Introduction and Background

When initially starting this adventure, the goal I had in mind was to investigate lamps that would be suitable for use in my tank when the time came to change metal halide lamps. At the time I found little information available on the 250-watt single-ended (SE) metal halide (MH) lamps that I was using. Prior to actually laying out the funds to buy the same type of lamps again, it seemed reasonable to research a wide variety of products that were available on the market. That way, I could be certain that the product that I purchased best fit the needs of my tank.

In the beginning, some fellow reef enthusiasts assisted my search for suitable replacement lamps. A few friends lent me some of their extra lamps, and I was also able to borrow a PAR meter from another associate. After using the new equipment, I took readings and shared my findings by posting an "innocent" picture or two on Reef Central. Thereafter, my project took on a life of its own. Soon, people were e-mailing me and asking me to post more information on other lamps. People offered to send me a lamp or two to test. While testing more lamps, I found some ways of testing to be better than others. I found myself with results that left me with questions that needed to be answered. I realized that I had lighting questions that I needed clarified so that I could better understand the results that I was seeing and, in turn, have the ability to explain my findings to others. I wondered about the reasoning as to why measuring photosynthetically active radiation (PAR) was better than measuring lux. In an effort to seek a fuller understanding, I spent a great deal of time researching, testing and discussing basic lighting with people more knowledgeable than I. This resulted in a massive thread on Reef Central encompassing some 50-odd pages, and boasting 1300 posts, wherein I discussed all of the lamps that I tested, along with my notations regarding each product and my final test results. Such a thread is definitely a large undertaking for any aquarist to comprehend. I decided to write the present article to summarize my findings in an effort to help reef aquarists better understand what I tested, and to share my results.

What Was Tested

I tested just about every lamp and ballast that I could get my hands on during the testing time. Whatever I received made it onto the final list. A total of 21 lamps and 8 ballasts were used, as shown in Table 1 below. Almost all of the lamps were tested on every ballast. I had to return a few lamps (and one ballast) to the person who lent them to me, as they needed to be used on the person's tank. Therefore, not every combination was tested. Still, each ballast/lamp was tested against enough lamps/ballasts that their performance tendencies are readily seen; whether being judged on its own or against other similar equipment.

Table 1 – Lamps and Ballasts Tested

Lamps Tested		Ballasts Used
Aqualine Buske 13000K	PFO 13000K	ARO (Hellolights) Electronic Ballast
AquaConnect 14000K	Radium	CoralVue Electronic Ballast
Blueline 10000K	Sunburst 12000K	Blueline E-Ballast Electronic Ballast
Blueline Super White	Sun Aquatics 10000K	EVC Ballast
CoralVue 10000K	Sun Aquatics 14000K	IceCap Electronic Ballast
CoralVue 12000K	Sun Aquatics 20000K	PFO HQI Ballast
CoralVue 15000K	Ushio 10000K	PFO Standard Ballast
CoralVue 20000K	XM 10000K	Reef Fanatic Ballast
EVC 10000K	XM 15000K
Hamilton 14K	XM 20000K
Iwasaki 6500K

Ballasts appeared to operate the same regardless of how long they had been used. I found no difference between one that was fresh out of the box and one that had been run for months, or even years. That led me to believe that there was no break-in period for metal halide ballasts.

How These Tests Were Conducted

I should state up front that I am a hobbyist; I'm not an electrical engineer, nor am I involved with the manufacture of lighting equipment. I conducted these tests solely to satisfy my own curiosity, out of my enthusiasm for the hobby. When these tests were started almost no information was available for 250-watt SE metal halide lamps. With that said, I tried to get as much pertinent information as I could to answer my question on how different 250-watt SE metal halide lamps performed. In turn, I wanted to share my understandings with other hobbyists so that we could make a better, more informed decision about what types of lighting we purchase to support our systems. As a side note, all of the lamps that I tested were burned in for at least 120-150 hours before being tested. Additionally, all lamps, excluding the Ushio, AB and Blueline Super White, were new at the time of testing. The remaining lamps had roughly six months of use on them (approximately 1600 hours).

