Here's a plot of a four bulb UV. Kill rate is best close to the quartz sleeve. Multi bulb designs allow other bulbs to kill when the pest is too far from the other. Bulbs are spaced so that in between the four bulbs intensity is still good for the kill. That goes the same for the chamber wall. It keeps the pest close enough to the UV ray.
This is our guidance to treating surface water. Good info on UV's....
http://www.epa.gov/ogwdw/disinfection/lt2/pdfs/guide_lt2_uvguidance.pdf
I run uvc pond with stingrays lots years never got some parasites what do you have to say abut that T1?
I do not think you will find a uv big enough for a 500 gal so I expect you need more than 1. I think the tmc ones are the best, just visit a stockist and I suspect the answer you seek is two of the biggest ones and a suitable pump.
I do not understand t1's logic...I.e his logic suggests. if the flow rate needed for 1 is say 1000 and your pump does 4000 then you need 4xuv....I think you will just have 4 uvs with too high flow rate for the kill to work as the unit says 1000....adding more just means the first ones do nothing and the end one will only work if it drops the flow rate to the designated 1000? Splitting 4000 between 4 would work as they all end up with 1000 each but that would need a 4 way split and equalising. The problem is we have no real way of testing but basic physics would suggest the 4 in a row is long.
I use 2 in series simply as what might get past the first may get killed by the next but I use a very slow pass rate (this also has a by product of heating the water)
I can speak to tmc and see what they say.
I do not understand t1's logic...I.e his logic suggests. if the flow rate needed for 1 is say 1000 and your pump does 4000 then you need 4xuv....I think you will just have 4 uvs with too high flow rate for the kill to work as the unit says 1000....adding more just means the first ones do nothing and the end one will only work if it drops the flow rate to the designated 1000? Splitting 4000 between 4 would work as they all end up with 1000 each but that would need a 4 way split and equalising. The problem is we have no real way of testing but basic physics would suggest the 4 in a row is long.
I use 2 in series simply as what might get past the first may get killed by the next but I use a very slow pass rate (this also has a by product of heating the water)