Author Topic: Uni-T UTi690A Review ( UTi690A UTi120S UTi712S )  (Read 15992 times)

0 Members and 1 Guest are viewing this topic.

Offline Paw85Topic starter

  • Regular Contributor
  • *
  • Posts: 60
  • Country: se
Uni-T UTi690A Review ( UTi690A UTi120S UTi712S )
« on: November 24, 2021, 06:23:19 pm »
This is my first review of a thermal imager, I will try my best and update the main post with more information as time goes to keep findings up to date.

The Uni-T UTi690A is a thermal camera with a resolution of 120x90 pixels in the budget range of ~180-190 USD.
You can find it both at AliExpress, and at Banggood here: https://www.banggood.com/custlink/DKvygTWWDM

Specification:
Resolution: 120x90 Pixels
Thermal range: -20°C to 400°C
Field of View: 38°(H)*50°(V)
Depth of Field: 10cm to infinity
Display: 320x240 2.4" TFT
Storage: Micro SD
Charging/Data: USB-C
Other features: Flashlight, Laserpointer, Tripod mount thread.

Review:
What you get with the thermal imager is an USB-C cable, a lanyard, a 16GB noname Micro SD card, a manual in mandarin and english, and a warranty/check card in mandarin.

The build quality seem very decent for the price.
It has a hard plastic shell with what appears to be TPE moulded on it to protect against impact.
It is rated as IP54 and says to be drop-protected to up to 2 meters.
At the top is a rubber cover that protects the USB-C port and SD-card slot.

The UTi690A does not have a normal camera so you're completely dependent on the thermal image shown on the display.
It does however have both a tiny flashlight and a laser pointer that help in low light situations and if you're completely unsure where you're pointing it at.
The flashlight is not very bright but OK for use at night time, and the laser pointer does not point exactly to the center of the image, but better than nothing.

Startup time is approximately 6 seconds from a cold boot.
The refresh-rate is not super smooth but works fine, the sensor may be ~25Hz as claimed but not the update rate of the display, unsure what refresh rate it actually have.

The menu system feels very responsive and even if the unit have less buttons than most thermal imagers, it's very easy and quick to nagive through the settings.
It have 6 different palettes to choose between. I do not know how they are on other Uni-T units, but they are not as good/clear as FLIR palettes, sometimes details is not easy to see with some palettes but easy with others.
Unfortunately there is no way to define the min/max temperature value of the palette yourself, so if something hot suddenly appear in the image, it'll adjust for that and leave small temperature difference difficult to spot.
When the temperature delta is below 10C degree of the recorded image, the palette will no longer be accurate: The min value will be correct but the max palette color will be 10C above that, no matter what it actually says on the display, leading to low contrast on the final image.
The temperature data do still seem to be correct thought, and one can try the different palettes as some will give better contrast performing under these circumstances.

When it comes to what information can be displayed on the screen, the following options are available: Center Point measurement, Hi/Lo Point measurement, Palette, Date&Time.
This is quite limited for some use, but for the price point it seem common with these options, and have the feeling that more advanced options may be too much for the processor to handle.

The image appear quite noisy but I only have a hacked FLIR E30bx to compare to, so not a fair comparison, when looking online at thermal imagers in the $200-300 range they seem to have similar performance when it comes to noise.
When it comes to sensitivity I dont know how to compare it to others, but holding a hand against the wall for 2 seconds, then it's very easy to see the full thermal mark of the hand.

For the use I purchased it for, to troubleshoot 3D-Printers it seem to work great.
You can easilly see single wires when they heat up as current pass through them, and the measurement at the center point gives the same readings as my FLIR E30bx does.
The work range does seem to be quite accurate, you seem to gain additional detail in the images down to a distance of ~10cm from the sensor.
I checked if possible to identify PCB-components and did so quite easily down to a size of ~3mm, smaller components than that and you would need an additional ZnSe lens or similar.

My conclusion:
Overall, happy with the thermal imager for the price paid, it seems to read measurements accurately and it's a unit I'm not afraid to handle or to let others borrow unlike the expensive E30bx unit.
For checking on 3D-printers and other objects of similar size it seems to work great.

