New Digital Phospate Test Kit

[cabrego]

Hi,

 

Just wondering if anyone has actually used one of the new phophate colormeters that are for sale. I believe Hanna makes them. What are your thoughts on it? There were mixed reviews on Reef Central.

[Origami]

I have it and I like it. I cannot comment on the accuracy, however, as I don't have a reference solution against which to measure it. Hanna's internal tests, however, were generally favorable according to the discussion over at RC. The device is convenient and, in the long run, cheaper. The reagents last longer (they have a 3-year shelf life) and are packaged in individual doses. It's also easier to read than color charts.

[johnnybv]

Hi,

 

Just wondering if anyone has actually used one of the new phophate colormeters that are for sale. I believe Hanna makes them. What are your thoughts on it? There were mixed reviews on Reef Central.

 

We are using them in store and they are great! Hanna individually tested each unit with a standard test solution that was made fresh every 2 hours, so I would have to say they are accurate. Untill now you couldnt touch a decent P04 test kit for less than $100 (Deltec) and even that one was pretty involved, I am pleased with both the ease and accuracy of the units.

John

Edited April 27, 2010 by johnnybv

[cabrego]

I have it and I like it. I cannot comment on the accuracy, however, as I don't have a reference solution against which to measure it. Hanna's internal tests, however, were generally favorable according to the discussion over at RC. The device is convenient and, in the long run, cheaper. The reagents last longer (they have a 3-year shelf life) and are packaged in individual doses. It's also easier to read than color charts.

 

In my view the value in having one of these meters is not just ease of use but it is also accuracy. If I don't have confidence in the measurement, then the meter is just giving an indication and 'normal' liquid test do a decent job of that. You don't know if you have .25 or .5 ppm, but you know you don't have 1 or 2 ppm. It would be interesting to see documentation from Hanna's testing if it public information. Just to be clear, I am not saying the product is not accurate, I am just wondering how you know if your meter is working accurately if you have one. Is there a calibration procedure with standard solutions?

 

Here is a link from RC that you might be talking about, posts 192 and 196 are particularly interesting. I am not sure if post 192's standard mixing methodology is accurate but his results are consistent. 196 is a comparison against the hanna photometer.

 

http://www.reefcentral.com/forums/showthread.php?p=16986538#post16986538

 

Anyone know how to mix a phosphate standard? I would be glad put together a test with someone.

 

We are using them in store and they are great! Hanna individually tested each unit with a standard test solution that was made fresh every 2 hours, so I would have to say they are accurate. Untill now you couldnt touch a decent P04 test kit for less than $100 (Deltec) and even that one was pretty involved, I am pleased with both the ease and accuracy of the units.

John

 

Glad to hear they are working well for BRK. Are these kits in stock for the masses, or do you have to be 'on the list'?

[rbonick]

Here's a review from reef builder.

 

http://reefbuilders.com/2010/04/20/hanna-phosphate-checker-review-notable-compromises-still-worth-the-resolution-and-portability/

 

According, to their test, the portable unit was close in accuracy to the 200 dollar unit

[cabrego]

Here's a review from reef builder.

 

http://reefbuilders.com/2010/04/20/hanna-phosphate-checker-review-notable-compromises-still-worth-the-resolution-and-portability/

 

According, to their test, the portable unit was close in accuracy to the 200 dollar unit

 

 

thanks for the link. Interesting results, one thing of note is a recipe to generate a phosphate standard using on the shelf products.

[johnnybv]

Again, Hanna individually tests each and every unit that leaves their facility with a test solution that is made fresh every 2 hours. This means that the unit is within the accuracy that is in its specs which are "Adaptation of Standard Method 4500-P E Accuracy 4%

[cabrego]

Again, Hanna individually tests each and every unit that leaves their facility with a test solution that is made fresh every 2 hours. This means that the unit is within the accuracy that is in its specs which are "Adaptation of Standard Method 4500-P E Accuracy 4%

[johnnybv]

Not even going to begin to argue.....

[Origami]

By the way, from what I understand about the meter - it is not accurate to 4%, it is accurate to 4% of the reading PLUS +/- .04 ppm, so if if you read 1 ppm you know you are somewhere between .92 and 1.08 ppm, which still seems pretty good. So in this case you could be as much as 8% off and if you are measureing down to .25 ppm you can be as much as 25% off. Regardless, for the price I think it seems like a good product. I do think it is important to understand how to interpret the results though.

