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.

TPS717: Output divider vs Ifb leakage current

Part Number: TPS717

I see that recommend lower resistor value for 3.3V output is 324kohms (with upper resistor @ 1.02Mohm referencing table 3 of the datasheet). That works out to ~ 2.5uA in the resistor divider chain (if there were no IFB leakage).

The leakage spec for the part is 0.02uA typical but can be as large as 1uA over temperature. Wouldn't this introduce a large error in the output voltage when using a 1.02M upper divider resistor?

Please reconcile this so that a circuit may be designed to maintain 3.3V output over temperature. If were to limit the impact of the 1uA leakage to say, a 33mV impact on the output voltage, then we'd have to use something like 33Kohms for the top resistor and 10.4K on the bottom. This is way outside the recommended resistor values (~ 10X less) - is that ok?. What output capacitance would be required in that case? 5uF? 10uF? How would we determine this?

  • Hi Steve,

    Yes, there is always a tradeoff between feedback network current and accuracy over temperature for external programmable devices. Generally speaking modifying this would not cause the device to not operate or be unstable. We normally recommend a feedback network total resistance of 100k-1M, some of the older devices, like this one, have much larger feedback resistor networks. Utilizing this feedback network you will see increased quiescent current in the system, but more robust performance on the output rail as the leakage current will have less of an effect. Modifying the branch however has the possibility to push the device into instability, by increasing the COUT minimum you should be able to avoid the instability, however I do not believe operating under such conditions will force the device to be unstable as the COUT + COUTesr is more dominant a factor. It would be a simulation or bench testing result that could give a confident answer to the question.

    Regards,

    John