Quirks:
The laser pointer is not on continusly when enabled, instead it's on when you hold the trigger. In this mode a picture is only taken if you do a quick pull on the trigger.
The unit came with Mandarin language by default, but quite easy to find where to change it to English.
Pressing the down button during normal use forces the sensor to calibrate.
« Last Edit: December 29, 2021, 07:41:39 pm by Paw85 »
 
The following users thanked this post: Fraser, edavid, LarryR, mrisco, redbook, impetus_maximus

Offline Paw85Topic starter

  • Regular Contributor
  • *
  • Posts: 60
  • Country: se
Re: Uni-T UTi690A Review (120x90)
« Reply #1 on: November 24, 2021, 06:43:24 pm »
Here are a few pictures taken with the UTi690A.
Some images were taken when trying out different palettes, those images do not have any measurements.

Picture 1: Wire to heated bed on a 3Dprinter.
Picture 2: Image of a thermal mark from touching a wall for 2 seconds.
Picture 3: Closeup of my daughter
Picture 4: Radiator (Distance 1 meter)
Picture 5: Radiator (Distance 0.5 meter)
Picture 6: Radiator (Distance 0.25 meter)
Picture 7: Radiator (Distance 0.1 meter)
Picture 8: Tubes/Wires at a car engine a few hours after use
Picture 9: Hand mirrored in stainless steel coffee machine.
Picture 10: Tool hanging grid, a few seconds after holding thumb against it to heat up. (Distance ~0.1 meter)
 
The following users thanked this post: mrisco

Offline Paw85Topic starter

  • Regular Contributor
  • *
  • Posts: 60
  • Country: se
Re: Uni-T UTi690A Review
« Reply #2 on: November 24, 2021, 07:03:15 pm »
I was not sure how to show the size of the device, but here is a comparison to the FLIR Exx for reference.
« Last Edit: November 24, 2021, 08:41:28 pm by Paw85 »
 
The following users thanked this post: mrisco

Offline mrisco

  • Contributor
  • Posts: 20
  • Country: pe
Re: Uni-T UTi690A Review
« Reply #3 on: November 24, 2021, 11:24:17 pm »
Excelent!
 

Offline LarryR

  • Contributor
  • Posts: 23
  • Country: ca
Re: Uni-T UTi690A Review
« Reply #4 on: November 26, 2021, 01:26:26 pm »
Thanks!  Great review.

Are you going to open it and take photos as well?
 

Offline Paw85Topic starter

  • Regular Contributor
  • *
  • Posts: 60
  • Country: se
Re: Uni-T UTi690A Review
« Reply #5 on: November 26, 2021, 04:18:12 pm »
Are you going to open it and take photos as well?
Unfortunately I will not, I was hoping that the small cover on the top would show some hidden screws or that screws would have been located behind the labels on the sides, but I can not find any screws externally.
The biggest reason I purchased this thermal imager was due to the IP-rating (and price/performance ratio) as I do not want to bring my expensive FLIR E30bx unit outside if it is for example raining, or in the garage where it can get very dusty (depending on what project is ongoing).
As it very likely uses a snap-fit construction I do not want to risk loosing the IP-rating if accidentally damaging any of the the snaps or seal during teardown.
That being said, I do hope someone with a unit will be brave enough to do a full teardown as I'm really curious about the construction!
« Last Edit: November 26, 2021, 04:20:39 pm by Paw85 »
 

Offline LarryR

  • Contributor
  • Posts: 23
  • Country: ca
Re: Uni-T UTi690A Review
« Reply #6 on: November 30, 2021, 03:17:03 am »
Not a problem. I just received mine today (ordered during the 11-11 sale). Quite a difference between my Seek compact and this unit but I did notice the uti120s uses a mechanical shutter like the Seek, for calibration as you can hear the clicking if you hold it up to your ear.

I'll see if I can figure out how to open the housing gently.
 

Offline mrisco

  • Contributor
  • Posts: 20
  • Country: pe
Re: Uni-T UTi690A Review
« Reply #7 on: November 30, 2021, 03:27:29 am »
Not a problem. I just received mine today (ordered during the 11-11 sale). Quite a difference between my Seek compact and this unit but I did notice the uti120s uses a mechanical shutter like the Seek, for calibration as you can hear the clicking if you hold it up to your ear.

