I've noticed a trend of capacitive dropper supplies not lasting.
This supply is super simple: you use a capacitor in series with the load to limit the current and voltage the device can consume.
I have fixed several such power supplies now where the only problem is that the safety rated X2 capacitor used for the implementation has dropped in capacitance.
This of course starves the circuit of power:
In many implementations there is a relay being driven, and thus when the load of the relays coil is added the device simply turns off or rather resets.
I have a question regarding the safety of actually using a X2 rated capacitor to begin with in such an implementation.
X2 caps may fail shorted, that would mean the little MCU or whatever the dropper is powering sees full line voltage, 230VAC, not cool!
The obvious solution is to use Y2 capacitors instead, but of course they are more expensive and not available in as large capacitance values generally.
Maybe that problem is academic at best, I don't suppose using inrush limiting with an NTC would help either.
If they can even be found in so small values, since these devices are generally never switched off.
But then any type of non-polarized capacitor could be used, if we anyway do not care about the safety of the failure mode...
So the problem is the capacitor itself, the "self healing" properties they have mostly just means when the layers get damaged they automatically burn off an area in a safe way:
A better name for them would be capacitors with a controlled failure mode.
I have even seen some datasheets now specifically mention to not use certain X2 capacitors for capacitive dropper implementations.
EPCOS / TDK has a new model that is branded as ok for this use: B3293 (305VAC) and also an older model called B3292*H/J (305VAC)
They also have models B3265* MKP and B3267*L MKP but they are without safety approvals.
Vishay has models F1772 and F339X1
Do you know of any others? I got tired of looking myself...
Or maybe some other solution to make the capacitors last longer?
This supply is super simple: you use a capacitor in series with the load to limit the current and voltage the device can consume.
I have fixed several such power supplies now where the only problem is that the safety rated X2 capacitor used for the implementation has dropped in capacitance.
This of course starves the circuit of power:
In many implementations there is a relay being driven, and thus when the load of the relays coil is added the device simply turns off or rather resets.
I have a question regarding the safety of actually using a X2 rated capacitor to begin with in such an implementation.
X2 caps may fail shorted, that would mean the little MCU or whatever the dropper is powering sees full line voltage, 230VAC, not cool!
The obvious solution is to use Y2 capacitors instead, but of course they are more expensive and not available in as large capacitance values generally.
Maybe that problem is academic at best, I don't suppose using inrush limiting with an NTC would help either.
If they can even be found in so small values, since these devices are generally never switched off.
But then any type of non-polarized capacitor could be used, if we anyway do not care about the safety of the failure mode...
So the problem is the capacitor itself, the "self healing" properties they have mostly just means when the layers get damaged they automatically burn off an area in a safe way:
A better name for them would be capacitors with a controlled failure mode.
I have even seen some datasheets now specifically mention to not use certain X2 capacitors for capacitive dropper implementations.
EPCOS / TDK has a new model that is branded as ok for this use: B3293 (305VAC) and also an older model called B3292*H/J (305VAC)
They also have models B3265* MKP and B3267*L MKP but they are without safety approvals.
Vishay has models F1772 and F339X1
Do you know of any others? I got tired of looking myself...
Or maybe some other solution to make the capacitors last longer?
Comment