Unfortunately this will not help you. This type of IR photography uses reflected IR, where the source of the IR is from the sun lighting up the objects. For temperature measurements, you need to detect long-wave IR actually emitted from (not reflected by) the objects you wish to study.
Mark
Yes, i have found that the LM35Z is not only within 1C but that it does not get too far off even up to 185C (3-4C higher). Not too bad for around a dollar.
You'll be lucky to get within several degrees C in boiling water, let alone 0.1°C. That's almost two orders of magnitude off.
You can also buy a "Class A" RTD sensor. It will be within +/-0.15°C at 0°C and +/- 0.35°C at +/- 100°C. You can calibrate your instrument to much better than that with ordinary precision electronic equipment.
It's difficult. The glass thermometer requires immersion to a certain depth in liquid in order to read correctly. You could conceivably (and with significant care) use that for calibration.
Forget about it.
Best regards, Spehro Pefhany
-- "it\'s the network..." "The Journey is the reward" speff@interlog.com Info for manufacturers: http://www.trexon.com Embedded software/hardware/analog Info for designers: http://www.speff.com
You can do pretty good radiometry at 3-5um, at least if you're working above freezing. I used to work for FLIR systems, going back to when the Portland operation made handheld cameras with PtSi cooled sensors. These days it's all done with microbolometer arrays at 8-12um.
-- Tim Wescott Wescott Design Services http://www.wescottdesign.com
You've already been acquainted with the difficulties inherent in this, but if you're nuts enough to continue your pursuit I just want to note one thing:
A long time ago in a job far, far away I worked with a cheap Chinese CO2 sensor. It worked by shining IR from a heated wire through a length of gas, then running three filters in succession in front of a thermopile to measure CO2 absorption, H2O absorption, and clear-path energy. It did OK when it worked but it was always breaking and my boss and I wanted nothing more than to replace it with one of our own manufacture.
After he ran out of money and laid me off it occurred to me that a better solution for the CO2 meter would have been to use a spinning diffraction grating; by timing when you read the sensor you would get full spectral information at the cost of having to keep it aimed at each spot for a good while.
-- Tim Wescott Wescott Design Services http://www.wescottdesign.com
They can do that.
A microbolometer array is a micromachined sensor array with each pixel having its very own itty bitty bolometer or other heat measuring device. Used to make imagers.
The last time I was paying attention the cool way to do it was with vanadium oxide resistors; these were sensitive enough and could be deposited on the silicon easily enough to work. The only unfortunate bit was that a given array could have 100mV of temperature signal riding on 3V of plain-old offset, so calibration was interesting, to say the least.
-- Tim Wescott Wescott Design Services http://www.wescottdesign.com
Forget thermistors. If you want cheap and accurate, the best thing is a plain old PN junction. Apply a constant current, and measure the voltage. Calibrate in ice water and boiling water. This approach is extremely linear, but like all approaches at this kind of accuracy requires extreme care. For instance, it will be no more accurate than the constant current source, and those tend to drift. The current can cause self-heating. And it is not as simple as it sounds to do a good ice water and boiling water calibration, since there will be impurities in the water that change the temperatures, etc.
The modern solid state approach uses a switched current through the junction, where the ratio of the higher current and lower current are well characterized (precision resistors controlling the current source, etc). This method is used to provide absolute as well as relative accuracy, so you don't typically need to calibrate. But you could probably use one of these chips and supply your own calibration to get the desired accuracy.
Best regards,
Bob Masta dqatechATdaqartaDOTcom D A Q A R T A Data AcQuisition And Real-Time Analysis
Not good advice. The extra sensitivity makes them a lot easier to use that Pt resistance sensors, and many interchangeable parts offer
+/-0.2C accuracy at a relatively low price. +/-0.1C and +/-0.05C interchangeability are available, at a price.Only if you wnat to calibrate it. And even then you are probably better off with something like the LM35. Calibration costs more than buying well-specified parts, particualry if you have to work up the procedures from scratch.
Make sure that you have a barometer to measure the atmospheric pressure above your boiling water. It makes a difference - strictly speaking, you want to calibrate you sensor when it is exposed to condensing steam at atmospheric pressure, rather than when it is immersed in boiling water, which can get super-heated ...
Not if you used distilled water - de-ionised water is usually fine too.
Serious precision temperature measurement around room temperature still seems to use platinum resistance sensors in an AC-excited bridge.
You need one low-temperature coefficient precision resistor in series with your sensor, but the other two arms of the bridge can be a thin film divider or a ratio transformer (which can be even more accurate, and stable). Check out national standards laboratory set-ups.
------------ Bill Sloman, Nijmegen
You can skip the precision constant curent with a radiometric measurement on a bridge with precision low tempco resistors.