I'll see if I can figure out how to open the housing gently.

Could you confirm that the UTi120S is the same unit as the UTi690A from this review?
 

Offline LarryR

  • Contributor
  • Posts: 23
  • Country: ca
Re: Uni-T UTi690A Review
« Reply #8 on: November 30, 2021, 02:19:54 pm »
From the specs, operation and outside appearance, the 690a and 120S appear to be identical.

You can remove two rubber covers to expose the lower housing screws in the handle but the two upper screws are hidden behind the display cover and I don't want to attempt to remove that just yet:
1336307-0

 
The following users thanked this post: mrisco, impetus_maximus

Offline Paw85Topic starter

  • Regular Contributor
  • *
  • Posts: 60
  • Country: se
Re: Uni-T UTi690A Review
« Reply #9 on: December 01, 2021, 11:05:28 am »
From the specs, operation and outside appearance, the 690a and 120S appear to be identical.

Out of curiosity, did the 120S come with Mandarin warranty card and Mandarin as default language on the device like the UTi690A does?
Too bad the two screws are under the screen cover, which most likely is attached with some kind of adhesive, will make reassembly quite a lot trickier.
« Last Edit: December 01, 2021, 11:07:50 am by Paw85 »
 

Offline LarryR

  • Contributor
  • Posts: 23
  • Country: ca
Re: Uni-T UTi690A Review
« Reply #10 on: December 02, 2021, 12:03:27 am »
Yes, both were in Mandarin. However, if you don't know about Google Lens, install it and play.....if you run Lens and point it at the warranty card then press the translate button, everything magically converts to your native language.

I've used Lens to match items to the billions of images Google has cached on its servers. It's a good utility to have. Not sure if it's available for iOS.  But I digress.

One thing I did notice is that there is not much of a dynamic range, at least to what is displayed. As you pan away from a view with lots of contrasting items, the display washes out rapidly.

There was talk in the Seek thread about using a Peltier device for cooling in order to artificially generate better contrast but it never went anywhere.

Any updated thoughts on cooling a cheap bolometer like this one?
 

Offline Paw85Topic starter

  • Regular Contributor
  • *
  • Posts: 60
  • Country: se
Re: Uni-T UTi690A Review
« Reply #11 on: December 02, 2021, 06:23:11 am »
I have a feeling as I was trying to write in the review that the palettes are not as good as on the FLIR E30bx that I can compare it to.
On the E30bx it feels like it adjusts the palettes really well between high and low values, giving a great degree of contrast even though there is only a very small temperature difference.
On the UTi690A, even thought it shows the min/max temperature at the palette, only the min value will be correct for the palette if the total temperature difference is below 10C degrees, where max value will not be at the max color the palette shown, giving a lower contrast.

I suspect this may be to "cover" how much noise there is in the images as higher contrast probably will make the noise very visible, just wish that it'd update the palette to the right as well as it becomes misleading when the temperature difference is very low.
Some palettes do seem to perform better than others thought in different scenarios when it comes to contrast, have you tried them all and see if you find one that works well where you need it?

I have attached a picture showing a cold spot where ceiling meet the walls in a room. On the grey palette it was impossible to see the cold spot on the screen (but easier seen on the computer), but with the other palette it was very easy to see it on the screen.
I do not know if this is how all Uni-T thermal imagers work or not, nor if lesser FLIR models do the same "trick" to cover up noise in the smaller sensors?
My best guess is that the palette will be accurate when the temperature difference is 10C or more, then the max value seem to line up with the palette.

That being said, the UTi690A is still able to show small differences in temperatures, and I'm still able to find where the temperature differences are as fast as with the E30bx, it is just as you say low contrast.
The recorded temperature values do seem to be alright, so whatever they are doing with the palette scaling does not seem to affect temperature readings.

I have updated the review with these findings.

EDIT:
I can probably confirm that the low contrast with the way the palette behave is intentional: I tried to open the images with the software for the UTi120B, and the images open just fine but once again do not scale to the graph, even when manually adjusting the palette endpoints within the software;
above a certain point the image no longer update the palette, just like on the thermal imager, so it's a design choice by the manufacturer, as both the thermal imager and the computer software do the same thing to the image representation.

