| Author | Message | ||
     Fatih KAHRIMAN (caucasus) Junior Member Username: caucasus Post Number: 6 Registered: 10-2010 |
Hi All Our instrument's problem have been found. Its optics was damaged whic is "grating" part of optics. If anyone meet like this problem who could be control this part of own intrument. This spectum shifts are based on not lamp or other parts of instrument. Regards Fatih | ||
| Gabi Levin (gabiruth) Senior Member Username: gabiruth Post Number: 43 Registered: 5-2009 |
Hi guys, This is a very interesting case - that exemplifies the disadvantages of a non=dual beam instrument, the sensitivity to several parameters that can affect the reference and the extreme care you need to take to keep it in good shape. However, in addition to the fact that there is an "increase" in gain - hence the "saturated" region from about 1300 to 1400 nm in the current reference trace - the position of the minimum on the trough shifted from about 1400 nm to about 1520. This is very strange and can not be just a "gain" issue. It does not seem to be a wavelength shift either because if you examine the position of the little "tweek" at about 1875nm you can easily see that this "tweek" is not shifted. Such "tweeks" are common in many optical systems and are the result of small absorptions from anti-reflective coating on the optical elements. Therefore before any more is done it should become understood why the signal reflected back to the sensors - has increased by quite a bit, but also what else has changed in it. The sad part of this situation is that from the information provided it appears that it will be extremely difficult to "reconstitute" the reference to exactly what it was. In a situation like that, most likely they will need to "rework" their calibrations after installing a new reference, some mighty job it could be. If Fatih contacts me directly I could help him to reduce the work load of recalibration. Gabi Levin [email protected] | ||
| Fatih KAHRIMAN (caucasus) New member Username: caucasus Post Number: 5 Registered: 10-2010 |
Hi Dave, Your �nformations and advices are very usefull and informative for me. Thanks for your suggestions. I will wait for a good answer from our customer service then I will other suggestions in this forum. I will dark-ref correction if it is not adjust then I will change the some aparatur of our �nstrument such as reference cup, lamp etc. Because, there is a research project based on this instrument. If �t �s not fitted, project will be cut off for some time. Thanks Best Regards Fatih | ||
| David W. Hopkins (dhopkins) Senior Member Username: dhopkins Post Number: 168 Registered: 10-2002 |
Hi Fatih, In my opinion, the detector is unlikely to be the problem. Further, I believe your scans suggest there has been a rather striking change in the reference samples. It is important that you not continue to use the reference without an adjustment of the gain of the instrument. Using a darker reference, as suggested by Venky, may be a useful diagnostic, but really does not address the basic issue that your scans of the reference sample have changed, and you need to determine why. You also have to decide how to proceed, so you can obtain good spectra from your instrument. I am not an instrument service expert. I can only offer my interpretations of what I see in your data. Best wishes, Dave | ||
| Fatih KAHRIMAN (caucasus) New member Username: caucasus Post Number: 4 Registered: 10-2010 |
Dear Hopkins, I dont know that how many micron is our reference standart. But it is well pressed. Also i didn't take it out of instrument when i cleaned its surface. Our customer service sent us the reference matter as exernal powder cup. I take some spectrum and send back to them. They give me an answer if this problem is possible from reference matter or not. I think that it could be from detector? Is it possible? I will wait for this day then try to other advices in this forum. Thanks a lot Best Regards | ||
| David W. Hopkins (dhopkins) Senior Member Username: dhopkins Post Number: 167 Registered: 10-2002 |
Hi Fatih, The reference sample cup may well be your problem. Powders can compress and change over time, and this compression will actually increase the amount of light reflected. The changes can be sudden, such as a fracturing and rearrangement of the material. What kind of powder are you using in your reference cup? You need to handle the reference cup very carefully, so that it is never dropped, as this would almost certainly affect its reflectance spectrum. It is important that your reference cup not change while you are collecting data for your calibrations. If it changes after a calibration has been made, it may only introduce a constant bias between your calibration results and subsequent samples. It would be a good idea to purchase a pressed sintered PTFE sample for use as a reference, so such a change could be avoided in the future. I would recommend a sample from Art Springsteen, who responded earlier in this series. It is possible that simply readjusting the gain to obtain the entire reference scan once more within the range of the AD converter will allow you to obtain reasonable scans again. You may need to contact Unity Scientific to find out how to adjust the gain on your instrument. Perhaps it is described in your Users Manual? Best wishes, Dave | ||
| Fatih KAHRIMAN (caucasus) New member Username: caucasus Post Number: 3 Registered: 10-2010 |
Dear Venkatarman Thanks for your advice. I will try to Sample-dark/ref-dark application. Dear Hopkins I cleaned our reference cup after we meet with this problem. It is not a ceramic. Our reference is powder and it is closed in a powder cup. Also, we study about quality parameters agricultural food samples especially maize flour. We analyze it for protein, oil, ash, moisture etc. in our maize breeding programme. Thanks Fatih | ||
| David W. Hopkins (dhopkins) Senior Member Username: dhopkins Post Number: 166 Registered: 10-2002 |
