TEXT AND DIAGRAMS BY SANJAY JOSHI AND DAVE MORGAN

Spectral Analysis of Metal Halide Lamps
Used In The Reef Hobby — Part 2

In the previous installment to this article (“Part 1”) we presented the spectral analysis of new 400-watt metal halide lamps. This month we will present the results of our analysis of used 400-watt lamps. Our purpose was to determine how the spectral distribution changes as a function of time, as well as the nature of this spectral shift for the various 400-watt metal halide lamps being used in the hobby.

TABLE I
PPFD and CCT for the
Iwasaki 6500 K lamps of various ages
Age PPFD CCT
New 203.7 6637
3 months 202.1 6326
6 months 190.4 6176
1 year 169.2 5800
1 year 166.1 5988
1 year plus 158.3 5939
2 years, 2 months 174.8 5977
2 years, 7 months 143.8 5394

Our initial goal was to track the changes of the same lamps over a period of time, in order to get an idea of how the spectral distribution changes over time. But, due to the difficulties of obtaining the data on the same lamp at various points in its life, getting a large enough sample, the length of time such a study would take and so on, we decided to take a sample of used lamps of varying ages, rather than track the same lamp through its lifetime.

We asked aquarists all over the country to send in any used lamps they were willing to loan for this study. Along with the lamps, aquarists provided us with the hours of usage for each lamp (according to their best estimates). Spectral analysis of was then performed using the same setup as described in “Part 1.”

We realize that not knowing the initial spectral distribution of each lamp precludes us from making strong statements about the deterioration of lamps over time. However, we feel that just knowing the range of spectral distributions of the lamps of different ages may provide some knowledge of how the spectral distribution looks over time to aquarists, and possibly some indicators on a replacement policy for the lamps.

Table II
Changes in PPFD as compared to a new 6500 K lamp
Lamp Percent drop in PPFD in waveband range
400 to 700
(PAR)
400 to 440
violet
440 to 490
blue
490 to 540
green
540 to 590
yellow
590 to 650
orange
650 to 700
red
3 months -0.69 13.35 1.07< td> -2.07 0.37 -19.54 -7.93
6 months 6.53 23.34 9.96 5.39 9.99 -14.13 -6.58
1 year 16.94 40.56 23.57< td> 16.59 18.61 -6.59 1.14
1 year 18.46 37.75 22.12 18.24 20.38 -2.54 6.66
1 year plus 22.29 38.63 28.82< td> 19.52 21.16 4.64 16.32
2 years,
2.5 months
14.19 30.06 15.59 14.92 15.54 -8.04 5.10
2 years,
7 months
29.41 55.96 42.00< td> 27.70 24.17 65.38 18.85

Iwasaki 6500 K Lamp

Figure 1 shows the spectral distribution of the eight Iwasaki 6500 Kelvin (K) lamps of varying ages analyzed in this study. From the figure one can easily see that as the lamps age there is a general decrease in the output across all wavelengths — indicated by a lowering of the intensity at the various wavelengths — and the decrease is more pronounced in the shorter wavelengths of the spectrum. As these lamps age they have more drop in the blue end of the spectrum as compared to the red end. Hence, the lamps tend to get “warmer” over time, which explains the nature of the spectral shift over time.

In addition to the spectral distribution, the “photon flux density” (PPFD) and the “correlated color temperature” (CCT) of these lamps were also calculated. The results are listed in Table I. As can be seen in the table, as expected, there is a drop in PPFD, as well as a shift in the CCT as the lamps age.

There is some variation in the data where you will notice that a lamp of older age has, in fact, “better” output than a lamp with less usage. This could have been due to higher initial values or a slower decay of the lamp halides. Because the initial spectral distribution of the lamps is not known this cannot be ascertained with any degree of certainty.

Although it can be visually determined from the graph in Figure 1, the lamps tend to have a higher drop in PPFD in the violet and blue ranges. Quantitative data showing the drops are shown in Table II.