I have uploaded a picture of when importing the image in the software and manually adjusted the max value of the palette to where one can see any difference, and it is very low value, giving false palette for low contrast images also in the software.
« Last Edit: December 02, 2021, 07:42:53 am by Paw85 »
 

Offline Bill W

  • Super Contributor
  • ***
  • Posts: 1102
  • Country: gb
    • Fire TICS
Re: Uni-T UTi690A Review
« Reply #12 on: December 02, 2021, 12:36:20 pm »

On the UTi690A, even thought it shows the min/max temperature at the palette, only the min value will be correct for the palette if the total temperature difference is below 10C degrees, where max value will not be at the max color the palette shown, giving a lower contrast.

I suspect this may be to "cover" how much noise there is in the images as higher contrast probably will make the noise very visible, just wish that it'd update the palette to the right as well as it becomes misleading when the temperature difference is very low.
Some palettes do seem to perform better than others thought in different scenarios when it comes to contrast, have you tried them all and see if you find one that works well where you need it?


All imagers will have a minimum span on the palette (temperature to colour mapping) for this reason.  It is going to be some arbitrary 'that is looking too nasty so stop now' decision at the design stage.

There are two other factors to consider though.
The eye-brain system is much more sensitive to colour changes than to brightness change, so mono or low colour change will look better.  Iron for example is a releatively restricted set of colour changes
The eye-brain is also doing averaging at around 10Hz, but evolutionarily trained to pick up on perceived movement.  So a noisy 30Hz update image will get good averaging while with a 7.5Hz update image (quite common for a '<9Hz' due to being 30/4 ) the noise is going to activate all our built in movement detection responses. 

Equally a still image is going to look far noisier than live video (at least at 30Hz)

Bill

Offline Paw85Topic starter

  • Regular Contributor
  • *
  • Posts: 60
  • Country: se
Re: Uni-T UTi690A Review
« Reply #13 on: December 02, 2021, 01:25:29 pm »
All imagers will have a minimum span on the palette (temperature to colour mapping) for this reason.

Do you know if this is how it is commonly done on thermal imagers?
I wish they would have scale the palette image so that what's at max temperature in the images were also the max "color" shown on the palette to the right, otherwise it doesn't add up correctly when viewing the image, which is a little confusing.
Me as a non-software-developer thinks this would have been very easy to do  ;) (Not sure I can explain it well by words so attached a picture, to the left is original with "false" palette and to the right with a corrected scaled one).

EDIT:
I have attached an image taken with the hacked E30bx to the left and UTi690A to the right, both using the same palette option showing the same corner of the room.
The cold spot was easy to spot with both thermal imagers but the E30bx give a better contrasting image, and from what I can see is that Bill W is very correct;
Even FLIR E30bx seem to scale the palette incorrectly when the temperature delta is low on the image and with "auto range" of the palette used as default, just not as much incorrect scaling as the Uti690A have, probably due to much less noise from the sensor.

Both devices also seem to do it the same way, use the min value of the scale correctly but the max value shown incorrect on the palette to the right.
It was however possible to manually adjust the min/max palette value of the E30bx to correctly scale and force it to use the full palette with more noise visible, an option that is not available with the UTi690A.

Best tip for the low temperature delta images is to use the palette named "Rainbow HC".
« Last Edit: December 02, 2021, 02:03:33 pm by Paw85 »
 

Offline Bill W

  • Super Contributor
  • ***
  • Posts: 1102
  • Country: gb
    • Fire TICS
Re: Uni-T UTi690A Review
« Reply #14 on: December 02, 2021, 07:32:10 pm »
All imagers will have a minimum span on the palette (temperature to colour mapping) for this reason.

Do you know if this is how it is commonly done on thermal imagers?
I wish they would have scale the palette image so that what's at max temperature in the images were also the max "color" shown on the palette to the right, otherwise it doesn't add up correctly when viewing the image, which is a little confusing.
Me as a non-software-developer thinks this would have been very easy to do  ;) (Not sure I can explain it well by words so attached a picture, to the left is original with "false" palette and to the right with a corrected scaled one).

