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

Solid state vibrator replacement for fluorescents likee Browning Porta Light

Solid state vibrator replacement for fluorescents likee Browning Porta Light

Click to view full size image

The schematic shows the complete circuit for the vibrator replacement. The T1 and T2 are the main chopping switches, the T3 eliminates the reverse diodes of T1 and T2 when they are OFF, so allows the voltage to swing negative and so generate the required flyback voltage to operate the lamp.
So T1 and T2 are controlled by the pulses representing the ON time of the respective contacts in the mechanical vibrator, the T3 is controlled so it is ON when either one of the T1 and T2 is ON and OFF when both T! and T2 are OFF (therefore it has a separate driving signal).
As the signal source for the transistors is used some "smart transistor" microcontroller, as that yields the simplest hardware circuit, so minimum chance for some fault.
Beside the main inverter functionality it may monitor the voltage (for that you need a microcontroller version with an ADC) and determine if there isn't some fault:
Battery under/over voltage: Take voltage samples during the ON time
Open load detection: Take a burst of samples at the OFF time and monitor, whether the voltage get clamped by something else than the MOSFET avalanche breakdown (so when the 50V overshoot disappear sooner than it would correspond to the complete core demagnetization at the 50V).

In case an under or overvoltage, switch OFF all transistors and keep them OFF until processor reset (by cycling the power switch), as the manual preheat is necessary for the successful restart. To prevent false trippings, respond only when the over/undervoltage is present for longer time continuously (so a random glitch won't shut it down)

With the open load detected, prolong the OFF time to about half second or so and then continue to the next pulse. When after the next pulse the open load is not detected anymore, resume normal operation.
With a real open load it will result into generation of test pulses twice a second, where all the energy in the core will be dissipated by the T1 or T2. Because of the just 1Hz, the average power will be very low. And once someone pushes the preheat, after first pulse it detect normal load condition and so resumes normal operation.
So even with an eventual lamp extinguish, from an user perspective it will operate all the time, but still do not stress the components during the real fault.

EnergizerPocket.GIF F4T5BLB_pocket.GIF ElectronicVibratorForFluorescent.GIF VibratorBallast.GIF

Light Information

Light Information

Lamp
Lamp Type:Fluorescent
Fixture
Ballast Type:push pull low frequency flyback, preheat
Electrical
Wattage:up to 40W
Voltage:12VDC

File information

File information

Download: Download this File
Filename:ElectronicVibratorForFluorescent.GIF
Album name:Medved / Reverse engineering
Keywords:Gear
File Size:19 KB
Date added:Jan 18, 2015
Dimensions:953 x 534 pixels
Displayed:159 times
URL:https://www.lighting-gallery.net/gallery/displayimage.php?pos=-103270
Favorites:Add to Favorites
Comments
Medved
Hero Member
*****
Offline

Gender: Male
Posts: 4211
View Gallery

View Profile Personal Message (Offline)
Jan 18, 2015 at 07:13 AM Author: Medved
For the transistors: Any "5V logic drive" MOSFET, Vds rated at least 100V and with low enough RdsOn will work well there, an example form European Farnell (you may copy the selection criteria and so find similar list with any of your favorite distributor).
As the microcontroller I have specified an Atmel ATTiny (because these I know, how to program), but other microcontroller makers have their equivalent. If the protection features are not required (then T1 and T2 need good heatsink to accept the open load overvoltage abuse), just the simplest type without the ADC would be sufficient.
Otherwise with the protection features the transistors will run cold all the time, so need no heatsink.

No more selfballasted c***

refridgedude1841
Jr. Member
**
Offline

Posts: 79
View Gallery

View Profile Personal Message (Offline)
Jan 18, 2015 at 11:59 AM Author: refridgedude1841
Nice! Would you be willing to make these for a fee + cost of components? This likely would work in the select a volt I have too right? Or would an automatic starter cause issues?
Medved
Hero Member
*****
Offline

Gender: Male
Posts: 4211
View Gallery

View Profile Personal Message (Offline)
Jan 18, 2015 at 01:10 PM Author: Medved
It would have to be components + materials + something for me + postage. The latest would not be so negligible, as it is from Europe to the US...
And the payment will most likely go over PayPal (I hope it will work, I do not want to share my bank credentials...)

Making the HW would not be that big problem for me, writing the SW functionality neither. Where I see problem is tuning the parameters (mainly the supply voltage compensation and the open circuit protection), when I do not have anything to try it on (problematic is the transformer, some F15T8 lamp and a power source I do have)...
So for at least some chance of success, I would need you to measure quite some parameters:
- The vibrator activity time scheme (take an oscilloscope snapshot of the primary voltage when operating the lamp)
- Oscilloscope snapshots of the primary voltages during operation in all three modes: Open load, preheat and normal operation. Be quick with the first two, they do stress the vibrator.
- Transformer wire resistances (take an Ohm meter and measure both primary and secondary winding)
- Transformer ratio (connect the secondary to ~12VAC/60Hz and measure the voltage on primary; Then do the same reversed, but that time use about 3VAC to feed the primary and again measure all the voltages)
- Transformer main inductance: When the secondary connected to the 12VAC source, measure the current.
- Transformer leakage inductance: The same as before, but short circuit the primary. Do this measurement fast, so it does not heats up and so changes the wire resistance...

Then I may at least simulate it to some extend...

And if it will work with an automatic starter? With glowbottle I'm not sure, it will depend on the exact vibrator timing. With electronic, I doubt. They require unclamped series inductance as the ballast impedance and that would not be the case.
But adding filament windings (or an extra small transformer; parameters would depend on the exact timing), it may be converted to RS...

No more selfballasted c***

Kappa7
Jr. Member
**
Offline

Gender: Male
Posts: 58
View Gallery

View Profile Personal Message (Offline)
Jan 18, 2015 at 02:03 PM Author: Kappa7
You can't simply use a "normal" electronic vibrator with two schottky diode in series with the mosfets drain to allow the voltage to become negative during the off phase?
It would not be the best for an efficiency point of view but in any case all the circuit with a low frequency trasformer it's not very efficient.
Medved
Hero Member
*****
Offline

Gender: Male
Posts: 4211
View Gallery

View Profile Personal Message (Offline)
Jan 18, 2015 at 02:34 PM Author: Medved
You can, but then the open load condition would be way more complex to make (here it utilizes the fact, than between the "local ground" and the battery positive supply you have the mirror of the secondary voltage).
Otherwise the reasoning for this topology was:
The drive scheme with the "gap" would need either a more complex logic, or a microcontroller for simple HW.
Once we have the microcontroller, it becomes very easy to drive the 3'rd MOSFET, so gain the efficiency. The efficiency is not that important from the overall efficiency point of view (indeed, the transformer losses are huge), but it means extra power dissipation to deal with on otherwise non dissipating board.
And with that configuration and the fact the simplest micro available here is the type with the ADC, the addition of the diagnostic is just "a cherry on the cake"...

No more selfballasted c***

© 2005-2019 Lighting-Gallery.net | Powered by: Coppermine Photo Gallery