After a few short tests that were done over my tank (results on this RC thread), it was painfully obvious that a separate environment to test the lamps was needed. This environment would need to be able to produce repeatable results. After much time, and many conversations with other people in the lighting industry, a solution was found. What I wanted was the raw output from the lamp only. Reflector efficiency varies greatly from one product to the other and many people use different reflectors and/or material to get extra output from the lamp, so I decided not to include any reflectors in my testing, as the results couldn't be universally applied if the variable of an assortment of reflectors was involved.

I finally settled on a 2' x 2' x 2' square box (seen above). Its inside was painted flat black to minimize the amount of light reflected back to the sensor. The mogul socket was mounted horizontally, approximately 5" from the ceiling of the box. The quantum sensor (a sensor used to measure PAR) was placed 8" below the center of the arc tube. I chose 8" because this seemed to be a common distance that people keep their MH lamps from the water's surface. To minimize voltage spikes and aberrations, a voltage regulator was used on power from the wall's electrical socket. A watt meter was plugged into the voltage regulator and the ballast plugged into the watt meter.

All lamps, before testing, were powered on and left to burn for at least 20 minutes before any measurements were performed, in order to allow the bulb's output to stabilize. After 20 minutes the sensor was turned on and time was given for a stable reading to be displayed from the meter. Some time (between 2-5 minutes) was given for the reading on the meter to stabilize for at least 30 seconds with no movement. At that time, readings were recorded.

Equipment Used:

• Apogee instruments QMSS-ELEC quantum meter with sensor: calibrated for electric lamps

• Apogee instruments leveling plate for the sensor

• APC Line-R 600 voltage regulator/power conditioner

• Electronic educational devices - Watts Up? power analyzer/watt meter

• Nikon 995 digital camera used with tripod for pictures

• BK Precision 530 multimeter

PAR vs LUX

Many people asked why I used PAR (Photosynthetic Active Radiation) as a way of measuring the lamps as opposed to lux (a photography and media term used to measure light values). Also, many people wanted to know if their lux measurements could be translated into PAR values. In an effort to be as concise as possible on this topic, I'll state that PAR measurement is useful for a number of reasons, the primary one being that it measures the band of light (400nm-700nm) where photosynthesis can occur. A secondary reason is that the light is measured uniformly. Lux readings are geared toward the way humans see light, so greens and yellows are given a higher weight than purples and blues. PAR readings measure nearly the whole visible light spectrum, 400nm-700nm, evenly.

Here are two graphs from Li-Cor, a manufacturer of instrumentation for environmental research. They show how accurate the company's sensors are in reference to an "ideal" curve. In our case, we are interested only in the curve represented by the dotted line.

A graph of how lux is measured:

A graph of how PAR is measured:

The graphs clearly show that PAR is a more accurate way to measure lighting for a tank housing photosynthetic organisms.

The graph also shows that getting a lux reading from one lamp is good for comparing it only to the same lamp, or to another lamp with the same spectral plot. Yet, a lux reading from a 6500K lamp could not be equally compared to the lux reading of a 20,000K lamp.

Apogee Sensor Accuracy

The PAR (or quantum) sensor and meter I used are made by Apogee Instruments. They make a reasonably priced sensor that a hobbyist can afford (i.e., under $300). Sanjay Joshi, in his lamp testing, has used a quantum sensor made by the Li-Cor company that costs well over $1000. Naturally, numerous questions were raised about the accuracy of the Apogee sensor used in the test.

The margin of error needed to be considered regarding this sensor. I contacted Apogee and discussed this with a representative, who agreed to send to me two Li-Cor sensors that the company uses for in-house testing. The sensors were first calibrated by Apogee and then shipped to me. After I received the Li-Cor sensors, I took measurements. Two Li-Cor sensors were sent to take measurements with, and to have an average baseline to work from so I could compare them to the Apogee. One measurement with each Li-Cor sensor was taken, and then the two were averaged to produce a value. While the lamp was still lit, a reading from the Apogee sensor was taken and compared to those obtained by the Li-Cor sensors. The table below shows the margin of error the Apogee had when compared to the Li-Cor. Also shown is an adjustment factor. If this factor is applied to the Apogee reading (take the initial reading from the Apogee and multiply it by the adjustment factor), it should correct for any deficiencies of the sensor. All lamps have a slightly different spectral curve, so the error is different for each lamp.

