victoly Posted February 19, 2015 Share Posted February 19, 2015 http://www.reefs.com/magazine/?do=wp_post&articleId=159 Great article by someone who knows his stuff. Three seawater standards were sent to triton and analyzed. Statistical analysis were performed. Some were close to standard, others not so much. TL;DR Precision? Yes. Accuracy No. Heavy marketing bias. Hell yes. Quote Link to comment Share on other sites More sharing options...
Juiceman Posted February 19, 2015 Share Posted February 19, 2015 Im lost, what does all that say? Quote Link to comment Share on other sites More sharing options...
FarmerTy Posted February 19, 2015 Share Posted February 19, 2015 Oh man! This looks like some good reading. Thanks Victoly! Is it wrong to get excited about lab data? Quote Link to comment Share on other sites More sharing options...
+brian.srock Posted February 19, 2015 Share Posted February 19, 2015 It's because you need the mittens on the kittens Juiceman. Quote Link to comment Share on other sites More sharing options...
victoly Posted February 19, 2015 Author Share Posted February 19, 2015 There were mittens, but they were not on the kitten. And the people selling the kitten with mittens PROMISED the mittens would be on the kitten. In fact, that's how they sold you on buying the mittened kitten. Quote Link to comment Share on other sites More sharing options...
+brian.srock Posted February 19, 2015 Share Posted February 19, 2015 3 Quote Link to comment Share on other sites More sharing options...
Juiceman Posted February 19, 2015 Share Posted February 19, 2015 .......... 1 Quote Link to comment Share on other sites More sharing options...
victoly Posted February 19, 2015 Author Share Posted February 19, 2015 Triton over promised and under delivered. Samples with known quantities of Calcium, Aluminum, Etc, were set to Triton for blind testing. The results came back with some relatively close, and others quite far off, enough that the information isn't helpful. Takeaway is that it *may* be OK for some parameters, but for others the information is biased and therefore not likely to be useful. Quote Link to comment Share on other sites More sharing options...
jolt Posted February 19, 2015 Share Posted February 19, 2015 And statistically the sample size is pretty small. So, for example std deviation of accuracy across calibrations could be even worse. 1 Quote Link to comment Share on other sites More sharing options...
victoly Posted February 19, 2015 Author Share Posted February 19, 2015 Yeah, i think triton is not coming off looking as bad as they could if they had sent in multiple sets of different standards. They only sent in one triplicate. If they had sent in a few different standards, this could be a much, much worse picture than is being painted right now,. Quote Link to comment Share on other sites More sharing options...
FarmerTy Posted February 20, 2015 Share Posted February 20, 2015 I'm using the results of the blind standard test as my correction factor for now. The precision is good enough for me for a ballpark number for most elements and I trashed the results for the one's that were not very precise. For accuracy, like I mentioned, I'm using it roughly as a correction factor for now. I'd feel more secure in doing that if there were other samples submitted on different days to test the variances on daily calibration but I'll work with what I got. The goal was to get a baseline reference annually for my parameters. If over time, perhaps it could be proven how accurate the results can be or not but the goal was precision. I just want to know a snapshot of what my levels are at X time and use that as a reference if my tank ever crashes to see what element was off and may have caused it. So, the accuracy isn't as important to me as long as the precision is there. And besides a couple of elements, the precision seemed well enough for me to trust for the time being. It's also the main reason I didn't do anything about my results. After discussion with Victoly a while back, I could be skeptic about the accuracy due to the limitations he informed me about but my hope was precision would be there. With the initial batch of results, I was interested in what they got as concentrations but wasn't planning on basing any decisions for dosing on those numbers. I'd have to look back at my numbers I tested with my own test kits prior to shipping off my sample to Triton but some of the applied correction factors are putting the results in more of the expected range I was anticipating before I shipped out the sample to be tested. I like them explaining the science behind the ICP... I was always curious but not enough to look it up myself. Quote Link to comment Share on other sites More sharing options...
ckyuv Posted February 20, 2015 Share Posted February 20, 2015 Ohhhhh good I was wondering what color the wammapel of the PB would be in 200 gallonen if the abwelchung was 5.63! 2 Quote Link to comment Share on other sites More sharing options...