Hi Fatih, You have apparently attached 2 scans of a single-beam scan of the reference disk that the instrument uses to calculate the ratio of the light reflected from a sample (I) to the intensity of the light incident on a sample (I0), which is an essential part of the calculation of the -log(I/I0) signal usually used to measure quality parameters such as the moisture and protein content of flour samples, and many other applications in food and agricultural samples. I agree with the previous responses that point out that the signal is apparently changed, and now is too high for the detector. Have you cleaned the reference disk? Does your unit use a ceramic reference disk? In most locations, it is a good idea to clean it at least weekly, and daily cleaning of the reference disk and the window protecting the optics (detectors and measuring beam) is probably a good practice. I think you should be able to readjust the gain to bring the signal somewhat lower, so that the readings at all wavelengths are less than 64000. I am surprised at the increase of the apparent width of the feature at approx. 1410 nm and its shift to longer wavelength. Hopefully, after readjustment of the instrument, you should obtain reasonable spectra from your samples. Out of curiousity, what kinds of samples do you measure? Best wishes, Dave | ||
| venkatarman (venkynir) Senior Member Username: venkynir Post Number: 121 Registered: 3-2004 |
kindly check back reference image .They look that it got saturation at 1300 ~ 1400 points. Please try do to the excerise of Sample-dark/refe-dark. You should few absortption peaks of polymer bands . | ||
| Fatih KAHRIMAN (caucasus) New member Username: caucasus Post Number: 2 Registered: 10-2010 |
Dear Howard, Art and Jerry Thanks for your interest again. Maybe, I must explain this file. Our instrument has polystyrene as internal reference matter.So, this spectra are same sample which are from polystyrene. Only they take at different time. Ind�cent figure from �nstrument installation and current figure from at this time. Ind�cent figure shows good spectrum of this matter. And current spectra must be as it. Our instrument use this spectra for baseline correction. We didnt cahange the lamp or other part of instrument. Only, I try to technical way for change the old spectrum file. I found the spc file of �nd�cent figure. And � change it with spec file of current figure. But it has never been changed. I contact the customer service of our country. But they dont give a good answer for solving my problem. Now, they contact to abroad service for this problem. Also � must say that � am not a professional person about NIR instruments. Only I am a basic user at level of calibration devoloper. So, my technicall skills is limited. Thanks again. Fatih | ||
| Jody Hoefs (hoefs) New member Username: hoefs Post Number: 3 Registered: 9-2010 |
Hi, we deal with Unity in NL, besides FTNIR of ABB, Unity has only reflectance measurments and the reference will be taken automatically as far as we know. Perhaps better ask Unity in the states, you can try Rob Gajewski [email protected] he knows everything. | ||
| Howard Mark (hlmark) Senior Member Username: hlmark Post Number: 356 Registered: 9-2001 |
Fatih - there are at least two separate things going on. It looks like there is too much signal in the region 1300-1400 nm, and either the detector or the electronics are being saturates, i.e., the signal is too strong (or a gain setting too high) and at some point (most likely either in an amplifier or the A/D) the system cannot respond lineariy to the signal level. To diagnose this, you can try partially blocking the beam of the instrument, to reduce the signal level. In addition, the apparent absorbance peak in the 1380-1550 nm region is broader than the one from the reference. This may be due to reduced resolution of the instrument, although that peak is so broad it seems unlikely that differences in the spectral resolution would affect it. What was the sample? The spectra you show appear to be single-beam spectra of water, except that there should also be a very stong band at ~1940 nm in this case. It's conceivable, but also unlikely, that the effect you're seeing is also due to non-linear response in the detector, compressing the background black-body spectrum at the higher levels and making the low-absorbance regions, corresponding to the high-signal regions, appear to be reduced and band broader in comparison. Have you asked the manufacturer about this? Their service department should be able to help you. Howard \o/ /_\ | ||
| Art Springsteen (artspring) Senior Member Username: artspring Post Number: 27 Registered: 2-2003 |
Hello, You don't mention if this is in transmittance or reflectance (not that it should matter). Are these the same sample? The reason I ask is they don't even look similar and a 100 nm shift in the absorbance peak is really strange. The other thing is there seems to be no absorbances at longer wavelengths. If the sample is polystyrene, there should be a strong absorbance at around 1650 nm, a multiplet of absorbances in the 2160-2190 nm range, and another strong absorbance around 2300 nm. (all locations are approximate-you can find exact positions on line). So it looks like the top graph is correct for the single absorbance but everything else is missing. Something is truily amiss. | ||
| Jerry Jin (jcg2000) Senior Member Username: jcg2000 Post Number: 36 Registered: 1-2009 |
Hi, Faith Welcome on board! It looks from the second spectrum that the detector is saturated around 1.4um. Did you change light source recently? Did you check the product support? Jerry Jin | ||
| Fatih KAHRIMAN (caucasus) New member Username: caucasus Post Number: 1 Registered: 10-2010 |
Hi all We have a Specrtastar 2400D NIR instrument. We interest to devoloping NIR calibration about quality parameters. But we have an unexpected situation about reference spectra of our instrument.There are some deviations from spectra of reference matter. I have been send as attached file which are good reference (incident spectra) and back reference (current spectra)images. What are possible problems about this case?Thanks a lot. Fatih   |