Table 2 - Apogee Sensor Accuracy

Lamp	Adj. factor	Error %
AB10K	1.042700	4.27
XM10K	1.000102	0.01
XM20K	1.054623	5.46
Radium	1.020256	2.03
CV10K	1.014444	1.45
Ushio10K	1.022647	2.26
PFO13K	1.066219	6.62
Iwasaki	1.004403	4.40
BLSW	1.035931	3.59
SUN10K	1.041306	4.13
CV15k	1.062583	6.26
CV20K	1.031709	3.17
BL10K	1.020969	2.10
HM14K	1.041438	4.14
AQ14K	1.027771	2.77
SUNBRST 14K	1.025221	2.52
CV12K	1.076923	7.69
SUN14K	1.026404	2.64

As the chart shows, for such a low price, the sensor is very accurate. On average, it deviates from the Li-Cor sensor by only 3.64%. After taking these readings I felt that, for hobbyists' purposes, the reading from the Apogee sensor was more than sufficient.

Lamp Heat Due to Ballasts

Many people, including myself, have always wondered how much heat is generated from adding metal halide lamps over a tank. I've also wondered how much of a difference fans make in keeping the lamp's heat from impacting the tank's water temperature. To answer these questions, I took measurements and found that different ballasts do vary the amount of heat produced by the lamp.

Temperature measurements were taken the same distance from the lamp as the light sensor was placed - 8". The ballast was turned on with a fan above the lamp, effectively pulling the heat out. After 35 minutes, the temperature 8" from the lamp was taken. To determine the contribution of ventilation to heat generation, we then turned off the fan and, after 10 minutes more (for a total trial of 45 minutes), measured and recorded the temperature again. Ambient room temperature during all the testing was 70°F. The results are found in Table 3 below.

Table 3

Ballast	Lamp	0 Min	35 Min	45 Min/No Fan
PFO Standard	Ushio	70.0°F	83.2°F	90.1°F
PFO HQI	Ushio	70.3°F	87.5°F	97.2°F
eballast	Ushio	70.5°F	81.4°F	89.5°F
Icecap	Ushio	70.1°F	78.6°F	85.2°F
CoralVue	Ushio	70.3°F	82.1°F	90.6°F
ARO	Ushio	70.2°F	79.6°F	88.3°F

It seems quite obvious that heat from the lamp can play a significant role in heating the water in the tank. After 45 minutes, the temperature with a fan was approximately seven degrees less than it was with no fan. In the case of the PFO HQI ballast, the ambient temperature in the test box rose almost 27 degrees Fahrenheit when not using a fan to vent excess heat! Proper ventilation, therefore, is a must in any lighting fixture/canopy.

Actual Power to the Lamp

Manufacturers of many of the new electronic ballasts report that they are better suited than magnetic ballasts to power aquarium metal halide lamps due to their sophisticated internal circuitry. In order to test this claim I decided to take some measurements on the wire between the ballast and lamp. The results are given below. While I'm unsure if these results speak better of electronic or magnetic ballasts, the interesting thing to note is that the electronic ballasts varied in their ability to deliver power to the lamp. These two specific lamps were chosen because they were ones that I could track down the power and current specifications as released by the manufacturer.

Voltage measurements were taken using a BK 5390 Multimeter. Current measurements are reflected in TRMS. (RMS stands for Root Mean Square, a technique used to measure AC voltages. True RMS, or TRMS, takes harmonic distortion into account to accurately measure non-sinusoidal AC voltage and current waveforms. On a sinusoidal waveform, RMS and TRMS are the same; therefore, TRMS is advantageous when measurements of non-sinusoidal AC waveforms are required.) The electronic ballasts seem to use the DC sine wave, so both AC and AC+DC measurements were taken. All measurements were taken on the wire between the ballast and the lamp. All ballasts were plugged into a PFO style plug that connected to the mogul socket for the lamp. Measurements were taken where the plug connects to the socket, so all ballasts were measured using the same wire. The lamps were first ignited by the ballast, then left to burn for 20 minutes. At that time, measurements were taken. The results are shown in Table 4 below.

The boldfaced first line in the table below shows the manufacturer's specifications of the lamp. Following are the readings taken while running the lamp on different ballasts. The Standard PFO ballast would not drive or start the Radium lamp, so no information is given for this combination.

