3 Watt inrush resistor will eventually fail?

[treez]

Hello,
We have designed a 60W offline (190-265VAC) Flyback LED driver (PFC’d). We have a 4R7, 3W inrush resistor near the mains input connector. (as in the attached schematic)
This is a 3W axial film resistor , type “ROX3S”
ROX3S resistor datasheet…
https://www.te.com/commerce/Documen...rtrv&DocNm=1773271&DocType=DS&DocLang=English

Do you think this resistor will eventually fail after a number of ON cycles, especially with some of them happening at the instant of mains peak voltage?

 

[betwixt]

I would guess, without going into complex calculations that it would be fine. The 3W rating is for continuous running which would equate to the lamp drawing about 800mA (~180W load) and the resistor is built to handle high surge currents. The surge would be brief and you already have a further current limiter after it. You only have about 5uF of capacitance to charge up.

Brian.

 

[c_mitra]

Resistors are robust beasts. If the rating is 3W nominal, it can survive 30W peak for several ms. (just guessing).

Fortunately, unlike semiconductor damages, resistor damages are not cumulative. The beast can heal with time (only partly but then...).

Film resistors are particularly strong but then I do not know the exact details.

 

[thunderdantheman]

What you want to look for is the pulse rating of the resistor.. Or if not specified for your current part, consider using a pulse-rated resistors.. these are designed and specified (among other things) for exactly what you are trying to achieve. This appears informative on the subject **broken link removed**

 

[c_mitra]

What you want to look for is the pulse rating of the resistor..

It may be not worth the trouble: a fast-acting fuse designed to protect an expensive transistor, was itself protected by the transistor getting blown up in a surge.

 

[Nik_2213]

What was the type of inrush catcher used on BIG linear power supplies ?

The sort of PSU with a pair of thousand-plus micro-Farad 'beer cans' ??
Plus a 'bleed resistor', lest their residual polarisation bites the unwary ?

I remember them as bespoke NTC devices, but such were out of my league...

 

[Relayer]

Hello treez,
4.7 ohm 3 and 5 watt resistors have been in use with virtually every bit of consumer
appliances over the years, and in those, it was extremely rare that it ever failed on its own.
Its main failure was due to excessive current from a shorted bridge rectifier diode/s or
chopper transistor, where the fuse took too long to blow.
In some instances when the degaussing thermistor goes S/C it can take it out as well.
If you were using a NTC thermistor to suppress inrush currents and spikes, that would be a different
story. The chemical composition of those devices changes over time and never seem to recover to
their normal state. They then eventually suppress less and less and allow more inrush current, to the
point of virtually becoming shorted and no longer do squat. Therefore they will fail after many years
with fairly violent consequences.
i.e. Taking quite a bit of other components with them.
Regards,
Relayer

 

[brushhead]

When I design something like this it's worth noting what the resistor's energy rating is as that is more relevant. You see a direct reflection of the Capacitor energy when you charge it. Saying that I have used NTC thermistors as other correspondents have pointed out.

Regards,
Rob.

 

[asdf44]

Yes you should find the joule rating of the device or use the pulse rating chart to extract one. Then compare to the energy in your cap.

A small subset of resistors have a proper joule rating. Other manufactures have app notes covering it generally.

An NTC would probably be superior for a non-bypassed inrush limiter. A PTC can be ideal for a bypassed application.

 

[c_mitra]

Most semiconductor device failures follow a Poisson distribution and is associated with a quantity "mean life expectancy". A design is robust when all the components used in the design have similar life expectancy at the beginning.

In your present application, the mean life expectancy is related to the related to the number of turn on/off events. You therefore need to estimate the mean number of turn on/off the resistor can stand. Poisson distribution is exponential in nature and the device can also fail before and after.

Perhaps some other component (perhaps a transistor used as a switch) may have a shorter expected life? In such a case, it is not worth worrying too much about a resistor. It may survive the life of the lamp...

 

[kekon]

Recently I've had this sort of problem. I made 20W flyback converter for 230V AC. The inrush resistor was 2 watts metalized film resistor. The filtering capacitor (after the bridge) was 22uF/400V. When the converter was turned on for the first time to the mains nothing wrong happened and the converter worked normally (just after power-up it was not fully loaded - only 50 mA was drawn from the output). When it was turned on second time, it worked for a while and then the inrush resistor blew up. The interesting thing is that there was damage only in one spot in the resistor - a tiny part of protective coating layer was thrown out forming a small black hole in the resistor's body. The strange thing is that the resistor was not damaged during the first surge applied but after second one and only after some delay when the inrush current completely decayed. It happened in almost every piece of the converters ! In my opinion, for the purpose of limiting inrush current a wire-wound resistor should be used. Wire-wound resistors have high surge current capability. I replaced my 4R7 resistors with wire-wound equivalents and now the converters work properly.