Yet another cheap thermal imager incoming.. Seek Thermal

[Fraser]

@Seeker,

I can state that the SEEK shutter appears less than ideal. My personal experience of such shutters has been that they are larger than the microbolometer and cover it fully during the Flat Field Correction event. I find it sad that the SEEK Thermal appears to have cut corners in areas that would not have been more expensive if done correctly and would have improved image quality. A small lens does not necessarily mean poor images and the tiny size of the SEEK was not a necessity, especially if that is the cause of the issues that we are witnessing.

The SEEK is a great idea and the microbolometer has potential. Sadly the camera design is flawed and needs revisiting. A few Dollars more would not have stopped buyers purchasing as the resolution is the best in its price bracket. I consider the SEEK camera to be little more than a prototype of technology that will hopefully follow.

Aurora

[Seeker]

Viewing the video, the shape and size of the shutter was the first thing that struck me as possibly significant, with respect to the gradient issue. Not so Mikestuff, though...he didn't refer to it. That's why I assumed it must be "obviously sufficient".
Prototype maybe Aurora...damn shame though, such a proto-preventable little fault preventing the unit from being a little gem (rather than just impressive). As stated, most buyers of this will be satisfied with what it can do, regardless--but it will irk me about the gradient and I intend getting to the bottom of it if I can (if my unit indeed arrives with the fault--it's yet to be agreed (it seems) that the fault is inherent in the design rather than a QC issue). So close to a gem!

[Uho]

I tried to increase the area of the gate-to no avail. Strong side illumination. If you set up tube - side illumination less. And the gradient is less. Poor design of the sensor. He sees the light side.

[Fraser]

Classic design of an industrial thermal camera has the microbolometer PCB mounted in free air but there is a large diameter cylinder mounted around the actual microbolometer. Length is dependant upon optical block design but I believe the cylinder offers a reduction in side sensitivity and internal illumination from the camera electronic/case. It acts like like blinkers to limit what the FPA can 'see' from off axis heat sources.

Some designs such as BST based thermal cameras do not have the cylinder fitted but they do have well designed optical blocks with less opportunity for off axis illumination. Its all about good design. IMHO the SEEK lens tube is too small and possibly too long relative to its diameter.

The SEEK microbolometer appears to have been designed as a low cost thermal detection FPA but we are not privy to the design objectives and its other intended uses. 12um microbolometers are not the most common and may have inherent problems due to the small die size and associated ROIC that also generates heat. The FPA may well meet the original design spec but needs some attention to how it is deployed to prevent thermal gradient and noise issues. I always thought it brave of SEEK Thermal to use direct bonding of of the microbolometer to a PCB that contained heat sources, rather than using a thermally neutral daughter board to isolate it from internal thermal sources.

SEEK thermal appear to have had miniaturisation as a key objective. Maybe if they had made the camera a little larger, and a little more expensive, they could have managed the issues that we are seeing. It is no good being the smallest, lowest cost solution if it is fatally flawed ! The SEEK is not fatally flawed but it does have issues that I would have expected to have been sorted out before release to the public.

Have SEEK Thermal been honest, open and transparent regarding the cameras limitations....... IMHO, no they have not met this criteria for managing customer expectations. I have used equipment that clearly states any known negative characteristics present in the design, such as frequency stability, real world accuracy, internal noise/signals that are to be expected etc.

I feel that SEEK Thermal should make it clear that the SEEK camera is a compromise solution providing thermal imaging but with a relatively high noise floor and a thermal gradient. The customer is then in a position to decide whether the cameras meets their needs. The publicity pictures certainly do not show the level of gradient and noise that we have witnessed in this thread.

I was and, to a degree, still am, a supporter of SEEK Thermal. I can't help feeling that they are 'fumbling the ball' though. A disastrous V1.6 software release that did not rectify the noise or gradient and produced daft temperature readings.....how the heck did that get out of development undetected ?  We are still waiting for a software from SEEK that deals with these issues. I do hope that this is not a 'we got your money now, so who cares' attitude.

SEEK Camera .....Such potential, as yet not fully released.

Sadly I cannot recommend the SEEK to others until they get it sorted out. Not great when the product was supposed to be so revolutionary. I am not losing any sleep over the $200 cost though. I spend more than that on thermal camera parts and lenses.

