| Author | Message | ||
     Tony Davies (td) Moderator Username: td Post Number: 167 Registered: 1-2001 |
Hello everyone! The choice of instrument is very difficult. It really does depend on the applications. I suggest that you need to start the exercise with an unbiased mind. On the question of dispersive v interferometer, Karl Norris posed an interesting question in an article in NIR news. The title was : Interactions among instrument bandpass, instrument noise, sample-absorber bandwidth and calibration error; [NIR news 9/4, 3 (1998)]. Karl�s question from his calculations was �Is it easier to construct a 10nm dispersive instrument with a noise level less than 50 mu au or an 8 wavenumber interferometer with a noise less than 100 mu au�? I am not sure if anyone has provided an answer. Best wishes, Tony | ||
| Michael C Mound (mike) Member Username: mike Post Number: 15 Registered: 7-2007 |
Thank you, Ken. | ||
| Kenneth Gallaher (ken_g) Senior Member Username: ken_g Post Number: 27 Registered: 7-2006 |
I was "on holiday" during this thread, but a couple of comments. Amongst spectrometers with moving components I doubt there is any more robust means of scanning that the "wishbone" type of interferometer that Bomem and Bruker use. There is basically no mechanical mode of failure for the interferometer. I worked for a sister company of Bomem for years making process analyzers and never saw an interferometer failure. I started out using moving grating analyzers and would never go back except for cost issues, or the very high absorption situations where dispersives apparently do better. The still high level of dispersive usage is history and habit, not best choice usually. As far as moving samples - it depends on what you want. If you really want highly time resolved results then yes you need diode array or AOTF. If you do not care about time resolution then a changing sample does not matter - especially if you are averaging spectra anyway. | ||
| Howard Mark (hlmark) Senior Member Username: hlmark Post Number: 134 Registered: 9-2001 |
Mike - for both reasons (possible wear and measuring incoherent parts of the sample, as David brought up) it sounds like a moving grating instrument isn't the best choice. If the sample is moving, then I think you want a technology that measures all the wavelengths simultaneously. Concern about wear can be alleviated by use of a "no moving parts" technology. Diode array, or possibly AOTF (which is sequential, but switching between wavelengths is fast enough that they could be considered "simultaneous", especially if you measure only a few carefully selected wavelengths, each one for a short time. Both of these have no mechanical parts. You have to be careful about using other technologies. FTNIR, for example, measures all wavelengths "simultaneouly" but still takes relatively long times to scan, and the sample will be changing during the mirror scan, which is as bad as changing during a wavelength scan. For a moving heterogeneous sample, what you would like is a "snapshot" of the spectrum at each place you measure it, and, if necessary, average together many such snapshots. The ability to do this will depend on the implementations of the technologies, as well as on the basic concept. It may still require a long overall time to get an average spectrum that is stable and representative of the overall composition of the process material. You should check with the manufacturers for all the different technologies available, to see what each one can do in practice, in the particular implementation that they provide. \o/ /_\ | ||
| David W. Hopkins (dhopkins) Senior Member Username: dhopkins Post Number: 116 Registered: 10-2002 |
Mike, You have a worse problem than the monochromator motor duty cycle. At that speed of sample going past the measuring beam, you are not getting a single scan from essentially the same area of sample, so you have a tremendous sampling problem. Your results depend upon the sample not changing much in the scan time. What is your scan time? You might do better to have a slip stream that moves much more slowly, so you can get representative scans of the material, and possibly take samples now and then for comparison to the reference lab method that can be compared to the spectral determination. Best wishes, Dave | ||
| Michael C Mound (mike) Member Username: mike Post Number: 13 Registered: 7-2007 |
Howard, Yep, it's a tricky calculation. However, the system is anticipated to run 24/7, as it is programmed to function in an automatic and continuous mode. There are no samples, per se, as the spectrometer duty task is to scan masses in the interrogation zone continously in process mode. As to heat generated by exposure to constant reflection, this is really not an issue since the analyte materials (bulk materials) are transported into the zone of the spectrometer at a rate of 2 meters/second. This means that at no time is the reflecting response an event that is more than milliseconds for any mass region...hardly sufficient time for heat build-up. If the process halts for any reason, the system shuts down and restarts when the process restarts. Thanks for your always welcome responses and suggestions. Best, Mike | ||
| Howard Mark (hlmark) Senior Member Username: hlmark Post Number: 133 Registered: 9-2001 |
Mike - That's a tricky calculation. 2,000 hours equates to four months if the motor is running continuously, 24/7, the whole time. If the motor is running only during, say, "normal business hours", that's 40 hours/week, then 2,000 hours is 50 weeks, or roughly one year. I gotta admit even that's kinda skimpy. On the other hand, does the motor run continuously? In several instruments of which I'm aware, the grating motor runs only during active scanning time, which is usually only a small fraction of the time that the instrument is turned on; most of the time it's waiting to be used: samples being prepared, notes being recorded, etc. To say nothing of coffee, lunch, and other breaks in the analytical day. Is active scanning time about 1/10 of the total work day? If so then the effective lifetime works out to 10 years, a reasonable amount of time. As I say, a tricky calculation, since it depends on the usage. Regarding Lez's comparison with FTNIR: there are all sorts of considerations that come in when you start comparing technologies: time for a scan, S/N ratios, whether multiple scanning is necessary or useful, pre- or post-sample dispersion, etc. Lez's comments about the effect on the sample is particularly pertinent: shining wideband sources on samples have been known to change and even burn samples because of the heating effects, even through an interferometer. So as with most questions involving the instrument technology, "it depends" on the application. \o/ /_\ | ||
