IASIM 2018 Advertisement

Calibration required

JohanG's picture

I am very new here. I have been used a NIR for Feeds and Raw Materials for a long time, but have not done any calibration myself. We have been acquiring our calibrations form FOSS for instance. (I would like to learn how to) At the moment I am looking for a calibration for molasses syrup that we would like to scan on the FOSS DS2500 using a slurry cup.
Has anyone done this before? Is there anything like that out there?  I would appreciate some feedback.
Thank you very much ([email protected])

khaled's picture

Hi Johan,
I would advise you to develop calibration for molasses by your self, it is not a complex job after all.
We have been developing this calibration in the past for Foss Nir 5000 model.
The parameter was Brix & Moisture content, RSQ of calibration was very good only for about 400 calibration sample.
You can use gold 0.5 mm transreflectan asseccories for scanning.
Good luck!
Best regards,
Khaled Sungkar

JohanG's picture

Thank you Khaled.
Did you use WinIsi for the development?

td's picture

Hello JohanG,
Welcome to the group!
Yes, there are some references to molases in NIRP publications:
K.J. Schaffler, Proc. S. Afr. Tech. Ass. 67, 2222-228 (1993).
A. Salgo JNIRS 6,A101-A106 (1998) from NIR97.
R. Mehrotra et al., Proc. 10th NIR Conf. 207-211 (2001). 
You can find them on this website.
NIR calibration
I have to disagree with Khaled
"NIR calibration is not a complex job" 
NIR calibration requires:, Knowledge, comprehension, practise, good software, hard work and diligence.  Ideally you should attend a training course; they are held at several national and international conference and by software providers. Otherwise you need a book. See the books on this website.
Good luck and best wishes,

khaled's picture

Hi Tony...
I am totally agree with you.... indeed NIRS calibration require knowledge and expertise.
My point was :  specific for Molasses - Brix and Moisture, it is not a complex calibration and I have done it well with RSQ 0.98 - 0.99
Compare to other calibration (other product, other parameter) where we are facing difficulty developing good calibration where most cases require more spectrum (could be thousands) and even can not get RSQ more than 0.8.
Hi Johan,
Yes, we use WINISI, I believe you can make good cal on it.
Pay more attention on the Precision of your Reference Method & your sample Range (should very big range). take sample from different shipment, supplier, source, etc.
Best regards,
Khaled Sungkar