Return to the thumbnail page Display/hide file information See previous file

Circuit to add EOL protection in self-oscillating ballast

Circuit to add EOL protection in self-oscillating ballast

Click to view full size image

About 8 years ago I've made many experiment on electronic self-oscillating ballast found in cfls. Two main problems I had when I was reusing ballast of EOL cfl for other fluorescent lamps:
1- The absence of an EOL protection caused many ballast to fail when the lamps go EOL or when the lamps is removed and the ballast continue to oscillate (due for example to the capacitance of the wires).
2- The ptc resistor used for preheating is matched with the cfl tube, if we connect a different tube(a T8 for example) very often the preheating current is not enought and the lamp starts in cold mode.

The image shows a circuit that can be added to add an EOL protection. The ballast circuits it's from an osram circolux, the EOL protection that I added is the part included in the blue line.
The EOL circuit measure the voltage present on the tube, after the current regulation inductor L4.
R7 and R9 form a voltage divider and can be adjusted to regulate the voltage that will trigger the protection. For example a T8 18W tube will run at about 50V, when it starts to rectify (one end starts to burn out) the voltage raise over 70-80V so the threshold can be adjusted @60-70V.
R11 and C6 adds a delay to avoid the trigger of the EOL protection when the tube is ignited.
Z2 is a Zener diode that help to avoid false triggering of the EOL protection, it's value will also set the trigger voltage with R7-R9.
T3 is the "switch" that shut down the self oscillating circuit by interrupting the base of the the bottom transistor of the self oscillating circuit.
R13, R6 Z1 and C9 create the voltage supply that switch on the T3 mosfet switch, the supply is (+U) is taked after the tube catode, so the ballast cannot run without a tube and changing the tube will reset the EOL.
R10, T4 and T5 form a latch: When the EOL protection is triggered the voltage of the T3 switch is kept low until the power is cycled or the tube is changed. Again C2 and C8 help to avoid false triggering of the EOL protection.

I've implemented this circuit with a few different ballast and it's work very well. I've made many try to run EOL tubes and the protection is triggered quite instantly or within a few seconds depending of the "grade" of EOL of a tube.

osramdulux.jpg youtube_MAArf1G1QVA.jpg STARTE2MOSSCH.JPG CIRCOLUX_CON_PROTEZIONE.JPG

Light Information

Light Information

Manufacturer:Kappa7 :-)

File information

File information

Download: Download this File
Filename:CIRCOLUX_CON_PROTEZIONE.JPG
Album name:Kappa7 / Electronic ballast circuits
Keywords:Gear
File Size:110 KB
Date added:Mar 19, 2011
Dimensions:899 x 623 pixels
Displayed:989 times
URL:https://www.lighting-gallery.net/gallery/displayimage.php?pos=-52810
Favorites:Add to Favorites
Comments
Medved
Hero Member
*****
Offline

Gender: Male
View Posts
View Gallery

View Profile Personal Message (Offline)
Mar 19, 2011 at 05:07 PM Author: Medved
The only drawback is, then the circuit loose it's simplicity...
Instead of voltage it may be tripped by current (drop on R12). With this and reconnecting the C4 to the "cold" side of the tube the circuit could be made even protected against shorts of lamp wires to grounded metal. As the inverter would form a kind of voltage doubler, it would trip the overcurrent protection, so stop the oscillation. Then the "doubler" effect would be very current limited by quite high impedance of C4 at 50Hz.

Coincidently i recently had another idea:
- Add two diode "halfbridge" (two series diodes, cathode of the top one to the )
- Split the resonance capacitor to two series connected, connect the middle point to the central point of the added diode halfbridge.
Splitted resonance capacitor would divide the lamp voltage. If the output Vpp reach the DC bus voltage, it start to be rectified and return the energy back to the DC bus while effectively limiting the voltage (the value is then controlled by the capacitor's ratio), so resonance current.
Up to now i only tried it in PSPICE (transistor power dissipation never exceed 2x the normal one for arc voltages from 50 to 450V), but dropped the idea as i wanted double lamp independent operation, what require or fixed frequency (or two complete inverters), what this circuit does not fulfill.
Because it is not the shut-down, such circuit should work to "finish off" dead lamps when filaments shorted (as it is popular here... )

Update: Meantime i upload the simulation schematic used to check that idea...

No more selfballasted c***

Kappa7
Jr. Member
**
Offline

Gender: Male
View Posts
View Gallery

View Profile Personal Message (Offline)
Mar 19, 2011 at 06:02 PM Author: Kappa7
Yes it loose it's simplicity but all the components needed are relatively cheap (the transitors can be low voltage).
I think that the current measurement should also work, but I never tried. Move C4 on the cold side will protect against shorts, but will loose the autorestart when the tube is changed.

I like your idea, but what's happen if the lamp is removed? The ballast will continue to oscillate and outputs the maximum voltage(without blown)?
Medved
Hero Member
*****
Offline

Gender: Male
View Posts
View Gallery

View Profile Personal Message (Offline)
Mar 19, 2011 at 06:26 PM Author: Medved
Yes. As the voltage is under control, there is no excessive stress on any component. Or at least this is the idea, up to now confirmed only by the simulator. I never tried it in real life, as i finally wanted something a bit different (two independently operating lamps on common inverter, so i moved to IR2153-based design, again with the same idea; even that work in the simulator pretty well, so today i was already busy with PCB layout...)

And @"all components are cheap": These days not anymore (those controllers and usable FETs are getting quite cheap as well). On Farnell the IRS2530 (i think the best for simple programmed start without active PFC, as it regulate the arc current) is for about 3Eur...

@"...will loose the autorestart"
Not, if you reconnect the +V to the "hot" lamp side (i forgot to note this before). The restart current would then flow via R5, chokes and finally ballast filament.

No more selfballasted c***

© 2005-2020 Lighting-Gallery.net | Powered by: Coppermine Photo Gallery