Every so-often, I drop into my local thrift shop because I just so happen to be passing by on another errand and have some time to spare. It’s often a good experience to sift through the random items for sale, and while the pricing is the least competitive it’s been in a long time, at least any money spent would be used in part for a good cause.
Among the things that catch my eye are the vintage electronics. This time, I came across a box of compact fluorescent tubes which were rather interesting to me. Priced at an attractive $5 for the box, I made the purchase and gave them a little more time before they reached their final destination (i.e. landfill).
A little CFL discussion and a Look at the Goods
At the moment, we are in the midst of a lighting revolution, and fluorescent tube lighting is on borrowed time. Regulations regarding energy efficiency are pursuing luminous efficacies unattainable with traditional fluorescent tubes, and hazardous material reduction legislation also looks unfavourably upon the mercury content within these tubes. As of last year, GE had already committed to shutting down its fluorescent tube manufacturing altogether. As a result, this is a timely find, especially when considering I did some reminiscing about fluorescent fixtures in an earlier post.
If you talk about compact fluorescent lamps (CFLs), the first thing that pops into most people’s minds (including the collective population of the internet) is that of the retrofit globes which screw/click into existing fittings, consisting of an electronic ballast and tube in a compact form.
But in reality, the term compact fluorescent lamp is broader than that. Technically speaking, the CFLs that are used in retrofit are better known as CFLi – the i standing for “integrated ballast/control gear”, but this term seems to have fallen out of popular usage. Instead, CFL merely means a fluorescent tube that is not in the traditional linear form, and is instead bent in a myriad of ways to reduce its size. Instead, now, such CFL tubes are often marketed as “Compact Fluorescent – Non Integrated” to avoid confusion.
The advent of CFLs were an important part of the history of the fluorescent lamp, as it made the high-efficiency technology more appealing to different use cases. This included higher power densities in smaller spaces to realize can-style downlights, as well as architectural “square” shapes enabled by 2D style tubes. In fact, there is a large array of CFL types and bases. The humble CFLi was a direct beneficiary of these developments, with some earlier CFLi bulbs carrying names such as PL (e.g. Philips PL) which reference the prismatic linear CFL tubes used to build them. I remember the earlier CFL bulbs also came in “separable” halves, where the electronic ballast had a PL base to accept a PL CFL tube and the standard Edison Screw/Bayonet Cap fixture at the other end to fit into a traditional fitting. Later bulbs integrated the two together for space savings, (likely) cost savings and reliability improvement (as the electronic ballasts had a limited lifetime as well).
While non-integrated CFLs have never been really popular in the majority of consumer fittings, they do see a certain amount of use in commercial fittings. I did see a good number of them even around the university I attended, it seems that the technology may still have a few years left. However, with the myriad of types, I suspect they will eventually be removed and replaced entirely as the tubes become harder to source and more costly, or the fittings begin to fail.
The tubes in question are Sylvania Lynx CFTE 42W 840 Long Life (278450LL) compact fluorescent tubes with a 4000K colour temperature, 3200 lumen output, 80-89 CRI and a GX24q-4 base. The luminous efficacy is about 76 lumens/watt, which is fairly impressive for fluorescent technology, but easily eclipsed by the best LEDs which reach 120-130lm/W. The units claim a 20,000 hour lifetime based on an 11 hours on/1 hour off cycle and are designed to be used on electronic ballasts only. The tubes are made in the UAE, although are an older tube, as presently available versions claim a 24,000 hour lifetime and use a blue base to distinguish the long life type.
The tubes use amalgam technology, where the mercury in the tube is made into a compound to regulate mercury vapour pressure and improve the high-temperature operating temperature characteristics of the tube, but as a consequence, the tubes have a significant warm-up time. This is necessary due to the high power density, as well as the significant bends in the tube.
The haul was a whole carton of 10, with a little water damage around the bottom, suggesting this was probably kept in a stock room somewhere and forgotten about. Then, possibly, the business closed and all the assets were sold, and this was part of the lot.