 

 

There's a good thread on this over at RC. It covers this topic.

 

http://www.reefcentral.com/forums/showthread.php?t=1804518

 

You'll see there that the worst-case accuracy is the specification at full-scale. It is much more accurate in close but, as you've observed, you can play with the math to make it seem worse.

 

For example, theoretically, you could get a 0.04 ppm reading for an RO/DI sample with 0 ppm phosphate. (The meters tested at Hanna, by the way, all come up dead on accurate at 0 ppm. My example is just to show how you can play with these accuracy numbers and tell a story that's less than illuminating.) Anyways, in that case, the % error is infinite, isn't it? Who would want to buy that?

 

Likewise, if the sample is really at 0.01 ppm, and the measurement given is 0.05 ppm, the meter error is 400%. The point is, the meter is still pretty darn good, even though it sounds horribly inaccurate if you couch accuracy solely terms of percentage error (because of the non-zero - the +/- 0.04 part - bias error that's being accounted for in Hanna's specification).

 

The accuracy specification is, by the way, the same as for their more expensive meters. See http://www.hannainst.com/usa/prods2.cfm?id=030003&ProdCode=HI%2093713 to see the specification for their more expensive 93713 (low range) model. It reads, "Accuracy ±0.04 mg/L ±4% of reading."

 

In my opinion, it's a really nice option to have and, over the long run, more economical than chemical tests that rely on color charts (which are often incredibly difficult to read, vary under different lighting, and still suffer from inaccuracies) to give you readings.

[cabrego]

There's a good thread on this over at RC. It covers this topic.

 

http://www.reefcentral.com/forums/showthread.php?t=1804518

 

You'll see there that the worst-case accuracy is the specification at full-scale. It is much more accurate in close but, as you've observed, you can play with the math to make it seem worse.

 

For example, theoretically, you could get a 0.04 ppm reading for an RO/DI sample with 0 ppm phosphate. (The meters tested at Hanna, by the way, all come up dead on accurate at 0 ppm. My example is just to show how you can play with these accuracy numbers and tell a story that's less than illuminating.) Anyways, in that case, the % error is infinite, isn't it? Who would want to buy that?

 

Likewise, if the sample is really at 0.01 ppm, and the measurement given is 0.05 ppm, the meter error is 400%. The point is, the meter is still pretty darn good, even though it sounds horribly inaccurate if you couch accuracy solely terms of percentage error (because of the non-zero - the +/- 0.04 part - bias error that's being accounted for in Hanna's specification).

 

The accuracy specification is, by the way, the same as for their more expensive meters. See http://www.hannainst.com/usa/prods2.cfm?id=030003&ProdCode=HI%2093713 to see the specification for their more expensive 93713 (low range) model. It reads, "Accuracy ±0.04 mg/L ±4% of reading."

 

In my opinion, it's a really nice option to have and, over the long run, more economical than chemical tests that rely on color charts (which are often incredibly difficult to read, vary under different lighting, and still suffer from inaccuracies) to give you readings.

 

I generally agree with everything you have stated! It looks like the link you posted to the RC is the same as I posted, so we are on the same page so to speak!

 

The curiousity has finally overcome me and I ordered my very own phosphate test kit from Marine Depot (there were only 4 left and i got it for 39.99), I also bought a bottle of of 100 ppm Hach Phosphate standard. The standard will be used to test the Hanna meter and it could also be used to illustrate the effectiveness of running GFO. I guess it is just part of my nature no want to know how/if things work.

 

Many of the marine societies are actively testing/documenting the Hanna units at meetings and social events and posting the results online, so why not us? anyway, i will keep you guys updated...

Edited April 29, 2010 by cabrego

[Origami]

I also bought a bottle of of 100 ppm Hach Phosphate standard. The standard will be used to test the Hanna meter and it could also be used to illustrate the effectiveness of running GFO. I guess it is just part of my nature no want to know how/if things work.

