2
\$\begingroup\$

I'm assembling a PCB and have only the power supply section installed, shown below. The input (VIN) is 9V from bench power supply and output (VCC) is 5V. During testing I measure 4.86V on the output and am seeing a ripple of ~140mV with a light load of 15mA through the LED.

I'm using an Oki 78SR switching regulator. Its datasheet says the output ripple at full load is under 20mV. What could cause increased ripple in my circuit?

schematic

simulate this circuit – Schematic created using CircuitLab

The Oki datasheet recommends the input and output capacitors should be 10uF or larger where I used 1uF accidentally. However the part doesn't require these to operate. I will replace them when the new parts arrive.

What's not shown are decoupling capacitors for approximately 30 more ICs on the board, mostly 0.01uF to 1uF ceramic and then six tantalum 15uF capacitors. This total is within the maximum capacitive load the Oki part can handle which is 300uF so I hope that isn't an issue. To be clear the ICs are not populated, but their decoupling capacitors are.

\$\endgroup\$
5
  • \$\begingroup\$ How are you measuring ripple? You have to measure right across the output caps without any long ground lead. Adding the additional output capacitance will help. \$\endgroup\$
    – John D
    Aug 6, 2018 at 15:41
  • \$\begingroup\$ Depending on how far the other decouplers are away from the regulator, they may not have much effect on the regulator ripple. \$\endgroup\$ Aug 6, 2018 at 15:42
  • \$\begingroup\$ I just put oscilloscope probe across ground and the regulator output directly. Ground lead used is the short one that is normally on a scope probe. The two 1uF caps are right next to the regulator, the others are distributed around the rest of the board (about 7x7 inches). \$\endgroup\$
    – user185972
    Aug 6, 2018 at 15:45
  • 1
    \$\begingroup\$ You're using a 1.5A switcher to provide smooth current to 15mA load without the right output cap. It's like asking a loaded tractor trailer to drive smoothly in stop and go traffic. Try increasing the load or even REDUCING the board capacitance. The control loop on those switchers can actually be thrown off with the wrong components. The ripple you see may be oscillations from the control loop instability (moved Zeros and Poles). \$\endgroup\$
    – MandoMando
    Aug 6, 2018 at 16:18
  • \$\begingroup\$ It could be that your inrush current is triggering the overcurrent protection and causing the supply to repeatedly go into protection mode. \$\endgroup\$ Aug 6, 2018 at 20:32

1 Answer 1

1
\$\begingroup\$

They used this circuit for their test condition:

enter image description here

So unless you have your caps going from +Vout to -Vout, and have 11uF of capacitance and have a resistor for a load, I would not expect the same result. It may be that the non-linear load of the LED could be creating the differences in the ripple, but also the lack of capacitance.

In either case increase the capacitance to reach the ripple values you require.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.