Thermal Camera below 1000 € with streaming

[Theolean]

Hi everyone,

for a small side research project I need a thermal camera. Since this is only a side project the budget is limited to about 1000 €.

In this project I want to monitor the temperature of a complex high current contact design. Until now this done by using Thermocouples (Type T). But due to the complex geometry it is not easy to fix the probes on the contact. Also they provide only one measurement point and I'm missing the whole picture.

Since the accuracy is not critical a non conductive measurement seems like a good option.

After a little bit of research the biggest challenge seems to find a camera that can be controlled with a PC (in an ideal world, the camera would provide an open Modbus TCP or serial interface). I also want to record the data over a long period (3 days or longer).

A resolution above 240x160 would also be nice. A display on the camera is not necessary, but I do not mind if the camera has one.

After a little bit of research the Seek Shot Pro seems to fulfill the requirements, but I'm not sure about the control / record part. Is it possible to control the camera over USB?

Do I have more options?

[Fraser]

As nobody else has answered, I will share my thoughts.

1. Buy a used camera as you gain a massive discount and often receive the required software included in the price.

2. Buy an industrial or science camera as they are equipped for what you need and work with image analysis software intended for process and target monitoring. Monitoring can have long recording periods, radiometric data association and temperature alarms. Exactly what you will need for your task.

3. Industrial and Science thermal cameras are designed for accuracy and high quality data collection. I presume this is important to you. They are often available in a ‘static’ box camera format with both local and remote control. Remote control can be via RS232, IEEE1394A, USB or Ethernet.

4. For your purposes I recommend a FLIR A20 with 160 x 120 pixels. If the budget permitted, the A40 is better with its 320 x 240 pixels. Both are high quality cameras. I have seen A20 cameras sold on eBay recently for £800.

5. A decent thermal analysis software will be required. The FLIR software for the A20/A40 was “Researcher” but the latest version is named ResearchIR. Try to buy a camera with the software included as it can be very expensive.

6. Ensure you understand Emissivity and it’s implications when carrying out temperature measurements. Contractors can have shiny contact surfaces and very low emissivity. It may be necessary to coat contact arms with a high emissivity paint to improve measurement accuracy. You cannot coat the actual contact surfaces though so they may be challenging to measure with any great accuracy. The thermal conduction to the arm may be your only truly accurate measurement.

7. Consider using an industrial non contact IR thermometer sensor from a company like Raytek. They produce focussed sensor heads that attach to logging systems, either self contained or PC based. They are used in production environments to monitor specific targets and log changes in temperature for preventative maintenance. Worth looking into and they appear on eBay often. The emissivity issue is still present however.

https://www.coleparmer.co.uk/i/raytek-raycmltv3-comapact-cm-series-infrared-sensor-voltage-output-with-10-cable/3966005?PubID=UY&persist=true&ip=no&keyword=&&gclid=CjwKCAjw-vjqBRA6EiwAe8TCk3fKT68OJ2Gb_ERUMwJbzx4Ga_jrCb8subDUTuHfr0p5MNInLGKx3hoCDrEQAvD_BwE&gclsrc=aw.ds


I have a logging industrial IR sensor system from AGEMA that would likely match your needs but I am not certain I want to sell it as it is very useful for calibration checking sources.

Fraser

[Fraser]

Have a look through the Raytek range of sensors ..... there are many ! Some have a specific focus distance for smallest spot size whilst others are the common Set FOV type.

https://www.raytek-direct.com/content/raytek-fixed-infrared-sensors

The unit’s I have are a mixture of K type thermocouple output (great to use with any logging thermocouple system) , focussed sensor head with logging panel mounted control and display unit, self contained logging sensor head with PC download and finally, the common ‘dumb’ sensors that output varying voltage or current that directly relates to temperature.... such as 10mV per Degree C etc. As stated, the sensors can have a focussed input or a wide field of view. The target must fill the field of view for accurate measurements so a Focussed type sounds more suitable for your application.

