|
Introduction
If you ask, “When should I replace my metal halide bulbs?” you will likely hear it authoritatively stated “every 12 months.” With high Kelvin bulbs, it may even be recommended that they be replaced as often as every six months. These recommendations, however, have very little research to back them up. While a considerable volume of research is available on bulbs' spectra and output thanks to dedicated aquarists such as Dr. Sanjay Joshi, et al2-19 and Joe Burger1, very little research has been done on the useful life of metal halide bulbs. Joshi and Morgan4 did examine the spectra and light output of some used 250-watt MH bulbs. This study included five different models of used bulbs. None of these used bulbs were double-ended. While this study provides some very useful data, there are some limitations to the study. The bulbs came from a variety of different sources and the authors admit that “often, the exact number of hours the lamp was used was not known.” The age was “given by aquarists in the total duration of time the lamp was in use.” They go on to note that “the operating environment for the lamps may be very different, leading to variation in the lamp output arising from the variation in operating conditions.” This introduces considerable uncertainty to the measurements. For example, one of the bulbs tested was a 10,000K Coralife bulb. They had three used bulbs that were described by their users as “one year old.” The PPFD values for these bulbs varied, however, by over 50%. It is impossible to know whether this variability was due to intrinsic differences between the lamps or to differences in time and the environment of use. They do conclude for two of the other tested lamps, though, that “these lamps could be used longer than a year” based on the limited drop in light output seen in their study. They also observe for the 20,000K lamp that “these lamps should probably be changed earlier than the lower CCT lamps” due to a greater loss of output. The authors published an additional article3 with similar evaluation of 400-watt MH bulbs. While these studies are quite useful, I do not know of any prior study of 250-watt MH aquarium bulbs prospectively over time utilizing multiple “identical” bulbs under “identical” conditions for a known duration of use.
Since I acquired a quantum meter for measuring PAR, I decided to document the output measurements of my metal halide bulbs beginning from the time I put them into service, in order to see how their output declines over time under similar operating conditions. I am currently using 250-watt Phoenix 14,000K double-ended (DE) bulbs, and I am unaware of any study on these bulbs after extended use.
Methods
My aquarium has a total of seven 250-watt double-ended (DE) metal halide bulbs, all in identical Reef Optix™ 3 (Sunlight Supply®, Inc.) fixtures and all running on identical Ice Cap™ electronic ballasts. In March 2006 I replaced all the bulbs, which were a mixture of 10,000K and 14,000K bulbs (see sidebar below). The new bulbs I chose were 14,000K Phoenix™ DE bulbs.
|
Figure 1. Light meter and holding device. |
Notes: When I replaced three old AB 10K and four old Aquaconnect 14K bulbs, I put the seven new Phoenix 14K lamps into service. Throughout the experiment I assumed that bulb #3 was an outlier, but at the end of one year I decided to replace that one bulb. I then discovered that it was, in fact, an Aquaconnect 14K bulb. It appears that I had accidentally mixed up a new Phoenix and an old Aquaconnect bulb, and replaced the old bulb into the fixture while discarding the new Phoenix bulb. This means that this bulb was actually about 15 months old when the study began and 27 months old at the study's conclusion. |
|
All light measurements were made with an Apogee Model QMSW-SS quantum meter (Apogee Instruments). In order to make the measurements consistent I built a simple device for holding the probe (Figure 1 above). This allowed the probe to be held a consistent distance from the glass, about seven inches, which brought the measurements into the meter’s range, as positioning the probe too close to the shield gave measurements above the quantum meter’s readable scale.
Before each set of measurements was taken the glass shields were cleaned of any salt spray. The holder and probe were moved along the glass shield to search for the highest reading (Figure 2 below). When this peak level was found, the measurement was recorded. The readings were reproducible within less than 1% for each measurement.
|
Figure 2. Taking a measurement. |
These measurements were made monthly for all seven fixtures for a period of one year. The measurements of photosynthetic photon flux density (PPFD) are reproduced in Table 1. I also calculated each bulb's output as a percent of that bulb’s initial reading. Those calculations are reproduced in Table 2. The bulbs go “on” and “off” on a staggered schedule, but each is on for eight hours per day. This equates to about 240 hours per month and 2920 hours for one year.