Table 4 – Power Profile of Ballasts Tested

Lamp	Ballast	AC AMP	AC+DC AMP	Volts	kHz	Line watts	Line volts
Radium	SPEC	2.80	2.80	95-100
Radium	HQI	2.63	2.63	113.50	59.98 Hz	308	117
Radium	ARO	3.20	4.79	112.37	76.80	249	117
Radium	eballast	2.21	1.95	110.65	80.73	242	121
Radium	IceCap	3.20	4.58	109.60	70.47	248	119
Radium	CV	2.29	2.01	109.63	67.52	264	119
Radium	ReefFanatic	2.28	2.80	131.90	42.03	246	119
Radium	EVC	2.37	2.88	115.80	44.68	247	119
Ushio	SPEC	3.00	3.00	100.00
Ushio	HQI	2.43	2.43	133.41	59.98Hz	339	118
Ushio	ARO	2.94	4.64	125.43	77.95	249	119
Ushio	eballast	1.80	1.59	123.04	86.68	243	119
Ushio	IceCap	2.87	4.40	127.78	73.43	254	119
Ushio	CV	1.84	1.62	125.39	72.84	254	121
Ushio	ReefFanatic	2.62	3.20	116	44.06	246	119
Ushio	EVC	2.10	2.42	132.20	41.98	246	120
Ushio	Standard	2.11	2.11	129.34	59.98Hz	273	118

The Meat of it All: Testing the Lamps

Table 5 below represents the main thrust of the testing that was done. Each lamp was tested on the ballasts available to me at the time. I was asked to return the CoralVue ballast a few months after testing, so fewer lamps are shown on that ballast than on the others. I received quite a few different lamps after returning this ballast to the manufacturer. The measurements for all of the ballast/lamp combos tested are shown below. PAR is the reading from the Apogee sensor after 20 minutes of lamp illumination. The columns marked "Watts" and "Amps" represent the power drawn by the ballast from the wall's electrical socket after being passed through the voltage regulator. Generally speaking, ballasts need to produce an initially high current pulse to ignite the lamp, and in some cases this shows up as a high amperage draw that is captured under "Max Amps."

This series of charts demonstrate in PAR the amount of light available for photosynthesis. In general, we are looking for higher PAR values. The data are sorted by PAR value ranking.