 

Don't forget that the range on the colorimeter is to 2.5 ppm so you're going to have to really dilute the phosphate standard that you've selected (1:50 to 1:100 or more). The reading that you get is going to be very dependent upon an accurate dilution. If you have access to a calibrated photometer, it may be better, once you've got a diluted solution, to compare the readings you get from the colorimter to those taken from the calibrated device. Even then, though, it would be statistically more relevant if you had a batch of colorimeters to test. The spec from Hanna must take into account not only the accuracy of individual units, but that of a large part of their production output in order to keep waste and production costs down.

[Origami]

LOL, cabrego. I came across this dialogue between you and Cliff: http://www.reefcentral.com/forums/showthread.php?p=17024239

 

Yes, it would have been better to get the 10 ppm standard. Marine Depot's pretty good about changing things up. Can you call them and intercept the order to make the change?

 

http://www.hach.com/hc/search.product.details.invoker/PackagingCode=2059703/NewLinkLabel=Phosphate+Standard+Solutions+as+PO4+%28NIST%29+3+mg&frasl%3BL+2.9+L

 

Hach also makes a 3 ppm standard. However, cutting a 10 ppm standard accurately will not be difficult using commonly available equipment.

[cabrego]

LOL, cabrego. I came across this dialogue between you and Cliff: http://www.reefcentral.com/forums/showthread.php?p=17024239

 

Yes, it would have been better to get the 10 ppm standard. Marine Depot's pretty good about changing things up. Can you call them and intercept the order to make the change?

 

http://www.hach.com/hc/search.product.details.invoker/PackagingCode=2059703/NewLinkLabel=Phosphate+Standard+Solutions+as+PO4+%28NIST%29+3+mg&frasl%3BL+2.9+L

 

Hach also makes a 3 ppm standard. However, cutting a 10 ppm standard accurately will not be difficult using commonly available equipment.

 

 

Well, I ordered directly from Hach so I don't know if I can change the order. Do you think it will be difficult to use the 100 ppm standard?

 

I am thinking 1 ml can be measured very accurately with a syringe and 100 ml can also be measured very accurately, so 1 ml of Standard in 100 ml of RODI give 100 ml of 1 ppm Phosphate solution. I will also have a fair amount left over in case I am not confident in my measuring techniques. One of the expensive meters would be good to have for this test too, but I don't know anyone locally that owns one.

 

 

Anyone have high precision pipettes or beakers I can borrow? Else I will probably start off with the syringes we have for our medicine and other aquarium test kits. I think my Mg test kit has a 1 ml syringe.

 

Any flaws with my logic?

[Origami]

If you can accurately measure 100 ml and 1.00 ml, you'll be fine. Just realize that the accuracy of your dilution efforts will factor into any measurement you make. For example, a 1 ml error in your 100 ml dilutant will yield a 1% error. The same problem exists when measuring out your 100 ppm standard. Accurately measuring these so that you get the desired ratio may be tougher than you realize. Read on. (By the way, you want to mix 1 ml of your 100 ppm reference with 99 ml (not 100 ml) to hit your 1 ppm target. Pretest your RO/DI to make sure that it's reading 0 ppm phosphates. It should, but it's amongst the first ions to be re-released when the DI resin depletes.)

 

Here's a quick sampling on what happens if your dilution is inaccurate (I've only carried things out to the 2nd decimal place):

 

1 ml 100 ppm standard + 98 ml RO/DI = 1.01 ppm standard

1 ml 100 ppm standard + 99 ml RO/DI = 1.00 ppm standard

1 ml 100 ppm standard + 100 ml RO/DI = 1.01 ppm standard

 

The same accuracy issues exist for the the standard, though:

 

1.01 ml 100 ppm standard + 99 ml RO/DI = 1.01 ppm standard

1.00 ml 100 ppm standard + 99 ml RO/DI = 1.00 ppm standard

0.99 ml 100 ppm standard + 99 ml RO/DI = 0.99 ppm standard

 

Notice how a very, very small amount of deviation from the 1 ml measurement yields an error on the order of the accuracy of the device that you're trying to test. That's where the problem lies. What happens if your syringe is just 4% off? Consider this: There are 20 "standard" drops in a milliliter. That means that if you're just 1 drop off that 1 ml measurement, you're off by 5%.

