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.

  • Resolved

TPS7A85A: Voltage sensing away from the IC

Intellectual 335 points

Replies: 9

Views: 99

Part Number: TPS7A85A

Hi,

How does the voltage sense feature (pin #2) work? I am planning to build an LDO board using this device and power my board with it. There will be a wire connection between the LDO board and my board. I want to ensure that the voltage at the power plane, which is on an inner layer in my board, is exactly (or very close) to the LDO output voltage. If I hook up the boards as shown in the picture below, will I achieve my goal? There was a zero ohm resistor connected between the Vout trace and SNS pin on TI eval board. In this case, the power plane acts like the zero ohm. Will it work? EVB stands for evaluation board.

Thanks,

Noman

  • Hi Noman,

    The sense pin connects the high side setpoint resistor of the feedback loop to the output of the LDO.
    As you can see in the image below there is a small impedance between Vout and the load.  By connecting the sense node very closely to the actual load, the feedback loop can account for this drop and maintain the load to a value very close to the expected Vout.

    It would be a good idea to keep 1 capacitor very close to the TPS7A85A output pin so that the stability of the part can be ensured.
    If there is a very long harness or significant inductance between Vout and the Cout, then the load impedance looks less and less like a capacitor and more like an LC circuit.  A capacitor right next to the TPS7A85A which meets the datasheet requirements for stability will eliminate this concern.

    Thanks,

    - Stephen

  • In reply to Stephen Ziel:

    Hi Stephen,

    My current schematic looks as attached. Do the capacitor values look okay to ensure stability? Pins 2 and 4 on connector J1 go to the connector on my EVB through wires as I showed in my last post. Thanks.

    Regards,

    Noman

  • In reply to Noman Paracha:

    Hi Noman,

    These capacitors look good and the LDO design also looks good.
    I expect it will be very stable.

    Thanks,

    - Stephen

  • In reply to Stephen Ziel:

    Hi Stephen,

    Since our last chat, I ordered the TPS7A85 eval board and did some measurements with my eval board. The test setup looks like the image shown below. The wire between the LDO board and my eval board is 4 inches long. I tried to keep it as short as I could. I disconnected resistor R2 on TI eval board and connected the top pad of that resistor to the yellow wire which goes to my eval board and connects to the power plane on the inner layer where I want to ensure 1.8V. I am able to see 1.8V on my board, but I have to set the voltage on TI eval board to 1.85V (i.e. 50mV higher than what I should be setting it to) via the jumpers on header J6. I don't understand the extra 50mV. Is the sense feature not working properly? Here are the voltage measurements at various points in the setup:

    • Vsetpoint (thru jumpers) on TI eval brd = 1.85V
    • Vout at TP5 on TI eval board = 1.91V
    • VDDPA at power connector, pin 4 on our EVB = 1.82V (no RF, Iquiescent=1.170A)
    • VDDPA at decoupling cap close to IC = 1.79V (with RF_IN = +5dBm, ITX ~2A)

    Can you tell me if I am missing anything? Thanks!

  • In reply to Noman Paracha:

    Hi Noman,

    The jumpers you have set in the image, from what I can tell anyway, would give 1.95V output.  Can you confirm that they are correct?  Also confirm that R3 and R4 are uninstalled.

    Where does the sense line tie to Vout, is it right at the connector input to your board?

    When you say you connected the top pad of R2, can you confirm that you are connecting to the node in the image below?
    I would do an ohm measurement from the yellow wire to the LDO pin 2, to confirm that they are still shorted and there is not some higher impedance there.

    It would not be hard for me to replicate a similar test here in our lab with this EVM, so if this is not resolved in a few days or less, I can look at it on the bench in our lab.

    Thanks,

    - Stephen

  • In reply to Stephen Ziel:

    Noman,

    It just occurred to me that the 4'' ground wire has current flowing through it, and that may not be getting canceled out by the remote sense.  Only the forward voltage drop would be getting canceled out. Typically in a system the ground plane is very low resistance, so this is not an issue, but here ground is also a harness.  So it's possible you are seeing some benefit from the remote sense but there is still some drop after all.  What wire gauge are you using?

    Thanks,

    - Stephen

  • In reply to Stephen Ziel:

    Hi Stephen,

    Thanks for a quick response. Here are my comments to your questions:

    • The jumpers you have set in the image, from what I can tell anyway, would give 1.95V output.  Can you confirm that they are correct? Yes, it is set to 1.85V. I took the picture before the measurement.
    • Also confirm that R3 and R4 are uninstalled. I did not make any hardware changes on the board. So, most likely they are DNP. Will double check.
    • Where does the sense line tie to Vout, is it right at the connector input to your board? It connects on the inner layer power plane. The force and sense lines are separate on the power connector. Then on our board, they both connect to an inner layer through blind vias.
    • When you say you connected the top pad of R2, can you confirm that you are connecting to the node in the image below? Affirmative. It is connected where you are pointing to in the picture.
    • I would do an ohm measurement from the yellow wire to the LDO pin 2, to confirm that they are still shorted and there is not some higher impedance there. Will do.

    Thanks again,

    Noman

  • In reply to Stephen Ziel:

    Hi Stephen,

    This could be the root cause. You are correct that there will be some voltage drop over the ground wire. I am using 28 AWG wire. Do you think with just 4" long wire, it can cause 50mV drop though?

    Thanks,

    Noman

  • In reply to Noman Paracha:

    Hi Noman,

    The resistance of 28 AWG is 64.9 ohms per 1000 feet.  At 4 inches that equates to 21.6 miliohms.  This neglects on board return resistance and contact resistance of the connectors.  In my experience it is common to see 5 miliohm contact resistance (maximum) for one mating end of a connector, or 10 miliohms maximum for the male and female ends of the connector.  You can review the datasheet for your connectors, the contact resistance should be specified.  If we just use 5 miliohms for both connectors, that gets us 10 more miliohms.  The PCB ground plane resistance will be some additional miliohms of drop.  It's not much, but when you are drawing amps of current it does add up and I can see 50 mV of drop there.

    You can try to increase the AWG wire thickness to reduce the voltage drop as the wire may be your largest contributor right now.
    You can also use multiple wires in parallel if your connector will allow that.

    This is unrelated but you will want to twist the forward and return wires together in the harness, if you have not done so, to help eliminate some of the harness inductance and improve the transient performance of the system.

    Thanks,

    - Stephen

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.