| Michael C Mound (mike) Member Username: mike Post Number: 12 Registered: 7-2007 |
Lez: I have had really good experiences with monochromators, so they are probably more hardy these days. Anyway, thanks again for your input. Mike | ||
| Lez Dix (lez_dix) Junior Member Username: lez_dix Post Number: 7 Registered: 10-2006 |
First let me admit I've skim read this thread. I have worked as an engineer on FTNIR and Monochromators. I do struggle today to find significant advantages to monochromators, they are mechanically more delicate and harder to clone. They can be more suseptable to vibration. However there are a lot out there and sometimes "better the devil you know". Also from the start of this thread, depending on what generation Monochromator if may be the encoder and not the motor. FTNIR tends to be post dispersive, look at what that may mean to your model and sample. Look at the robustness of the deck. Look at the sampling systems available etc. That's about as useful as I can be. If it was useful at all. | ||
| Michael C Mound (mike) Member Username: mike Post Number: 11 Registered: 7-2007 |
Howard, et al, Thanks once again. My concern is that the manufacturer of the spectrometer warrants the spectrometer motor on the grating for a maximum of 2000 hours service. This equates to only around four months, which seems an unlikely kind of assurance for an around-the-clock device usage. In any case, your caveats and suggestions are welcome and deserve consideration on my part. Thanks again, Mike | ||
| Ian Goodyer (zinir) New member Username: zinir Post Number: 2 Registered: 5-2006 |
Opening this out a bit: what do all you experts see as the main advantages (and disadvantages) of motorised grating based instruments over the other systems that are available? | ||
| Howard Mark (hlmark) Senior Member Username: hlmark Post Number: 132 Registered: 9-2001 |
Mike - as I said, there are several NIR technologies that are comparable to the fixed channel XRF/XRD types. But without detailed knowledge of exactly what happenned to the grating instrument that had the problem, it's unwise to jump to conclusions about the cause, or to assume that a more "robust" motor wouldn't have had the problem. The problem may not have been in the motor at all, it may have been in the optics, somewhere else in the grating drive or mounting, or in some other component; in any of these cases no amount of beefing up the motor would make a difference. Even if it was the motor, it could have been a factory defect, either in the motor or in the instrument or in the mounting or other related part; again, no amount of extra robustness built into the motor design would have helped. Presumably the instrument manufacturer selected a motor that would stand up to the expected usage; if not, they couldn't have stayed in business for all the warrantee repair they would have had to make. Don't forget that gratings are used because they have advantages of their own. When a different technology is appropriate, then that should be used. But regardless of the technology, any of them can "break"; and in all cases, for all technologies, the "fix" is the same: a call for field service, and the instrument operation is restored. To the user, this part could almost be a "black box", since what's important is accurate answers, not what's inside the box. \o/ /_\ | ||
| Michael C Mound (mike) Junior Member Username: mike Post Number: 10 Registered: 7-2007 |
Howard, Thanks. I would be the last person to condemn anything or anyone...just being cautious. If one were to stay with the motor grating version, I would think that a more robust motor component(i.e., more expensive, probably) would stand up to constant and long-term service. Thanks, once again, Mike | ||
| Howard Mark (hlmark) Senior Member Username: hlmark Post Number: 131 Registered: 9-2001 |
Mike - I don't recall the message you're referring to, but I think you should be careful not to condemn an entire technology because someone had a problem with one particular instrument. Every technology is subject to having breakdowns of various types. It would be like saying that automobiles are no good because someone had a flat tire once. But there are also many thousands of diffraction-grating-based instruments in use today, successfully cranking out results. That said, there are several other technologies other than motor-driven diffraction gratings. The oldest (and therefore most widely-used) of these is to use interference filters. These are probably best when you already know the wavelengths you wants to use for your analysis; since then you need only use those filters. These instruments tend to be limited, in that they generally do not measure the entire spectrum. Several other technologies exist, both without and with moving parts (which may or may not include motors), that form the basis of NIR instrumentation that allow measurement of the full spectrum; these include diode-array instruments, acousto-optic tuned filters, Fourier-transform and others. The variety of technolgoies makes it almost impossible to create an exhaustive list, or to try to compare all their advantages and disadvantages. \o/ /_\ | ||
| Michael C Mound (mike) Junior Member Username: mike Post Number: 8 Registered: 7-2007 |
Somewhere in the past discussions, someone had a problem with errors ascribed to motor gratings in the span above 1000 nm. Would it be possible to avoid a motor grating to handle the ranges from 250-2500 nm? If one were to identify the specific ranges of interest, as can be done with XRF and XRD with fixed channels (monochromators), mightn't this work with NIRS? These fixed channel monochromators are used in lieu of a motor-driven scanning goniometer in both instrument typss. Comments most welcome. Thanks, Mike |