Sorry to be a bit off-topic, but I would be interested to know what your specification is going to be for photometric stability (i.e. what is the baseline noise level at a single wavelength measurement), how frequently time points can be taken, and whether it will be possible to have a magnetically stirred cuvette.

Wow! Thanks Sir for introducing me to these problems. It would be great if you could elaborate bit on magnetically stirred cuvette version? I just have a 10mm cuvette placing space in the spectrometer in the current design. Also thanks for asking the right questions. I will put those on the specs sheet. 

Looking forward to more queries like these..

I would take inspiration from the old agilent chemstation. They got it right then: https://remchowiki.science.oregonstate.edu/lib/exe/fetch.php?media=protocol_agilent_1100_hplc_chemstation.pdf

You see the different parts of your instrument and it is clear what to click to change the different things. Made to be understandable by the user. Not the engineer.

Hi Jan

Point Noted. Thanks !

You mention noise but is it on absorbance or intensity ratio?. The uncertainty in quantitation is optimal (minimal in absorbance) at about A = 0.4 for a given noise on intensity. Linearity at high A depends on stray light. The user may be warned to use absorbance less than a limit value (to be determined, stray light may depend on wavelength) or look carefully at the calibration curve.

Sir Gal

You are right. I should have mentioned clearly that I was talking about noise level in INTENSITY. 

Your thoughts regarding warning in case absorbance exceeds a threshold makes complete sense. I will add the feature.

Regarding the fact about minimal 'noise' at A=0.4, what is the application of this fact. Do you mean that I should take readings at A=0.4 and put that into the specs sheet as "best case noise = x% ". 

Thanks and please drop any other suggestion that might come to you. 

The exact optimum value is 0.4343, and transmittance  T = 0.368. These values come from the minimum of the dA/A curve as function of T. This was discussed and criticized and amended by several authors, in papers and in books (the optimum values depends in fact from the type of noise). I attached an old paper on that mater. Anyway it seems to be advisable to work above A = 0.1 and below A = 1, but additional dilution or concentration steps may also introduce more errors ! At high A values non linearity occurs. Stray light, which produces curved calibration plots at high A, is abundantly treated in didactic documents or booklets from manufacturers.

The spectrometer you mentioned seems fine if it is comparable to that produced by Ocean Optics, USA.

However, I would like to know how good is the intensity calibration covering the entire spectral region; 370 - 1000 nm? It will be interesting to address this issue that is wavelength and intensity calibration. You did not mention about it resolving power.

I will be interested in this spectrometer when  you get it finally configured.

best of luck,

Aslam 

Sir Jean,

Thanks for the literature. I will study and write its essence here.  

Sir Baig

Thank you for your interest. I am able to resolve all the peaks of CFL which ocean optics resolves. I have added two spectra - one from my device and other from ocean optics.

Regarding intensity calibration: I have this thing in mind but I don't have a well calibrated spectrometer or known intensity source. They are expensive! May be I will visit a lab nearby sometime and request them to let me use their stuff.

Regarding wavelength calibration: I did it sometime back, but then had to move some components, so that calibration is no longer valid. But that's an easy task!

Dear All

I have added new photos for spectra comparison and a ppt file showing the screenshots of the software to be made. Please have a look and advice.

Thanks

Dear Shubham Rathore,

The comparison spectra  you showed look very good. However on the Y-axis you mentioned absorbance which starts at 0.4. What reference have you used for Zero and above absorbance? Is it relative or absolute?

with best wishes,

Aslam 

Its not absorbance value sir. Its actually intensity. The freeware software that I am using right now needs to be fixed for my device. Please ignore the labels on the y-axis. 

The emission intensity (count) from a lamp cannot be translated into absorbance. Anyway the spectrum looks good (resolution, noise...). At first sight the software seems OK, but finally only a user may judge it.

Yes. For absorbance i'll have to subtract the dark noise and divide the spectrum of blank by spectrum of sample.

Yes I need to take feedback from end-user to fix the bugs and add functionalities. 

Regarding the user interface, a few observations:

1.  Is everything actually going to be that large?  It's easy to see, certainly, but also comes across as a bit cluttered. 

2.  The odd shapes by the buttons..  Perhaps try a sideways "play" icon by "analyze", and try icons that intuitively "make sense" for the others?  For example, a partially filled test tube or flask for concentrations and an hourglass or some sort of "clock" for the time. 

3.  You want your interface to be intuitive and user-friendly, but take care that you do not oversimplify it to the point of it having a "toy" appearance.  An under-sophisticated interface can give the impression that your program is of lesser quality, or lacking in functionality. 

4.  Metadata.  Metadata metadata metadata.  Any analytical software package needs to have the ability to capture myriad data whether the user has specified it or not.  Imagine a time that you performed an analysis and later thought, "Oh!  I wish I had thought to capture the dPressure while that was running!  That would help me better understand whether what I'm seeing is "real" or due to a hiccup in the system!"  Or something along those lines.  Not only should you be able to capture metadata, but you should provide a space within the software to manipulate it.  A reporting/report building capability where the user can fully customize not only WHAT information they want recalled, but also how it should appear.  This should apply to reporting, but also in the way data is displayed in the software as well.  The default layout may make sense to you, but another user might prefer a more linear layout, or a more compact one, or might want to minimize the amount of information on display to avoid being distracted by superfluous data fields. 

In my personal and mildly limited experience, Waters' "Empower" system and SDMS are BY FAR one of the best systems out there.  It's not a spectroscopy system in the "cuvette reading" sense, but it IS a chromatography system that can also be used to capture data from other systems and work in tandem with a LIMS, eLN, import/export interface, etc. 

What an exciting time for you!  Best of luck!  

The software should have the ability to export the data as a text file or Excel spreadsheet for offline processing and production of publication quality graphics.

I would implement the possibility to evaluate Raman spectra: there should be an option to convert the wavelength to wavenumbers as well as some kind of fluorescence removal algorithm. Even though your devices might not be directly intended for this kind of application, you can prepare at least the software for future developments. Very simple Raman spectrometers are being described in the literature. For more info and examples please check the attached material.

http://www.flintbox.com/public/project/1956

Article Automated Autofluorescence Background Subtraction Algorithm ...

Article Inexpensive Raman Spectrometer for Undergraduate and Graduat...

Thank you Sir Adam and Andrea for your response. 

Yes it has the ability to export.

Hello Andrea

Wavelength-wavenumber conversion is one feature I'm going to give. Not sure about the fluorescence subtraction and raman spectra evaluation at this point of time. But in future, I'll definitely give it a thought.   

Dear Sara mam

Thanks for such an detailed advice.

Thanks a lot for your wishes. :)

Dear  Shubham Rathore !

Your "trial and error method" is not applicable for your problem.

I recommend you to develop the project and to follow the main steps. E.g.,

 https://www.teamgantt.com/guide-to-project-management/how-to-plan-a-project/

So you can avoid a lot of problems which I had many times.

E.g., I have develop some algorithms for NMR portable spectrometer. But I did not know the parameteres of  this spectrometer.  The resolution of spectrometer was low and therefore my algorithms were not good enough. 

From discussion I see the interface problem. But you do not know what

interface you need! And so on...

Regards, Joseph Dubrovkin

Thank you Joseph Sir for your reply. 

I read that article and have some idea of your point now. I will definitely consider them in my design.