Yet another cheap thermal imager incoming.. Seek Thermal

[callipso]

@Aurora,

what's your RealWork^TM, that you get to play with such awesome gear?

[Fraser]

My thermal cameras are all owned by me and not my employer

My work 

Aurora

[callipso]

Judging by the  and by the sand colour of the Mickey... It's the military, isn't it?

[miguelvp]

I'm getting a friend to 3d print the lens adapter

Now to choose between 4", 3", 2.5" 2" 1.5" and 1" focal length, I think I'll start with the 4" one (101.6mm)

Edit: link to adapter thingiverse post:
https://www.eevblog.com/forum/testgear/yet-another-cheap-thermal-imager-incoming/msg542224/?topicseen#msg542224

[jaybeez]

Received my android unit today, and I am very dissappointed. the gradient makes the unit almost unusable within a minute.

My gradient is hot at the bottom left and cold atvthe upper right. With a cold camera and the lens facing down on a flat surface, the screen shows a temperature thats about 10 degrees hotter than ambient, and a colder gradient gradually mives in over several shutter clicks, ultimately creating a colder spot that is about ambient temp.

I've watched Mike's video several times and I am not sold on his solution (no offense Mike). I just can't comprehend how lens position would be the cause of a gradually apearing gradient.

I'd be more inclined to believe that the shutter was acting like a piston or a fan blade and displacing air inside the lens housing, cooling one portion of the sensor.

Mike removed a portion of his lens housing, and colored a portion of the inside. maybe the extra space around the shutter opening and slight adjustment of the housing is reducing the piston effect when the shutter moves.

[Fraser]

We have asked SEEK for comment on the cause of the gradient as they believe that they have identified it. So far no comment from them 

If we know the cause it provides users with the opportunity to modify their camera. I may well scrap the lens tube as that is what I believe is the cause of the issue. My other thermal cameras are of an open lens holder design with free air flow around the lens elements, shutter and microbolometer. The SEEK is the first TIC that I have seen with the tube type lens holder. There may well be a very good reason why this is the case !

I attach a picture of a typical TIC optical block. No enclosing of any part of the optical path. I just found a picture of a Talisman Wasp. Note the open optical block design. Sorry it isn't a clearer picture.

Aurora

[frenky]

What could be done (to void warranty  ) is to remove the shutter flag, fill in the gap with some hot glue and use signal from original shutter to trigger new diy shutter in front of the lens.
I would be willing to try that if they would ship the damn thermal module to EU...   

[mikeselectricstuff]

I've watched Mike's video several times and I am not sold on his solution (no offense Mike). I just can't comprehend how lens position would be the cause of a gradually apearing gradient.

Could be change in temp of the lens housing (relative to the sensor and/or lens)?
The fact the lens was glued and I had no way to accurately hold & adjust everything meant it was difficult to conclusively prove anything -  having to manually hold things may also have contributed some thermal effects.
It is possible that the temp was different between my original "before" and "after" results, though I did leave it off for a while to stabilise after I'd been handling stuff.

Must remember to try the lens form the Audi camera on it some time.

 

[eneuro]

@miguelvp

Probably our new EEVBlog member @slackaz which showed also realtime video how useless this thing can be is looking for information there when it will be improved, but if I were him and in the country where this company is located of course sended this device back and requested money back since there is no other option to avoid this gradient issue for the moment.

I got my camera a few days ago. I agree the gradient issues basically ruins the device.

Such thing should be found by Seek team before they send it to customers.
Now talking about magic software updates and that they found why this issue is, but can't tell even people in this EEVbog forum only means this is their public relations oldshool game 

The question only is if are there any Seek Thermal cameras which are not affected by this gradient issue? 
I'm not talking about perfect conditions created to avoid it for short period of time.

[efahrenholz]

I opened my camera at work today, did some experiments, even cracked the glue free of the lens housing. My imager was free of any residue at all. I attempted to reposition the lens which made no appreciable difference. I tried all different ways to prove it was a thermal issue. I simply couldn't make a difference in the gradient. I also may have destroyed the sensor because I think the lens housing pushed some bond wires together and now I have a black image output. I have to get under a microscope and look, might need to just bump them free of each other. I hope. Otherwise it's $200 in the garbage.

They did mention a temporal drift algorithm. I'm also leaning towards it being poor comoensation. We can play with all the electronics all day, if there is a bug in the software, it won't make a bit of difference.

[Fraser]

@eneuro,

A very good question.

I was not seeing a lot of unhappy posts on Facebook regarding the gradient issue which surprised me (remember they did not delete my question regarding such) so the user base appears unaware of a problem. From pictures shown here, it would appear to be obvious BUT I wonder......

