• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Subscribers 62 subscribers
  • Views 925 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

PMIC with integrated DDR3L controller, looking for something like the NXP MC34VR500

Bob C
Intellectual 585 points
Other Parts Discussed in Thread: TPS65086, MSP430G2121, TPS65023, TPS51206, TPS65023B, TPS650231

Hello,

I'm looking for a  PMIC that integrates a DDR3L power controller (1.35V, VTT, and REF).  


The NXP MC34VR500 is close, but the voltage level outputs on its bucks are too low, since one of my rails need to be 2.5V. 


I need 2A of 1.1V and 1A of 2.5V plus DDR3L and a couple of LDOs.   I'm assuming I have a VIN (wall wart) to 3.3V buck regulator in front of the PMIC, but maybe I could also integrate VIN to 3.3V. 


I have been striking out on the TI site trying to search for PMIC with DDR3L. 


Thanks

Bob

  • 0
    TI__Genius 17380 points

    Hi Bob,

    We don't have many PMICs with integrated VTT. The TPS65086 has an integrated VTT LDO but most other options would use a discrete VTT. For the TPS65086, you can use a microprocessor like MSP430G2121 to set the defaults on power up or if you have a sufficient business case, a custom OTP can be preprogrammed by us.

    BUCK2 can do conversion from wall wart to 3.3V and then 3.3V can be used for input to converters (3.575 V is preferable if you don’t require 3.3V since it will give more headroom for 2.5V)

    BUCK3 for 2.5V (already has this as default)

    BUCK4 for 1.1V

    BUCK5 for 1.35V (DDR3L voltage)

    LDOA1, LDOA2, and LDOA3 are available for LDO requirements

    BUCK1 and BUCK6 would be unused so no need to buy the FETs [note: BUCK6 feedback needs to be connected to whatever rail is VDDQ, in this case BUCK5]

    SWA1, SWB1, and SWB2 are also available.

    Alternatively you could possibly use a TPS65023 + a discrete VTT LDO.

    VREF can be done with a resistor divider (~3 kΩ) from the 1.35V or if you are worried about it, an op amp.

  • 0 in reply to Kevin LaRosa
    Intellectual 585 points
    Thank you Kevin.

    After a first pass of the datasheet, the TPS65023X looks like a good fit. I would use VDCDC3 for DDR3 1.35V.

    I still need to come up with a solution for VREF and VTT.

    If you have any recommendations for particular parts or app notes, please share them with me. I'm only powering a single 16-bit DDR3 DRAM.

    Bob
  • 0 in reply to Bob C
    TI__Genius 17380 points

    Hi Bob,

    It's a bit outside of my expertise, you may have better luck posting on the non-isolated DC/DC forum (https://e2e.ti.com/support/power_management/non-isolated_dcdc/ ). Something like the TPS51206 could work, though it may be overkill. Alternatively, you may be able to get away with simple resistor termination; we've got a reference design with two DDR3 here: http://www.ti.com/tool/TIDEP0012

  • 0 in reply to Kevin LaRosa
    Intellectual 585 points
    Thanks again Kevin.

    I'm probably going to order one of the eval boards for this PMIC today, but I need to decide on 65023, 65023B, or 650231. Are any of these on a newer process than the others? I suppose the B revision is newer than the non B part?

    I'll want to use the QFN package.

    Is there a compelling reason why I would choose one part over the others? I'm not seeing anything meaningful.

    Bob
  • 0 in reply to Bob C
    TI__Genius 17380 points

    Hi Bob,

    The difference between TPS65023 and TPS56023B is shown in Table 17 in the datasheet; namely the TPS65023B I2C communication is faster. 

    The differences between TPS65023, TPS65023B, and TPS650231 are shown in Table 17 of the TPS650231 datasheet, namely TPS650231 changes the LDO1 [011] setting and removes DEFDCDC2 options.

  • 0 in reply to Kevin LaRosa
    Intellectual 585 points
    Thank you Kevin for the reply and pointing out Table 17 in the 650231 datasheet. I hadn't noticed that one.

    Boy, that's a very subtle difference between the 65023 and 650231. I'll guess it was needed for compatibility with a particular processor.

    I'll probably order the 65023B EVM a little later today.

    Thanks again

    Bob