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.

Part selection advice sought for -ve input, -ve output, switching DC-DC regulator.

Dear Sirs,
I have a live product which needs a DC-DC power supply re-design, and I would like some assistance with selecting the correct TI parts and power-supply circuit architecture(s).

The unregulated DC inputs are (17V, 0V, -17V), and the required regulated outputs are +12V@6A, +5V@3A, +3.3V@3A, and -12V@6A.

The old design used linear regulators, which were electrically very low-noise, but were very power inefficient, and dissipated a lot of unwanted heat.

We would like to develop new higher-efficiency DC-DC switching regulators using TI parts.
Please bear in mind that the regulated outputs should be low-noise as they are feeding some very sensitive analog circuitry.

Regulating from +17V down to +12, +5 and +3.3 is easy thanks to the WEBBENCH software, however, we can't seem to get the WEBBENCH software to offer a (non-inverting) solution for going from -17V to -12V. It says this is not (yet) supported.

As such, we need help selecting the correct TI part number, and an app-note, to do the -17V to -12V converter schematic.

Regards,
Nicholas Lee

  • Hello Nicholas,

    Converting a -Vin negative voltage to a -Vout lesser negative voltage is fairly easy as the topology goes.
    A buck or synchronous buck will work fine.

    The problem is sensing the negative output voltage.
    Since the IC and converter circuit ground is actually at -Vin, the Vout feedback voltage is the wrong polarity.
    The feedback must be level shifted.

    See the attached paper on negative converters.
    buck-negative buck implementation.pdf

    Negative Buck

    Negative Buck Feedback level shifter


    Use WEBENCH to run a +17Vin to +12Vout @ 6Amps.
    You will see many results, controllers with external fets.
    At this power level a controller with external fets is the best approach, thermally, performance, and cost.

    Then use the level shifter to correct the feedback polarity.

    Seldom we get requests for such high current negative supplies, thus we don't have examples to match your exact needs.
    Nor does our Webench model this, yet. 

    Please let me know your thoughts.

  • Dear Ed,

    Thanks for responding. I think it is the voltage translator in the feedback path which is the trick to making this work. I will try and simulate some of the TI options in Spice as a sanity check, and to test the performance.

    Plenty of applications need split-rail regulation, but  finding information for designers is nearly impossible.

    Regards,

    Nicholas Lee