Metal Halide PAR map - DIY measurement

[karimwassef]

I decided to use my Apex PMK PAR meter just before changing out my year old MH bulbs to see the before and after distribution of light.

My tank's top aperture is ~32" x 96" and I use 3 x 400W MH pendants with the bulb center 12" over the tank lip surface (~1" above the water overflow surface).

I used two different kinds of MH. On the two side MHs are generic horticulture pendants running single sided bulbs 10K XM 400W. Looking from the back, the leftmost pendant is 2" higher ~14".

In the middle is a Hamilton Cayman Sun pendant running a double ended bulb 14K Hamilton 400W. I keep most of my special SPS under this light.

I knew a year was WAY too long, but I decided to see how bad it was. I also expected that the Hamilton pendant would be better than the generic horticulture pendants.

Here's the setup:

<a href="http://s1062.photobucket.com/user/karimwassef/media/LED%20experiments/IMG_5358_zpsjm7cvdnb.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/LED%20experiments/IMG_5358_zpsjm7cvdnb.jpg" border="0" alt=" photo IMG_5358_zpsjm7cvdnb.jpg"/></a>

I use supplemental blue LED, but here's what it looks like without the LEDs. The two views are back and front, so the leftmost pendant from the back is the rightmost lit area from the front.

<a href="http://s1062.photobucket.com/user/karimwassef/media/IMG_5537_zpsvrdkkv0u.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/IMG_5537_zpsvrdkkv0u.jpg" border="0" alt=" photo IMG_5537_zpsvrdkkv0u.jpg"/></a>

Here's some more detail on the pendants and bulbs

<a href="http://s1062.photobucket.com/user/karimwassef/media/LED%20experiments/IMG_5355_zpsuzgrzi1n.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/LED%20experiments/IMG_5355_zpsuzgrzi1n.jpg" border="0" alt=" photo IMG_5355_zpsuzgrzi1n.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/LED%20experiments/IMG_5356_zpsoms838sa.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/LED%20experiments/IMG_5356_zpsoms838sa.jpg" border="0" alt=" photo IMG_5356_zpsoms838sa.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/LED%20experiments/IMG_5357_zpsddqdtiwf.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/LED%20experiments/IMG_5357_zpsddqdtiwf.jpg" border="0" alt=" photo IMG_5357_zpsddqdtiwf.jpg"/></a>

Yes - I know I can get better PAR by cleaning the glass, but I wanted a real measure of what I was getting and I don't get to clean the glass constantly.

[karimwassef]

I created an eggcrate grid and a PAR sensor on eggcrate so I can keep the pitch identical between measurements.

<a href="http://s1062.photobucket.com/user/karimwassef/media/LED%20experiments/IMG_5285_zpsddoen5uu.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/LED%20experiments/IMG_5285_zpsddoen5uu.jpg" border="0" alt=" photo IMG_5285_zpsddoen5uu.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/LED%20experiments/IMG_5354_zpstqukwgg2.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/LED%20experiments/IMG_5354_zpstqukwgg2.jpg" border="0" alt=" photo IMG_5354_zpstqukwgg2.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/LED%20experiments/IMG_5287_zpskcqp1yg2.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/LED%20experiments/IMG_5287_zpskcqp1yg2.jpg" border="0" alt=" photo IMG_5287_zpskcqp1yg2.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/LED%20experiments/IMG_5297_zpsg4q8c9ss.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/LED%20experiments/IMG_5297_zpsg4q8c9ss.jpg" border="0" alt=" photo IMG_5297_zpsg4q8c9ss.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/LED%20experiments/IMG_5345_zpsubpddcjw.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/LED%20experiments/IMG_5345_zpsubpddcjw.jpg" border="0" alt=" photo IMG_5345_zpsubpddcjw.jpg"/></a>

I created an image on the grid of the pendant outlines. This is to help line up the map with the pendant locations.

<a href="http://s1062.photobucket.com/user/karimwassef/media/LED%20experiments/IMG_5310_zpslj4fyggl.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/LED%20experiments/IMG_5310_zpslj4fyggl.jpg" border="0" alt=" photo IMG_5310_zpslj4fyggl.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/LED%20experiments/IMG_5316_zpsxfjcf61w.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/LED%20experiments/IMG_5316_zpsxfjcf61w.jpg" border="0" alt=" photo IMG_5316_zpsxfjcf61w.jpg"/></a>

This method allowed me to create a uniform map of PAR. So here's the PAR map before and after the bulb change

<a href="http://s1062.photobucket.com/user/karimwassef/media/LED%20experiments/Capture_zpsqflx1soi.png.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/LED%20experiments/Capture_zpsqflx1soi.png" border="0" alt=" photo Capture_zpsqflx1soi.png"/></a>

This is from the back of the tank, so the left of the map is the right lit area from the front of the tank. The bottom of the map is the back of the tank.

