Calibrating The Seneye Free Ammonia Sensor

[Dan_P]

I am currently engaged in a project with @taricha that needed free ammonia measurements. The Seneye device seemed like a good low cost option. What I want to share is the calibration data.

The Seneye ammonia sensor uses a gas permeable film that is embedded with a pH indicating dye. When free ammonia diffuses into the film and reversibly reacts with the dye, the film turns blue. The more ammonia there is in solution, the bluer the film becomes. The Seneye device, like a Hanna Checker, has a photometer that measures the color intensity of the film. I needed accurate measurements of free ammonia and calibrated the device. This involves submerging the Seneye device in water with a known amount of total ammonia, measuring the PH and temperature of the water. From the total ammonia, pH and temperature, the free ammonia is calculated and compared to what the Seneye measures. The first graph below is probably the free ammonia concentration range that the average aquarist is concerned about. For an off-the-shelf device, it seems to do a good job quickly responding to the presence of ammonia, though it underestimates the amount. Not bad though.

The second plot shows how the Seneye readings respond over a very large free ammonia range and consisting of calibrations performed over 30 days. While a first impression might be “look how inaccurate it is”, what I see is a well behaved instrument with some drift over time at the very high range. Very little drift was detected when working below 0.2 ppm free ammonia (<10 ppm total ammonia) and none in the range relevant to aquarists.

When I finish the current project, I need to see how my aquarium and experimental systems respond to ammonia doses with the lights on versus off. Then, I think my aquarium sand needs to be interrogated for ammonia consumption and oxidation to nitrate

 

[Garf]

I am currently engaged in a project with @taricha that needed free ammonia measurements. The Seneye device seemed like a good low cost option. What I want to share is the calibration data.

The Seneye ammonia sensor uses a gas permeable film that is embedded with a pH indicating dye. When free ammonia diffuses into the film and reversibly reacts with the dye, the film turns blue. The more ammonia there is in solution, the bluer the film becomes. The Seneye device, like a Hanna Checker, has a photometer that measures the color intensity of the film. I needed accurate measurements of free ammonia and calibrated the device. This involves submerging the Seneye device in water with a known amount of total ammonia, measuring the PH and temperature of the water. From the total ammonia, pH and temperature, the free ammonia is calculated and compared to what the Seneye measures. The first graph below is probably the free ammonia concentration range that the average aquarist is concerned about. For an off-the-shelf device, it seems to do a good job quickly responding to the presence of ammonia, though it underestimates the amount. Not bad though.

The second plot shows how the Seneye readings respond over a very large free ammonia range and consisting of calibrations performed over 30 days. While a first impression might be “look how inaccurate it is”, what I see is a well behaved instrument with some drift over time at the very high range. Very little drift was detected when working below 0.2 ppm free ammonia (<10 ppm total ammonia) and none in the range relevant to aquarists.

When I finish the current project, I need to see how my aquarium and experimental systems respond to ammonia doses with the lights on versus off. Then, I think my aquarium sand needs to be interrogated for ammonia consumption and oxidation to nitrate

So, it's good straight out of the box then, that's good. But what about all those people that have "calibrated" their devices to read lower, as suggested on a few work threads on this site. Very odd behaviour in my view.

 

[Randy Holmes-Farley]

I did not realize it was a simple pH response.

That opens up the possibility of a false positive by any solvent soluble base that may be present.

Seafood can contain quite a few of these, and at least right after feeding, the values may not be reliable.

Folks dosing amino acids may also get false responses.

Histamine from bacterial degradation of histidine, Putrescine, spermine and spermidine from arginine, etc.

These are all going to be at low levels, but I do not know how low and may interfere at the very low end.

 

[taricha]

nice @Dan_P
That is one workhorse slide running up and down the ammonia scale like that!
How much equalibration time do you find the slide needs for you to feel good about it?

 

[Dan_P]

I did not realize it was a simple pH response.

That opens up the possibility of a false positive by any solvent soluble base that may be present.

Seafood can contain quite a few of these, and at least right after feeding, the values may not be reliable.

Folks dosing amino acids may also get false responses.

