-
Seek Compact Pro temperature readout issue
Posted by
cascade
on 09 Jan, 2018 05:20
-
-
#1 Reply
Posted by
cascade
on 09 Jan, 2018 06:19
-
After about 20 mins without unplugging but not all that time in the Seek app the cabinet is below freezing and my hand is still typical body temperature. Noise level is significant now.
-
#2 Reply
Posted by
cascade
on 09 Jan, 2018 06:28
-
So just after posting the last reply the Seek went unresponsive and i had to disconnect it and plug it back in. The app picked it up but now the cabinet is freezing -24C.
-
#3 Reply
Posted by
cascade
on 09 Jan, 2018 06:32
-
Another quick replug and the cabinet is now at -1C but my hand still at 36C.
-
#4 Reply
Posted by
Jarrod
on 09 Jan, 2018 15:38
-
Have you reported the issue to Seek support? If not that is probably a good idea.
-
#5 Reply
Posted by
cascade
on 09 Jan, 2018 16:48
-
Submitted my issue to Seek and looks like they are slammed:
Please note that we are experiencing an unusually high volume of inquiries and it may take us a while to work through each one. This may result in response times of up to 10 business days, so please review the links below, as some may contain the answer to your questions. Please do not submit repeat inquiries for the same topic, as it will only slow things down further.
-
#6 Reply
Posted by
Fraser
on 09 Jan, 2018 17:30
-
If you can return the unit for another, I recommend you do that. What you are seeing is very unusual behaviour that I believe to be a fault in the camera and not the host or software.
Fraser
-
-
The numbers you are getting are certainly disconcerting, but they are all at one pixel. Isn't there a mode where the app displays 2 markers at the min & max temperatures? If you do that, do you see the same sort of differences for the cabinet temperature? As you pan the camera around the scene, the markers will probably move to different pixels thereby enabling you to see if it is an individual pixel problem or "all over the place".
Seeks need to warm up about 20 minutes to get stable temperature values, but they should not be that varied; they just start out high and settle into their "correct" values. If Seek calculates temperatures anything similar to the way I do, referenced to the shutter temperature, then temperatures close to the shutter temperature will be more accurate than those very cold or very hot because they will be less affected by errors in the gain compensation. But I don't think the amount they are off could change like you are seeing.
Another thing, can you hear the shutter click every few seconds? The increasing noise suggests that it might not be operating. And they adjust things on each shutter frame. If you can hear the shutter and hold the thing steady and wait 10-20 seconds, does the temperature change significantly a few times? I have seen one of my non-pro units jump a few degrees Fahrenheit after a shutter frame (and then back on a later one), but certainly not 20 degrees; maybe your unit has a similar but much more severe issue.
-
#8 Reply
Posted by
cascade
on 09 Jan, 2018 21:50
-
I used spot mode to make it easier to describe the issue Im seeing. Regardless of the mode the temperature readout drop is there.
I can hear the shutter click every few seconds when moving and then it takes a little longer when in stationary position. It makes like a 2 step springy click. The readouts increase by a fraction after each click and the clarity of the picture gets better. Then the noise starts fading in and the temperatures get less accurate. Also at some point the picture can get really blotchy and with lots of noise.
Looks like the temperature readout is more stable (albeit being lower than what my IR thermometer is showing) when i keep the phone stationary. As soon as i start moving it and keep coming back to the same spot the readouts start dropping and then when i get out of the app for several seconds and come back i can see much lower values. It looks like it is able to catch up by a few degrees by cycling the shutter. By going in and out the app i can make it show below freezing temperatures on object that have a significantly lower temperature. Like going into a cold basement vs just looking at a warm tv. I can point it at the tv and get some high temperature readout then get out of the app and go to the basement, point it at the floor and turn the app on, hear the shutter refresh for the first time and get a below freezing values without seeing them climbing back up to a more realistic numbers.
The picture was taken from a desk in the office where the phone has been stationary pointing at a metal stool about 7ft away and a wall another 4ft behind it. The best temperature I get when taking a fresh measurement is about 60F which is way below 68F set on the thermostat. After letting it sit for about 10 minutes and doing some minor fiddling I get about 48F. Also the noise gets to an undesirable level.
-
#9 Reply
Posted by
cascade
on 09 Jan, 2018 21:53
-
And after posting the last message the picture just froze. The app is still working but I am unable to make the Seek cycle the shutter. And after disconnecting and plugging it back its working.
-
-
Okay, it's not just the one pixel, the thing is sick. I'm sitting here wanting to figure out what is wrong with it but you shouldn't have to do that. You have my sympathy for getting stuck with a "lemon". I wonder if their "unusually high volume of inquiries" is due to other similar complaints; maybe they made a bad batch.
Your description of the shutter sound and image shortly after shutter action is normal, but the ongoing and extreme degradation afterwards is not.
-
#11 Reply
Posted by
Micz
on 10 Jan, 2018 21:03
-
Same issue on mine... What emissivity you set ? When I set glossy it shows 330 C and real temperature is about 0. When I set to matt firstly shows about -25 C and after a while it shows -15 C
And sometimes I get some kind of fog in right corners
-
#12 Reply
Posted by
cascade
on 10 Jan, 2018 23:46
-
In fact observed similar extreme high when changing emissivity to glossy when outside in -5C degree weather. I think all issues you described I can find similar with mine too.
