Hanna Checker PO4

Are you having problems with algae? If not, I wouldn't worry about it. If you wanted to get it down even more you could try possibly vacuuming your sand bed or just do more frequent water changes.

 

http://reefkeeping.com/issues/2006-09/rhf/index.php

 

I wouldn't ignore that amount of PO4 if you are keeping SPS corals. PO4 is a known inhibitor of the calcification process and can be harmful long term.

Several scenarios are possible, you have a significant source of phosphate coming from someplace (rock, ca rxr media, carbon, decaying matter, food certainly contributes). Your RO system just isn't doing a good enough job, make sure your DI resin isn't exhausted. Your source water may be contaminated with high levels of PO4, especially if you're drawing from a rural well. Or your PO4 could have previously had elevated levels of PO4 (due to any one of these reasons) and although it's lower now, you are still exhuasting the GFO quickly and it needs to be replaced already even if you recently changed it.

That RC article is a great read also!

 

Jim, have you thought about carbon dosing? (bio pellets)

 

+1

 

Keep in mind those checkers can have a miss reading of +-0.05 so your actual level can be anywhere between 0.14-0.04

 

The accuracy of the device is reported as the mean standard deviation. I'm pretty sure the reported mean standard deviation is actually +/-0.04ppm as indicated by the literature from Hanna on the device itself. Indeed this level of accuracy is the same as the $240 dollar model, the differences are probably percision based.
Note that this mean standard deviation is the square root of the variance between many data points and the actual value, we can therefore determine that the mean variance= std-dev^2= 0.04^2= 0.0016 This mean variance (remember variance is a non-discrete measurement of difference between data points and a discrete actual value) is derrived from many trials, where mean variance= (n1 + n2 + n3)/n This low variance could be the result of, for example (4 + 1 + 1.2 + .8 + 1.1 + 1)/6= 1.51 Of course the same value can be derrived from a pool of much smaller variations ex; (1.51+1.61+1.41+1.71+1.31+1.51)/6= 1.51 and you still derrive the same mean variance.
This second example is a display of much higher degree of percision with few radical outliers as in the first example. The expected deviation for any given measurement= std-dev/n^1/2. (n^1/2= square root of n) in this case each subsequent repeatable trial reduces the expected deviation. Again this assumes that each trial yields the same result and the instrument has a high degree of percision.
That's your statistics minute, I'm Andy Long signing out! haha

 

Great thread about these meters. Start at post #68 and see the results he posts a little later.

http://www.reef2reef.com/forums/general-reef-discussions/48866-lets-talk-about-hanna-meters-5.html

 

I was erroneous when I said before that the accuracy of +/-0.04 was = mean standard deviation. The std-dev is actually 2x 0.04 (both the + 0.04 and the - 0.04)
Again let me stress the importance of analyzing the accuracy statment and not jumping to conclusions about the capabilities or utility of the measurements provided by this instrument:
Let's say the generally agreable preferred alotment of PO4 for an SPS reef tank is something like 0.06ppm, if your mean standard deviation (stated accuracy) is 0.08ppm then the mean average variance of any measurement is= std-dev^2= 0.0064. meaning, you are likely to have an variance of just 0.0064 per reading. Within the realm of this 0.06ppm window, 0.0064/0.06*(100)= 10.666% error. The percent deviation is even less considering your 0.09ppm reading (just 7%!). In other words, your value is likely to actually be between 0.0837-0.0963ppm.
If you're suspicious of incorrect results, repeat the measurement several times, until you arrive at a proficient degree of certainty. Not only is reproduction of values a display of the intruments percision, but it also can provide confidence in the accuracy of the result (expected std-dev= std-dev/n^1/2). Of course at 0.08 std-dev it takes 16 trials to reduce this expected deviation down to 0.02 with a reported value of 0.08=std-dev, and while reagents are cheap, they're not that cheap!
I could go on, but suffice it to say these instruments are reliable, and specific enough to be valuable to the hobbyist, afterall they were developed with us in mind!