These IR sensors are not as ‘fun’ as a thermal camera but it is all about capturing the required data, right ?

Fraser

[Theolean]

Thank you a lot for your reply!

The Flir A20 look like a very good match, with one (big) exception - the price. A new A20 is well above the limit, and used equipment is always a problem. I mean not in general, but I'm working in a company where it is a problem to buy used equipment...

A normal sensor is of course also an option and I have this option in my mind if I can not find a camera. But of course you just have one or a few data points - you do not see the whole frame. But it is definitive an option.

Emissivity is a good point. For most parts I can use black heat shrink, which should have an epsilon = 0.95 or something in this area.

That said, if the project would be bigger / more relevant I could / would buy an industrial grade camera, but unfortunately it is not. With these limitations, should I go for some IR sensors or a Seek Shot Pro in combination with a Software like Thermovision from JoeC?

Theolean

[Cat]

How about a low resolution thermal imager from Optris?
They also have pyrometers and "normal" thermal imagers, with your description it seems like the optris Xi 80 could solve your problem.
80x80 Pixel 50Hz thermal imager, motorized focus, USB and PoE interfaces. Software included.
https://www.optris.de/optris-xi-80
https://www.optris.com/optris-xi-80 (in english)

Disclaimer: I'm not associated with Optris but had a ThermaCam E45 to play with a few years ago 

[Fraser]

Cat,

That Optris product looks really good  You can pay more than that unit’s cost just for industrial image analysis software !

Fraser

[Theolean]

Interesting... never heard of this company. Looks really interesting, the price is also decent.
Thanks for the suggestion.

I will have a deeper look at this model.


A question from a beginner in the infrared measurement world. The optris has a resolution of just 80x80px for about 1k. The Seek Shot Pro has a resolution of 320x240px for 700$.
What is the reason for this price difference? Is it precision or the software / interface? I try to understand what the cost driving factors are.

[Fraser]

Price can depend upon many factors....

1. Consumer Vs Industrial market segment

2. Build quality and ruggedness.

3. Quality of Microbolometer

4. Quality of lens

5. Stability, repeatability and accuracy of Radiometric measurements.

6. MTBF (Reliability and longevity)

7. Capabilities of system. More tends to cost more !


Basically industrial systems often cost more than consumer grade products. Industrial equipment is often built to high standards and is encased in a rugged housing. Consumer grade equipment can be a very mixed bag when it comes to quality, performance ruggedness and longevity.

Full disclosure.... I have owned the SEEK Thermal Classic Camera dongle and I quickly sold it. I found its image quality very disappointing. It was not intended to be a high performance radiometric thermal camera and it showed in both design and performance. It was intended to be the VW Beetle of the thermal camera world.... i.e. “The People’s thermal camera”. Is a later version of the SEEK thermal camera a better performer ? I will leave that for you to determine with comments from owners of such on this forum. Just remember it will be built down to a price and not intended to be an accurate radiometric camera, especially in varying ambient temperatures.

Personally I would buy a Therm App or Thermal Expert in preference to any SEEK Thermal product. Those two camera options do have their own limitations though and neither are intended to be highly accurate radiometric measurement systems.

Fraser

P.S. What did I not like about the SEEK Thermal Classic Camera .....

High noise content in image
Low contrast in images
Poor PCB/Lens/FFC flag design leading to thermal gradient across the image (corrected in software fix)
Poor accuracy
FFC shutter operating every 2 Seconds to counter the microbolometers appalling stability.
Dongle format was not to my liking compared to the ‘all in one’ cameras that I use.
Lens of mediocre quality
Build quality highly suspect so long life not expected

I hope the newer models are better !

[frogg]

A question from a beginner in the infrared measurement world. The optris has a resolution of just 80x80px for about 1k. The Seek Shot Pro has a resolution of 320x240px for 700$.
What is the reason for this price difference? Is it precision or the software / interface? I try to understand what the cost driving factors are.