Rene
I read in sci.electronics.design that Tim Wescott wrote (in ) about 'need a temperature sensor with .1 degree accuracy', on Fri, 14 Oct 2005:
Langley devised the bolometer It's really a kind of thermometer Which can measure the heat From a polar bear's feet At a distance of half a kilometre
Author unknown
Presumably a microbolometer is for measuring teddy bears' feet at 50 cm.
-- Regards, John Woodgate, OOO - Own Opinions Only. If everything has been designed, a god designed evolution by natural selection. http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk
A well stirred bath of boiling distilled water (beaker on a magnetic stirrer hot plate) and corrected for local barometeric pressure should get you very close to 0.1°C accuracy.
Right, and the fine print probably states that the glass thermometer is guaranteed good to 0.1°C only at one specific temperature, IIRC the one I have used covers -10°C to 110°C and was NIST tracable certified accurate to 0.1°C at 37°C when immersed to the line in a stirred bath and exposed to 20°C air above the line. The catalog may omit these details but the cal sheet will not.
This gives you 3 cal points (ice, thermometer cal temp and boiling) within 0.1°C if carefully done and corrected for pressure. When I last did this ~15 yeqrs ago the glass thermometer agreed within 0.1°C at ice point and boiling point also, and 3 of 4 types of RTDs used (SST jacketed immersion probes for liquids, small glass bead RTDs for air and spring loaded ceramic packaged types for surface temp) when adjusted at 0 and 100°C (using Analog Devices 5B isolated RTD module zero and span adjustments) would all track the thermometer within
0.1°C over the full range from 0 to 100 quite reliably; way better than needed for the intended purpose (diesel generator testing) but IMO you should calibrate with a reference having at least 4x required accuracy whenever possible. The six 10 ohm copper RTDs embedded in the 3 MW generator windings were not amenable to immersion in the stirred bath, however :-).This is getting, confusing, complex and interesting!
In the same line as the Omega hint, and in the same cole palmer catalog, I saw a Digi-sense Digital Handheld thermistor thermometer, which has range of -40 to +125 Degree C and claims accuracy of +- 0.2 Degree C. Also offers an optional probe. Now is there any fine print here?
It appears to me till now, that the process of measuring "true" temeperature is more complex than measuring temperature of the sensor accurately? I mean this friend person just wants to go outdoors (with the sun shining and all) and go to a spot (a longitude/latitude that he identified from his satellite thermal IR image) and take/sample the temeperature there. and then he wishes to relate the satellite thermal imagery with the on-ground samples that he thus takes, and come to some conclusion about this urban heat phenomenon..and sometimes he cannot go INTO dense vegetation, or a brick kiln, so he infers the temeprature at THAT site by pointing the IR thermometer at it.
so what I mean is, Can he use this 30$ Digi-sense thermometer at the site and just take a sample and note it and claim that AT THAT POINT ON EARTH AT THAT TIME the temperature was X degree C +- 0.2 deg C? or it just dosen't work like this?
But I shall, at least in my spare time and slowly, wish to interface and calibrate the sensor, because he takes PDA computer, and it has a RS 232 port, and I'd love it if he just pressed a button and the temperature was directly logged with the time and the GPS position (yes the GPS is fairly inaccurate!)
It also occurs to me that I need to better understand this "temeprature" term!
Many thanks, Suraj
P.S. As a bonus, I experimented with my digital camera and used my TV remote to illumate a target, and gave a 5 second manual exposure and the image came rather decently bright! this leads me to making a BIG IR torch and do some "stealth" photography!
snipped-for-privacy@worldnet.att.net wrote:
Look at Omega,
Standard disclaimers apply; I don't get money from Omega.
Matt Roberds
The satellite surely can't give remote sensing data accurate to tenths of a degree. Is your friend directly involved in evaluating the satellite or is there some other motivation? Then there is an issue in time and spacial resolutions. Seems like something about this project is either over done in specifications or under described.
Don't assume that marketing copy writers understand the differences between "digits displayed to the right of the decimal point" and "precision" and "accuracy."
As a minimum, read the manufacturer's datasheet, not a blurb on a web page. If they are indeed making an assertion with respect to absolute accuracy then I would expect to see:
A number for the instrument's accuracy as a standard deviation or a statement that the stated accuracy is +/- so many std devs. There should be a statement regarding operating and/or environmental limits.
AND a statement that the instrument calibration is traceable to a standard maintained by NIST or an equivalent national standards bureau.
AND a signed and dated seal on the instrument (preferred), or a similarly signed and dated document with the instrument's serial number, stating by whom and when it was last calibrated and including a graph or table showing the residuals between the measurements of that instrument and a maintained standard throughout the instrument's range.
