Oil testing kit Beta programme - Sample scanning notes
Contributors: @TedF and @cindy_excites
Date: 4/Nov/2015
Here we document our experience using the Public Lab Spectrometer 3.0 (assembled using these instructions) to scan our Oil Testing Kit Beta samples, which arrived with our OTK packages.
1) Scanning Mineral oil
We begin by sampling mineral oil. First without laser attenuator:
Mineral Oil similar matches (by clicking on the "Find similar" button):
Then tested the Mineral oil with the attenuator:
NOTE: Unfortunately, we cannot accurately tell which gradation of attenuation we used as the attenuator slider is slid into the black cuvette frame. But perhaps this does not matter?
Then compared the two - without and with attenuator (by clicking on the "Compare" button):
Mineral Oil similar matches (by clicking on the "Find similar" button):
This exercise was most useful in familiarising myself with the different functions of Spectral Workbenck [ https://spectralworkbench.org/ ] and the different parts of the spectrometer. I also realised that the spectrometer had some technical issues, which I fixed using Lego blocks. [ HYPERLINK HERE ].
2) Scanning unknown sample # 26
We proceed to scan sample # 26 and we carry out three repetitions of the scan. For each rep I turned the cuvette around (to scan from the other side).
Note: For (nearly) all samples that follow I have created screen clippings for the capture page, the annotation pages, and the similar spectra page of each rep, as well as the compare all reps page.
Scanning of sample # 26, rep a:
Sample # 26 similar matches for rep a (by clicking on the "Find similar" button):
Scanning of sample # 26, rep b:
Sample # 26 similar matches for rep b (by clicking on the "Find similar" button):
Scanning of sample # 26, rep c, capture page:
Scanning of sample # 26, rep c - sample annotation:
Sample # 26 similar matches for rep c (by clicking on the "Find similar" button):
Comparison of all 3 reps (by clicking on the "Compare" button):
3) Scanning unknown sample # 79
We then scan sample # 79 and we carry out three repetitions of the scan. Scanning of sample # 26, rep a:
Sample # 79 similar matches for rep a:
Scanning of sample # 79, rep b, capture page:
Scanning of sample # 79, rep b - sample annotation:
Sample # 79 similar matches for rep b:
Scanning of sample # 79, rep c:
Sample # 79 similar matches for rep c:
Comparison of all 3 reps for sample # 79:
4) Scanning unknown sample # 111
We continue with scanning sample # 111 and carry out three repetitions of the scan
Scanning of sample # 111, rep a, capture page:
Scanning of sample # 111, rep a:
Sample # 111 similar matches for rep a:
Sample # 111 similar matches for rep b:
Sample # 111 similar matches for rep c:
Comparison of all 3 reps for sample # 111:
4) Scanning unknown sample # 5w-30
Next we scan sample # 5w-30 and carry out three repetitions of the scan.
Scanning of sample # 5w-30, rep a, capture page:
Scanning of sample # 5w-30, rep a:
Sample # 5w-30similar matches for rep a:
Scanning of sample # 5w-30, rep b:
Sample # 5w-30similar matches for rep b:
Scanning of sample # 5w-30, rep c:
Sample # 5w-30similar matches for rep c:
Comparison of all 3 reps for sample # 5w-30:
5) Scanning unknown sample # 20w-30
Next we scan sample # 20w-30 and carry out three repetitions of the scan.
Scanning of sample # 20w-30, rep a, capture page:
Scanning of sample # 20w-30, rep a:
Sample # 20w-30 similar matches for rep a:
Scanning of sample # 20w-30, rep b:
Sample # 20w-30 similar matches for rep b:
Scanning of sample # 20w-30, rep c:
Sample # 20w-30 similar matches for rep c:
Comparison of all 3 reps for sample # 20w-30:
6) Scanning unknown sample # 80w-30
Next we scan sample # 80w-30 and carry out three repetitions of the scan.
Scanning of sample # 80w-30, rep a, capture page:
Scanning of sample # 80w-30, rep a, annotation:
Sample # 80w-30 similar matches for rep a:
Scanning of sample # 80w-30, rep b:
Sample # 80w-30 similar matches for rep b:
Scanning of sample # 80w-30, rep c:
Sample # 80w-30 similar matches for rep c:
Comparison of all 3 reps for sample # 80w-30:
Note: I thought I had forgotten to compare all three reps for this sample, but it was not possible to go back and select them after I had done many other readings; only recently captured spectra apprear in the compare window and the search function does not return results outside the listed recently selected spectra
7) Scanning the Diesel sample
Next we scan the Diesel sample and carry out three repetitions of the scan.