Histamine from bacterial degradation of histidine, Putrescine, spermine and spermidine from arginine, etc.

These are all going to be at low levels, but I do not know how low and may interfere at the very low end.

I think the permeability is limited to unionized molecules, which I assume you alluded to as solvent soluble compounds. I think methylamine is a possible interfering compound. I don’t recall seeing that amino acids are a problem, but as you say, metabolites might. Sounds like this has potential as a study.

 

[Dan_P]

nice @Dan_P
That is one workhorse slide running up and down the ammonia scale like that!
How much equalibration time do you find the slide needs for you to feel good about it?

I found 15 minutes is more than enough time to thoroughly equilibrate the film, though the bulk of the response is achieved in a few minutes. Recovery time is quick, minutes if the sensor is close to zero, maybe 10 minutes to reach zero if the sensor saw 1 ppm free ammonia. For aquarium use, I thought the response-recovery times met our needs.

The color change for this film is weak, not like the Seachem ammonia sensing films. That’s the trade-off for these films. Strongly colored films don’t have agile response-recovery times. Also, the weakly colored films actually stop changing color, while the strongly colored film like the Seachem film tends to continue to accumulate color for awhile. I can’t remember the brand name, but I found a competitor to Seachem that works like the Seneye film.

 

[Dan_P]

So, it's good straight out of the box then, that's good. But what about all those people that have "calibrated" their devices to read lower, as suggested on a few work threads on this site. Very odd behaviour in my view.

I haven’t seen those posts. Maybe the lot of film they were using was hypersensitive or the Seneye device was programmed differently, reading high instead of low.

 

[Randy Holmes-Farley]

I think the permeability is limited to unionized molecules, which I assume you alluded to as solvent soluble compounds. I think methylamine is a possible interfering compound. I don’t recall seeing that amino acids are a problem, but as you say, metabolites might. Sounds like this has potential as a study.

Yes, only a solvent soluble base will penetrate the membrane, and in this context that really is limited to uncharged amines.

 

[brandon429]

the trim setting=moving the nh3 baseline to a guesstimate by the owner based on lots of data about what nh3 runs in seneye-tracked reef tank displays even if the current reading on the slide doesn't generate an average read.

this is how I look at trim setting: if we sampled 200 uploaded seneye nh3 logs from any poster, the vast vast majority run .001-.005~ nh3 as an average

so, if there's an install of seneye on a 7 year running nano reef and it only has one fish, and the machine registers .04 vs .00x thousandths, there's no reason to be a hundred times out of ammonia spec control, so a trim adjustment to move that calibrated safe/.04 down to .005 as the running baseline now wouldn't make it incompatible with hundreds of data sources.

the precision changes it can show, from any base setpoint when put through various load tests in the reef tank, was what mattered to me all this time. I thought there were significant calibration proofs made between api ammonia and seneye when one of our comparison threads showed a running .04 nh3 for a seneye on a suspected cycled display, and also when that slide was placed into a great aged test nano reef stocked full of ammonia users.

the fact his setup said .04 I thought made it a good trim candidate but if he didn't change it, seeing .04 as the baseline allows one to still use it to track minute load changes, or ability to starve a biofilter in test fallow setups, or the ability for a sandbed full of waste to truly spike ammonia etc.

 

[Randy Holmes-Farley]

the trim setting=moving the nh3 baseline to a guesstimate by the owner based on lots of data about what nh3 runs in seneye-tracked reef tank displays even if the current reading on the slide doesn't generate an average read.

this is how I look at trim setting: if we sampled 200 uploaded seneye nh3 logs from any poster, the vast vast majority run .001-.005~ nh3 as an average

so, if there's an install of seneye on a 7 year running nano reef and it only has one fish, and the machine registers .04 vs .00x thousandths, there's no reason to be a hundred times out of ammonia spec control, so a trim adjustment to move that calibrated safe/.04 down to .005 as the running baseline now wouldn't make it incompatible with hundreds of data sources.

the precision changes it can show, from any base setpoint when put through various load tests in the reef tank, was what mattered to me all this time. I thought there were significant calibration proofs made between api ammonia and seneye when one of our comparison threads showed a running .04 nh3 for a seneye on a suspected cycled display, and also when that slide was placed into a great aged test nano reef stocked full of ammonia users.

the fact his setup said .04 I thought made it a good trim candidate but if he didn't change it, seeing .04 as the baseline allows one to still use it to track minute load changes, or ability to starve a biofilter in test fallow setups, or the ability for a sandbed full of waste to truly spike ammonia etc.