 

So, let's say that your syringe measurement is just 2% off (less than +/- 1/2 drop of error). Let's also say that your ability to accurately measure 100 ml is off by just 1 ml (100 ml +/- 1 ml). The result is that, because of the relatively high concentration of the reference, that you're "standard" solution, obtained through dilution, can be off as much as +/- 3%. That's not very good when you're trying to test a meter that's supposed to be accurate to +/- 4%. (When you do measure an error, how much do you attribute to the device and how much to the "standard?")

 

The best thing to do, if you trust the accuracy of the standard that you purchase to be accurate, is to buy a standard that's close that requires minimum dilution and start from there. Remember, these "standards" are calibrated against calibrated meters which, if done right, ultimately trace their accuracy back to NIST standards.

 

Another alternative is to test against a calibrated meter (I've already mentioned this) and compare the results.

[cabrego]

If you can accurately measure 100 ml and 1.00 ml, you'll be fine. Just realize that the accuracy of your dilution efforts will factor into any measurement you make. For example, a 1 ml error in your 100 ml dilutant will yield a 1% error. The same problem exists when measuring out your 100 ppm standard. Accurately measuring these so that you get the desired ratio may be tougher than you realize. Read on. (By the way, you want to mix 1 ml of your 100 ppm reference with 99 ml (not 100 ml) to hit your 1 ppm target. Pretest your RO/DI to make sure that it's reading 0 ppm phosphates. It should, but it's amongst the first ions to be re-released when the DI resin depletes.)

 

Here's a quick sampling on what happens if your dilution is inaccurate (I've only carried things out to the 2nd decimal place):

 

1 ml 100 ppm standard + 98 ml RO/DI = 1.01 ppm standard

1 ml 100 ppm standard + 99 ml RO/DI = 1.00 ppm standard

1 ml 100 ppm standard + 100 ml RO/DI = 1.01 ppm standard

 

The same accuracy issues exist for the the standard, though:

 

1.01 ml 100 ppm standard + 99 ml RO/DI = 1.01 ppm standard

1.00 ml 100 ppm standard + 99 ml RO/DI = 1.00 ppm standard

0.99 ml 100 ppm standard + 99 ml RO/DI = 0.99 ppm standard

 

Notice how a very, very small amount of deviation from the 1 ml measurement yields an error on the order of the accuracy of the device that you're trying to test. That's where the problem lies. What happens if your syringe is just 4% off? Consider this: There are 20 "standard" drops in a milliliter. That means that if you're just 1 drop off that 1 ml measurement, you're off by 5%.

 

So, let's say that your syringe measurement is just 2% off (less than +/- 1/2 drop of error). Let's also say that your ability to accurately measure 100 ml is off by just 1 ml (100 ml +/- 1 ml). The result is that, because of the relatively high concentration of the reference, that you're "standard" solution, obtained through dilution, can be off as much as +/- 3%. That's not very good when you're trying to test a meter that's supposed to be accurate to +/- 4%. (When you do measure an error, how much do you attribute to the device and how much to the "standard?")

 

The best thing to do, if you trust the accuracy of the standard that you purchase to be accurate, is to buy a standard that's close that requires minimum dilution and start from there. Another alternative is to test against a calibrated meter (I've already mentioned this) and compare the results.

 

Thanks for the input, those are all factors that must be considered. IMO the only way around those factors is to not cut the standard. As soon as you cut any concentration of the standard down (regardless of the concentration) you allow measurement error to creep into your methodology. You seem to have more confidence in the 10 ppm solution, is there a way to minimize measurement error with the more dilute phosphate standards? I can give hach a call and see if they can hold my order if it makes sense to do that..

[Origami]

Thanks for the input, those are all factors that must be considered. IMO the only way around those factors is to not cut the standard. As soon as you cut any concentration of the standard down (regardless of the concentration) you allow measurement error to creep into your methodology. You seem to have more confidence in the 10 ppm solution, is there a way to minimize measurement error with the more dilute phosphate standards? I can give hach a call and see if they can hold my order if it makes sense to do that..

 

The link I provided above was to a 3 ppm standard that Hach makes. Cutting it 50% gives you a 1.5 ppm solution that is nearly midway into the range of the Hanna colorimeter.

