• State TI Thinks Resolved
  • Locked Locked
  • Replies 11 replies
  • Answers 2 answers
  • Subscribers 62 subscribers
  • Views 349 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.

TPS51200-EP: About the Vref precision

Matth. B.
Expert 1135 points
Part Number: TPS51200-EP
Other Parts Discussed in Thread: TPS51200, TPS51206

Hello,

On the datasheet of TPS51200, I have trouble about two table (I'm using DDR3L and DDR4) :

By reading value, it seems that Refout could vary by +/-2% for Vrefin=0.6V.

Regarding datasheet of DDR4 component from Micron for example, they requested no more than 1% of VDD (1.2V) for Vref variations so +/-12mV.

If I add the resistor tolerance on Refin, I will be out of specification using TPS51200.

Same thing for Vo, +/-15mV for TPS51200 which is 2.5%, requested no more of 3% by Micron. Using 0.1% resistors could be fine but I'm afraid about the +/-25mV of output voltage tolerance to REFOUT.

So, what I'm missing to say that we can use TPS51200 for DDR4.

Thanks

  • 0
    TI__Mastermind 24075 points

    Hi Mattieu, 

    The TPS51200 REFOUT tolerance to REFIN is +/-12mV at +/-1mA load. This spec is in datasheet page 6. 

    In DDR3 or DDR4 application, the Vref current is only a few uA, therefore the REFOUT variation to REFIN will be much smaller than +/-12mV.

    If 0.1% resistor is used for REFIN, REFOUT tolerance (0.5*VDDQ) to should be within +/12mV which is +/-1% of VDDQ. 

    The VTT tolerance to REFOUT is +/-25mV with +/-2A load. 

     The total VTT tolerance to (0.5*VDDQ) is 12mV+25mV=37mV. 

    Most DDR4 memory allow VTT tolerance larger than +/-37mV. 

    Therefore TPS51200 can be used in DDR4 application. 

    Thanks

    Qian

  • 0 in reply to Qian Chen
    Expert 1135 points

    Hello Qian,

    Thanks for your feedback.

    So, do you have the tolerance of REFOUT at a constant load (ie process variation and temperature variation only)?

    Also, do you have datas of variation with aging on these outputs?

    Thanks

    Matthieu Baque

  • 0 in reply to Matth. B.
    TI__Mastermind 24075 points

    Hi Mattieu, 

    The +/-12mV REFOUT tolerance in datasheet spec table already includes process and temperature variation. 

    TPS51200 was released many years ago, and it is being used by many customers in huge quantity. By now no customer reports there is aging problem  on REFOUT and VTT. 

    Thanks

    Qian

  • 0 in reply to Qian Chen
    Expert 1135 points

    Hello Qian,

    I can't say it is OK to use it for DDR4 without number. If I take the one in datasheet this is not OK (12mV is the max allowed for DDR4 chips) but as you said the real load variation is less than 1mA.

    So, my question is do you have datas without load variations or with a load variation smaller than 1mA?

    Thanks

    Matthieu Baque

  • 0 in reply to Matth. B.
    TI__Mastermind 24075 points

    Hi Mattieu, 

    With +/-1mA load, the actual REFOUT variation should be below 10mV. 

    with 0.1% tolerance resistor for REFIN,  the total REFOUT tolerance should meet +/-1% of VDDQ.

    Thanks

    Qian

  • 0 in reply to Qian Chen
    Expert 1135 points

    Hello Qian,

    In order to minimize the refout variation, I would like to decrease the value of the refin resistor divider. Goal is to have less impact with the Refin input current. In the datasheet, value are 10 KOhms.

    What are the minimul value for these resistors?

    Also, for only two DDR3L components, in order to have less initial tolerance than the 12mV of the TPS51200, is it not preferable to use bridge divider directly for Vref pins and not the VRef buffer of TPS51200. I'm thinking about the use of resistors for Vref and TPS51200 for VTT.

    Thanks

  • 0 in reply to Matth. B.
    TI__Mastermind 24075 points

    Hi Matthieu, 

    If it is only for two DDR3L chip, I would suggest to use TPS51206. 

    No REFIN resistor divider is needed by TPS51206. 

    Thanks

    Qian

  • 0 in reply to Qian Chen
    Expert 1135 points

    Hello Qian,

    Thanks for your feedback but for a lot of reason (technical and planning), change is complicated for us.

    So, do you think it's better to use TPS51200 or resistor divider for Vref?

    Thanks

  • 0 in reply to Matth. B.
    TI__Mastermind 24075 points

    Hi Mattieu,

    Yes, you can use resistor divider for Vref. 

    The Vref current for DDR3L should be very small (a few uA each DIMM chip). 

    The related information is available on TPS51200 datasheet page 14. 

    Thanks

    Qian

  • 0 in reply to Qian Chen
    Expert 1135 points

    Hello Qian,

    Again some request:

    About the +/-25mV Vvotol and +/-12mV Vrefouttol, these data include initial tolerance + temperature variation + load variation.

    Is it possible to have the three independent parameters (ie: +/-X mV for initial tolerance + Y mV for temperature + Z mV for load)?

    Thanks

  • 0 in reply to Matth. B.
    TI__Mastermind 24075 points

    Hi Mattieu, 

    Sorry, I don't have independent data. 

    Thanks

    Qian