Results
As Table 1 shows, six of the seven bulbs' initial intensity readings were at least 90% of the highest bulb's reading. Bulb #3, on the other hand, had an “initial” output of only 67% of that of the highest bulb. This bulb was, in fact, already over one year old at the start of the study rather than new as were all the other bulbs.
After six months, the six bulbs' intensity was still 78-84% of their initial output (average, 81.5%). Even after 12 months their output had only fallen to 72-79% of the initial level (average, 75.7%). In fact, their intensity at 12 months was about 93% of what it had been at six months. Bulb number 3 had an output that was 97% of its initial reading at six months and 92% at 12 months into the study. The intensity curves had flattened considerably by 12 months, which suggested only a very gradual further decline in intensity beyond a year.
Discussion
The light output of the six Phoenix bulbs shows an initial decline over the first three to four months and then levels off at about 75% of their initial output. From these data I raise the question: “If a bulb is not replaced at six months, when it has lost 20% of its output, then why would we replace it at 12 months when its output has fallen only another 6%?” It may be more cost effective to plan the lamp number and placement based on 70-75% of the bulbs' initial output levels. Initially, we could shield or raise the lights and then gradually lower or unshield them over the first four months or so, when the intensity is at its highest but falls most quickly. Beyond six months the bulbs' light output remains fairly constant at about 75%, so only minor adjustments would be needed for at least a year. Doing this may considerably extend the bulbs' effective life.
As I indicated earlier, bulb #3 was actually an old Aquaconnect 14,000K bulb that had mistakenly been retained when the other bulbs were replaced. That means that this bulb was actually about 15 months old when the study began, and 27 months old at the study's conclusion. The Phoenix and Aquaconnect bulbs are thought to be the same bulb sold with different labels. If this is correct, then this bulb would appear to still maintain 70% of its output at 18 months, 66% at 24 months and 63% at 27 months of use.
I have been in the habit of replacing my bulbs about once a year as is so often recommended. The measurements suggest that this may not be necessary, though. In my aquarium setup, my lighting is suspended by a pulley system, which allows the lights to be easily raised and lowered (see Figure 3 below). Lowering the lamps a few inches can easily compensate for a 25 to 30% drop in lamp output to maintain a consistent light intensity for the corals by decreasing the distance from the bulbs. Having a reliable light meter, I can measure the output of all seven of my bulbs in fewer than five minutes. I now target a PAR level of 400-500 for the most light-demanding corals, and adjust the light pendants’ height accordingly. For now, I will continue using the same bulbs and perhaps take a quick follow-up reading after 18 months and see if the other six bulbs have held up the same as the Aquaconnect bulb.
|
Figure 3. My adjustable light rack. |
Of course, light intensity is only part of the story. I had no way of accurately assessing the changes, if any, in the bulbs’ spectral shift. The light spectrum, however, has been found to be far less important to corals' growth than total PAR levels.20 Changes in spectrum definitely affect the aquarium inhabitants' apparent color, though, and conceivably have a longer term effect on the corals' pigmentation. These effects are somewhat subjective, though. It may be that you would want to change the bulbs before their light output requires it simply because you don’t like how they appear. So far, though, in my tank I have not seen a noticeable change in the bulbs’ color, nor has the corals’ apparent coloration deteriorated.
Conclusion
I have shown how the output of one particular metal halide bulb diminishes initially and then slows its decline over time. Knowledge of the bulb’s output curve could affect the decision of when to replace the bulb. I have suggested a method for extending the bulbs' useful life by adjusting for this drop in their light output over time. This is only a single bulb and ballast combination, but it does suggest that at least some bulbs have a longer useful life than is commonly believed.
|