Figure 1 Figure 2
Figure 4 Figure 3

Age graphs of the various lamps tested. Clockwise from upper left:
Figure 1 — 6500 K Iwasaki lamps; Figure 2 — 10,000 K Aqualine Buschke lamps;
Figure 3 — 20,000 K Osram/Radium lamps; and Figure 4 —6000 K GE lamps.

The spectral output of each lamp was compared to the new lamp. Lamps with about a year of life at 10 to 12 hours of use per day seem to lose about 17 to 22 percent of their intensity in the “photosynthetically available radiation” (PAR) range (400 to 700 nanometers; nm). Interestingly, the largest contributor to this drop is in the violet range, almost 40 percent for year-old lamps.

TABLE III
PPFD and CCT for the Aqualine Buschke
10,000 K lamps of various ages
Age PPFD CCT
New 135.0 9946
700 hours 89.7 8143
800 hours 122.8 9198
3600 hours 90.0 8381
1 year 100.2 7827

A negative number indicates that the lamp output in that range was higher than that of the new lamps used for comparison. It does not necessarily indicate that the lamp output in these ranges tends to increase. It could very well be that these lamps had higher outputs in the respective ranges to begin with, than the new lamp.

A conclusion that could be drawn from this data is that the rate of decrease in PPFD during the second year is less than that during the first. The worst two-year lamp tested only had a 29.4-percent decrease in PPFD. If we assume that, similar to the one-year lamps, 17 to 22 percent of this occurred during the first year, then only 7 to 12 percent of decrease occurred during the second year. One implication of this to the aquarist may be that instead of changing these lamps every year, they could be used for at least 18 months.

Aqualine Buschke 10,000 K Lamps

Figure 2 shows the spectral distribution of the five 10,000 K Aqualine Buschke lamps we tested, and Table III presents the PPFD and CCT of these lamps. Table IV presents the drop in PPFD output when compared to the new lamps.

Table IV
Changes in PPFD as compared to a new 10,000 K Aqualine Buschke lamp
Lamp Percent drop in PPFD in waveband range
400 to 700
(PAR)
400 to 440
violet
440 to 490
blue
490 to 540
green
540 to 590
yellow
590 to 650
orange
650 to 700
red
700 hours 33.57 29.35 49.75< td> 52.14 8.04 45.76 37.58
800 hours 9.04 11.67 11.84 11.63 3.31 8.07 8.73
3600 hours 33.33 45.64 29.62< td> 32.97 55.22 -0.35 5.15
1 year 25.78 47.46 22.13 13.74 43.54 -6.04 -6.65

The 700-hours lamp seems to be an anomaly. As seen from the data, the initial PPFD output is significantly less than the 6500 K lamps, and the output tends to drop 25 to 33 percent within the span of a year. There were no two-year lamps available for testing, as most aquarists tended to replace them on a yearly basis. The output in the violet and blue range tends to fall more rapidly than the 6500 K Iwasaki lamps.

TABLE V
PPFD for the Osram/Radium
20,000 K lamps of various ages
Age PPFD
New 116.2
5 months 80.18
7 months 87.21
1 year 70.61
1 year plus 55.40
6500 hours 84.65
6500 hours 26.09

Osram/Radium 20,000 K Lamps

Figure 3 shows the spectral distribution of the various 20,000 K Osram/Radium lamps tested, and Table V presents the PPFD. Table VI presents the drop in PPFD output when compared to new lamps. Here again, one of the lamps at 6500 hours of life seems to be an outlier (an anomaly). As seen from the data, these lamps tend to lose their output at a much higher rate compared to the 6500 and 10,000 K lamps. Year-old lamps had lost about 40 to 50 percent of their intensity. Unlike most of the other lamps, these lamps show an almost uniform reduction in intensity across the range.

GE 6000 K Lamps

These lamps have recently shown up for use in reef aquaria. They are currently mislabeled in the hobby as 10,000 K lamps. According to the manufacturer they are, in fact, 6000 K. The recommended ballast for these lamps is the ANSI M-135 or an M59 with an igniter. Because they are being sold by the hobby as replacement lamps for use in existing metal halide fixtures, we tested them using the standard metal halide ballast. Given the low usage of these lamps in the hobby, their availability for testing was limited.

