Transient response 50% step change

AN1344

12/14

7

Transient response 50% step change

The result is 322 mV or 1.8% dynamic regulation with a settling time of 500 microseconds.

8

Output ripple

The ripple was measured using an HP probe socket attached after the output connector.

This minimizes stray noise being picked up by the scope probe ground lead, which shows up

as high frequency noise.

The top trace shows the reduction in cost from eliminating L4 and C11. This gives a ripple,

at 6 A load, of 125 mV peak to peak. With the low pass filter the ripple is reduced to about

13 mV excluding voltage spikes.

9

EMI consideration

When dealing with EMI, it is always best to reduce noise at its source. Figure 11 shows FCC

class B plots comparing EMI at 6 amps load with snubber R3 and C4 in and out. The blue

trace, or lower trace, has the RC snubber across the diode. The EMI is reduced by 4 to 8 db.

Adding a 2W resistor and a capacitor here is much less expensive than adding across the

line capacitors and inductors in the EMI filter. This unit passed both EN55022 class B and

FCC class B.

10

Performance and cost consideration

This design has been optimized for performance. Cost can be reduced by substituting a 17V

zener for the TL431. The output regulation falls to the +/- 5% voltage set point, plus a

+0.084/°C temperature drift of the zener. The cost of the TL431 and 3 other passive

components can then be eliminated. If more output ripple voltage can be tolerated, than L4

and C11 can be eliminated.

11

Conclusion

This design delivers over 100 W for both voltage ranges by utilizing the VIPer100A-E with a

voltage doubler in the front end. The power supply has excellent regulation, current limiting,

short circuit protection, meets both EN55022 and FCC class B and best of all is from

STMicroelectronics.