Non-isolated AC / DC power supply 0.5W with a long service life may be LinkSwitch

[Jadeit]

I need a power supply that would supply about 0.5W with solid efficiency and last for several decades.
I look at a non-isolation source with LinkSwitch-TNZ and I have a few questions.


1. Fusible resistor.
The constructor used a 8,2 Ohm 1W fusible resistor, it has two functions
- protects capacitor C6 and "circuit breaker" from high current when the power supply is switched on
- in the event of a failure, it should act as a fuse and burn out

What's wrong with me?
It seems too big for both dimensions and insurance current.
If you look at the datasheet, the resistor will burn only at a power dissipation at a 16W resistor and size is 100mm x 3,5mm.
Any other more suitable and smaller solutions?
a) use 8.2 ohm SMD 0603 resistor, burn at lower current, the flame that it would bother me
b) use 8.2 ohm SMD 0603 resistor and SMD fuse 70mA, I don't like it, it's two parts
c) anything else?


2. Zero cross detector
Z1 and Z2 pins are configured to provide a lossless (<5 mW) zero
crossing detection (ZCD) circuit
If I understand correctly, ZCD is a separate function block inside the LinkSwitch-TNZ IC and is not required for the IC function.
In other words, if I don't need a ZCD, can I omit the components crossed out in orange?

3. Feedback?
The FB description is somewhat ambiguous, plus in the DER-874 there are errors in the marking of the components (the figure and the description replace the resistors)
If I come up with the plug in the picture, the resistor divider R5 / R4 should keep the FB at 2V at the output voltage 6V. In DEE
This will only happen if the current to the FB is about 39uA , in DER874 is Resistors R3 (probably thinks R5) and R4 divides the voltage such that
the FB pin is set to 2 V. When the current delivered into this pin exceeds IFB (49 μA), a
low logic level (disable) is generated at the output of the feedback circuit.
If I want an output voltage such as 4V use R5 44.2k R4 320k ?

 

[KlausST]

Hi

you can´t expect a tiny standard 0603 size resistor "burns open" while it can withstand 300V AC.

There is a good chance that it "burns" but the spark/fire stays on. The heat will burn the PCB, overheat and even melt the plastic case. And there is a good risk for a big fire. (PCB -> plastic case -> house...)

I strongly recommend any device that is specified as fuse for 300V AC. Maybe you have a sealed metal case with limited oxygen inside.

For sure there is no guarantee that the "worst case" happens. But the designer is responsible ...

Klaus

 

[carpenter]

I'm not sure if anyone expects 300V endurance.
The only way to have a 300V resistance at 8.2 Ohm is the current over 36A.
The problem is elsewhere
Small SMD resistors have such a small distance between the ends of the resistor that an arc can burn between them even if the resistor itself burns out.
I use resistors 0603 as a lighter in electric igniters
Test experience says.
Pure metallized 0603 burns and does not create a permanent arc
But a layer of dust or grease is enough and the arc already tends to burn in these impurities

Therefore, the fuse should be so large that it cannot occur even in dirt.

3. Feedback?
If I want an output voltage such as 4V use R5 44.2k R4 320k ?
No for output voltge 4V wiil be for R5=44.2 calculated R4=21k and R3 NA becouse for output voltage under 5,2V is used internel 5V Regulator.
PI offers PIExpert software to design its resources, it will calculate the values of the components.

 

[KlausST]

Hi,

I'm not sure if anyone expects 300V endurance.
The only way to have a 300V resistance at 8.2 Ohm is the current over 36A.

I did not mean that there is 300V across the resistor during normal operation.

But the member wants to use the resistor instead of a fuse. And when a fuse is "open" then the full input voltage is across it´s leads. And the input voltage is given with 90 ... 300V AC.

A fuse needs to comply safety regulations. It nees to burn without damaging parts nearby. And it needs to suppress arcing (within specified operating voltage range).

A usual resistor when burns may overheat parts nerby. The PCB is the most closest (unless it is additionally coated somehow). High temperature may carbonize the PCB and carbon is conductive. So if the distance is short and the voltage is high enough, then the "carbon current" will be enough to keep temperature high ... even if there is no arc anymore.

In worst case a standard resistor used as a fuse does not "improve" safety, it may even cause fire.

My recommendation in short: Don´t do it.

Klaus

 

[mtwieg]

What's wrong with me?
It seems too big for both dimensions and insurance current.
If you look at the datasheet, the resistor will burn only at a power dissipation at a 16W resistor and size is 100mm x 3,5mm.
Any other more suitable and smaller solutions?
a) use 8.2 ohm SMD 0603 resistor, burn at lower current, the flame that it would bother me
b) use 8.2 ohm SMD 0603 resistor and SMD fuse 70mA, I don't like it, it's two parts
c) anything else?

It's possible that you might find a smaller substitute fuse to replace that fusible resistor, but you still have to use a component which is certified for use as a fuse (look for the appropriate UL or IEC standard). For example something like these.

But there's a lot of other factors to consider when picking a fuse, such as its speed. If this is for a commercial product, you need to consult the appropriate certification body/test lab.