Teardown: Philips Tornado 24W Spiral ES CFL

I’m technically on hiatus, and I am piled up with work, but some days I just find it hard to go without doing something to break the monotony, so I’m back with a few seemingly random posts.

We kick off with a teardown of a Philips Tornado 24W spiral CFL globe, which recently failed a few days ago. I was quite a fan of these globes as they seemed to be of good quality, were easily available in local supermarkets, and had the highest power rating for the physical size. They pretty much start instantly, and warm up quite quickly despite the thin “long” tube geometry. Generally, quality CFLs over 20W are hard to find, and even more difficult to fit into oyster fixtures, so we did end up recommending these to others and buying a second lot which is only just on its way out after … gosh, I don’t know how many years!

20151001-2217-5691 20151001-2218-5692

The base itself is compact, but so is the spiral shaped globe, which makes it a good fit for more restrictive luminaries. There is a warning not to use them in enclosed holders due to heat build-up, but we’ve seen no issues ourselves. If anything, this globe was well due for replacement. The globe has a rating of 1550 lumens, and an efficacy of 67 lumens per watt – some currently available LED globes claim similar efficacy ratings.

Getting this globe open wasn’t too difficult – many years of high temperature operation left the plastic brittle, and a little prying around the lips was enough to separate the body.

20151001-2220-5693

A quick post-mortem check-up seems to show both filaments are still electrically conductive – i.e. they are not open circuit. They both measure identically at 2.5 ohms.

20151001-2223-5695 20151001-2223-5694

This leads to the logical question – what caused the lamp to fail? A closer look at one of the ends seems to show significant heating, and a change in the end “seal” looks like it had cracked at one stage due to heat.

20151001-2224-5696

This suggests to me that the globe failed due to age-related failure – a depletion of thermionic emission mix at the filament would have lead to increasing voltage drop across the tube, which probably would have led the ballast to apply more current to the filaments to try and “pre-heat” them, which eventually may have caused the end to heat up enough that the glass cracked and the tube itself became “open to the atmosphere”. I’ve seen this sort of failure mode on many lamps which are at their end of life, and often they will “strobe” slightly in the final few hours as the ballast switches between running and “warming” the filaments, before finally giving out in a “swirl” of light, sometimes with a pop and small cloud of smoke.

But can we find further evidence to support this theory? Actually, it seems like we can. An examination of the rest of the globe found the fusible resistor (about 4.7 ohms) was still intact, meaning the ballast was powered. A look at the ballast circuitry doesn’t reveal any components that look like they’ve failed. Ballast failure is normally catastrophic, and obvious – the quiet lack of functioning may simply be due to the tube itself being filled with air, and hence an arc cannot be struck due to the insulative properties of air.

20151001-2230-5697 20151001-2230-5699 20151001-2230-5700 20151001-2230-5698

It’s interesting to note the presence of a Aishi capacitor, which seems quite popular in Philips globes. Despite the abuse, it measured just as expected on the LCR meter, whereas cheaper globes have occasionally been found to have vented capacitors which eventually do short out and blow a primary side fuse. The main transistors are marked EKA F6 13003, which seems likely they are equivalent to the Fairchild KSE13003 NPN high voltage silicon transistor. Sadly, no understandable date codes were found.

Of note is that this PCB aims to reduce the volume of the ballast section. It does so by being cleverly “cut out” where the tube ends would be, allowing it to nestle in-between. The capacitor is also made to stand proud of the board so that it can sit into the “cap” at the end. All of this helps to reduce the Z-height, and allow the globe to fit in more compact spots.

20151001-2230-5701

Another aid is the use of SMD components on the rear. On the left, we can see the input diode bridge configuration, for example.

We also make some surprising discoveries – the markings on the board made from breaks in the solder resist appear to say TP220-25MSL(T3) V1.0 and 134397. From the code itself, it seems likely that this globe was OEMed for Philips by Hengdian Group Tospo Lighting Co. Ltd. and was originally a 25W globe design. You wouldn’t even know if you didn’t take it apart. The product code isn’t on their website anymore, although a gander through the Archive.org archives seems to result in frustration due to incomplete/broken website structures. Regardless, it is noted that all their globes use TP as an identifier – so I think I’m right in assuming this is a Tospo product.

The globe has since been replaced with an LED unit instead, in a slow and steady march to higher efficiencies and lower maintenance.

About lui_gough

I’m a bit of a nut for electronics, computing, photography, radio, satellite and other technical hobbies. Click for more about me!

This entry was posted in Electronics, Lighting and tagged , , , . Bookmark the permalink.

2 Responses to Teardown: Philips Tornado 24W Spiral ES CFL

  1. jitter says:

    You wrote that you didn’t find any understandable datecodes, but there is one I understand.
    In the lower right corner of the label it says “C8”.

    A long time ago I was explained by a Philips employee that their lamps carry a letter and a numeral as datecode. The letter indicates the month and the numeral the last digit of the year.
    Every decade the year and the month are swapped around.

    This decade, it’s year first, e.g. 1A for January 2011. That puts the production datecode of your lamp, C8, in March 2008.

    Curiously, I have two 15 W Tornados that uses an un-Philips-like datecode in the YYWW style.

Error: Comment is Missing!