Table 5

ARO (Hellolights) Electronic Ballast:
Lamp	PAR	Watts	MaxWatts	Amps	MaxAmp
Iwasaki	605	250	253	2.11	2.19
XM10K	526	250	251	2.16	2.37
BLSW	465	250	254	2.16	2.20
EVC10K	463	249	256	2.16	2.19
SUN10K	461	248	256	2.14	2.23
AB13K	434	252	254	2.15	2.51
Ushio	408	251	269	2.14	2.61
AQ14K	391	253	258	2.23	2.33
CV10K	360	249	251	2.11	2.15
HM14K	338	249	253	2.17	2.63
Radium	327	249	252	2.18	2.23
SBURST12K	302	250	270	2.18	2.50
CV20K	286	249	265	2.16	2.55
XM20K	270	250	253	2.16	2.41
SUN14K	259	249	255	2.15	2.37
CV12K	258	250	273	2.17	2.48
BL10K	253	246	250	2.14	2.19
CV15K	243	250	252	2.16	2.22
SUN20K	227	246	278	2.17	2.53
PFO 13K	206	251	252	2.16	2.19
XM15K	171	252	256	2.2	2.27
CoralVue Electronic Ballast:
Lamp	PAR	Watts	MaxWatts	Amps	MaxAmp
Iwasaki	569	247	266	2.00	2.36
XM10K	541	257	270	2.12	2.50
BLSW	480	252	270	2.02	2.20
CV10K	430	262	265	2.10	2.13
AB13K	368	224	268	1.80	2.39
Ushio	363	255	265	2.08	2.38
HM14K	363	258	268	2.09	2.49
Radium	337	260	269	2.13	2.89
CV20K	316	260	265	2.09	2.22
CV15K	311	260	265	2.10	2.16
BL10K	265	256	261	2.09	2.12
XM20K	257	248	269	2.01	2.19
PFO 13K	203	244	268	1.97	2.17
Blueline E-Ballast Electronic Ballast:
Lamp	PAR	Watts	MaxWatts	Amps	MaxAmp
Iwasaki	558	244	247	1.99	2.14
XM10K	521	243	244	2.00	2.03
BLSW	474	246	250	1.99	2.11
EVC10K	468	245	249	1.99	2.05
SUN10K	468	244	245	1.99	2.00
Ushio	430	245	247	2.00	2.03
AB13K	428	246	249	1.99	2.09
AQ14K	390	246	247	2.00	2.03
Radium	343	243	244	1.98	2.00
CV10K	341	242	243	1.95	1.96
CV20K	340	242	244	1.93	2.00
HM14K	323	242	243	1.95	2.05
SUN14K	266	240	241	1.93	1.95
BL10K	261	240	242	1.93	2.71
CV12K	243	242	268	1.98	2.52
CV15K	238	242	243	1.95	2.21
SUN20K	234	240	240	1.92	2.64
XM20K	228	244	250	1.99	2.12
XM15K	194	248	249	2.0	2.24
PFO 13K	178	244	245	1.97	1.99
EVC Ballast:
Lamp	PAR	Watts	MaxWatts	Amps	MaxAmp
XM10K	559	246	249	2.03	2.80
Iwasaki	517	247	250	2.03	2.87
AQ14K	438	250	254	2.03	2.1
SUN10K	425	247	248	2.01	3.29
BLSW	423	247	249	2.00	3.20
EVC10K	412	246	248	2.01	2.08
Ushio	394	247	248	2.02	3.65
HM14K	350	246	250	2.01	2.06
Radium	294	247	248	2.03	2.59
SBURST12K	291	246	251	2.01	3.66
BL10K	290	247	248	1.98	2.37
SUN14K	261	246	248	2.00	3.33
CV12K	257	246	294	1.97	2.02
XM20K	254	246	247	2.00	2.59
SUN20K	241	247	248	2.01	2.14
XM15K	211	252	256	2.07	2.12
PFO 13K	184	246	248	2.00	2.40
IceCap Electronic Ballast:
Lamp	PAR	Watts	MaxWatts	Amps	MaxAmp
Iwasaki	651	252	275	2.17	2.45
XM10K	530	252	254	2.20	2.82
SUN10K	494	252	255	2.14	3.86
BLSW	494	252	254	2.15	2.83
EVC10K	492	252	254	2.15	2.96
AB13K	422	254	257	2.17	2.30
Ushio	414	252	258	2.19	3.41
AQ14K	406	252	264	2.17	2.32
CV10K	387	250	251	2.17	2.46
HM14K	352	251	252	2.17	2.29
Radium	330	252	275	2.19	3.56
CV20K	307	251	253	2.20	2.25
SBURST12K	304	253	284	2.17	2.61
XM20K	270	252	301	2.17	2.81
BL10K	263	248	250	2.18	3.68
CV12K	259	251	253	2.17	3.32
SUN14K	253	251	308	2.18	2.79
CV15K	242	251	252	2.20	3.85
PFO 13K	227	251	268	2.18	2.86
SUN20K	224	247	249	2.20	3.00
XM15K	187	252	254	2.17	2.83
PFO HQI Ballast:
Lamp	PAR	Watts	MaxWatts	Amps	MaxAmp
Iwasaki	950	355	384	3.02	4.79
XM10K	835	340	341	2.93	4.07
BLSW	702	349	359	3.00	5.72
EVC10K	700	345	350	2.99	3.87
SUN10K	699	342	346	2.95	3.36
AB13K	659	341	349	2.96	4.55
Ushio	617	349	360	3.00	7.42
CV10K	457	333	335	2.87	3.25
AQ14K	453	311	313	2.85	4.74
HM14K	420	317	361	2.79	6.86
Radium	395	311	315	2.75	3.28
SBURST12K	378	311	698	2.79	14.27
CV20K	355	321	330	2.79	5.18
CV15K	336	317	322	2.76	4.55
CV12K	328	310	314	2.73	4.45
XM15K	318	312	847	2.79	16.33
XM20K	314	327	337	2.85	3.94
SUN14K	314	327	490	2.92	8.69
PFO 13K	308	311	418	2.76	8.08
BL10K	292	310	385	2.83	10.03
SUN20K	232	292	393	2.81	10.65
PFO Standard Ballast:
Lamp	PAR	Watts	MaxWatts	Amps	MaxAmp
Iwasaki	691	286	331	2.53	6.86
XM10K	600	268	269	2.47	3.39
AB13K	566	301	303	2.64	3.81
BLSW	514	273	274	2.49	5.47
EVC10K	512	271	278	2.50	3.12
SUN10K	510	270	277	2.51	2.69
Ushio	485	284	285	2.53	4.57
CV10K	324	231	233	2.32	4.03
SBURST12K	309	269	269	2.51	6.01
Radium	283	231	232	2.38	6.81
CV20K	275	226	261	2.30	7.93
XM15K	264	265	267	2.55	7.42
PFO 13K	248	257	258	2.50	6.80
CV15K	242	227	239	2.35	4.14
XM20K	242	285	285	2.57	3.75
SUN14K	228	224	228	2.32	4.08
AQ14K	228	212	274	2.33	8.29
HM14K	214	219	220	2.30	7.12
CV12K	190	230	230	2.38	3.54
SUN20K	148	198	199	2.25	6.11
Reef Fanatic Ballast:
Lamp	PAR	Watts	MaxWatts	Amps	MaxAmp
XM10K	571	247	248	2.03	2.77
Iwasaki	523	247	248	2.03	3.30
BLSW	454	246	247	2.00	3.75
AQ14K	441	247	249	2.03	2.62
SUN10K	420	246	249	1.98	3.68
EVC10K	417	247	248	2.01	3.44
Ushio	383	246	248	2.04	2.67
HM14K	350	247	248	2.03	2.60
BL10K	294	246	247	1.94	2.70
SBURST12K	283	247	250	2.01	2.21
SUN14K	257	246	247	2.01	2.15
CV12K	252	246	248	2.00	2.92
XM20K	246	246	334	2.03	3.31
SUN20K	239	246	247	2.00	2.02
XM15K	209	251	253	2.07	2.09
PFO 13K	185	246	248	2.00	3.35