The reason I suspect this is because on the top of the package is a label for Mc Grath’s Hill Auctions, in Mulgrave. This suggests the whole package was probably auctioned, but didn’t find a bidder, and was then donated to the thrift shop by some commercial arrangement. Being in your average consumer thrift shop is probably one of the worst places for this sort of lamp – most consumers just wouldn’t have the fittings or know what to do with them! Eventually, after sitting on the shelf, it would have probably gone into the bin if it were not for me!
The EU energy efficiency scale and the icon with two dolphins on the front, suggests this globe is really not that old – possibly mid-2000’s by my reckoning. I suspect the dolphin logo is an indication of reduced mercury content, due to various requirements in different parts of the world with mercury doses reduced to no more than 5-6mg per tube.
The globe itself looks very similar to the picture on the box, featuring three interconnected U-shaped tubes in a delta shape, and a plastic base with four pin connections (to be suitable for electronic ballasts).
Markings are made on the body of the plastic base in blue print, with the ratings on one side, and the “LONG LIFE” indication marked on the opposite side. The rectangular base post has protrusions which allow for the socket to retain a “grip” on the lamp once seated.
The tubes are physically fairly long due to the 42W rating, so the tube seems to advertise a “crush protection” in the form of a silicone spacer which is pre-installed towards the end. This is in contrast to the consumer globes which often have glass hollows glued to the ends to maintain spacing, and may be a sign of high temperatures being a design issue.
Let there be Light? Maybe. Maybe not?
What good is having some tubes if we can’t make light with them? Well, rather sadly, these seem to have a base which wasn’t really that common, and generally speaking, a matching ballast is required to run them. Such fittings aren’t easy to come by at all. But there’s nothing wrong with a little improvising …
A while back, I salvaged a Vossloh Schwabe ELXe 236.523 T8/TC-L Electronic Ballast from the university and found it to be perfectly operational. Might this be a close enough match? Lets do some thinking.
The ballast is rated to supply:
- 60w to 2 x 30w linear tubes
- 64w to 2 x 36w linear tubes
- 68w to 2 x 38w linear tubes
- 62w to 2 x 36w bent tubes
- 80w to 2 x 40w bent tubes
While it’s not got any rating for a 42W “bent” lamp, it seems to give the full 40W to a 40W TC-L tube, so it is at least capable of a similar power. Unlike some other ballasts, this one seems to have some tube-voltage dependent power delivery – notice only the 40W bent tubes get the full rated power delivery.
Doing a little more digging, it seems 40w TC-L tubes operate around 126V so have a current of around 317mA which is practically where we need it (i.e. 320mA). But the 42W CFTE is expected to have a higher operating voltage of about 131V. It’s a minor difference, which the ballast should be able to handle (after all, as tubes age, the voltage drop increases), but we need to test to see just how much power will be delivered.
Even though it claims not to be useable on inductive control-gear, trying it with a common 4-foot 36W choke could be unwise because it may regulate the current somewhere near the 430mA nominal current which is about 34% above the 320mA tube rating. This might reduce slightly because the operating voltage of 131V is higher than that of the 4-foot 36W tube which operates around 106V, but how the ballast might react depends on its design. If I crudely assume an inductive ballast to be a resistor, at 230V it has an effective impedance of ~288.372 ohms. Thus the calculated run current of the 42W tube on the ballast is about 343mA or a slight overdrive of 7.2% which it could probably handle. This is likely also to affect the performance of any power-factor compensation as well. I could probably try it … but then again, I don’t feel like needlessly modifying an existing fitting of mine with potential destruction of the tubes. The electronic ballast was “free”.
The reasoning behind not supporting inductive control-gear despite the four pin base is likely to be that the filaments aren’t particularly robust against a starter-based ignition which is more variable in the pre-heating of the filaments. They might be over-heated quicker due to their more compact size, or fail much sooner as they aren’t designed for frequent cycling as each starter-based start-up is like having 5-or-so on/off cycles on the lamp. But that’s really just a guess on my part. Some of the issues can be easily mitigated by swapping over to an electronic starter instead, but these aren’t that easy to obtain. One thing that’s for sure is that you won’t get as much brightness for the same energy input on an inductive ballast because of the gas re-ionization losses due to the low operating frequency resulting in more “dead time” where the tube is not conducting.
