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.

TPS7A30: TPS7A30 and TPS7A49

Part Number: TPS7A30
Other Parts Discussed in Thread: TPS7A49, , TPS7A39

I am using the TPS7A49 and TPS7A30.  Input voltages are +/- 15 VDC respectively, and output voltages are +/- 10 VDC respectively.  Output current draw is very low-- about 12 mA.  Here is a snippet from my schematic showing their connections.  As far as I can tell, they are connected by the book.  The pins labeled as pins 9 refer to the power pad.  Input and output filter capacitors are X7R and are placed close to the parts.  These parts are hand soldered to several first-article PCBs for testing.  As such the power pad is not soldered.  However, power dissipation is so low that the ICs don't get even the slightest bit warm.  My problem is that both ICs will sometimes, but not always, show some wandering in their output voltages.  Specifically the output voltages will drift around by 1 to 3 mV in a frequency band of about 0.1 Hz to a bit below 1 Hz.  This wandering can easily be seen on a scope.  Other than that, the outputs look quiet with no visible oscillation.  This wandering is a critical fault in my application and we are desperate for a fix.  I have noticed that these parts are hard to get.  Do you think it's possible we have received bootleg parts?  Has TI seen that problem with these ICs?  Your earliest response would be hugely appreciated.  Thanks in advance for any light you can shed on this situation.

  • HI Roy,

    This is the first I have heard of this type of issue. the TPS7A49 and TPS7A30 have been on the market for many years now. If you can, please provide a scope plot of what you are seeing. While the power pad is not an ideal GND is is connected to GND so may have some impact on performance. Are you able to see the same issue on our EVM?

    Do you know if adding a artificial load of a few more mA helps to resolve the issue? This can be easily done by using smaller value resistors in the feedback divider.

    I am curious as to what application is sensitive to this low frequency drift. Please share along with a schematic of the LDO if you can.

    Note: We do have the new TPS7A39 which combines these two devices into one part and adds start-up tracking. However, this is based on the same topology as the TPS7A49/30 so if these are in fact our parts and you are seeing the issue, then you might see it in the TPS7A39 as well.

    I hope this gives you some ideas.


    Regards

    Thanks,

  • Hi Roy,

    Please refer to the following on how to insert images on e2e.  Unfortunately the forums are a bit picky on how images must be attached to a post.

    Very Respectfully,

    Ryan

  • HI Roy,

    Did decreasing the resistors in the FB loop help resolve the issue or are you still seeing it?

    Regards,

    John

  • Thanks for you reply.  The parts are putting out about 20 mA in this application.  The Feedback resistors I am using are 149k and 20k, which produce 10 volts out from the regulators  The 149k is rather high, and its Johnson noise may not be helping matters.  I could change to 15k and 2.0k resistors, just as an example.  The 10x reduction in resistor values should reduce resistor noise by sqrt(10), or 3.15X.   I will try this change out this afternoon--another engineer is using the equipment for other tests this morning.

    It may be worth mentioning that given the 10V output, the noise gain through the regulators is 8.5X what it would be if the feedback resistors were set for no gain, i.e., an output voltage equal to Vref.  I'm sure this is part of my problem.  Of course, at very low frequencies, the noise-reduction capacitors don't help, and one is firmly in the 1/f region.

  • As a follow up, decreasing the resistors did help.  Given the noise gain of the devices when they put out 10 volts, I may just be at the limit of their noise performance.  1/f noise is always the bugaboo of voltage regulators--I hope TI continues to try to find new processes and topologies to address this problem.

  • Hi Roy,

    I am glad that helped. Since this worked, I believe the root-cause is the fact that all LDO's transition modes between no-load to some-load.

    In no-load operation, the LDO is trying to make sure that parasitic leakages do not allow VOUT do drift up. So in this phase it is actually sinking a very small current to GND.

    When a load is present, the LDO is now modulating the pass element to regulate VOUT as designed. During this phase (which is commonly in the range of 1-20mA) depending on the LDO you might a minor (i.e. mV change in VOUT). So by decreasing the FB resistors, you raised the no-load current enough such that the LDO was always in regulation mode, thus avoiding the (wandering) you were observing earlier.

    If I can understand more about how this low frequency oscillation was impacting your design I can better work with our designers on how to improve.

    Feel free to ask if you have any other questions.

    Thank you,
  • The regulator was always under load, so I think decreasing the resistors simply reduced the Johnson noise of the resistors as well as reducing the effect of any input-referenced current noise of the converter..  I was concerned about the wandering due to the somewhat unusual design of the circuitry that was powered by these regulators.