Few observations...

1. The PAR change over a year is ~ 40%-50% (half light lost). I was expecting 20%-30% so this was more significant than expected.

2. The different distribution due to the pendant type & bulb was also significant. I can't tell for sure if it's the pendant or dual/single ended or the color, but here's what I think... the Hamilton is a more focused reflector vs the horti pendants that are really intended to cover a much larger area (4' x 4'). So the horti pendants are basically wasting a lot of light outside the tank. They may suit a larger surface tank better. Also, the special pattern on the Hamilton creates a quad-mirror effect with four "hotspots". That's good to know and explains some growth variation.

3. The 2" height difference between the two horti pendants also seems too large (40%??). I use the same generic ballast company "Apollo" but purchased at different times. One is shorter than the other so there may be a change there too.

I just though this was cool for a hobbyist to be able to do and wanted to share. I'm going to do the same for my DIY LEDs, so this is basically my baseline. :D

[karimwassef]

I also wanted to capture the effect of height, so I made a little eggcrate scaffold and elevator. The little white zipties create hooks to hold the elevator in place.

<a href="http://s1062.photobucket.com/user/karimwassef/media/LED%20experiments/IMG_5361_zpsgio57gf7.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/LED%20experiments/IMG_5361_zpsgio57gf7.jpg" border="0" alt=" photo IMG_5361_zpsgio57gf7.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/LED%20experiments/IMG_5360_zpsdignrlcx.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/LED%20experiments/IMG_5360_zpsdignrlcx.jpg" border="0" alt=" photo IMG_5360_zpsdignrlcx.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/LED%20experiments/IMG_5363_zpsduygpi4z.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/LED%20experiments/IMG_5363_zpsduygpi4z.jpg" border="0" alt=" photo IMG_5363_zpsduygpi4z.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/LED%20experiments/IMG_5365_zpszxjs5jhc.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/LED%20experiments/IMG_5365_zpszxjs5jhc.jpg" border="0" alt=" photo IMG_5365_zpszxjs5jhc.jpg"/></a>

I can't get any closer than 3" (for obvious reasons)

<a href="http://s1062.photobucket.com/user/karimwassef/media/LED%20experiments/0_zpso6htw8jw.png.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/LED%20experiments/0_zpso6htw8jw.png" border="0" alt=" photo 0_zpso6htw8jw.png"/></a>

The curve fits the square law (where x is negative denotes distance down from the bulb center)

PAR (x) = 626953 / (9.8 - x)^2

I only ran this on the Hamilton bulb.

[sirreal63]

Nice work. Obviously the reflectors are a problem. 4-250 waters will give better coverage, more light and less heat.

[karimwassef]

I really like the Hamilton 400. I'll probably swap out the two horti for two Hamiltons and then use one of the wider horti pendants for my large quarantine Rubbermaid tub (48" x 36"). I might use the second for a mangrove patch 48" x 48".

It's not that they're bad reflectors. They're just not right for the job. They have a wide slant so their rays are not even directed straight downwards. They're made to sit several feet up and cover a 6' x 6' area.

The Hamilton is also more compact giving me more options to add supplemental LEDs around them.

[shifty51008]

that's weird, there has been studies on MH lights and they shouldn't be loseing that much light after a year. I change my bulbs every 18-20 months and never have issues.

[shifty51008]

http://www.advancedaquarist.com/2008/7/aafeature1

[karimwassef]

I remember Sanjay's work. I'm nowhere as sophisticated, but these are my results.

[karimwassef]

I'd like to hear from other PAR owner, maybe we can get some group data going.

Do you measure your PAR regularly?

How do you know you're getting equivalent readings (probe location)? I found out that moving 1/2" could be significant, so I mapped out the whole area.

Do you see the same hotspots if you have a Hamilton?

Did you find that dual vs single ended made a difference? How about choice of ballast? Or color (10k vs 14k)? How about manufacturer (XM, Ushio, Hamilton, Radium,...)

Anyone out there with LED PAR maps already? I'll post mine but would be great to get some references from others. To make it easy, we can use 12" as a reference (shortest distance source to probe).

With the availability of home PAR mapping, we could crowd source a huge amount of data.

[oreo57]

http://www.advancedaquarist.com/2008/7/aafeature1


That study is a bit of an anomaly...
A similar study by Allen Chantelios [3] using seven 250W DE Phoenix 14000K with Icecap electronic ballasts shows a similar trend of gradual decay over time - of an average of 25% over 12 months


And a bunch of others..
http://classic-archived-site-47330.web4.hubspot.com/Portals/47330/images/estimated%20lumen%20maintenance-resized-600.jpg
http://classic-archived-site-47330.web4.hubspot.com/hovey-electric-power-blog/bid/77000/Metal-Halide-The-Shocking-Truth-In-The-Life-Of-A-Metal-Halide-Bulb

http://www.lrc.rpi.edu/programs/nlpip/lightinganswers/mwmhl/characteristics2.asp
All lamps undergo some lumen depreciation, or decrease in light output, as a lamp is operated.

http://www.pnnl.gov/main/publications/external/technical_reports/PNNL-22727.pdf

[karimwassef]

So my results are just about right on the red curve.

[karimwassef]

Also - I'm using PAR instead of Lumen. Not sure how much effect that has.

[oreo57]

So my results are just about right on the red curve.
Well still a little extreme.. ;) Current "wisdom" is MH will hit 50% at
EOL, whatever the manuf believes that is.. ;) but is usually longer than 1 year in "hours"..
In the above case 2 1/2years (approx) @ 10hrs/day

Also - I'm using PAR instead of Lumen. Not sure how much effect that has.
Well output is output, but if there is a spectrum shift (lumens weighted in green, PAR overall in vis. spectrum) you could have a minor difference.
Based on Dr. Joshi's chart, spectrum decay from the higher energy phosphor decay (assumed) is there but not 25% more there but it would need to be better quantified.
http://www.personal.psu.edu/sbj4/aquarium/articles/MetalHalideLamps2_files/photo4_files/f2fig4.jpg

Losses in the 350-450 would be under-reported in "lumens" but the shift from 600 to 575 may up the figure.. Anyways PAR "is' what counts.. ;)
http://www.personal.psu.edu/sbj4/aquarium/articles/MetalHalideLamps2.htm

to be honest, if you would have asked me to guess at the output loss over a year 25% would have been my answer. What good is guessing though.. ;)

[karimwassef]

Hmm.. So another datapoint. I ran my halides from 7am to 10pm ~ 15hrs a day.

I was able to contain temperature and algae effects so the longer duration drove greater growth... I did have massive calcium uptake though.

[oreo57]

Hmm.. So another datapoint. I ran my halides from 7am to 10pm ~ 15hrs a day.

I was able to contain temperature and algae effects so the longer duration drove greater growth... I did have massive calcium uptake though.

so 5460hrs would put you at 35% lumen loss..Sort of corresponds, and well within the realm of possibility.. all things considered..;)
Others would put you at 20-25% "lumen" loss:
http://www.hazardouslocation.info/images/figure-4.jpg

I suppose running them "harder" shortens lifespan.. ;)

[karimwassef]

I'm using electronic ballasts though.

[oreo57]

I'm using electronic ballasts though.
huh, go figure.. According to this the advantage of electronic Ballasts is better lumen (PAR by extension) maintenance..
https://www.1000bulbs.com/pdf/Advance-HID-brochure.pdf

Improved lumen maintenance – the lamp and ballast system’s
ability to minimize light output depreciation over the life of the lamp
– is the most fundamental and significant benefit of electronic
HID ballasts, especially medium wattage (250 – 400W), high
frequency ballasts such as Advance’s DynaVision® ballast. At
8,000 hours of lamp life, DynaVision delivers a 30-50% improvement
in lumen maintenance over conventional HID systems (magnetic ballasts driving probe start metal halide lamps) and a 20% or more improvement over
pulse start systems. Conventional HID systems typically experience
a 50-60% fall-off in light output over the published life of the lamp

[karimwassef]

Yes. I think making our own PAR map measurements could help separate marketing hype from reality ... data is the answer.

[karimwassef]

Has anyone else considered creating a PAR map of their lights? If you collect the grid data on an eggcrate, I can show you how to make the color map in excel (or just do it if you need me to).

<a href="http://s1062.photobucket.com/user/karimwassef/media/LED%20experiments/Capture_zpsqflx1soi.png.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/LED%20experiments/Capture_zpsqflx1soi.png" border="0" alt=" photo Capture_zpsqflx1soi.png"/></a>

[shifty51008]

Has anyone else considered creating a PAR map of their lights? If you collect the grid data on an eggcrate, I can show you how to make the color map in excel (or just do it if you need me to).

<a href="http://s1062.photobucket.com/user/karimwassef/media/LED%20experiments/Capture_zpsqflx1soi.png.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/LED%20experiments/Capture_zpsqflx1soi.png" border="0" alt=" photo Capture_zpsqflx1soi.png"/></a>

I do like the looks of that map, great job. I'm deff gonna do that when I borrow a par meter again.

[karimwassef]

awesome. let me know how I can help.