Aurora

[Seeker]

Uho, well that saves me going through widening the gate (for a bigger shutter) and being disappointed, good man.

Aurora--"A disastrous V1.6 software release that did not rectify the noise or gradient and produced daft temperature readings..." does that mean the "update" was intended to moderate the gradient, but failed? Or wasn't intended to address it, anyway? If the former, then that's hopeful, they consider it an issue and are working on it.

[Uho]

Painted black lacquer verge sensor. Gradient decreased! Now it evenly over the edges of the image. Still try to paint the inside of the tube.

[Seeker]

Aha! Let us know the results.

[mikeselectricstuff]

Painted black lacquer verge sensor. Gradient decreased! Now it evenly over the edges of the image. Still try to paint the inside of the tube.

I wonder if having the inside of the lensholder black may make it more dependent on ambient temp, due to higher emissivity.
There has to be a reason they went to the expense of painting it & not leaving as bare metal

[Seeker]

If the main problem turns out to be uneven (sensor-seen) emissivity of the lens housing, maybe the final solution could be as simple as a shaded paintjob...black segueing to white, pattern arrived at by experiment. It's a thought.

[Fraser]

Table of emissivities, including paint.

http://www.thermoworks.com/emissivity_table.html

It is not always colour that dictates emissivity....the material from which paint is made is an important factor too,

I use car VHT matt black paint for high emissivity. Gloss paints should be avoided to avoid reflectivity issues.

Aurora

[Seeker]

Like the gloss paint used on the Seek housing? Yep, not sure why they used that. I would have used matt black, but what do I know. Anyway, quick and cheap experiments (when i get mine, shipping imminently) will determine if coatings have any bearing on our problem.

[Uho]

Paint inside the tube is not improved gradient. I do not have a matte paint. Perhaps this is the problem.

[Seeker]

BTW...just having a quick look through previous pages...I see the product name displayed as "SeeK". That presumably means "See^oK"...I had wondered about the odd name. Duh. 

[barehill]

If the shutter is a pivoted (or sliding) blade placed in front of the sensor then as the blade opens and closes there will be uneven obscuration of the sensor vs. time. Since the shutter is there to block radiation for baseline calibration the shutter may not dwell long enough to compensate for the asymmetrical obscuration. Only a variable-iris shutter at the exit pupil would give perfectly uniform obscuration vs time.

[Fraser]

If SEEK thermal released the promised SDK and provided the ability to control the FFC event, it would be much easier to determine exactly what is going on with the thermal gradient issue.

SEEK Thermal, if you are still reading this thread, how about helping us a little by supplying engineering access to the SEEK camera or the promised SDK ? There are some very clever people on this forum who could assist you in ironing out the teething issues with this first generation camera.

Aurora



 

[Uho]

Minimum gradient that could be done. A lens with a short focal point of GE. The original lens is broken.

[Fraser]

Regarding the use of a Germanium lens on the SEEK. Be aware that the Ge lens has a  transmission characteristic that decreases as the lens temperature increases. The original SEEK lens is a form of Chalcogenide moulded glass and does not suffer from this effect to the same degree. The camera would need to contain a calibration table and thermal sensor on the lens housing to compensate for the Ge thermal characteristics. Not a bug issue here but something to bear in mind when replacing a Chalcogenide thermal camera lens with one made from Germanium.

http://www.thorlabs.de/newgrouppage9.cfm?objectgroup_id=1780

http://eom.umicore.com/en/infrared-optics/featured-articles/current/

Aurora

[Uho]

I know it. In the case of Seek it does not matter. Worse than it already has not done. Exact temperature readings can only dream of.

[WS-PI]

Dear All,
I am following this discussion now for a couple of days and I am wondering why everybody complains the Seek is not as good as it should be or could be. I own one for some days now and I am totally satisfied with the quality. I am working as a reliability engineer for a semiconductor company and use a FLIR for my daily work.  But under no circumstances I would compare the FLIR with the Seek. We use the FLIR as a measuring tool and it is calibrated every 12 month. Why should I expect the same accuracy from the Seek. The Seek is not offered as a high precision measuring tool. It is offered to see the unseen. And also you could do some measuring but with the same limitations other consumer products have. With an accuracy of +- 2 degrees for the first 5 minutes of operation I could do a lot of useful things in my home. If I operate the Seek for more than 5 minutes a failure increases because the Seek gets warm. But because I know about this behavior I can take care when interpreting the results. From my experience with the FLIR and the Seek I know that uncertainty about the right emission factor is very often the major factor that determinates the measurement accuracy.
Therefor let’s be fair to Seek. I think you get a great product for your money. If you have higher requirements then you have to spend at least five times more money and then you can of course complain about gradient, accuracy, optics and so on.

