• TI Thinks Resolved

TIDA-020008: Simulation, hardware & system design tools forum

Intellectual 340 points

Replies: 8

Views: 127

Part Number: TIDA-020008

Hi,

I’m trying to determine the trace widths for my power traces for a motor driver circuit. I'm referencing TIDA-020008 design and I have a few questions: 

  1. In TIDA-020008 layout, how much current can go through reverse protection diode D1? I see that it only has a trace width of 10 mil. Just want give it sufficient trace width.
  2. What’s the recommended minimum trace width for VBAT or VIN+ traces?
  3. What’s the recommended minimum trace width for 3.3V traces?
  4. What's the recommended trace width for signal lines?

 My motor driver circuit's application has a motor RMS current of 3.5A and stall current of 9A.

 

Thanks,

John

  • John,

    1. D1 doesn't carry the main current, even under reverse battery conditions.  D1 and Q2 ensure that the gate of Q1 is pulled low under reverse battery conditions.  So the current through D1 is on the order of 12V/43kOhms (when reverse battery conditions occur).

    2. VBATT through Q1 to VIN+ to the motor drive FETs Q3A and Q3B is the high current path.   For TIDA-020008, the width is 100 mils (top and bottom layers), which is probably very conservative for your motor current specs.  We use IPC 2221 as a guideline for trace width; there are some on-line calculator tools (e.g. https://www.4pcb.com/trace-width-calculator.html or http://referencedesigner.com/cal/cal_06.php ) which do the calculations.  In the TIDA-020008 design, each 100 mil width supports a 10A current with 45 degree temp rise assuming 2 oz copper on the exterior layer plated to a 3 oz equivalent.  Note that the peak (stall current) should not occur for any significant length of time, and even the running current of 10A is typically only active for a few seconds in these applications.

    3. For the 3.3V traces, I used 20 mil width for the long trunk, but then branched out into 10 mil traces for individual devices.  The design is a bit constrained in that it fits into an existing form factor, and that the MCU is on a separate (LaunchPad) board. 

    4. In general, 10 mil is a default for signal line width.  If space is limited, 8 mil can also be used.  One constraint for a cost-effective 2-layer board is that we want heavy copper thickness for current carrying capability, but this means that very thin traces can be problematic due to the inconsistency in etching the heavy copper.  It is best to check with your board fabrication house to see if they have additional guidelines if traces below 10 mils are needed.

    -Clark

    -- Clark Kinnaird

  • In reply to Clark Kinnaird:

    Hi Clark,

    Thank you so much for the feedback. For my layout, the high current path is from VBATT through Q1 to VIN+ which then goes to a motor driver DRV8873. My copper thickness is 2 oz for top and bottom layers and 1 oz for the mid 2 layers. 

    For high current path, my trace width is 80 mils, is this ok? Per IPC-2221, for 2 oz, temp-rise of 30 degC, seems like greater than 65 mils should be fine. Do I need to run traces both on top and bottom like you did in TIDA-020008?

    Thanks,

    John

  • In reply to John Manyala:

    John,

    For your motor current levels, 80 mils on 2 oz copper sounds more than adequate.  Regarding the top vs top and bottom layers, I routed in parallel on the bottom layer due to my much higher motor current level, but in your case it isn't necessary.

    -Clark

    -- Clark Kinnaird

  • In reply to Clark Kinnaird:

    Hi Clark,

    Thank you so much. One final question, if I want to add a heat sink for thermal dissipation on the bottom layer (GND), how should I define it within the layout?

    Thanks,

    John

  • In reply to John Manyala:

    John,

    I typically use the heatsinks in our Altium library, and add this to the schematics as a new component.  Then update the layout based on the schematic, and edit the heatsink properties so it is on the bottom layer.  You will want to avoid any through-hole pins from components on the top side, and probably tent the bottom-side vias to avoid shorting any signals through the heatsink.

    -- Clark Kinnaird

  • In reply to Clark Kinnaird:

    Hi Clark,

    Thanks again. I cant't find any heatsink within Altium's library. Do you have the part number for the heatsink you used for the motor driver (DRV8873)?

    Thanks,

    John

  • In reply to John Manyala:

    John,

    In the TIDA-020008 design, I did not use a heatsink for the DRV8873-Q1.  Because the motor is active for relatively short durations, it isn't necessary to add a heatsink. If you think a heatsink is necessary for your application, the choice would depend heavily on the mechanical size constraints and thermal path available.  For the TIDA-020008, the board was simply mounted in a plastic enclosure, which is typically found under the automotive seat.  No forced air or separate heatsink components were needed.

    -Clark

    -- Clark Kinnaird

  • In reply to Clark Kinnaird:

    Hi Clark,

    Thanks. I totally agree, I doubt that I'll need an heat sink too for my application since the motor is expected to be active for less than 15 seconds at a time.

    Do you happen to have worst case circuit analysis for the DRV8873-Q1?

    Thanks,

    John