A while back, in one of the rooms I was sitting in at the uni, there was a flickering 2 x 36w fluorescent fitting. This particular fitting was troublesome, as contractors had come and gone and relamped the unit, only for it to continue flickering away.
One day, another set of contractors came, took the fitting down, removed the ballast and replaced it. All was well, and the flickering stopped. The contractors discarded the ballast in a nearby office waste bin, and left to their next job, satisfied that they had fixed the problem.
The sight of the ballast sitting in the bin caught my eye, so I decided to snatch it and give it a temporary home. Call me a hoarder, but I just wanted to learn more about it.
The unit in question was not a cheap iron choke style ballast using starters, but was a German made instant start electronic ballast. It was very light, which is one advantage! The unit comes from Vossloh Schwabe, a vendor which even GE recommends for some of their fluorescent tubes. The unit had a model number of ELXe 236.523, and was designed for use with T8 or TC-L tubes in a twin configuration.
The allure of electronic ballasts include higher lamp life, less perceivable flicker and more efficiency. This can be seen on the nameplate, where the ballast provides:
- 60w to 2 x 30w linear tubes
- 64w to 2 x 36w linear tubes (8w saved)
- 68w to 2 x 38w linear tubes (8w saved)
- 62w to 2 x 36w bent tubes (10w saved)
- 80w to 2 x 40w bent tubes
This has to do with the way the ballast is designed, and to do with the fact that tubes run at higher frequencies than the standard mains 50/60Hz produce more light for the same power.
The unit is designed for usage on 220-240v lines and can operate from DC or 50-60Hz supplies with a power factor of 0.98 (great) and case temperature of 70 degrees Celsius maximum. Ambient temperature can range from -20 to 50 degrees Celsius.
It doesn’t seem Vossloh Schwabe have any literature for this model anymore, although the ELXe series is summarized in their catalogue as an instant start ballast. It has an ignition time of about 0.5s by applying enough voltage to break down the gas and initiate the discharge. As this is harsh on the electrodes, it is designed for areas where lamps are turned on-and-off infrequently (five times a day). The service life is claimed to be 50,000 hours with failures <=0.2% per 1,000 operating hours.
This seems rather conservative, or even a little low, considering a tube is often claimed to have a service life of about 10,000 to 15,000 hours, it means the ballast would only see 3 to 5 lamp changes before its life had technically “expired”. At the end of the 50,000 hours, you would expect about 1 in 10 units to have failed. I suspect many of the units around the uni have fared better than that.
The insides look quite bare. The unit is made with a paper-type PCB as you would find in many power supplies. One thing you will notice is many magnetics within this unit – transformers and chokes are relatively numerous. I suspect there is an isolating transformer on the input, with an RFI suppression choke on the left side. There seems to be a transformer or a choke of some sort with a feedback winding nearer the middle, and components seem to be of high quality – with a Nippon Chemi-Con capacitor in the middle and International Rectifier ICs. The output transformer style configuration seems very similar to what you will find in a laptop LCD backlight inverter – and that’s pretty much expected because this unit seems to operate the lamps cold cathode style.
The underside is covered with solder resist and mostly surface mount resistors, diodes, capacitors and transistors with one bridge rectifier (to make high voltage DC from the input) and some unidentified IC on the underside. The design appears to be V03, designed in 2002 – over 10 years ago.
I got a little curious when I got home. Not seeing any obvious signs of failure internally, and no weak solder joints, I didn’t see any good reason why this unit wouldn’t be functional.
I scrapped together some insulated wire, and terminal connection blocks and directly connected to the pins on two 36w T8 fluoros I had left over.
Applying power, the ballast made a quick “chirp” and the lamps came on. Initially, the lamps exhibited some snaking, as the discharge was somewhat unstable as the lamps hadn’t warmed up yet. Once the lamps had warmed up – the output was stable, and all seemed to be well.
So why didn’t the unit fare very well? I think it could have been a loose wire. The connections to the ballast are made with push-fit connections which can oxidise or be made poorly.
Another reason could be the energy saving occupancy sensor it was connected to. The occupancy sensor looks at the room and tries to detect movement to reset a time-out. Once the sensor times out, it turns off the lamps to save energy. With very aggressive time-outs, it can result in high numbers of switching cycles within a single day. This could have led to the destruction of the electrodes which causes unstable discharge and flickering. So I suppose it makes sense to check your lighting configuration before you mindlessly connect it to occupancy sensors. The ballast seems to have fared perfectly fine though.
This has been the first electronic fluorescent ballast I had laid my hands on, and the quality of the components used in this German-made design seems to be quite evident. It is light, and more energy efficient than magnetic ballasts and produces better quality light with less flickering once warmed up. The ballast itself seems to be perfectly functional under testing, and the rapid-start nature seems to be hard on fluorescent tubes, as it operates them like cold-cathode tubes. It contained much fewer components than I had expected.