Another Aliexpress special, the AC DC eliminator

[5U4GB]

Specifically it's a DC Elimination Power Supply AC DC Blocking Power Supply Board.  Not sure why they go to all that effort when this will do the same thing much more efficiently.

[Someone]

Not the most insane thing in the world:
https://sound-au.com/articles/xfmr-dc.htm

[Circlotron]

I wonder how one of those things would go if put in series with a fluorescent tube. As the tube ages, and particularly toward end of life, one of the cathodes often gets less emissive than the other, causing both a flicker and asymmetrical conduction because of a certain amount of half wave rectification. With one of these devices attached, the resulting DC component would put a DC offset across the tube and maybe brute force symmetrical conduction again. Or maybe make it worse. Not sure.

[Andy Chee]

I wonder how one of those things would go if put in series with a fluorescent tube.

An electronic fluorescent ballast would already contain such a DC blocking capacitor, though given their high-frequency operation, the capacitor can be much smaller.

But in a traditional magnetic ballast circuit, I suspect that aliexpress capacitor will form a tank circuit with the magnetic ballast, resonate to high voltages, and destroy the tube.

[johansen]

I wonder how one of those things would go if put in series with a fluorescent tube.

An electronic fluorescent ballast would already contain such a DC blocking capacitor, though given their high-frequency operation, the capacitor can be much smaller.

But in a traditional magnetic ballast circuit, I suspect that aliexpress capacitor will form a tank circuit with the magnetic ballast, resonate to high voltages, and destroy the tube.

No to both of those.

The capacitance needed to allow for a 1vac (dc built up in two separate dc caps) voltage drop under load is of such magnitude that there is no way to make it resonate at anything more than just a few hz, certainly not a less than 2 Q inductor of an old ballast.

Electronic ballasts are, these days, full power factor corrected ac to dc to ac converters, drawing better than .99 pf.

A capacitor in series with the input will simply drop the pf slightly and increase the current, having no effect on the tube.

[Andy Chee]

I refer to the attached electronic ballast circuit:



Note that C17 is effectively a DC blocking capacitor.  The value of C17 and L2 must resonate at the correct frequency for the tube to operate.

If the aliexpress capacitor is installed in a similar position in a magnetic ballast system, who knows what the resonant frequency is. 

Using the aliexpress value of 20mF and ballast inductance of 400uH, the resonant frequency is 56.3Hz.  Uncomfortably close to mains frequency.

I don't know what the inductance of a magnetic ballast is.

[Circlotron]

I refer to the attached electronic ballast circuit:

I was actually thinking of traditional iron cored ballasts with no electronics.

[Andy Chee]

I refer to the attached electronic ballast circuit:

I was actually thinking of traditional iron cored ballasts with no electronics.

Same principle applies.  The traditional iron ballast and aliexpress capacitor, would be the equivalent of L2 & C17 in my posted schematic.

If they resonate anywhere near mains frequency, you'll burn out the tube or ballast or both.

[Kleinstein]

The series capacitor in the DC block is huge to only drop a little, while the tube balast is made to drop at least 10s of V in a 110 V country and more like 50+ V with a 230 V design. So the resonance would be at a frequency much below the mains frequency. In addition the usually DC blcokers have additional diodes to limit the voltage at the capacitor - with too much voltage the diode(s) get active and would dampen a resonance. So at most (if really in a resonance) maybe 1-2 V in addition to mains and this should not be an issue, at least most of the time.

[SeanB]

I have used a capacitor as a current limit in a flourescent lamp circuit, and for small tubes it works well enough, though it was running the tube slightly high in current, so it did suffer from sputtering on the ends rather rapidly. However you also got them used as part of the ballast in lag lead use, where you use a single ballast for 2 tubes, and the capacitor provides a phase shift on the second tube, so that with the combination you get absolutely no flicker in the light output from the fixture, though unfortunately you have to use T12 tubes, as T8 have too high a strike voltage to operate like this, and you also need a grounded reflector close to the tubes to keep them lit reliably as they age, or use the special tubes with a starter strip on the glass that is used with a contact touching the caps, to provide an electric field to ensure reliable striking every cycle.

[johansen]

I refer to the attached electronic ballast circuit:

(Attachment Link)

Note that C17 is effectively a DC blocking capacitor.  The value of C17 and L2 must resonate at the correct frequency for the tube to operate.

If the aliexpress capacitor is installed in a similar position in a magnetic ballast system, who knows what the resonant frequency is. 

Using the aliexpress value of 20mF and ballast inductance of 400uH, the resonant frequency is 56.3Hz.  Uncomfortably close to mains frequency.

I don't know what the inductance of a magnetic ballast is.

Please examine the rest of the circuit...

[Gyro]

I have used a capacitor as a current limit in a flourescent lamp circuit, and for small tubes it works well enough, though it was running the tube slightly high in current, so it did suffer from sputtering on the ends rather rapidly. However you also got them used as part of the ballast in lag lead use, where you use a single ballast for 2 tubes, and the capacitor provides a phase shift on the second tube, so that with the combination you get absolutely no flicker in the light output from the fixture, though unfortunately you have to use T12 tubes, as T8 have too high a strike voltage to operate like this, and you also need a grounded reflector close to the tubes to keep them lit reliably as they age, or use the special tubes with a starter strip on the glass that is used with a contact touching the caps, to provide an electric field to ensure reliable striking every cycle.

I have an 8ft 125W florescent lamp in the garage that has a traditional inductor ballast and a capacitor in series. That possibly does form a tank circuit to improve striking voltage, I don't know. The configuration was in the old Philips lighting catalogue for long tubes. Circuits for other (shorter) tubes just had the series inductor, with the capacitor across the mains input for PF correction.


P.S. The starter strip tubes were also used for dimmable installations, where the heaters were powered from separate, well isolated, secondaries of a transformer. They were pretty rare though. Winding a length of fuse wire along the length of the tube and tying the ends to the end caps (or even a pin at one end!) was an old bodge for getting tubes to strike reliably when the supply was a bit on the low side.

[Circlotron]

I have an 8ft 125W florescent lamp in the garage that has a traditional inductor ballast and a capacitor in series. That possibly does form a tank circuit to improve striking voltage, I don't know. The configuration was in the old Philips lighting catalogue for long tubes. Circuits for other (shorter) tubes just had the series inductor, with the capacitor across the mains input for PF correction.

Maybe this? -> https://maker.pro/forums/threads/capacitor-in-series-with-fluorescent-ballast.55257/

[madires]

However you also got them used as part of the ballast in lag lead use, where you use a single ballast for 2 tubes, and the capacitor provides a phase shift on the second tube, so that with the combination you get absolutely no flicker in the light output from the fixture, though unfortunately you have to use T12 tubes, as T8 have too high a strike voltage to operate like this,

There are special starters for in-series setups, e.g. Osram ST151 or ST172.