FarmerTy Posted February 20, 2015 Share Posted February 20, 2015 Ohhhhh good I was wondering what color the wammapel of the PB would be in 200 gallonen if the abwelchung was 5.63!That one my friend hit my funny bone for some reason! [emoji23] 1 Quote Link to comment Share on other sites More sharing options...
victoly Posted February 20, 2015 Author Share Posted February 20, 2015 Meh, baseline correct to your hearts content. Bottom line is your hanna phosphorus ULR is closer to truth than Triton was. Enough for me to know that Triton is a good sales job and that's about it. Quote Link to comment Share on other sites More sharing options...
victoly Posted February 20, 2015 Author Share Posted February 20, 2015 Ohhhhh good I was wondering what color the wammapel of the PB would be in 200 gallonen if the abwelchung was 5.63! You keep your nazi stuff out of this. Quote Link to comment Share on other sites More sharing options...
FarmerTy Posted February 20, 2015 Share Posted February 20, 2015 Meh, baseline correct to your hearts content. Bottom line is your hanna phosphorus ULR is closer to truth than Triton was. Enough for me to know that Triton is a good sales job and that's about it. I'll take Triton's lack of accuracy over API. That's not based on anything but it was fun to say! Quote Link to comment Share on other sites More sharing options...
jolt Posted February 20, 2015 Share Posted February 20, 2015 All I really want is a reliable alkalinität sensor that monitors mein wasser real time. Is that so much to ask for ?!?!? 2 Quote Link to comment Share on other sites More sharing options...
Reburn Posted February 20, 2015 Share Posted February 20, 2015 I almost think I'd rather take the API. At least you know API is a pinto. They aren't a pinto in Mercedes clothing like Triton. The thing that get me is Triton pushes it proprietary blend of additives to correct imbalances that they think you have. Anytime I hear the word proprietary without scientific explanation on why it's so ground breaking it turns me off. 1 Quote Link to comment Share on other sites More sharing options...
Gig 'em @ NDstructible Posted February 20, 2015 Share Posted February 20, 2015 Is API American Petroleum Institute, or the Teledyne - Advanced Pollution Instrument? There are too many API's in my line of work for these acronyms! 1 Quote Link to comment Share on other sites More sharing options...
jolt Posted February 20, 2015 Share Posted February 20, 2015 The problem is that TLAs* cannot cover the space well enough. In my opinion we need to all move to using EFLAs** so there is no confusion. * Three Letter Acronym ** Extended Four Letter Acronym 2 Quote Link to comment Share on other sites More sharing options...
FarmerTy Posted February 20, 2015 Share Posted February 20, 2015 Triton IS the pinto as far as laboratory tests are concerned. Most accredited labs will charge you $25-50 for one analyte, Triton will give you 33 for that price. Granted most of the results are going to be inaccurate but at least a majority are precise. Don't get me wrong, I don't consider this a pro-Triton or a anti-Triton debate, I'm purely aiming to see what usable data I can mine from the apparent rubbish of data. If there is precision in the data, than I honestly don't care if you tell me my calcium is 5,000,000 ppm. As long as everytime I submit my sample with my 420 ppm of calcium, you always tell me it's 5,000,000 ppm... then I just correct it using a correction factor and call it a day. I don't want to jump to conclusions yet as mentioned before, I'd feel more secure in correction factors I derived from reefs.net's article if this test was also reproduced with multiple days involved to account for daily calibration discrepencies and possibly a couple other control test variables that I currently can't think of right now. I just can't honestly look at the data in front of me and say that there is NO usable data. There is to me at least. It may not be the golden ticket all of us reefers were wanting with precision laboratory results for our saltwater for a reasonable price but surely with the results that reefs.net produced that there is at least some usable data to be mined from the Triton results. Obviously, use at your own risk. For me, I won't be basing any dosing or corrective action on the results from my Triton test but I will be submitting an annual sample to track trends. For me, it's ideal at least because I don't do water changes so I can see what my current regiment of calcium reactors, biopellet reactors, GAC, chaeto, amino acid additions, Ceramico block, and feedings add to/remove from my system annually. It also gives me a good fingerprint in time so that if my tank crashes in 2018, I can look back at 2016, when my tank was humming along perfectly and looking it's best and see if I notice any discrepencies in concentrations that could allude as to what happened. Oh, when my Al was at 5,000,000 ppm in 2016 and then it shot up to 6,000,000 ppm in 2017, and that's when I started noticing problems... you get my overly facetious point. [emoji16] I don't know much about their supplements or corrective products but if people are actually basing their decisions on the Triton results and using their products to "cure it", then I worry about how those miracle solutions are going to really affect people's tanks. For everyone's sake, I hope it's just bottled NSW with a fancy label. [emoji4] Quote Link to comment Share on other sites More sharing options...
