TPS73701DRVEVM-529: Minimizing current in the resistor divider network

Hi Anand,

The datasheet recommendation is to allow the customer to easily ignore the nonideal feedback current, which adds an error term to the Vout equation you see in your image from the datasheet.  We can run statistical analysis software to assess the tolerance when you need to increase the feedback resistors to lower the quiescent current, as you are describing.  Please provide me your Vout and the Vout steady state tolerance, and I will run the statistical software to give you your best resistor combinations.

Thanks,

- Stephen

Hi Stephen,

Appreciate your quick response,

Kindly consider the Vout = 3.2V and steady state tolerance of 1%.

Hi Anand,

Keep in mind that the internal reference error is at best +/- 1% but only at room temperature.  The addition of the external feedback resistors will always lead to a slight increase in the tolerance. 

I ran 100,000 monte carlo simulations for every resistor combination from 100 k-ohms to 10 megaohms using E196 series resistors.  Each resistor is 0.1% tolerance.

While the reference voltage and resistors are typically +/- 6 sigma, most customers do not have the manufacturing volumes to require this in their specifications.  +/- 6 sigma inside a tolerance spec of +/- 1% is more costly than +/- 3 sigma inside the same tolerance spec.  A more common requirement is +/- 3 sigma which is 99.7% of all production units passing.  So individual parameters are modeled as +/- 6 sigma but I recommend using +/- 3 sigma for your Vout requirement.

Unfortunately the feedback current is probably too high to choose very large R1/R2 feedback resistors.  The error caused by this current begins to dominate the output voltage at large R1/R2 values.  If these results are insufficient then what you need is a different LDO with smaller feedback current.

Thanks,

- Stephen

Hi Stephen,

Thank you for your response.

Our output voltage is 3.2V and we may have up-to 3% tolerance.

Could you please give us more details on how you have tabulated these resistor values.

For Ex, considering R1 and R2 to be 301k and 105k, we are observing around 3.48V output voltage in our board(Higher than the required 3.2V). But the simulated value provided in the table (previous response) it is around 3.2V.

The output voltage equation given in the datasheet : VOUT = (R1 + R2)/R2 ´ 1.204. However, with the given values(for ex: 301k and 105k), Vout tends to be greater i.e around 4.6V(which is not feasible in our case, as input is around 3.7V). 

Kindly share your thoughts regarding the same. 

Also, could you please let us know the quiescent current of the IC

Thank you.

Hi Sushruta,

I have received your reply and I need up to 2 business days to provide a response.  But yes, I can go into as much detail on this topic as you would like.  Please give me some additional time to provide a response and we will go from there.

Thanks,

- Stephen

Thank you Stephen.

We will be waiting for your response.

Warm Regards,

Sushruta

Thank you for you patience Sushruta. 

Best,

Juliette