 

The nice thing about a 50% dilution is that you can use the same measurement device to measure out the standard and the dilutant. This helps to bring the error in the measuring system under control because the same error is propagated into both the standard and the dilutant. (You didn't have that when you were using the syringe and the beaker or graduated cylinder.) For example, 5 ml of standard + 5 ml of dilutant, yields the same concentration as adding 5.1 ml of standard to 5.1 ml of dilutant. Both result in a 50% dilution. So, if you've got a reasonably precise 5 ml syringe or laboratory pipette, you should be able to do pretty well as long as you're consistent in the technique that you're using to measure both components of your diluted standard.

[cabrego]

The link I provided above was to a 3 ppm standard that Hach makes. Cutting it 50% gives you a 1.5 ppm solution that is nearly midway into the range of the Hanna colorimeter.

 

The nice thing about a 50% dilution is that you can use the same measurement device to measure out the standard and the dilutant. This helps to bring the error in the measuring system under control because the same error is propagated into both the standard and the dilutant. (You didn't have that when you were using the syringe and the beaker or graduated cylinder.) For example, 5 ml of standard + 5 ml of dilutant, yields the same concentration as adding 5.1 ml of standard to 5.1 ml of dilutant. Both result in a 50% dilution. So, if you've got a reasonably precise 5 ml syringe or laboratory pipette, you should be able to do pretty well as long as you're consistent in the technique that you're using to measure both components of your diluted standard.

 

very good point. Hach also carries a 1 ppm solution maybe I will go with that one instead. I can use it uncut for 1ppm and use a 50/50 dilute for .5 ppm. The questions about interpreting test results are usually in the sub 1 ppm range, so 1 ppm solution might be a good compromise between cost and usefulness.what do you think?

[Origami]

As long as Hach can trace their references back to a NIST reference, that's good. I'm sure they can, by the way.

 

The next thing to consider is that you're going to have a single data point. That is, you'll only be testing a single meter and, worse, a single meter at a single point. That will give you some confidence in the general accuracy of the meter in the range of 1.0 ppm, but it won't give you much of an understanding of the accuracy of device over it's measurement range.

 

Let's take a simple device that we all use: A refractometer. Some people say we can calibrate these with RO/DI and dialing the meter into 1.00 density. But that's not quite right. It's better to use a calibrated solution that's close to what you intend to measure. In other words, use a 1.0264 calibration fluid (35 ppt) and calibrate your refractometer against that instead. This compensates for many error sources in the local vicinity of the measurement while ignoring the accuracy in places you're less concerned with.

 

For a measurement system using a linear approximation, neglecting non-linear errors, you have a bias (or constant) error and a slope error to contend with. The bias exists at all measurements while the slope error grows as the magnitude of the measurement grows. Thus, the error at any particular point of the curve can be different than at another point. By calibrating with a known source at a particular point, you can minimize errors in subsequent measurements around that point.

 

 

For a "real" assessment of these devices, you really need a sample size (that is, many of these devices) that will allow you to have a statistically relevant measurement of the accuracy of the meter design, and calibrated standards against which to take measurements.

 

(I've got to run. I have to be home by 3 PM today....)

[cabrego]

As long as Hach can trace their references back to a NIST reference, that's good. I'm sure they can, by the way.

 

The next thing to consider is that you're going to have a single data point. That is, you'll only be testing a single meter and, worse, a single meter at a single point. That will give you some confidence in the general accuracy of the meter in the range of 1.0 ppm, but it won't give you much of an understanding of the accuracy of device over it's measurement range.

 

Let's take a simple device that we all use: A refractometer. Some people say we can calibrate these with RO/DI and dialing the meter into 1.00 density. But that's not quite right. It's better to use a calibrated solution that's close to what you intend to measure. In other words, use a 1.0264 calibration fluid (35 ppt) and calibrate your refractometer against that instead. This compensates for many error sources in the local vicinity of the measurement while ignoring the accuracy in places you're less concerned with.

 

For a measurement system using a linear approximation, neglecting non-linear errors, you have a bias (or constant) error and a slope error to contend with. The bias exists at all measurements while the slope error grows as the magnitude of the measurement grows. Thus, the error at any particular point of the curve can be different than at another point. By calibrating with a known source at a particular point, you can minimize errors in subsequent measurements around that point.

