Another long weekend, so a good excuse to add some more posts. This post has a double serving of failed Philips CFLs – a Genie 14W and an old style Tornado 20W. These CFLs were used by the strata for the common area lighting, and wandering out on a mail-check run, I bumped into the contracted electrician servicing the lighting. After asking what he did with the dead CFLs, he gave the expected answer that he just turfed them. I asked him rather nicely if I could have them because I like to tear things apart and blog about them, he was comfortable with giving me three different types. Because one of the three was still fully functioning, we will be focusing on the two failed units.
Philips Genie 14W
The Genie is a CFL with three connected U-shaped tubes. Previously, I had torn down a Philips Genie 11W, but that one had a slightly different body shape that had less ridges. It seems that the Genie line may be made of a few different designs, so I decided to tear this one down.
This unit has a cool daylight colour temperature and a luminous efficacy of 56 lumens/watt. This is relatively good, given some lower end LED bulbs today are at this level. The date code appears to be 0E, so it might be May 2000.
When power is applied, nothing happens, although the globe does draw a small amount of power. This implies the fuse hasn’t failed, and so a catastrophic failure did not present itself.
As with Philips globes, the single-sided ballast PCB is “clipped” into the plastic base and the lamp electrodes are wire-wrapped to pins on the PCB.
Already, we can spot an issue – the Aishi 3.3uF 400V capacitor has vented and failed. Surprisingly, even in this condition, the amount of goop vented was relatively limited, and even on application of line voltage, the capacitor didn’t “arc over” and short out.
The circuit doesn’t appear to be particularly different, but the PCB definitely is different compared to the other Genie I tore down. The fuse is mounted in-line with the incoming life lead. This single-sided board has some surface mount components on the underside, but they are a bit crooked. Because of the outdoor usage of the globes, there is some dead insect residue as well.
A slight melting can be seen on the side of the capacitor’s heatshrink. It seems this may be when the active lead was soldered to the Edison screw cap and the heat travelling up the lead may have caused that melting.
There didn’t seem to be many markings on the PCB which identifies who manufactured the PCB or the assembly.
The two filaments were still intact at the end, but their resistance seems to have gone up to fairly high levels.
The capacitor only registered 6.183nF of a 3.3uF rating – so about 1/500th of its original rating.
It seems that this globe probably failed almost simultaneously in two ways – filaments appear to have been worn out over time, while the capacitor also failed. It seems likely that the capacitor may have failed first – whether this contributed to earlier degradation of the filament is not known. There is slight darkening of the filaments, and the increased resistance seems to suggest the filaments were on their way out anyway.
Simultaneous failures are a good thing – it shows that they didn’t over-engineer any one component too far so as to be “wasted”.
Philips Tornado 20W
This seems to be an “original” Tornado. This Tornado is nothing like the newer slimline sort, which I had torn down a 24W version earlier, and has a body more similar to the Genie 11W I tore down earlier. It has a warm white colour temperature and achieves 70 lumens/watt – well into the same territory occupied by some present-day retrofit LED bulbs. The date code A7 decodes to January 2007, making it 9 years old.
The PCB is clipped in again, making it quite similar to the Genie above, but the fusible resistor is not in-line with the active – instead, it is mounted on the PCB. The Aishi capacitor is just fine as well.
Another single sided PCB, but this time, there are no surface mount components on the underside. There seems to be some text in the solder resist which says VA10-1E, I think.
On the whole, the circuit looks just fine, and it probably was, because it appears this one failed because one of the filaments went open. Maybe it was a circuit failure causing it to burn out the filament, but it seems unlikely because that might blow the fusible resistor instead. Maybe it’s just a freak filament failure, or the tube itself may have opened to the atmosphere.
I didn’t test for ESR or working voltage, although that could be where it fails.
It’s still consist with my experience in regards to Aishi capacitors – they may be Chinese, but they’re almost always fault-less. This post was the first instance I’ve seen an Aishi vent!