Really wanted to make a use out of this product but after reading around and getting a first hand experience with general quality and its issues not sure if that brand is worth my money.
To people that own Seek devices how is the longevity of these products? Would you say that with limited use it can last years or its one of these gizmos that you try to turn on after a year in a drawer and its dead or develops a major flaw for no reason?
-
-
To people that own Seek devices how is the longevity of these products? Would you say that with limited use it can last years or its one of these gizmos that you try to turn on after a year in a drawer and its dead or develops a major flaw for no reason?
I bought my first XR back in June of 2015 and the new XR and non-XR one year ago. They have spent quite a few days and nights out in the summer heat and freezing cold (as low as -10F) and as far as I can tell still work as well as when new. I have brought them in when rain is forecast to hit that side of the house, though, so they have not gotten really wet..except the non-XR got thunderstormed once. I also have had 2 of them apart & violated ESD precautions with them & they still work.
You guys are making me scared to hook this Pro to a phone again lest it get "upgraded" to a broken state!
Regarding that emissivity setting, though, if you tell the app you are looking at stuff with low emissivity when the scene is actually high emissivity it will cause you to get higher temperature readings because it thinks you are looking at stuff that doesn't give off much heat even if it is hot. Reflective (shiny) surfaces are like that & "glossy" makes me think that is what that setting is for. Shiny solder on a circuit board will normally show up as a cold spot even if it is boiling hot An emissivity setting that would compensate for that will make everything else look even hotter.
-
#14 Reply
Posted by
cascade
on 18 Jan, 2018 18:27
-
Just an update with the feedback from Seek support:
(Response time 9 calendar days)
If you are experiencing inaccuracies, it could be due to being used in extreme cold temperatures. Extreme temperatures (hot or cold) can affect the sensor and cause it to not be as accurate. Also, surface temperature measurements are affected by reflection and other factors. All Seek imagers are generally accurate within 5 degrees C above or below the reading. In some cases the Compact may have a short warm-up period (only a minute or so) during which the accuracy will improve from startup.
Please keep in mind that the units have been designed to provide thermal images and notional temperatures, not to perform as finely calibrated temperature devices.
Please let us know if you have any other questions.
Seek support
Based on that I love the Amazon prime no hassle return option.
-
-
I'm glad you are able to return the unit, but do you dare try another one?
As for the Seek response:
" In some cases the Compact may have a short warm-up period (only a minute or so) "
HA! More like in ALL cases and no less than ten minutes.
"All Seek imagers are generally accurate within 5 degrees C above or below the reading"
Why isn't that the spec. then, instead of +/-2C, or maybe they don't state it anymore.
"the units have been designed to provide thermal images and notional temperatures,"
I don't recall seeing that on their web site either. If it's "notional", they should explain that on their web site.
-
-
I realize this is an old thread but I just received a SEEK Thermal Compact PRO and wanted to post this for anyone who comes along later.
Buying from SEEK is a real game of Hit or Miss. Too often you miss. Meaning you will get a unit with issues.
I bought two. The first one was horrible in situations with low temperature differential where everything was close to the same temp or where there was not much heat.
The screen was basically noting but noise. Couldn't make out a THING. BUT....where the temp was above 80 or so, BEAUTIFUL images.
On the other one, there is more noise in the images ALWAYS. It's just not as sharp and clear. HOWEVER, with the 2nd one, I CAN see things better where there is low temperature differential or lower temps. Instead of totally unusable in those situations, it's marginally usable.
However, the 2nd one is shutting off CONSTANTLY. It's driving me nuts.
My conclusion is in the world of Thermal Imaging, you get a Toy for under $1000. Anything you get for under $1000 will be kinda sorta useful and kinda fun, but seriously lacking in quality.
I have a 30 day return on this 2nd one and I'm seriously about to pack it back up but will give it a few more days since I have 30.
If you buy one, just be aware it really is a cool toy and you can actually do some stuff with it, but so far, they all seem to have at least one or more of a basketful of inherent potential problems.
-
#17 Reply
Posted by
DaneLaw
on 16 May, 2018 22:05
-
It would make it so much easier to relate if you could back up your issues with some examples, after all its extremely easy to share from these product video or photos.
and since you don't write what host / or even standard (IOS or Android) you using it on, it gets even more difficult to relate.
Thermal cameras and particularly budget smartphone one's where their so extremely many varieties hence the host (phone) it not just plug and play, there are often many factors in play to use such product and get the optimal result like you fx see from others and particularly with Seek's product there is all dependent on the sensor and doesn't have a standard camera to lean on and use this gimmick as a make-up, like edge detection etc in regards to enhancing the result.
What system are you using and which phone? and have you tried on other mobile devices?
Have also taken the plunge on a Seek pro, and purchased it privately (socalled new) så that certainly also a risk, but that's pretty much the only option from here Denmark, as a Seek pro cost 800 to 850USD here on regional stores, after taxing and another +25%VAT on top and up there in price there is better alternatives..