The Optris has a much more sophisticated internal assembly due to its motorized focusing mechanism. It's also IP67 sealed in a steel case, and has a much more sophisticated IO system. The Seek Shot pro is nowhere near the same thing.

The Optris NETD is not that great, however. The Optris is designed for industrial process lines and harsh environments. If you don't need that and you can just have a significantly less robust plastic construction with no IP rating, then there are other options.

Being honest, is your requirement "please give me as much frame rate and resolution for as cheap as possible"?

Do you need radiometric data? If you don't need radiometric accuracy and just need to see temp differences, then things get a lot cheaper. The radiometric requirement is a huge cost adder.

[Max Planck]

In this project I want to monitor the temperature of a complex high current contact design.

Could you provide some more details about the geometry. Not only the contacts but also the elements around.

Since the accuracy is not critical a non conductive measurement seems like a good option.

If there is need for temperature measurement, there is always a critical accuracy level. It will only be different in different cases. What will by your acceptable error level?

Emissivity is a good point. For most parts I can use black heat shrink, which should have an epsilon = 0.95 or something in this area.

Not a good idea, since it has a non negligible thickness and is not a good heat conductor. Another question is the geometry around. Do we have a cavity effect and/or reflections?

I also want to record the data over a long period (3 days or longer).

Does it mean you want to close the circuit and let it operate for a few days, recording temperature distribution?

Max

[Theolean]

Thank you all so much for replying.

@Fraser
it seems like there is much more to a thermal camera, than I can just find in a data sheet. Like optics/lens, contrast and MTBF. Therefore, I am very thankful for your help and input!

@frott

The Optris NETD is not that great, however. The Optris is designed for industrial process lines and harsh environments. If you don't need that and you can just have a significantly less robust plastic construction with no IP rating, then there are other options.

I do not need any robust construction nor a special IP rating. The camera will be used in a laboratory and will never be near water. The temperature and humidity are controlled and should also not influence the camera.

Being honest, is your requirement "please give me as much frame rate and resolution for as cheap as possible"?

The frame rate is my least concern, because the heating is relatively slow process. Even one picture per second would be more than fast enough. Right now I use Thermocouples to measure the temperature and I record the data every 10 s and this is fast enough.
Resolution is also not my mayor concern, but because I do not have any experience with thermal cameras, I want to understand why certain cameras cost so much more and how I can determine the quality of a camera. Since I'm new in this field, I do not have experience and don't know: Oh yeah, company A is known for high quality. And better be careful with company X.

Do you need radiometric data? If you don't need radiometric accuracy and just need to see temp differences, then things get a lot cheaper. The radiometric requirement is a huge cost adder.

Do I need radiometric data... good question. I want to measure the temperature distribution at the surface of the contact system. It should be possible to tell if the upper part of the contact is 1K, 5 K or 10 K cooler.

Could you provide some more details about the geometry. Not only the contacts but also the elements around.

I cannot show pictures of the contact system, but the contacts are pressed together in an open frame (no closed housing). The diameter of the contacts is about 50 mm. The height overall is about 150 mm.

If there is need for temperature measurement, there is always a critical accuracy level. It will only be different in different cases. What will by your acceptable error level?

The temperature will be between 20°C and about 150°C. If the absolute value is within +-3 K everything is fine. But the camera should me capable to measure a temperature difference of 1 K.

Not a good idea, since it has a non negligible thickness and is not a good heat conductor. Another question is the geometry around. Do we have a cavity effect and/or reflections?

As mentioned before, the observed process is pretty slow, so I assume, that the thermal conductivity of heat shrink has only a small influence on the measurement. From my experience: the temperature will change 1-2 K (max) in 10 s.

Since it is an open frame, I don't expect cavity. Reflections can be partially expected: two small metal rods are in some distance to the contact system, but they can be shielded.

Does it mean you want to close the circuit and let it operate for a few days, recording temperature distribution?

