Just like an image intensified NVD a thermal imager has quite a few parameters that dictate performance (detection range). If you are looking for a thermal imager these are important because you are fixing to spend several thousand dollars on any TI and a buyer should start with three simple questions: 1) what do I intend to do with a thermal imager? 2) what do I want the thermal imager to do? 2) how much do I have to spend? That seems like its over simplifying but by answering these questions you can figure out why a 320x240 thermal imager costs anywhere from $2.8k to $15k+. By far the major cost of the inside of thermal imager comes from the detector, lenses (germanium glass) and the display. The housing can be expensive as well but I am speaking about the internal parts at this point. DOD wanted thermal imagers that were lighter, smaller (and cheaper). The size and performance is usually tied directly to the cost. A small, high end thermal is going to cost a lot. A $12k 320x240 Insight MTM is not the same thermal imager as a $3k FLIR 320x240 PS-32. Typically speaking again, if you buy a small TI that is low cost it will not perform like a larger model in the same price range- you'll be giving up something. One way to look at it is that dollars buy smaller packages as much as they buy performance. The PS-24 has a lower resolution like 240x180 the reason its a weird resolution is because FLIR did the same thing that L3 did years ago- it was a cost saving measure. When L3 or FLIR manufacturer these detectors some of them come off the line with bad pixels in the array, usually at the edge of the detector. So instead of tossing them in the trash can they built a unit around them- al la the PS-24. Price point tends to be a big part of any sale and FLIR's price makes it attractive, it also makes it harder for vendors to explain the advantages of a more expensive system- there is a big difference between the low cost units and the more expensive TI's but the cheap ones have their place- if you only need to look a short distance the PS-24 or the PS-32 might fit the bill. A lot of our customers need to look farther out here in Texas so we tend to sell longer range systems. Here's a really brief overview when it comes to TI specs- this is not all encompassing and other factors come into play but its good to look at these on a spec sheet before you buy. Thermal imaging units can look very similar on paper side by side and operate very differently when out in the field- I think its at very least difficult to compare systems solely based on spec sheets. Resolution- Reso ranges between 100x80 and 640x480. 1024x768 detectors are out but you will need to rob a few liquor stores before you buy one. 320x240 is probably the most common with 640x480 growing as it gets less expensive. 640x480 detectors seem like they would be twice as nice on paper but looking though them the user notices that objects closer to the user tend to look better than 320x240 reso and then typically flatten out and only look marginally better at longer ranges. That seems counter intuitive and could be subjective but that is what most users see when you are comparing two units that have similar specs and magnification. The reason for this is that the higher resolution units can get away with smaller (cheaper, slower) objective lenses so manufacturers use it to keep costs down. If you buy a 640x480 unit with big expensive glass or add a $2k magnifier to it you will see bigger differences but now your at $15k again. Refresh Rate- Faster is better, you can see about 24 frames a second. Anything slower can be annoying and the reason you see high end systems with 60Hz is because it allows for better scanning and viewing from moving vehicles. The FLIR PS series are 9Hz and that's a big gripe about the PS series. Magnification- This is important and like a camera you need to know the difference between optical mag and digital. A 2x digital zoom basically cuts your image quality in half verses optical zoom that will maintain the image quality but affect the field of view, like any daytime optical system would, like a rifle scope. I won't go as far as to say that digital zoom suck but its not very good, even on high res thermal imagers. The bad part about optical zoom is that the lens will get bigger and goes back to the cost. That lens on the W1000's you have seen online lately? It cost almost $4000 to produce that 100mm lens and you can smash it on the ground and the metal scrap yard will still give you about $1000 for the little pieces. Germanium is expensive, really expensive. This is where buyers make a mistake a lot of times. They buy a nice handheld TI that doesn't have enough lens. Sometimes you can add an afocal lens but the lens is very expensive as well. High resolution systems not withstanding you can usually assume that if the objective lens is smaller you can expect shorter detection ranges. Don't go buy a $8k COTI clip-on if you want to scan long distances across fields, buy one if you want a small, fused NV and thermal and look around short ranges. FLIR got the PS-24/32 cost down considerably by learning to pour and polish the germanium lenses. They aren't a diamond point turned lens like larger lenses but hey the PS series still work, just at shorter ranges and slower refresh rates. Detector pixel pitch (measured in microns) Don't stress over it- its more about manufacturing and less about performance. You can make the detector smaller and it also relates to smaller cheaper glass again but it doesn't mean a lot in relation to what a user sees looking through the unit. A 17 micron pitch (about the smallest commercial pitch currently) will get you a smaller detector and this could be put in a smaller housing- that's really the primary advantage after you pay for it. If FLIR can fab more detectors on a single wafer and then they can sell them cheaper. Big Army agrees- while new 17 micron thermal weapons sights are fine thay won't pay even a nominal amount to upgrade their current 25 micron PAS-13's despite being hounded buy BAE, DRS and Raytheon. Sensitivity- This is important but you have to be careful because its a two edge sword. Sensitivity is important- it means most to the user when the humidity and ambient temperature is higher, allowing for better detection ranges. Bad part is a detector trades response time for sensitivity and now you have a new set of problems to deal with, you can end up in the same boat or even worse in some cases if you don't have a clever way to mask it with software. You FLIR M18/M24/307 owners will notice this, the sensitivity is good on all those units but at times the background will 'wash out' at times and this its whats happening, a NUC calibration will sometimes clean it up, and sometimes won't, depending on conditions. Honestly, 30mK to 80mK is fine as long as the firmware is written well along with some other considerations. (ATN THOR customers should be reading this twice because ATN simply bought a detector [just the IC chip itself] from FLIR, in order to keep costs down and are now learning that those boards around the detector are important too.) Detector type- This is a fight between FLIR, Raytheon, DOD, BAE and L3. All current and detector types work pretty well with small/minor differences. If any of the smart guys from above are reading this then I will get crappy PMs but in reality AS, VOx, BST all look pretty good when comparing them side by side. These different technologies do use different methods of non-uniformity calibration (recal) so you could argue about choppers and shutters vs manual recalibration but really the only topic a typical end user talks about is that the most widely used device- the shutter- drops in and freezes the picture frame about every 2 degrees or after a specified time. People have figured out that this happens a lot right when you don't want it to, this isn't Murphy's law the unit is doing it on purpose. The military has noticed this too and now you have systems that you manually have to re-calibrate. This may be even MORE annoying because you have to remember to recal it every few minutes or the image slowly degrades with you finally remembering after you wonder what you have been missing because you forgot to cal the unit for the last 15 minutes. Weapon Sights- this is especially important side note. If you want a weapon sight you should really take care in choosing a quality unit, the vast majority of commercial thermal weapon sights simply use an off-the-shelf detector and board set, add a reticle and a housing that will mount to a rifle, shoot it for a while and pray that they don't break under recoil. Even the detectors themselves are designed different and you blow up a thermal imaging weapon sight on a rifle you just broke the part that costs the most in the system. There are some commercial units out there that are holding up to decent recoil but they are generally the more expensive units. A thermal weapon sight is great because its extremely fast at detecting and getting on target and its by far easier to video your night time pig killing rampage but unless you by a 3x optical unit or better you are going to be shooting at relatively short (~120 to 150 yards at 2x)ranges. 640x480 is pretty much the top resolution and it still lags behind a night vision device but again the thermal imager is extremely fast when it comes to target acquisition. All in all, try to look through the system you want to buy, or talk to someone that has looked through a lot of different thermal imagers so they can relay the differences between the systems. I didn't discuss the displays, the user menus and other features that also can affect your viewing pleasure (or displeasure).