We need your help with a survey of instrument noise.
Submitted by td on 29 May 2013 - 2:30pm
Karl Norris and I invite you to take part in a survey of instrument noise. To do so you will be asked to measure repeat spectra of three or four samples and send the data to Tony Davies.
We intend to make the following calculations:
- Actual noise levels on each instrument.
- PCA study to characterise the major noise sources.
- Compute differences between instrument types for the same samples.
- Search PCA loading plots for common loadings in the PCA data.
- Explore the use of the fourth derivative with different gap sizes to measure the random noise.
- Anything that we have not thought of but should have!
We think we can handle data from about 100 instruments but in order to make the survey really useful it needs to have the widest possible range of instrument types. So for this reason we will restrict the numbers of instruments of the same model to 16 and contributors will be selected to maximise the number of instruments from different manufacturers, models and locations.
Having made the selection, we will send participants detailed instructions and invite them to make the measurements within a short time-scale and send them electronically to Tony, in original instrument and J-CAMP formats.
We hope the results will be sufficiently interesting to be published in JNIRS with a summary in NIR news. The identity of participants and the manufacturer of their instruments will be coded. Participants will be given the code(s) that apply to their results.
If you would like to take part, please e-mail Tony ([email protected]) giving your instrument’s manufacture and model and your location.
We welcome a discussion on this forum about this proposal.
Best wishes,
Tony
knnirs
29 May 2013
We need your help with a survey of instrument noise.
Tony Davies and I invite you to take part in a survey of instrument noise. To do so you will be asked to measure repeat spectra of three or four samples and send the data to Tony Davies.
We intend to make the following calculations:
1) Actual noise levels on each instrument.
2) PCA study to characterise the major noise sources.
3) Compute differences between instrument types for the same samples.
4) Search PCA loading plots for common loadings in the PCA data.
5) Explore the use of the fourth derivative with different gap sizes to measure the random noise.
6) Anything that we have not thought of but should have!
We think we can handle data from about 100 instruments but in order to make the survey really useful it needs to have the widest possible range of instrument types. So for this reason we will restrict the numbers of instruments of the same model to 16 and contributors will be selected to maximise the number of instruments from different manufacturers, models and locations.
Having made the selection, we will send participants detailed instructions and invite them to make the measurements within a short time-scale and send them electronically to Tony, in original instrument and J-CAMP formats.
We hope the results will be sufficiently interesting to be published in JNIRS with a summary in NIR news. The identity of participants and the manufacturer of their instruments will be coded. Participants will be given the code(s) that apply to their results.
If you would like to take part, please e-mail Tony ([email protected]) giving your instrument’s manufacture and model and your location.
We welcome a discussion on this forum about this proposal.
Regards,
Karl
Permalink
venkynir
01 Jun 2013
Hi Karal & Tony Davies
My first question is
Actual noise levels on each instrument.:
How to select operting condition here .Some instrument work within four wall under controlled temp.Some should work in fied where EMF noise in predominate ? Some time country like ours India the volatge varies from 190 AC to 230 AC( Freq 45~51 Hz). Some work under TE cooled condition ?. Optical noise ? EMF noise due wide usuage of mobile phone ?
Explore the use of the fourth derivative with different gap sizes to measure the random noise.
Why fourth derivate ? Do you think signal will with stand up to that level in the case samples have very low S/N ratio.
Is good move but done with collective groups .I Like that .We are suffering at Indian Industrial condition .
venkynir
Permalink
td
10 Jun 2013
Hello Venkynir!
Thank you for your comments and questions.
1) I think we are interested in noise levels in the environment where the instrument is being operated. We will record whatever participants tells us.
2) If S/N is low the instrument is not going to be much use for NIR analysis but with good S/N Karl has demonstrated some very useful applications of 4th derivatives.
Best wishes,
Tony
Permalink