Yes, more or less. I will start the measurement, wait until I reach equilibrium. Switch the current off, wait until it cools down. Then repeat the hole process many times. I expect that such a measurement will need a couple off days, (more than three days). So it is important to have a stable and reliable system, that can run for days.

[frogg]

I want to understand why certain cameras cost so much more and how I can determine the quality of a camera. Since I'm new in this field, I do not have experience and don't know: Oh yeah, company A is known for high quality. And better be careful with company X.

Thermal cameras are not luxury watches or automobiles. Being focused on subjective qualities like "high quality" is, unfortunately, uninformative. "Brand" is usually only useful for after-sales support, and you must look at every single product within a brand, because some products in brand X are not good, and some products in brand X are great. There are also "brands" and then "detector manufacturers" and "brands that are detector manufacturers." You have to determine which of the three you are dealing with.

Cameras cost more based on:
1. Resolution
2. Frame rate
3. Optics
4. Features
5. NETD
6. Physical Construction
7. Sensor size / Sensor type
8. Noise reduction algorithms
9. How much profit they want to make.

Quite frankly, the only features you "need" are: manual min-max, low NETD, close focus, remote shutter activation, and time lapse capture. We could suggest cameras from $300 to $200,000 based on your requirements. Virtually any thermal camera on the market can satisfy your reliability requirement.

To detect a 10 deg K gradient across a contact, you need a very close focus lens. The Seek Shot Pro won't do that without some custom optical help.

[Max Planck]

Not a good idea, since it has a non negligible thickness and is not a good heat conductor. Another question is the geometry around. Do we have a cavity effect and/or reflections?

As mentioned before, the observed process is pretty slow, so I assume, that the thermal conductivity of heat shrink has only a small influence on the measurement. From my experience: the temperature will change 1-2 K (max) in 10 s.

Heat shrinks are usually made of poor thermal conductors with a thermal conductivity around 0.1-0.2 W/mK, so adding one will degrade contacts cooling and increase their steady state temperature.
Their wall thickness will be around 0.5 mm - just guessing - so you will have a temperature drop (heat srink wall will be a thermal resistance).

Personally, I would go with black paint or soot (I know, it is old fashion, dirty, etc. but high temperature paints are often using resin binders, thus are difficult to remove after, whereas soot is not).
I would go with a paint, with some tests first on a metallic sample.

If, as you wrote, the process is slow, what you need is a comparison at steady state, not after just 10 s.

Max 

[Theolean]

Thank you again for all the information, that you share.
This topic is much more complex than I thought it would be.

Quite frankly, the only features you "need" are: manual min-max, low NETD, close focus, remote shutter activation, and time lapse capture. We could suggest cameras from $300 to $200,000 based on your requirements. Virtually any thermal camera on the market can satisfy your reliability requirement.

To detect a 10 deg K gradient across a contact, you need a very close focus lens. The Seek Shot Pro won't do that without some custom optical help.

Could you name 1 or 2 cameras below $1000, that would fulfil theses requirements?

Heat shrinks are usually made of poor thermal conductors with a thermal conductivity around 0.1-0.2 W/mK, so adding one will degrade contacts cooling and increase their steady state temperature.
Their wall thickness will be around 0.5 mm - just guessing - so you will have a temperature drop (heat srink wall will be a thermal resistance).

Personally, I would go with black paint or soot (I know, it is old fashion, dirty, etc. but high temperature paints are often using resin binders, thus are difficult to remove after, whereas soot is not).
I would go with a paint, with some tests first on a metallic sample.

If, as you wrote, the process is slow, what you need is a comparison at steady state, not after just 10 s.

Max 

This is interesting, because I have seen heat shrink in different measurements. (All measurements were from big companies, that produce switchgear)
I will keep it in mind. It is relatively easy to paint the parts, so this seems like a good option.

And I will not only make one measurement. Right know I have TC and measure every 10s the temperature, so that I can see the development and change in temperature. Steady state is usually reached after 45 min or longer.