WS-PI

[Uho]

Possible for the same money to do better. What prevents the  balance manual mode? Add one point in the program a couple of lines of code. But Seek closed code and are doing nothing. Instead of expensive and beautiful box boxing to make the sensor separately. On a separate electronic board. Personally, I have a gradient of 9 degrees is not satisfied. I'm trying to eliminate the shortcomings of engineers Seek.

[Seeker]

WS-PI, your opinion is appreciated, it's a budget unit more for consumers rather than tradesmen, but the gradient (as discovered on buying the unit) should not be there. There's no mention of it in the ads....
I personally would not attempt to improve the performance of a budget unit of anything, but I would be irked if I bought it flawed (mine not arrived yet).
There would be zero chance of a buyer not noticing that one end of a supposed flat field is apparently shown hotter than another. I believe the complaint is "it's got a bug" rather than "it's not accurate enough".
Maybe it's just too small for present technology, as aurora said.

BTW by "flawed" I mean there is one huge distortion present, across the screen. That sounds to me more a design error than a limitation of cost.

[Fraser]

@WS-PI,

Early on in this thread I made a similar comment and, to a degree, you are right. For $200 the SEEK breaks new ground in terms of bangs per buck. HOWEVER over the months I have come to realise that the SEEK design is flawed in a way that is hard to accept. Any TIC developer knows about microbolometer noise management and the issues of poor optical path/FFC shutter design. SEEK Thermal was started by two gentlemen who worked for the well respected Indigo TIC OEM,who were taken over by FLIR. They left FLIR and started SEEK Thermal with a lot of Indigo design knowledge.

People like myself are disappointed on several levels

1. The product has noise management and temperature gradient issues that were not expected on a finished product with its background pedigree. SEEK Thermal appear to claim that they were unaware of the Gradient issue ! What ? how could their UAT have missed such a obvious issue ?
2. The software was originally issued as sort of a BETA release that would be updated to a better version soon. Such upgrade and associated improvements have not occurred to date.
3. The latest revision of software (V1.6) was so seriously defective that it was pulled. UAT? what UAT ? Poor processes from SEEK Thermal. We are now back on V1.4 until further notice.
4. Communications from SEEK Thermal are very poor and they are not even trying to manage users expectations. Where in their documents have they admitted performance is limited due to noise and thermal gradient.
5. SEEK Thermal have not released the promised SDK.

As an engineering blog it is inevitable that the membership will provide comment on a design, especially one that breaks new ground like the SEEK Camera. However our comments should be taken in context. I have no issue with what SEEK Thermal are attempting to do, I support it. The camera is flawed however and it needn't be, even at the low price point. Some could say that SEEK Thermal have squandered the new 12um microbolometer performance by using it on a sub standard platform.

For info, I am an industrial user of professional FLIR thermal imaging equipment that costs tens of thousands of Dollars, so I do understand the value for money aspects of the SEEK product. What you are witnessing here is disappointment in a product that could have been so much better if its development had been completed before release.

Aurora

[Fraser]

I stumbled upon a recent Youtube video showing the Hema thermal camera and a discussion of the challenges facing both it and the SEEK camera. Erik has a Seek and is using it as a comparison piece for his development work on the Hema. Erik is also developing a stand alone thermal camera as a next step in development. He also stated that his designs will be fully open source, including schematics and source code. Interesting times ahead I hope. I wish Erik well in his endeavors.

Nice video, definitely worth a look:



I like healthy competition in a marketplace as it can lead to faster development of products. Let us hope that the Hema reaches the production stage and is a success. Such competition could lead to improvements in the SEEK and possibly the development of more affordable TIC's.

Aurora

[Lurchbox]

Hey all, great job on all your development for interfacing the seek with PC. I have been using it with a $60 micro center TW700 windows 8 tablet and it performs quite well. Are there any programs compiled that can take capture pics directly?

[miguelvp]

On my Dell Venue Pro Windows 8 tablet, pressing the Windows button and the volume down takes a screenshot. Not sure about your tablet.