Yet another cheap thermal imager incoming.. Seek Thermal

[efahrenholz]

@mike,

Try manually moving the shutter over the image sensor slowly, to check the orientation of the shutter as it moves over the image sensor. I'm very curious to see if the dark area in most images also happens to be the point where the flag moves into and out of the image area.

[marshallh]

Maybe the software is using a "too-early" frame from the sensor that contains afterimages of a blurry moving shutter for compensation
Is it possible that the speed of the shutter action is a variable that depends on mfg batch, phone battery voltage, or something else?

[efahrenholz]

Considering there is a regulator, the battery level wouldn't make a significang difference, even with the minimal draw on the coil to stay open. I couldn't imagine it would pull the regulation down enough to slow the shutter...but it is a possibility.

[ricksastro]

Interesting thought on the shutter shape...certainly a good theory.   But I think miguelvp's experiment (post 398) isn't explained by this.   Maybe a combo?

[eneuro]

Is it possible that the speed of the shutter action is a variable that depends on mfg batch, phone battery voltage, or something else?

With decent scope maybe it will be possible make own DIY using only this part cut from whole PCB - sensor itself and put there MPU we like

This above it not of course thermal image of sensor part of PCB, but just for fun applied this quite nice iron256 LUT which is very similar or maybe even based on FLIR's one mentioned many posts above 
 
While using even basic image processing methods I was able get decent output based on this  USB raw sensor data provided in this thread earlier by @marshallh, the most interesting thing is simply hack or learn its protocol between MPU and sensor, and maybe even make... own DIY shutter and forget about this builtin in-simply remove it 

In my Linux app added some adjustable gamma and can controll how smooth output image is, so with this iron256 LUT it looks much better than those very noisy looking Seek Thermal outputs and it is time to make OpenCV version of this software, while probably I was able to find quite good way to fill those hexagon black dots and other probably bad pixels, so it could be nice make this code more efficieint and accelerated by Nvidia CUDO technology supported in OpenCV.

Sample PNG RGB output after 4x resizing oryginal sensor 208x156 size to output size below to compare with those images taken from Android devices

[mikeselectricstuff]

I've now established that it is definitely nothing to do with the shutter.
 It is either the lens itelf, or the alignment of the lens. 

[mikeselectricstuff]

With decent scope maybe it will be possible make own DIY using only this part cut from whole PCB - sensor itself and put there MPU we like

I think the only viable way will be to write new code for the MCU - there are 18 bond-wires onto the die and some disappear onto vias inside the lens housing

[ricksastro]

I've now established that it is definitely nothing to do with the shutter.
 It is either the lens itelf, or the alignment of the lens.

Definitely look forward to hearing about this!   thanks for investigating.

[Fraser]

Mike,

Thanks for the news. So we are back to the possibility of issues in the optical block. Misalignemnet of the lens should not cause the viewing of a flat field to have a temperature gradient unless wildly off axis causing only partial illumination of the microbolometer. Such would be very obvious when looking at a normal target scene.

There is the possibility of narcissus effect between the microbolometer and the rear of the lens. The rear of the lens element should be convex which tends to reduce the chances of narcissus effects in the optical block. I have not seen such effects inside the optical block of a TIC before so have no idea of the real world visual effect of such.

From the images there appears to be a cooler minority area surrounded by a warmer majority area. In order to cause a cooler area on the microbolometer face a cooling influence would be needed and we are seeing a 9 degree differential !. I know of no such cooling source around the SEEK optical block. That leaves the possibility that the majority warm area is in fact the error and not the nominal. Wow this is making my head hurt ! The shutter was a good candidate for causing the cooler area anomaly as it had the potential to cause a negative offset in the FFC table but as you have discounted that we are kind of back to square one looking for the physical source of the error.

I look forward to any discoveries that you make regarding possible causes of such an anomaly.

Aurora

[callipso]

How did you manage to rule out the shutter?

[IanB]

I haven't seen it mentioned yet, but there is a solenoid that operates the shutter. I recall that the solenoid is continuously active and holding the shutter open while the camera is operating, rather than the expected reverse situation where the solenoid activates momentarily to close the shutter when commanded. If the solenoid is continuously activated, could this not be a source of heat close to the optical assembly? Could it even be warming up the shutter in an uneven way?

[Fraser]

OK blue sky thinking time.......

Unlike my other TIC,s which tend to have very open optical blocks, and reasonable air circulation space, the SEEK uses a tube with the microbolometer at one end and the lens at the other. I know that a microbolometer +ROIC generates some heat (usually 30C nominal) so could we be seeing a the effects of heat accumulation inside the lens tube ? The effects of such warm air containment in front of the microbolometer are unpredictable and could potentially cause hot and cold spots on the images. The cooler area of the images could be where the heat is able to escape due to the shutter lever openings ? It would be useful to establish exactly where the 'blob' of cooler readings is physically located in relation to the layout of the PCB and lens block etc. I am still not certain on this orientation issue.

