Post Go back to editing

ADP1607 Vout not meeting datasheet

We got our first samples back that has our device with the ADI regulator (ADP1607ACPZN-R7) on there. We are seeing something that I don't understand and may make this regulator unusable.

We are trying to get an output voltage of 1.875V with a max voltage is 1.95V and our min voltage is 1.83V.
So, we put R1=137k and R2=280k. We used 1% tolerance resistors.

Also, the spec sheet says that Vfb = 1.259 +/- 2%.

The maximum error due to the feedback current is 0.25uA*137k = +34mV.

Adding all these things together, we are supposed to see an output voltage the never exceeds 1.96V and never goes below 1.83V.

We measured 10 parts and here is what we see:

  1. 1.97V, 1.95V, 1.91V, 2.00V, 1.95V, 1.92V, 1.87V, 1.87V, 1.87V, 1.90V

These are all with very light load currents (on the order of 10mA).

What is going on? Why am I seeing 20% being bad in only a sample size of 10? This means in high volume things are going to get way worse.

  • Your test result is reasonable as the IC your were testing is the ADP1607ACPZN-R7 whose operating mode is AUTO, which means the part
    operates in PFM ( skipping pulses) at light load and in PWM mode with heavier load. As it says in the datasheet, "In PFM mode, the converter only switches when necessary to keep the output voltage between the PFM comparator high output voltage threshold and the lower sleep mode exit voltage threshold." For more information, please refer to page 11 and figure 24, 25 in the ADP1606/ADP1607 datasheet. At 10mA load, the part is supposed to operate in PFM mode with large output voltage ripple as shown in Figure 24.

    Please note that the +/2% tolerate for the VFB is specified for PWM mode operation.

    For higher accuracy at the output at light load, please use the PWM only part ADP1607ACPZN001. Thanks.

    Best regards,


  • Thanks, Hong, that makes a lot of sense.

    A follow-up question:

    (1)  What would be the quiescent current on the PWM-only part?  Is there an efficiency curve for the PWM-only part at 1.8Vout?

    (2) I'm curious if I can still use the standard Auto part even with the voltage output variability.  What is the tolerance on VFB in PFM operation?  Is there any way to improve accuracy to within 20-30mV?

  • You are welcome! Thank you too.

    1) we don't have the quiescent current data for PWM operating mode. It would be higher than that in Auto mode. There is no efficiency figure for PWM operation either. The efficiency for PWM operation at light load would be much lower compared to Auto mode operation due to the FETs switching loss and inductor coreless which dominate the total power losses at light load.

    2)  You could try increasing the output capacitance to reduce the output ripple, therefore improve the output voltage regulation at light load. Also please refer to the following equation and fine tune R1 and R2 for the desired output voltage setting.