Capacitive Power Supply

[gauravkothari23]

Hi all,

I am designing a capacitive power supply to drive the 2835 1 Watts 6V x 6 LED's. (Circuit Attached)

For voltage dropper from 220V AC, I am using 0.1uF 1206 Ceramic x 3 = 0.3uF Capacitor(C1, C2, C3).
But i am bit confused what has to be the voltage of that capacitor.

As i have tested with 1206 SMD 50V capacitor 3 in parallel as shown in circuit, it works well, but can i use it for my applications to drive the 1W x 6 LED's, or the capacitor has to be higher voltage. I have to use SMD because i dont have much space on the PCB to use bigger one or through hole capacitor.

 

[Easy peasy]

630VDC rated - else it will go pop eventually ...

--- Updated Oct 14, 2022 ---

R1 & R3 should be 33 ohm each then they will have the same watts - and you can use the same part in both.

I see you are going for ~ 20mA ...

 

[FvM]

But i am bit confused what has to be the voltage of that capacitor

I don't see a source of confusion, quite obviously 220 VAC is dropped at the capacitor. 250 VAC rating required, 630 VDC typically works, I would check AC rating though.

--- Updated Oct 14, 2022 ---

630VDC X7R capacitors are typically not suited for 230 VAC operation, expect capacitance drop to 50% due to nonlinear characteristic. There are a few dedicated AC ceramic capacitors, e.g. https://content.kemet.com/datasheets/KEM_C1099_250_VAC_SMD.pdf

 

[gauravkothari23]

I don't see a source of confusion, quite obviously 220 VAC is dropped at the capacitor. 250 VAC rating required, 630 VDC typically works, I would check AC rating though.

--- Updated Oct 14, 2022 ---

630VDC X7R capacitors are typically not suited for 230 VAC operation, expect capacitance drop to 50% due to nonlinear characteristic. There are a few dedicated AC ceramic capacitors, e.g. https://content.kemet.com/datasheets/KEM_C1099_250_VAC_SMD.pdf

Thanks for the datasheets.
But why is it that as per the circuit, C1,C2 and C3 which are 0.1uF/50V dont blow even when it is connected to 230V AC, even i have checked the temperature it quite normal and still at room temperature after continues operation of 4 to 5 hours,

 

[FvM]

Yes, can happen. May be it's exploding after 6 hours, or 6 days.

Did you measure the current in operation? Effective capacitance will be unlikely near to 100 nF at this voltage level.

Also consider that grid connected devices need to withstand 1500V surges in the lowest overvoltage category.

 

[gauravkothari23]

Yes, can happen. May be it's exploding after 6 hours, or 6 days.

Did you measure the current in operation? Effective capacitance will be unlikely near to 100 nF at this voltage level.

Also consider that grid connected devices need to withstand 1500V surges in the lowest overvoltage category.

So, can i use 3 capacitor of 0.3uF/100V 1210 package in series to make 0.3uF/300V. The problem in INDIA, its extreme difficult to get SMD ceramic Capacitor greater then 100V. upto 100V MLCC are easily available.

 

[KlausST]

Hi,

3 capacitor of 0.3uF/100V 1210 package in series to make 0.3uF/300V.

3 capacitors of 0.3uF each in series gives 0.1 uF in total not 0.3uF.

Klaus

 

[gauravkothari23]

Hi,

3 capacitors of 0.3uF each in series gives 0.1 uF in total not 0.3uF.

Klaus

ok, sorry, i have to use 1uF/100V capacitor 3 in series to get 0.33uF/300V.
Right..?

 

[KlausST]

Hi,

correct.
But mind to use some high ohmic resistors in parallel to each capacitor avoid DC runaway.

Klaus

 

[Easy peasy]

Even 300V will fail at the first mains surge - for a product that is to be sold - do you really want this ?

Any ripple frequencies on the mains will cause higher currents, which may over-heat and kill the safety resistors - or the caps themselves - we often see this type of fault.

Also a fuse or proper fusible resistor a good idea

 

[gauravkothari23]

Even 300V will fail at the first mains surge - for a product that is to be sold - do you really want this ?

Any ripple frequencies on the mains will cause higher currents, which may over-heat and kill the safety resistors - or the caps themselves - we often see this type of fault.

Also a fuse or proper fusible resistor a good idea

Thanks, But yes, definitely i am planning to add a 500mA of fuse.

--- Updated Oct 15, 2022 ---

Recently i have done one more testing for the SMD 1206 Capacitors which i am using in my circuit.
i have connected the capacitor 0.1uF/50V straight to 220V AC as a load without any additional load. Nothing happened to the capacitor. i was expecting it might blow. but i kept it for 30 minutes, it still running and not eve n getting hot. and later ON i removed the capacitor to check if has gone OPEN, but no, have tested the capacitor in LCR meter, it shows proper value.
why is it not bowing.

 

[Easy peasy]

pure luck, also the capacitance is very small at higher voltages - however if you leave it on for a week it will be dead ...

 

[cupoftea]

here is a cap dropper in LTspice..you can play with it to see your situation. (change it to yours)
Sorry, i think you made mistake...you cant put 50v MLCC on mains and have it survive.
Must have been higher rated cap, or you didnt connect it properly.

...mind you , we had cracked 630V, 100n MLCCs get attached to rectified mains...and they went bang....but even after that , they still showed 100nF!!! (they were cracked by rough depanelisation, and also rough assemblers screwing the pcb to a heatsink and the PCB and cap getting bending stressed).

What it is C = E0.A/d.....and you can blow up quite a bit of it, and still have A/d ratio the same, so it still shows 100nF (or whatever)