Using an NTC device instead of RF1

4 posts / 0 new
In our design using a LNK304 for a sub 1W supply, we have a separate fuse and so RF1 does not need to be fusible. We are thinking about using an NTC thermistor to limit inrush current. do have reccomendations for specifying that device/ also, if we choose to use a wirewound, non-fusible resistor for RF1 is there any reason for it to be a 2W resistor? it seems like it will only dissapate a small amount of heat and i assume the 2 watt rating is to keep the fuse from blowing? thanks! charlie
my understanding of RF1 is ro limit inrush to the front-end caps. why does the power output of the supply affect our need to limit inrush? for that matter, why do we need to limit inrush? thanks!

Hello Charlie

The link between power and inrush is not direct so I understand your question.

In low power designs (applications like mobile phone chargers up to 10 W) it is generally lower cost to replace the fuse with a resistor. The function is primarily to disconnect the AC supply from the power supply in the event of a catastrophic failure (for example bridge diode failing short circuit).

At these power level the inrush current is less of a concern as the impedance of the EMI filter, input capacitors etc limits the inrush current to below the maximum rating of the bridge rectifier diodes. Adding a fusible resistor only improves this by lowering the inrush current further. It also acts as part of the differential EMI filtering.

As the input power increases the loss in the resistor is unacceptable so a fuse is used instead. The increased input capacitance and lower impedance of the EMI stage now requires inrush limiting (to protect the diodes and other input stage components) and this is where an NTC would be used.

You asked earlier about resistor selection. We generally recommend 2 W wire wound types to handle the instantaneous dissipation seen when AC is applied. For example at 265 VAC (375 Vpk) and a 10 Ohm resistor the instantaneous peak power is 14 kW (V^2/R). Film construction resistors eventually fail in 230 VAC application. Some years ago I built a stress tester and attached is a photo of a 2 W film resistor that failed after about 20 cycles of AC being applied into 10 uF of capacitance (notice the black spots where the outer layer is damaged).

We always recommend flameproof types to prevent the possibility of fire (and meet safety requirements) when the resistor fails during a fault.

Cheers

PI-Chekov