Legend: AB13K = Aqualine Buske (AB) 13000K, (also labeled 10000K in some stores, but they are the same lamp), AQ14K = Aquaconnect 14K, BL10K = Blue Line 10000K (also called 10K+), BLSW = Blue Line Super White MH, CV10K = Coralvue 10000K, CV15K = Coralvue 15000K, CV20K = Coralvue 20000K, EVC10K = EVC Technologies 10000K, HM14K = Hamilton 14000K, Iwasaki = Iwasaki 6500K MH, PFO13K = PFO 13000K, Radium - Radium also called Radium 20000K, SBURST12K = Sunburst 12000K, SUN10K= Sun Aquatics 10000K, SUN20K = Sun Aquatics 20000k, SUN14K = Sun Aquatics 14000K, Ushio = Ushio 10000K, XM10K = XM10000K MH, XM20K = XM20000K MH

A graph showing the average PAR value for each lamp from all the ballasts.

Bulb Appearance

The picture on the left is with actinics off. The picture in
the middle is with two 140-watt URI VHO actinics on.

Conclusions

When reading through all these data, the first thing to become apparent is that they are voluminous. Because of this, we need to have a goal in mind when viewing the data. The information can mean different things to different people. Looking at the data to simply find the best lamp will not yield good results. Rather, think long and hard about what you are keeping, what color lamp you enjoy, and what kind of ballast you are looking for.

In my case, I did change my lighting system after doing these tests. After starting the testing, I wanted to find a lamp/ballast combination that put out as much PAR as the Ushio/HQI ballast combination I had been using. Finally, I settled on the combination of the XM10K lamp and the Icecap electronic ballast. The Ushio on an HQI ballast still puts out about 15% more PAR than the XM10K running on an Icecap, but I felt that it was close enough. After making the switch my colors and growth were just about the same as before. The main differences are less heat produced and less electricity consumed.

Someone else's goal might be different, but it will make your life much easier if you have an idea of what you are trying to achieve when interpreting the data.

Acknowledgements

I would like to thank PFO Lighting, DIY Reef, Premium Aquatics, HelloLights, Ocean Encounters, Icecap, Champion Lighting, Sun Aquatics, Sanjay Joshi and all the hobbyists who lent me lamps and let me ask them questions about how to do these tests. Without all of you, I never would have been able to gather this information and pass it on.