That being said, my choice of ballast isn’t particularly suitable either as the ballast appears to be an instant start type. The ballast chirps and starts the lamps quickly, with no noticeable pre-heating. The connection to one filament is just a single wire, suggesting it does not do any pre-heating current, thus it’s not a programmed-start unit with pre-heating that would be gentle on the filaments. It will shorten the life of the lamps, but I’m just happy if I get them to light even a few times.
No base? No problem. A few terminal blocks, a few random-coloured wires of mains rating or higher, a knife to cut off a little of the shroud, and judicious use of a screwdriver and the connection to the lamps is sorted.
As I had already wired the ballast up prior for test purposes, I thought I’d “bling” it up by adding an in-line switch to the cord, and give the input terminals some strain relief using a cable tie to grab the cable near the earthing eye. This particular ballast doesn’t seem to require an earthed reflector to start reliably according to the data I’ve been able to obtain – unlike some other rapid start ballasts.
The moment of truth was to flip the switch and see what happens.
The ballast chirped as it does, and the arc was struck successfully. The filament area can be seen to be slightly darkened, probably because of the harshness of instant starting them. The low brightness, including the complete lack of light from the areas away from the filament, is a normal behaviour for amalgam lamps.
After four or so minutes of operation, the brightness was even amongst the tubes and a lot of light was being produced. Rather alarmingly, the temperature around the filament area seemed quite hot, so I checked the power on the Tektronix PA1000 (which was in-line at all times to detect anomalous power usage) which indicated that the whole unit was consuming 83W, or just about the right amount. After three hours of continuous burn time, no failure occurred, although the IR thermometer was claiming the filament area was above 100 degrees C. This might not be unusual, given the high power density of the tube.
Not satisfied with just a few loose tubes and wires on a table, a small drill bit, a few bulldog clips, a few cable ties and an IKEA storage container later – I have something that’s a little neater.
Because of the random inspiration, I’m also in the process of grabbing a few electronic ballasts to add to my collection. Maybe I’ll end up converting one preheat fixture into an electronic programmed-start fixture just for the sake of it.
Bonus: My First Fluorescent Fixture
Since I did mention my first fluorescent fixture (which was a salvage) in a previous posting, but didn’t have the chance to show much of it, I’ll add it as a “bonus” part here.
This fitting was quite an old one, and it came with a Philips PowerMiser 18W T8 tube which was somewhat aged. Because I wanted an even older look, I actually went to Kmart and purchased a Mirabella branded T10 20W tube. At the time, the T10 tubes were actually pretty limited in number, so I thought I’d buy it because it was “special”.
I’m kind of glad that I did that, because the T10 tubes are even harder to come by nowadays. They really are visibly “fatter” than the energy saving T8’s that replaced them. They are less efficient, but the nostalgia gets the better of me. Nowadays, walking into one of the stores, it’s quite likely that you can’t even buy fluorescent tubes anymore.
It seems that the cost of the phosphor coating may be the reason why they didn’t coat the phosphor right to the ends of the tube. This particular tube has significant gaps in coverage at both ends which poses a slight risk in terms of increased UV emission from the tube. Rather interestingly, the light emission is influenced at the end by the filament position.
The fitting itself has an Atco EC20 20W inductive ballast, although the marking that says 600×38 suggests to me that it was probably designed for a T12 tube originally. I’ve never actually held or knowingly seen a T12 tube first-hand.
This is where the PFC capacitor that failed was. Because the neutral was looped into the PFC capacitor terminal and then onto the terminal block, I joined the ends of the wire with a small terminal block. The unit seems to be made by SUIV lighting industries, of which I cannot find a trace of anymore.
It’s amazing to think that I grew up watching the educational campaigns pushing home users to move from incandescent to fluorescent lighting, and now, we’re witnessing the death of fluorescent lighting in preference for LED technology. It feels sad to see it go, but as far as ‘fitness for purpose’ goes, the LED lights have many technical advantages which cannot be denied.
Despite this, having stumbled across a random batch of CFL tubes, I decided to save them from eventual scrappage and then improvise a system to get them to light successfully using a ballast that was not rated to drive the lamps in the first place. It did an acceptable job, although the instant start is hard on the filaments.
I also had a chance to show you what the insides of the first fluorescent fitting I salvaged looked like, and the T10 tube I purchased to go with it. Good memories.