Scanning of the Diesel sample, rep a, capture page:
Scanning of the Diesel sample, rep a, annotation:
Diesel sample similar matches for rep a:
Scanning of the Diesel sample, rep b:
Diesel sample similar matches for rep b:
Scanning of the Diesel sample, rep c:
Diesel sample similar matches for rep c:
Comparison of all 3 reps for the Diesel sample:
7) Scanning the Crude sample
Next we scan the Crude sample and carry out three repetitions of the scan.
Scanning of the Crude sample, rep a, capture page:
Scanning of the Crude sample, rep a, annotation:
Crude sample similar matches for rep a:
Scanning of the Crude sample, rep b:
Crude sample similar matches for rep b:
Scanning of the Crude sample, rep c:
Crude sample similar matches for rep c:
Comparison of all 3 reps for the Crude sample:
My favourite functions in Spectral worksbench in this exercise was the "Find similar" button. I will in the future go back and write a bit more about some of the interesting 'similar' spetra that I found. It would have been useful to have a "how to dispose of your samples" and "how to clean your cuvettes" guides. I first used acetone but it damaged my cuvettes (I lost two in the experiment). I then used dishsoap but to no effect. I then used window cleaner and that seemed to do the trick!
Related research notes:
Oil testing kit Beta programme - Package content
Oil testing kit Beta programme - Assembly: cuvette frame
Oil testing kit Beta programme - Assembly: spectrometer
Oil testing kit Beta programme - Calibration
Oil testing kit Beta programme - Technical adjustments (hack!)
5 Comments
Hi, Cindy - wow, you had spectacularly well-lit scans! Your laser must be quite powerful, or your camera especially sensitive. I wonder what accounts for this variation, as I know not everyone's been able to get such clear bright spectra.
I was also worried that your crude sample would be too dark to see the fluorescence, but it looks like it worked fine -- huh! Interesting! I was about to suggest diluting in mineral oil but i guess it's not necessary.
I was especially interested to see the "find similar" function match to @ethanbass's diesel samples. Great!
Reply to this comment...
Log in to comment
Cindy the flat spots on your spectra are the blown out blue channel: https://publiclab.org/notes/mathew/09-23-2014/graded-oils-using-uv-fluorescence#Questions+and+next+steps
the note is in the wiki but clearly we need to be clearer about this: https://publiclab.org/wiki/oil-testing-kit#Illuminate+the+sample
Reply to this comment...
Log in to comment
Hi @mathew! So you mean that in all the scans above where I see the intense blue channel had overexposure? As in the post https://publiclab.org/wiki/oil-testing-kit#Illuminate+the+sample ?
In this image for example,
is it that the blue channel should also be at the 'ideal exposure range'?
Is this a question? Click here to post it to the Questions page.
Reply to this comment...
Log in to comment
I'm trying to figure out what this flat spot is, its just soo flat! I associate it with a blow out.
Digging in to the spectra themselves on your user profile, I'm not finding a blown out blue channel, its just very high.
I think you did everything right on the collection end. I want to see if this is a camera issue-- not sure. It could be the right shape.
Reply to this comment...
Log in to comment
Hi, Cindy - can you link to your actual data on SWB?
The average should be in the "ideal exposure range" but the important part for the 3 color channels is just that they don't hit the top of the graph and 'get smushed' -- that'd mean we're losing the data from the top of that peak. The three color channels are averaged to get the "main" line. The "ideal exposure range" is really for two reasons: to prevent clipping (smushing) and to encourage people to get enough brightness that the line is not as affected by random noise, which could happen with a very dim spectrum. A spectrum with greater than 10% or so should have good enough dynamic range to use, but 25% or more is ideal.
In the second image from your last comment, for example, the blue peak is not in the ideal range, but it's also not flattening at the top, so it's OK, I believe.
Is this a question? Click here to post it to the Questions page.
Reply to this comment...
Log in to comment
Login to comment.