If you point a thermal camera at a normal mixed temperature target scene, and the span is auto set to 'wide' by the camera, the gradient may not be obvious to users. When you point a thermal camera at a relatively uniform target of approx one temperature, like a table top, the camera may be auto setting its span to the smallest available (normal behaviour in a TIC). A small span will show the gradient far more than a wider span.

Not saying that this is the case or acceptable, but it may well be that the scientific minds of the EEVBlog have carried out UAT that a casual user would not. If the SEEK provides a decent picture under everyday circumstances it will likely be adequate for most consumers. There have to be compromises to achieve such low cost. This may be one of them but I wish SEEK would talk about it. A hardware solution is IMHO better for my needs than a software 'botch'. I am happy to do the work so its not as though I am asking SEEK to sort it out.

Aurora

[mikeselectricstuff]

I opened my camera at work today, did some experiments, even cracked the glue free of the lens housing. My imager was free of any residue at all. I attempted to reposition the lens which made no appreciable difference. I tried all different ways to prove it was a thermal issue. I simply couldn't make a difference in the gradient. I also may have destroyed the sensor because I think the lens housing pushed some bond wires together and now I have a black image output. I have to get under a microscope and look, might need to just bump them free of each other. I hope. Otherwise it's $200 in the garbage.

If it's unfixable, don't sling it.... I could do with a sacrificial unit to teardown the sensor and trace out the PCB.

[frenky]

I was not seeing a lot of unhappy posts on Facebook regarding the gradient issue which surprised me (remember they did not delete my question regarding such) so the user base appears unaware of a problem. From pictures shown here, it would appear to be obvious BUT I wonder......

People just don't notice it.
Check this video from 5:58 thermal gradient on his module is awful and he doesn't notice it...

[Fraser]

Yes I see what you mean. If SEEK Thermal admit the issue is serious they may be worried about returns and recall requests.
Aurora

[eneuro]

I would be willing to try that if they would ship the damn thermal module to EU...   

Maybe there are also other reasons than ITAR limitations?
EUR-Lex Access to European Union law
In the case of our country when normal person not company fill such complain form, than wait I'm not sure 30 days I guess, than if the problem is not resolved and make lawsuit in the court than when unsolved and not repaired issue was found  within 6 months than in the case of court lawsuit manufacturer have to proove that there is no issue at all, while after 6 moonths customer have to proove it which in the case of modern technology it can extremelly difficult and probably one had to pay for expert in this area.
I didn't gone througth court with such issues, but in many cases one email and official letter is fine, while company lawyers knows that one can make lawsuit vs them if they dismiss official document (not email).

So, talki that something will be done sometime by Seek Thermal maybe also simply game for time... to pass this 6 month pwriod-but ok I'm not a lawyer so in the case of broken thermal device maybe layers offize could help in US...

There is also another issue with Seek Thermal-for me is not acceptable there is no even end user manual available, which as I know is required in EU - I'm not usr ehot it looks like in US, so maybe if someone destroyed this device becouse of tried somehow "tune" this device to avoid this gradient, maybe there is a chance even without any lawsuits that will get his money back, because of there is NO SEEK THERMAL USER MANUAL available   

[mikeselectricstuff]

I would be willing to try that if they would ship the damn thermal module to EU...   

Maybe there are also other reasons than ITAR limitations?
EUR-Lex Access to European Union law
In the case of our country when normal person not company fill such complain form, than wait I'm not sure 30 days I guess, than if the problem is not resolved and make lawsuit in the court than when unsolved and not repaired issue was found  within 6 months than in the case of court lawsuit manufacturer have to proove that there is no issue at all, while after 6 moonths customer have to proove it which in the case of modern technology can extremelly difficult and probably one had to pay for expert in this area.
I didn't gone througth court with such issues, but in many cases one email and official letter is fine, while company lawyers knows that one can make lawsuit vs them if they dismiss official document (not email).

So, talki that something will be done sometime by Seek Thermal maybe also simply game for time... to pass this 6 month pwriod-but ok I'm not a lawyer so in the case of broken thermal device maybe layers offize could help in US...

There is also another issue with Seek Thermal-for me is not acceptable there is no even end user manual available, which as I know is required in EU - I'm not usr ehot it looks like in US, so maybe if someone destroyed this device becouse of tried somehow "tune" this device to avoid this gradient, maybe there is a chance even without any lawsuits that will get his money back, because of there is NO SEEK THERMAL USER MANUAL available   

There are no ITAR issues (well at least not as far as the finished product is concerned  )
If they ship from the US to the EU, then it's a personal import and not subject to EU consumer law

I have no doubt that the only reason is they're still struggling to fulfill US orders.