Absent all three of those, any claimed "accuracy" is, at best, a dumb mistake by the English major who wrote the blurb and, at worst, a deliberate attempt to mislead the customer.
-- Rich Webb Norfolk, VA
I looked at the Cole-Parmer web site and tried to find the specific unit you saw. I'm looking at their model 93210-00 thermistor thermometer. And there is fine print. If you took the time to download and read the instruction manual for that model, you would see that it uses a YSI 400 series calibrated thermistor. These thermistors are manufactured and specified to very close tolerances, and can be used to good advantage in temperature measuring equipment. The model 93210-00 thermometer is specified to have an accuracy specification of
+/- 0.2 deg C, but look at the specs a bit further. This accuracy is only valid over the temperature range of 18 to 28 deg C. The "useful range", or the range of temps that are deemed to be "in the ballpark", is a bit larger at 0 to 40 deg C. Everything else is meaningless. This same method of specification of themometers applies to most other consumer-level (cheap) instruments. You have to read the complete spec.. not just the one that the marketing folks use to catch your eye.I couldn't find any of the Cole-Parmer models for $30 US... Which model did you see for that price?
I did see a model EW-93210-50 logging thermometer that boasts much better accuracy, but it's also twice the price of the other unit. It offers you an RS-232 port for connection to your laptop or handheld computer. These units can offer the higher accuracy because of the tightly specified thermistors used in the probes, and by use of a microprocessor-based measuring system. The processor uses a lookup table that was created at the time of calibration to compare the thermistor reading to the table of temperature values. The temperature from the lookup table is the value displayed. When these units are calibrated, they have to be calibrated as a probe-instrument set. In order to maintain the stated accuracy, they have to be used as a set.
I'm not sure that "temperature" is your problem. I think you might need to understand how to read and interpret manufacturers specs, which should be a required course toward an engineering degree. {:>) Just be sure that you look at the complete spec before you buy (or recommend) an instrument.
Cheers!!!
-- Dave M MasonDG44 at comcast dot net (Just substitute the appropriate characters in the address) Never take a laxative and a sleeping pill at the same time!!
Not to disagree too strongly with your comments, Dave, but "temperature" is his big problem -- or at least understanding the principles of temperature measurement.
Go back, Suraj, and reread my other post on this subject. There are fundamental matters that you have to understand and take care of before you even start reading manufacturers' data sheets. Keep in mind that all these devices maintain their specifications only under certain definitely specified conditions, and if you don't use the equipment under these conditions, you don't get the accuracy in your measurement, even if the equipment meets its specifications perfectly!
I haven't looked closely at the satellite measurements' techniques and specifications, but I do know that they get their fractional degree data by averaging many, many readings over rather wide areas. There is no such thing as a city with a temperature of 22.7C. There will be a city whose average temperature at a particular time is 22.7C, with a range of individual measurements at that time maybe +-5C or more.
The satellite measurements can quote this kind of resolution and accuracy because their sensors are stable and known to this kind of resolution and accuracy. So if yesterday's temperature for Rome is
22.7C, and today's is 23.2C, you can be sure that the average of the readings of many places in Rome is 0.5C higher today than that average was yesterday. That's all. Your friend is not going to verify the satellite readings with a handheld, or IR, or any other kind of ground-based measurement. I doubt seriously he could get an equivalent measurement with the satellite's own equipment from a balloon!The best he could do is verify that his location is some temperature near the satellite's reading. But then, maybe his location is in one of the extreme places :-). If he had dozens or hundreds of sensors spread all around the city, he still might not be able to couple his sensors to their surroundings well enough to get an equivalent reading (remember, you can never measure the temperature of an object -- only the temperature of the sensor). IR readings won't help, either, since they are made from different angles at different distances over different areas.
Now, if your friend has access to a satellite that he can use to rendezvous with the weather satellite, and take the same kind of measurement at the same time, from the same location in orbit, with sensors calibrated to the same specifications under the same conditions, then he might have a case for saying reasonably that he's verified or disputed the satellite's measurement.
John Perry
This is true, that would even be much better solution than trying to apply precision constant current circuit.
Bridge is easy and cheap to get 0.1% tolerance resistor for the bridge and relatively good instrumentation amplifier with high spec DC performance (low offset and drift). I think there is chopper stablised (or auto-zero) op-amp. Depending on your end product application.
In addition, you do not need calibration to verify the circuit, you may switch to 100R fixed resistor and check for zero readout if you insist. on calibration.
Good lucks.
Riscy.
Lemme see....will two different types of thermocouples and a glass thermometer, all agreeing with each other within 1C worst case be good enough?
I understand that digital camera constructionhas been changed to include an IR blocking filter. Why? So you cannot use it to take tittie pictures thru (transparent) tshirts....
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.