CFL cap not enough ripple current rating.

rating of capacitor used in cfl

hello

I have just taken apart a (240V rms, 50Hz) 15W Compact fluorescent bulb and done tests on it. The Input DC Bus has 110V of ripple on it (DC Bus goes from 220V to 330V with 10ms period.

It's DC bus (following the diode bridge) has one 3.3uF , 400V electrolytic capacitor on it. It is Aishi GA series cap:-

**broken link removed**

This cap has 99mA of ripple current capability.

However since the DC bus has 110V of ripple on it this means that the RMS charging current alone into this cap is 82mA......The discharging current will add to that, definetely taking the total RMS capacitor current to above 100mA.

This means that this cap is not properly rated for its ripple current.

Do you know why they have put in a cap thats not properly rated ?

Won't it fail prematurely ?

(Strangley, despite the bulb's packet saying "15W", .....on the bulb itself it is written "230V, 144mA"....which is 33 Watts )

cfl voltage rating

The capacitor usually fails. It is trapped in a hot enclosure. They have it spaced as far away from the hot light tube as possible but it still fails.

The circuit has a poor power factor so its apparent power is much higher than its real power. A home power meter reads only the real power.

capacitor ripple current rms

The capacitor isn't operated above it's maximum rating. I see from your description "the RMS charging current alone" that you apparently don't know how to determine the true RMS current. Assuming a load current proportional to bus voltage, I get about 93 mA ripple current. That is still rather high, so I agree with the assumption, that this part may fail first, but it's within specifications.

As another point, the small capacitance/high ripple voltage is simply a means to improve the power factor. It's still poor with about 0.63, but would be worse with a larger cap. Cause you most likely already heard of the term apparent power and understand it's meaning, you shouldn't be surprized about the current specification. The said 0.144 is however a maximum value and shouldn't be observed in normal operation, 0.11A is the expectable value.

what does capacitor in cfl do

Audioguru, thankyou for your replying
FvM:- Thankyou i appreciate your reply.....

please may i speak again regarding this capacitors rms current....as this is an area of mystic qualities to myself.

Looking at the capacitors voltage waveform, and considering the bit where it goes from 220V to 330V (charging) i calculated the charging current during this interval using i = Cdv/dt

...i then integrated the square of this over pi radians and took the root of the answer to give the RMS of the charging current.

Is this an incorrect methodology to calculate the RMS value of the charging current into this cap ?

When the cap voltage gets to about 330V (the peak), the cap is then discharged by the switching converter which also exists in this bulb.

...This discharge current will then add to the charging current, to give the proper full value of capacitor RMS current, which i am taking as the ripple current experienced by this cap.

.....As i stated above, the charging current alone comes out at 82mA.

...So when i include the discharging current as well it will surely take the total over 100mA,(?) and violate the ripple current rating of this cap ?

Added after 1 hours 44 minutes:

sorry to reply again...

in fact , i would be most grateful if readers could peruse these waveforms representative of the cap voltage and current

This image shows what this caps voltage (green) and current (blue) would look like...



This shows roughly what the capacitor voltage looks like on its own ........



This shows what the capacitor current looks like on its own....



Regarding the current waveform, due to its complex from, is it true that the only way to calculate its RMS value would be to do a piecewise integration ?

best ripple current capacitor

I see a problem in determining the switch frequent share of the capacitor current. I assumed (as previously said), that
the switch frequent component is not directly loading the electrolytic capacitor. If it actually does, the RMS current is
higher in fact, but I see no chance to estimate it from your measurement.