EDIT:
I have attached an image taken with the hacked E30bx to the left and UTi690A to the right, both using the same palette option showing the same corner of the room.
The cold spot was easy to spot with both thermal imagers but the E30bx give a better contrasting image, and from what I can see is that Bill W is very correct;
Even FLIR E30bx seem to scale the palette incorrectly when the temperature delta is low on the image and with "auto range" of the palette used as default, just not as much incorrect scaling as the Uti690A have, probably due to much less noise from the sensor.

Both devices also seem to do it the same way, use the min value of the scale correctly but the max value shown incorrect on the palette to the right.
It was however possible to manually adjust the min/max palette value of the E30bx to correctly scale and force it to use the full palette with more noise visible, an option that is not available with the UTi690A.

Best tip for the low temperature delta images is to use the palette named "Rainbow HC".

I think 'all'  is true as any imager that did not would turn to a black / white noise mess if put the lens cap on.

As for your problem with the scale, it is a screen design issue.  The two numbers are scene max / min temp and very badly put by the scale, as most of the time this is the same as the scale max/min.  What you very reasonably want is scale top / bottom and a separate reading for scene max / min.
That would however cover up even more of our precious paid-for thermal pixels.
 If the designer had done that several other options are possible, allowing small areas to saturate hot or cold in colour but still be measured for instance.

You are also seeing that the camera is doing a scene scaling of (min) to (min + least_span) rather than say (average) +/- (least_span/2).




Offline Paw85Topic starter

  • Regular Contributor
  • *
  • Posts: 60
  • Country: se
Re: Uni-T UTi690A Review
« Reply #15 on: December 21, 2021, 02:13:28 pm »
Have received a second UTi690A now and while they are the same product model, I find that the LCD screen differ quite a bit.
The "old" one have higher contrast on the LCD and the "new" one lower contrast but brighter image.

When taking a picture of the same subject and framed the same way, there is certainly a visible difference on the LCD screens, but when looking at the images on a computer, they both look identical.
The "old" seem to be more accurate when comparison it to the computer display, but the "new" display is with some palettes slightly easier to see things in very low temperature delta scenes (low contrast, outlines slightly easier to spot)

Not quite sure which of them I prefer really, both works just fine. Just wanted to report a difference that was found in two copies of the same thermal imager.
Both was reported to have been produced in August according to the certification card, and both use the same firmware version.
« Last Edit: December 21, 2021, 02:34:52 pm by Paw85 »
 
The following users thanked this post: impetus_maximus

Offline mrisco

  • Contributor
  • Posts: 20
  • Country: pe
Re: Uni-T UTi690A Review
« Reply #16 on: December 21, 2021, 03:39:22 pm »
There are differences in the limits also, I also found the same when I did the comparison between the 690A and the 120S  ???
 

Offline Paw85Topic starter

  • Regular Contributor
  • *
  • Posts: 60
  • Country: se
Re: Uni-T UTi690A Review
« Reply #17 on: December 21, 2021, 03:55:40 pm »
There are differences in the limits also, I also found the same when I did the comparison between the 690A and the 120S  ???

Would you mind explaining what you mean with difference in limits?
Both seem to perform within specifications when I test them.
 

Offline mrisco

  • Contributor
  • Posts: 20
  • Country: pe
Re: Uni-T UTi690A Review
« Reply #18 on: December 21, 2021, 04:17:12 pm »
Maximum and minimum temperatures registered in the display area.
69.7 vs 68
21.3 vs 20.9
« Last Edit: December 21, 2021, 04:19:17 pm by mrisco »
 

Offline Fraser

  • Super Contributor
  • ***
  • Posts: 13148
  • Country: gb
Re: Uni-T UTi690A Review
« Reply #19 on: December 21, 2021, 04:23:33 pm »
It is not uncommon to see minor differences in LCD panels used in equipment as OEM’s often buy from more than one source. They are normally pretty similar in performance however. I recall that FLUKE actually upgraded the LCD panel in one of their camera series because of poor contrast and colour rendition. I suspect that the two 690A cameras are fitted with LCD panels from different manufacturers. At the budget end of the market there could be quite a lot of parts supplier variation due to market prices and stock availability. In these troubled times of component shortages this may be even more common.