Gig 'em @ NDstructible Posted February 20, 2015 Share Posted February 20, 2015 I agree with Ty that these results are really only good for precision testing over time. I would not put much faith into the accuracy of the results and start messing with my dosing to achieve a goal for a certain concentration of a given compound/element. The problem I see with the precision testing over time and applying a correction factor to the results is their instrumentation may/probably will drift over time. Especially if you're only sending in a sample once every year, you can bet that their instruments will have been re-calibrated since the last test or simply just drifted over time. I know that with the chemiluminescence instruments I work with that they drift daily and need to be re-calibrated with a span and precision check every night. Some of my instruments can be within ±5% of the target concentration one day, and ±15% the next day depending on the calibrations that were run and any issues that may have developed. I would be curious what Triton's results would be if you sent in 3 identical standards 3 months apart from each other. If their results were within ±10% of each other every 3 months, then I would consider it safe to assume it's precise to track your annual trends. 3 Quote Link to comment Share on other sites More sharing options...
victoly Posted February 20, 2015 Author Share Posted February 20, 2015 Good point. The bigger picture problem is that giving the same sample over three separate analysis STILL wouldn't give you an accurate correction factor. The big picture reason that ICP-OES isn't a good fit for marine analysis is that high TDS samples (i.e., salt water) plug up the nebulizing portion of the machine and may not disperse all constituents equally. What that means is that just because sample A is precise, if sample B has a different constituent makeup/salinity/TDS, it might let different proportions of chemicals through. That's why I was alluding to sending in different reference standards to see if the correction factor holds true for multiple standards. Standard A: Alk 100, TDS 10,000 Standard B: Alk 150, TDS 9,000 Standard C: Alk 200, TDS, 9,500 What also leads me to believe that this is BS, is that allegedly Triton is pursuing legal action against the authors of this paper. If they were legit they'd show WHY they were legit, not try to put the squeeze on people who are trying to independently verify Triton's claims at their own expense. Quote Link to comment Share on other sites More sharing options...
FarmerTy Posted February 20, 2015 Share Posted February 20, 2015 Good points fellas. Let's see if they follow suit with perhaps different standards and also intervals between the sample submittals. I'll work with what I got for now. That's pretty lowly business practices but maybe that's the norm in Germany? Not saying that Germany has lowly business practices but perhaps there are different business norms over there that we may see as hostile here. Quote Link to comment Share on other sites More sharing options...
Gig 'em @ NDstructible Posted February 20, 2015 Share Posted February 20, 2015 Yeah, that's a bit sketchy that they're do defensive instead of just owning up and giving a sound explanation. Water testing at this level is not my area of expertise, but I bet there are a lot of factors that go into the accuracy of these tests. How the elements react with each other during transport, if they get absorbed or altered during transport, the effect of pH on the results, the effect of temperature on the results, the effect of salinity/TDS on the results, etc etc Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.