Aurora

[Fraser]

IanB,

Some TIC's have a normally closed shutter when off in order to protect the microbolometer from prolonged exposure to high input signal levels such as a fire or radiator in the FOV. The designs pull the shutter out of the optical path when the unit is switched on. TESTO came up with a shutter actuator that is latching in both open and closed positions and changes state when a pulse is applied to the solenoid. This was done to save power in their 880 series cameras and is patented.

The solenoid heat output was considered but we think it too little for the effect that is being seen. I proposed possible magnetic field interferance issues but that also looks unlikely. 'MiguelVP' has also eliminated heat and magnetic field influences in his experiments of yesterday.

Keep thinking about potential scenarios pepes, we may yet solve this puzzle 

Mike has a disassembled SEEK on his work bench and likely has the best chance of discovering the cause through good deduction and testing. He also has an E4(E8) to inspect the unit with and to see the heat around critical areas such as the optical block.

Aurora

[mikeselectricstuff]

I haven't seen it mentioned yet, but there is a solenoid that operates the shutter. I recall that the solenoid is continuously active and holding the shutter open while the camera is operating, rather than the expected reverse situation where the solenoid activates momentarily to close the shutter when commanded. If the solenoid is continuously activated, could this not be a source of heat close to the optical assembly? Could it even be warming up the shutter in an uneven way?

No - heat is negligible. I have repeated the gradient using a substitute shutter comprising a strip of plastic manually removed from in front of the sensor.
I think it is either the lens (marginal field of view)  or lens aligment

[miguelvp]

I don't know why Mike ruled out the shutter.

After being on for a while, if I force the shutter open and point it down, the blob disappears.

After being on for a while and the shutter forced open if I force it close, the image is less even but not terrible:

After forcing it close for a while if I release it and let it operate normally and point it down, it's as bad as it gets.

[marshallh]

Using the PC application by sgstair I've noticed mine does not have any gradient on startup. After aobut 2 minutes it starts to become noticeable:


So this rules out compensation by the android app.

I have some thermal pics of its opreration as well.

[marshallh]

[Fraser]

Thanks for the pics Marshall.

That last picture is interesting..... I am seeing 38C through the SEEKs lens but that could be narcissus effect from your F30 ? Have you a picture of the rear of the microbolometer as it is bonded to the PCB and so should show up on the rear of the PCB if hot.

Aurora

[Fraser]

Ok more conjecture.....

If the SEEKs new microbolometer runs hot (possibly 38C looking at Marshalls thermal images)  then any shutter that is placed in front of it will start at ambient and quickly heat towards 38C.

The shutter is not designed to be heated ! It is supposed to normalise at ambient and its behaviour when heated could be unpredictable, especially if the shutter flag thickness, and hence thermal mass, varies across its surface. The shutter arm may also conduct away heat from that part of the flag.

It would be very useful to know what temperature the microbolometer is running at 38C would be very warm when compared to measurements that I have made on other microbolometers.

Aurora

[mikeselectricstuff]

I don't know why Mike ruled out the shutter.

After being on for a while, if I force the shutter open and point it down, the blob disappears.

- yes it's seeing a static scene, so when it thinks it's done a cal, it's subtracting the image form itself, giving  ablank field.

After forcing it close for a while if I release it and let it operate normally and point it down, it's as bad as it gets.

the issue is that the "blank field" of the shutter-closed image is different from that of a blank field through the lens.
From what I've seen so far I think it's the lens rather than the shutter, but it is rather hard to test.
Other minor possibilites are IR coming through the shutter aperture, or reflections from the inside of the lens housing

Brain getting a bit fried now but my current best guess is lens alignment ( edge of field of view)

[sgstair]

I'm not sure how LWIR-reflective the shutter material is - but could we possibly be seeing a reflection of the sensor's heat?

[marshallh]

taken a coulpe minute apart as it warms up - definitely more noise:


Also you can see that each frame is compsed of 4 subframes... I'd bet the LPC is doing average internally at 8.5*4fps\



Back of sensor

Shutter is visible here, it's the hook shape that's green in front of the lens housing. Obviously much warmer than the lens housing



Off-axis with shutter open


offaxis with shutter CLOSED

[mikeselectricstuff]

Shutter is visible here, it's the hook shape that's green in front of the lens housing. Obviously much warmer than the lens housing

Bear in mind shutter is black plastic, lens housing is metal with white paint on it, so at least some of the difference could be down to emissivity.

[Fraser]

That microbolometer still looks hotter than I would have expected but interestingly there is no evidence of a hot spot on the rear of the PCB under the microbolometer.

It will be very interesting to hear SEEKs thoughts on this matter, hopefully tomorrow.

Aurora

[mikeselectricstuff]

That microbolometer still looks hotter than I would have expected but interestingly there is no evidence of a hot spot on the rear of the PCB under the microbolometer.

It will be very interesting to hear SEEKs thoughts on this matter, hopefully tomorrow.

Aurora

Don't hold your breath. They've not exactly been very communicative to date