Fraser
If I have helped you please consider a donation : https://gofund.me/c86b0a2c
 

Offline Fraser

  • Super Contributor
  • ***
  • Posts: 13148
  • Country: gb
Re: Uni-T UTi690A Review
« Reply #20 on: December 21, 2021, 04:45:31 pm »
mrisco,

That is normal variation for this level of device and falls within the stated measurement tolerance. Of course there could be a situation where both cameras are within +/-2C of each other in terms of measurements, yet both cameras are actually way off calibration when checked against a known accurate Black Body source  ;) The best way to test a thermal cameras measurement accuracy is by the use of a target at a known temperature. This can be done at room temperature for a quick check.
To do this, I recommend sourcing a 120mm x 120mm x 6mm Aluminium plate and painting it with a high emissivity paint such as a very Matt camouflage paint. An accurate thermometer is used to measure the ambient temperature and that of the metal plate. You can now do a rough measurement accuracy test but there will be small emissivity errors if you are not certain of the plates paint coating emissivity.

Fraser
« Last Edit: December 21, 2021, 04:56:47 pm by Fraser »
If I have helped you please consider a donation : https://gofund.me/c86b0a2c
 

Offline Paw85Topic starter

  • Regular Contributor
  • *
  • Posts: 60
  • Country: se
Re: Uni-T UTi690A Review
« Reply #21 on: December 21, 2021, 06:14:00 pm »
Maximum and minimum temperatures registered in the display area. 69.7 vs 68 21.3 vs 20.9
The pictures I took was hand-held trying to have hold the imagers at the same position pointing at the same spot after one another, but it's not identical, so minor temperature differences is likely due to this, and variance between the sensors, but both is within margin of error.

I suspect that the two 690A cameras are fitted with LCD panels from different manufacturers.

I do agree with you that it's likely from different manufacturers, having used them both a bit more this evening and testing palettes it seems they are displaying the colors slightly different as well, usually don't see that much of a difference in batch to batch from one manufacturer.
But both do perform well, no details being lost when viewing and viewing angles seem to be close as well.
 

Offline impetus_maximus

  • Contributor
  • Posts: 31
  • Country: us
Re: Uni-T UTi690A Review ( UTi690A UTi120S UTi712S )
« Reply #22 on: May 06, 2022, 12:39:53 pm »
many thanks Paw85 for the review.  :-+
seriously considering grabbing one. although the washed out look of your second unit has me reconsidering.

i'm curious how difficult swapping the '2000mAh' battery will be at it's end of life. :-/O Chinese cells are not the best.
guessing it's an 18650 cell in the grip.
 

Offline Paw85Topic starter

  • Regular Contributor
  • *
  • Posts: 60
  • Country: se
Re: Uni-T UTi690A Review ( UTi690A UTi120S UTi712S )
« Reply #23 on: May 07, 2022, 05:07:40 pm »
The washed out look looks worse on the pictures than in reality, if I had not had two of the same unit I would probably not have noticed.

Changing the battery will probably be a bit difficult, as far as I know no one have dared to open the UTi690A or it's siblings as it seem that you have to remove the transparent LCD cover to get access, and it appears to be glued.
But hopefully the battery will last a while, and if having to change it later on, a few marks on the LCD cover is probably something we can live with ;)
 

Offline Fraser

  • Super Contributor
  • ***
  • Posts: 13148
  • Country: gb
Re: Uni-T UTi690A Review ( UTi690A UTi120S UTi712S )
« Reply #24 on: May 07, 2022, 06:35:37 pm »
The display plastic is relatively easy to remove without damage :) Just warm the plastic with a hair dryer whilst making sure you do not go mad and melt the plastic ! Then I use a PVC sucker to lift a corner of the display plastic. If it does not lift easily, a plastic prying tool, used for mobile phone display removal, may be used to lift an edge of the display and gently work around to release all sides. The adhesive is double sided tape that softens with the application of heat. It is often re-usable if care is taken lifting the plastic. The case screws are then visible.

Once access is gained to the interior of the camera, the battery change is a relatively simple task as the cell is nothing unusual.

The cell in the camera should last a few years anyway :)

Fraser
If I have helped you please consider a donation : https://gofund.me/c86b0a2c
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf