T5's Spectrum shift

jezzeaepi

New member
When I started researching what lighting I was going to use, I ran accross the idea of bulbs "shifting spectrum". No matter where I looked, I could not find any real data on how fast a t5 bulb shifts its spectrum. All I could find was anecdotal evidence that supported both sides of the camp. On one hand we have some seriously succesfull t5 users*cough* Iwan *cough* who recomend changing bulbs every 6-8 months. On the other hand we have people who have corals that are doing fine under 12month+ t5's.

The only way to solve this for sure was to do spectral readings on the bulbs as they age. Keep in mind I am no scientist, just a dude with a fish tank. Thankfully my roomate is going to school for a science related career so I talked him about my quandry with the flouro spectral shifts. He just happens to do lab work in one of the labs at the local University that has a spectrometer. After explaining our idea for the experiment to the professor, we got permission to run the tests.

The tests were all run on brand new bulbs. I am using a brand new 36inch aquactinics fixture. An initial spectral reading was taken to block out any background light. The readings were takin once the intensity leveled off. This seemed to happen after about a minute. I wanted to leave them running for longer before taking the reading, but we had some techinical difficulties in the beginning and the building was about to close, so we had to make some compromises. If you ahve any questions about the techincal aspect of the testing let me know, Ill pass them on to my roomate.

This is my first attempt at making graphs from data in excel so if you have any suggestions let me know.

My original spectra:
104807uvlsa.JPG

104807uvlaquasun.JPG

*Note that UVL hasnt published their spectra yet, so those are new info.
104807atiblueplus-med.JPG

104807atiaquablue.JPG


I noticed a difference in relative intensity between the published ATI bulbs, and what I got. Next time I will run the bulbs longer to see if that relative intensity changes at all.

I will take more readings once the bulbs have finished their 100 hour burn in which is next Monday. After that, I will take readings once a month and post them here. Unfortunetly my access to the spectrometer will run out in 9 months. Im hoping to have some meaningfull data by then. Im also going to see waht I can do about getting some older t5's to test. Even better would be if I could get some used overdriven bulbs as well.


Peace all
 
I have the info from the 100 hour run sitting at home, I just havent had time to post it yet. For the few of you who are interested I should have it up by tomorrow.
 
Hi jezzeaepi,

I did a simple test myself using lux meter as i don't have the luxury of getting my hands on a spectormeter. I notice that using different light set, the readings can be vastly different even thought i am using back the same tubes.....

could that be the reason why your meaurement would be different from their published value?

Which brings about an interesting point. Has anyone done any studies or comparsion of different light sets (using the same tubes) before? for example, Tek vs ATI powermod? this will be interesting
 
Great work for a basic look at the bulbs over time.

The one thing I would add (not likely easy or cost effective for yoi guys) is clean power and measurement to ensure that the bulbs are being run with the same parameters each test.

Keep up the good work... this is interesting to say the least.
 
Also... can we get all the graphs posted the same size? Let me know if you need someplace to host them ( I own my own servers).

Bean
 
I scanned the data in the xls format and every single peak that I looked at was in the exact same place. Keep in mind that the machine measures the relative intensity approximately 3 times for each nanometer of the spectrum. Therefore a shift should be apparent fairly quickly. The xls that is outputted by the machine contains 2050 points of data between the spectrum of 179.49 to 868.41.

Ill update the graphs at a more timely hour =P
 
I think the graphs would be easier to compare if they all had the same scale gradations and relative size between those gradations...

That way the could be overlaid in photoshop or similar.
 
Back
Top