[eneuro]

If they ship from the US to the EU, then it's a personal import and not subject to EU consumer law

Really? I have to investigate this in local customers lawyers office.
For the moment this product is not officially sold by Seek Thermal to EU, so if someone from US bought it and shipped as person to person than ok,
but what if US customer bought this device via Seek Thermal and is US citizen which got not tis what he expected to be in thermal camera, however will notice it lets say after a few months?
There is no user manual requirement in US? 

Update: On this video above  there is some sort of Seek Thermal black papers, but no google'able electronic version?
Ok. I do not complain no more-simply it is not a time to buy this device right now....let others find bugs in their software and hardware 

[jaybeez]

I've watched Mike's video several times and I am not sold on his solution (no offense Mike). I just can't comprehend how lens position would be the cause of a gradually apearing gradient.

Could be change in temp of the lens housing (relative to the sensor and/or lens)?
The fact the lens was glued and I had no way to accurately hold & adjust everything meant it was difficult to conclusively prove anything -  having to manually hold things may also have contributed some thermal effects.
It is possible that the temp was different between my original "before" and "after" results, though I did leave it off for a while to stabilise after I'd been handling stuff.

Must remember to try the lens form the Audi camera on it some time.

Im just trying to limit the variables before digging into this. im convinced that the problem went away after your modifications, your demonstration at the end shows that. my question is why?

lens position or housing modifications

with my unit, from a cold start it works perfectly, and the gradient slowly appears over several calibrations. if its turned off and turned back on immediately the gradient remains. if its turned off and left along for 5-10 minutes the gradient is gone at start up. that almost rules out anything except a source of heat or cooling.

I think aurora might be onto something with the openness of the holder.

Mike, how does your unit perform with real use now that you've made the modifications? im interested in the quality of the pictures and video.

[mikeselectricstuff]

There is also another issue with Seek Thermal-for me is not acceptable there is no even end user manual available, which as I know is required in EU

There is help in the app. What's the big deal about a manual?
As they haven't published a spec, image quality is a subjective issue.
Compared to the nearest equivalent, Therm-App, cost is 20%, so it could be argued that if quality is at least 20% as good, that's reasonable....

[mikeselectricstuff]

Mike, how does your unit perform with real use now that you've made the modifications? im interested in the quality of the pictures and video.

As the lens holder is no longer stuck down it can't really be used at the moment!

Something that might be interesting to try is heating/cooling the lens holder.

[jaybeez]

Mike, how does your unit perform with real use now that you've made the modifications? im interested in the quality of the pictures and video.

As the lens holder is no longer stuck down it can't really be used at the moment!

Something that might be interesting to try is heating/cooling the lens holder.

I was thinking more along the lines of adding an opening of the same shape and size on the opposite side of the lens housing, placing it back in its original location and retesting it.

[Fraser]

This may sound bit harsh to some but $200 is peanuts for a unit that is capable of producing a medium resolution image. Take a look at the resolution and price of a Lepton or a Fluke VT02. SEEK have achieved an amazing retail price even when considering the current issues being discussed here. I think we have been spoiled by the Flir E4 achievements. I am still fully behind SEEK. Early adopters expect teething troubles.. Now the M u thermal camera is an example of a real dodgy TIC
Aurora

[miguelvp]

Preliminary observations done yesterday with an IR thermometer and my wife's Galaxy S4 with the temperature set to measure the center spot.

Paint 0.96 emissivity at 12 inches (IR thermometer is a 12/1 so to measure one inch diameter I have to be at 12 inches distance) it read within one degree (Fahrenheit)

Skin 0.98 emissivity it was under 2 degrees Fahrenheit.

Other random measurements but leaving the emissivity at 0.96 on both devices the Seek was around 1 to 2 degrees Fahrenheit off and always reading lower than the IR thermometer (UNI-T UT301A I believe it is that model I'll double check when I get home)

Also note that people that have been playing with the raw data are not really seeing that problem but maybe it's because between compilation and testing even if the camera is plugged it, it's not really running.

For example look at bktemp images after cleanning up the signal.
Image 7 to 8 he does quite a bit of filtering but I think from 7 we can even get rid off those remaining lines.

[Fry-kun]

How often does it calibrate in normal operation?

I don't have a working cellphone (mine is CyanogenMod 11, which doesn't work with Seek) -- so I'm trying it with the computer USB connection...

[jaybeez]

How often does it calibrate in normal operation?

I don't have a working cellphone (mine is CyanogenMod 11, which doesn't work with Seek) -- so I'm trying it with the computer USB connection...

every few seconds. more frequently if the image is changing.