Table VI
Changes in PPFD as compared to a new 20,000 K Osram/Radium lamp
Lamp Percent drop in PPFD in waveband range
400 to 700
(PAR)
400 to 440
violet
440 to 490
blue
490 to 540
green
540 to 590
yellow
590 to 650
orange
650 to 700
red
5 months 31.0 24.47 23.67< td> 39.09 35.75 40.87 35.54
7 months 24.95 25.54 25.83 26.57 24.32 24.12 22.04
1 year 39.23 41.44 38.93< td> 39.90 39.47 39.83 33.38
1 year plus 52.32 49.82 52.26 52.53 53.02 53.46 53.24
6500 hours 27.15 36.24 30.08< td> 26.82 24.12 23.17 91.93
6500 hours 77.35 80.45 79.08 77.44 75.76 74.30 73.02

Figure 4 shows the spectral distribution of the GE 6000 Ks we tested and Table VII presents the PPFD of the various. Table VIII presents the drop in PPFD output when compared to new lamp. Interestingly, the new lamp tested was “worse” than the three-month-old lamps.

Conclusion

In this article we presented the results of spectral analysis of used 400-watt metal halide lamps in an effort to provide some indicators of spectral differences that occur over the lamps’ age. We agree that the lack of information about the initial spectral distribution of each lamp limits our ability to make accurate comparisons, but we feel that this data will at least provide some sense of how lamps of different ages look under a spectroradiometer and how the spectral characteristics differ. Clearly as indicated by the data, there are larger (faster) drops in the intensity in the violet and blue end of the spectrum for all the lamps. Hence, lamps with higher CCT tend to have a higher rate of intensity drop. Can this data be used to suggest a replacement policy for the lamps? The answer is not all that simple. Before this question can be adequately answered we need to define at what point does a lamp reach the end of its useful life. Should this useful life be based on (a) an absolute threshold based on the total PPFD in the PAR range independent of the lamp or, (b) a specified amount of reduction in the individual lamp’s PPFD in the PAR range, or (c) a specified amount of PPFD in a given radiation waveband?
TABLE VII
PPFD and CCT for the
6000 K GE lamps of various ages
Age PPFD CCT
New 140.7 6217
700 hours 150.9 6010
700 hours 145.5 5833
1 year 87.05 6834

The answers to the questions involve many aspects and are far beyond the scope of this article. One large factor in answering these questions involves the specific needs of the corals themselves. Many corals are known to show some photoadaptive ability.

Furthermore, by changing the distance between the lamp and the water surface we can increase the amount of light that enters the tank. So, as the lamps age and drop in intensity, we can counteract this drop by reducing the distance between the lamp and the water surface, thereby extracting more life out of the lamp as long as we are convinced that the spectral distribution is still acceptable. It is not hard see how this can complicate the issue of setting rigid guidelines as to when a lamp is good or bad.

Table VIII
Changes in PPFD as compared to a new 6000 K GE lamp
Lamp Percent drop in PPFD in waveband range
400 to 700
(PAR)
400 to 440
violet
440 to 490
blue
490 to 540
green
540 to 590
yellow
590 to 650
orange
650 to 700
red
700 hours -7.25 -12.17 -0.64< td> -7.50 -1.97 -8.98 -12.62
700 hours -4.41 -4.79 4.75 -1.65 1.50 -7.55 -12.34
1 year 38.13 12.62 44.76< td> 37.24 28.62 49.63 48.79

Acknowledgements: We would like to thank several hobbyists, without whose help this study would not have been possible. These aquarists were kind enough to loan us their new and used 400-watt lamps and help reduce the financial burden of having to acquire them all. Thanks to: Richard Harker, Dana Riddle, Wayne Shang, Mike Fontana, Jack Chernega, Eric Borneman, James Wiseman, Dan Wohls, Ed Roan and Kevin Carpenter. Finally, we would like to thank Dr. Paul Walker of Penn State University and LiCor Inc., for the use of their spectroradiometers and the darkrooms for testing of the lamps.

HOME Table Of Contents FEATURE

©Copyright, Aquarium Frontiers and Fancy
Publications Inc. All rights reserved.