Arbitrarily adjusting a reef tank reading to match some claimed average reef tank readings will soon have every reef tank reading the same, whether it has any basis in reality or not.

 

[ingchr1]

So, it's good straight out of the box then, that's good. But what about all those people that have "calibrated" their devices to read lower, as suggested on a few work threads on this site. Very odd behaviour in my view.

When I was using a Seneye, there were slides where NH3 was "stuck" on 0.001. Those I had trim UP to see response on the slide.

One other thing I noticed with my Seneye, was that the pH would drift up the last week or so of slide life.

There are several posts I made in the following thread on the unscientific observations I made.

My last two slides needed to have NH3 trim to get off of a 0.001 reading. I had the trim set at +0.004 to get a 0.005 reading, what I deemed a "normal" reading for my tank based off past slides.

The new slide I put in last night is off of the 0.001 reading and appears to be reading 0.007 or so as "normal". So I have set the trim back to 0.000.

Not sure what could be different with this slide from the other two, all three came out of the same box. They are 2018 slides. My next batch of slides were purchased last fall. Will see how those go.

When I put this slide in I wiped down the sensor windows, again not sure if that made the difference or not.

7:16AM - Slide expired
6:14PM - New slide installed. Note at this point the +0.004 trim was still applied, I did not remove that till this morning.
7:56PM - Did one of my two daily doses of Dr. Tims NH4Cl. Four drops or so.

As a side note, I only soaked this slide for 24hrs prior to use.

For pH, slide behavior has been consistent through slide changes. The last week or so of a slides life the pH value drifts up. Install new slide and it's back to reading close to what my controller reads. I set the trim for pH several slides ago and have not had to touch it.

 

[brandon429]

Randy

to not adjust it ever means we expect all slides to function within perfect spec, that may be a mechanical q.a. stretch goal per 1000 meters sold. and, this is the first meter with staying power that citizens can wield (mindstream=gone) I'm sure the masses will hone this thing down nicely as steam picks up.

I'm not sure I believe it when it says a running reef tank is .04, there's times to challenge any test kit it seems. I spend orders more time challenging nondigital kits that's for sure, most seneyes land in spec on a cycled tank.

I can understand the value of simply leaving it where it lands and measuring the changes from the stated baseline for sure. I can't wait to see what Dan does with this

his findings and recommendations will be very influential can't wait to see

 

[Randy Holmes-Farley]

Randy

to not adjust it ever means we expect all slides to function within perfect spec, that may be a mechanical q.a. stretch goal per 1000 meters sold. and, this is the first meter with staying power that citizens can wield (mindstream=gone) I'm sure the masses will hone this thing down nicely as steam picks up.

I'm not sure I believe it when it says a running reef tank is .04, there's times to challenge any test kit it seems. I spend orders more time challenging nondigital kits that's for sure, most seneyes land in spec on a cycled tank.

I can understand the value of simply leaving it where it lands and measuring the changes from the stated baseline for sure. I can't wait to see what Dan does with this

his findings and recommendations will be very influential can't wait to see

If one needs to adjust them, it should be with a standard, not a supposition of what a reef tank value “should be”.

 

[Dan_P]

When I was using a Seneye, there were slides where NH3 was "stuck" on 0.001. Those I had trim UP to see response on the slide.

One other thing I noticed with my Seneye, was that the pH would drift up the last week or so of slide life.

There are several posts I made in the following thread on the unscientific observations I made.

Thanks for the links.

The zero point on my current sensor is also 0.001 and I am pretty sure that is because the film is slightly more colored than the one used to program the photometer. Trimming sounds like a baseline correction and should not affect whether the photometer senses a color change in the film.

The Seneye pH function works like the ammonia sensor but it has a membrane that permits water to diffuse through it. The pH measurements of my unit are off by tenths of a pH unit. I used a lab pH meter to measure pH and temperature in my work. The temperature measurements were also off.

But all this is OK. It is normal for analytical device to need calibration.

 

[Dan_P]

the trim setting=moving the nh3 baseline to a guesstimate by the owner based on lots of data about what nh3 runs in seneye-tracked reef tank displays even if the current reading on the slide doesn't generate an average read.

this is how I look at trim setting: if we sampled 200 uploaded seneye nh3 logs from any poster, the vast vast majority run .001-.005~ nh3 as an average

so, if there's an install of seneye on a 7 year running nano reef and it only has one fish, and the machine registers .04 vs .00x thousandths, there's no reason to be a hundred times out of ammonia spec control, so a trim adjustment to move that calibrated safe/.04 down to .005 as the running baseline now wouldn't make it incompatible with hundreds of data sources.

the precision changes it can show, from any base setpoint when put through various load tests in the reef tank, was what mattered to me all this time. I thought there were significant calibration proofs made between api ammonia and seneye when one of our comparison threads showed a running .04 nh3 for a seneye on a suspected cycled display, and also when that slide was placed into a great aged test nano reef stocked full of ammonia users.

the fact his setup said .04 I thought made it a good trim candidate but if he didn't change it, seeing .04 as the baseline allows one to still use it to track minute load changes, or ability to starve a biofilter in test fallow setups, or the ability for a sandbed full of waste to truly spike ammonia etc.

The safest way to deal with the baseline correction is submerging the device in a sample with 0 ppm ammonia. I haven’t tried it, but RO/DI should be such a sample that is accessible by everyone.

A sensor that gives a reading of 0.040 seems like a bad one to me because it corresponds to a very high total ammonia concentration, one that the API test kit could measure. Maybe you meant 0.004 ppm.

 

[taricha]

One more Q, did you ever run a slide high enough to damage it? If so, how high and how did the damaged slide behave?

 

[stoney7713]

The safest way to deal with the baseline correction is submerging the device in a sample with 0 ppm ammonia. I haven’t tried it, but RO/DI should be such a sample that is accessible by everyone.

A sensor that gives a reading of 0.040 seems like a bad one to me because it corresponds to a very high total ammonia concentration, one that the API test kit could measure. Maybe you meant 0.004 ppm.

I used my last slide and don't have the instructions, but I'm positive they state not to use RODI water as it may damage the slide.

IS there any merit to that?

 

[Dan_P]

One more Q, did you ever run a slide high enough to damage it? If so, how high and how did the damaged slide behave?

I went up to 30 ppm total ammonia and it recovered just fine. I didn’t look for the limit where permanent staining occurred. By the way, the response curve continues to flatten above 20 ppm (expected from literature) and needs a different calibration curve for high free amine levels.

 

[Dan_P]

I used my last slide and don't have the instructions, but I'm positive they state not to use RODI water as it may damage the slide.

IS there any merit to that?

I didn’t know this. Thanks.

Is there merit? When I have time, I will put this warning to the test. Can’t think of a reason why it should matter, especially for short exposures.

 

[brandon429]

New Cycle - Seneye vs API

Hello All, Have just started to cycle a new DD ReefPro 900. Am using Fritzzyme Fishless food and Turbo Start and am 2 days in. I have dosed twice with Fishless food on the back of the seneye readings which are as follows: But the API result seems to be showing a different story. I`ve seen a...

www.reef2reef.com

.04 here's the thread. the nano reef it was baselined on with one tiny fish and packed wall to wall in surface area was not running truly at .04 but that suffices for a baseline, and it suffices there to show api will show dark green in a tank that has no ammonia issues.

anyone who read or saw that api reading/color in that thread above, taken on it's own, would in unison agree the tank isn't cycled. but paired with a calibrated seneye, we didn't even need to trim it into the thousandths to make use of the seneye reading portion.