Went for IOS as that was the only option for purchase at 290USD + minor shipping fee from US (no tax, private low declared) and had to go out and purchased a freaking IPAD 2018 as I didn't have any Items with lightning port.
Anyway just got it and here is some examples of the results I'm getting from a noob point of view and 290USD investment.
https://www.youtube.com/channel/UCTJtb7oyUVo0CVfd0RQ5rsw?Im getting this level of native focus.(video just made and don't have much to compare too, hence the first thermal/IR cam in my regi, but overall think its decent from a sub300 USD value. )
]
Wicked framerate (all over the place)
Frame rate : 17.074 fps
Minimum frame rate : 1.563 fps
Maximum frame rate : 600.000 fps
Full mediadump-info & values in the YT text section. all native' no thirdparty encoding or enhancement..
-
#18 Reply
Posted by
pauledd
on 01 Jun, 2018 06:12
-
Hi there
I am quite new to the thermal topic but regarding the temperature precision I have a question...
Assuming I can get raw data from the Seek Thermal Pro and I have a body in the scene with a known
temperature, would it be possible to relate the pixelvalue of the body's surface to any other pixel in the scene
and thus get a fairly close temperature value for any pixel?
-
#19 Reply
Posted by
Fraser
on 01 Jun, 2018 10:43
-
A reference emitter is not uncommon in Thermal camera deployments at Airports for the detection of Bird Flu etc. The Reference is of known temperature and Emissivity so may be used as a test target for the purposes of dynamic calibration.
To use a similar technique you would need to know the Emissivity of your known temperature reference and that of your target subject. Emissivity differences between reference and target will introduce an error into the situation that must be compensated for.
Thermal cameras of reasonable quality are normally specified as +-2C or 2% tolerance on measurements. You cannot rely on such a tool to provide greater accuracy than this unless you get into laboratory experiment levels of temperature references and measurements.
A thermal camera is normally calibrated using 2 point system. As such, any temperature that is not at those two points of calibration is liable to some level of error. The microbolometer does not exhibit a perfect linear response over its coverage. For maximum accuracy, a thermal camera needs to be 2 point calibrated with one of the points very at the temperature of the target. An alternative to provide greater accuracy is to use multi point calibration over the whole temperature range of the camera. Such methods are expensive and not available to the user of simple thermal cameras like the Seek products.
So in précis, a temperature reference will provide an absolute temperature accuracy indicator and confidence check, but it will not guarantee highly accurate temperature measurements beyond the specification of the thermal camera. Running a side by side temperature comparison using a variable temperature reference of known accuracy and Emissivity is a possibility but remember to compensate for the different emissivities involved.
Fraser
-
#20 Reply
Posted by
pauledd
on 01 Jun, 2018 10:51
-
cameras of reasonable quality are normally specified as +-2C or 2% to,earn even on measurements.
Okay, that explains seeks +/-5°C range for that price...
-
#21 Reply
Posted by
DaneLaw
on 01 Jun, 2018 11:44
-
Are most of these mobile attachment-cameras not way of the mark when it comes to temp accuracy.
one thing is specs and another thing is real life accuracy.
recall that Therm APP TH model that Opgal cherry-picked for MJlorton.
that was a 2000 USD product,
and it was completely off the mark both in temp accuracy and spot ratio.
it was something like -14% off at a basic 50* degree example and where it showed 43*. (and some of Thermp Apps product are with relatively low degree portefolio, but something like 50 degrees should be right in the ballpark where one should expect relatively good accuracy, particularly on a close subject under ideal circumstances.
The sum of it all' was that "temp-accuracy" varies a lot amongst different temperatures, the Opgal Thermp App TH was way better at high temp, then it was at basic temp like' 50degrees with an -14% off reading.
-and the spot ratio was listed at 100/1 and where MJlorton tests showed more like 37/1.
If you seek (sorry for the pun) high accuracy, then it's not these mobile attachments you shall look at..(at least the impression I'm left with)
-
#22 Reply
Posted by
Fraser
on 01 Jun, 2018 12:30
-
Sorry for all the typos in my reply. Writing quickly on an iPad is fraught with such dangers
Typos corrected.
Fraser
-
#23 Reply
Posted by
pauledd
on 01 Jun, 2018 15:43
-
I just did a quick test on my room wall with my cheapo PIR-thermometer (which has ε 0.95 hardcoded). It measured 27°C. Then with the Seek and cold start it initially showed 25°C on the same spot (ε set to 0.97). But then it starts to drop to 20°C after ~7 Minutes and stays relatively constant. If I continue to measure other surfaces and add 7°C to the displayed value I get quite close to the PIR-thermometer values. It would be nice if I could add some kind of software offset to the values displayed in the seek apk... Is that possible somehow? Or is this hardcoded in the "libSeekware.so" files?
-
#24 Reply
Posted by
Fraser
on 01 Jun, 2018 17:04
-
I will be honest with you, if I had a thermal camera, Self contained or Dongle, that displayed a 7 degree measurement error, I would return it to the seller or OEM for correction. That is a crazy error at ambient.
Fraser