 

 

For a "real" assessment of these devices, you really need a sample size (that is, many of these devices) that will allow you to have a statistically relevant measurement of the accuracy of the meter design, and calibrated standards against which to take measurements.

 

(I've got to run. I have to be home by 3 PM today....)

 

I certainly will only have a measurement from one unit and at best I will have two measurements that I can have a reasonable degree of confidence in. I am certainly not trying to conduct a "real" assessment of these meters-I would expect Hanna to have done this prior to putting them on the market. Especially with the new batch of units they had to shove out the door to meet the unexpected demand, right? My test will be closer to a QC check on my personal meter-after all the value for me is to really understand the number I get out of the unit. I am also sure Hanna appreciates users giving them feed back too. If the meter is within the stated range of 4% +/- .04 ppm at the two points I measure I think I will be satisfied.

 

Your comment regarding measuremnets of linear measurement device, linear errors, and statistical signficance is a whole other topic that would require us to ensure we have our nomenclature strait with regard to accuracy, precision, standard error, etc. I am glad to see there is at least 1 person on WAMAS willing to delve into the details with me...

 

p.s. I changed my order to the 1ppm standard, should be receiving it next week.

[Origami]

I certainly will only have a measurement from one unit and at best I will have two measurements that I can have a reasonable degree of confidence in. I am certainly not trying to conduct a "real" assessment of these meters-I would expect Hanna to have done this prior to putting them on the market. Especially with the new batch of units they had to shove out the door to meet the unexpected demand, right? My test will be closer to a QC check on my personal meter-after all the value for me is to really understand the number I get out of the unit. I am also sure Hanna appreciates users giving them feed back too. If the meter is within the stated range of 4% +/- .04 ppm at the two points I measure I think I will be satisfied.

 

Your comment regarding measuremnets of linear measurement device, linear errors, and statistical signficance is a whole other topic that would require us to ensure we have our nomenclature strait with regard to accuracy, precision, standard error, etc. I am glad to see there is at least 1 person on WAMAS willing to delve into the details with me...

 

p.s. I changed my order to the 1ppm standard, should be receiving it next week.

 

I'm glad that you were able to switch over to the 1 ppm standard. For the most repeatable, accurate test, you should use the same cuvette and try to orient it the same way between the reference and the measurement run.

 

I certainly agree with you on keeping a standard around for the occasional QC of test devices. You can't blindly trust that devices are always giving you the right reading. For example, I returned from a week's vacation (with my father-in-law watching the house) to find that the pH in my display was lower than usual. Over the next several days, it continued to drop. Everything in the tank looked fine, so I recalibrated the probe. But still, the measurement continued to drop. It went from 8.2 to 8.0 to 7.8 to 7.2 ... I knew enough to know that I couldn't trust the probe and a regular old chemical test proved that the pH was just fine. Still, I wanted to see how far the probe would go until it stabilized. It took more than a week for it to stabilize, but when it finally did, the ACIII was reporting a pH of 2.65! The moral of the story is to always cross-check results when the data seems contradictory or doesn't match observations. Kept cool and in the dark, the phosphate standard should be stable and last a good long while.

 

By looking at the test protocol and the limited amount that I've read about the colorimeter, it's my opinion that it integrates and measures the transmitted light of a specific frequency or frequency range (according to the LED used), comparing the final measurement to the reference measurement, and mapping the difference to a measurement using either a mathematical equation, a fixed reference (look-up) table, or possibly a post-manufacturing, calibrated reference table. That's why little things, like ensuring the cuvette is clean and even oriented similarly between the first measurement, that the solution is not full of microbubbles (from overly aggressive shaking, for example), or any other thing that might alter the overall transmittance besides that which indicates the phosphate level, are important to ensure repeatability. I guess what I'm saying is that error can come from a lot of places, internal, external or procedural. Ultimately, attribution and control of the measured error attributed to the device is what the manufacturer is after when publishing a specification if they're concerned about maintaining quality. In this regard, the specification needs to be set loose enough to allow decent yields through manufacturing, but not so loose that it doesn't meet the customer's need.

 

Let us know how your test comes out.