for the /RESET pin description.
    Generally, a reset with a pulse width between 100 μs and 1 ms is recommended after 3.3-V VCC reaches its 90%. The clock signal must be active during the last 60 μs of the reset window.

    Notice that the 100us to 1ms is AFTER  the 3.3 V supply reaches 90% of its value or 2.97V.  Given that the clock must be active for a minimum of the last 60us of the RESET window, a 200us Reset pulse only leaves 40us of wiggle room for ramp of the 3.3V.  When you start taking into account the tolerance of the R and C.  If they are both 5% and at the bottom of their spec then your reset pulse drops to 180.5ms and only leaves you around 20ms of ramp for the 3.3V.

    Increasing to 10K (or the 20K that Elias suggests) gives you plenty of margin for the device to see the reset correctly.