Author Topic: High speed thermal imaging cameras - When 60fps is just not enough !  (Read 17410 times)

0 Members and 1 Guest are viewing this topic.

Offline railrun

  • Regular Contributor
  • *
  • Posts: 113
 

Offline FraserTopic starter

  • Super Contributor
  • ***
  • Posts: 13149
  • Country: gb
Re: High speed thermal imaging cameras - When 60fps is just not enough !
« Reply #51 on: March 14, 2019, 01:12:32 pm »
Hi Railrun,

Wow ! That was unexpected!  I had not searched for the file yet but you have saved me the trouble. I will download it and anything else that looks useful on that page. I am unsure whether more of the SDK is needed, but at least I can try this file now  :-+

Thank you again

Fraser
« Last Edit: March 14, 2019, 02:57:36 pm by Fraser »
If I have helped you please consider a donation : https://gofund.me/c86b0a2c
 

Offline FraserTopic starter

  • Super Contributor
  • ***
  • Posts: 13149
  • Country: gb
Re: High speed thermal imaging cameras - When 60fps is just not enough !
« Reply #52 on: March 14, 2019, 05:36:51 pm »
Thanks to all who have tried to help with the GUI software issue  :-+

I need to do further instigation as the SCx000 series SDK is a chargeable software and it seems unlikely the GUI would require the use of such. I would expect the GUI to have been created with the SDK and to contain al required files to run independently. I now have the ISC_camera.dll file but as Ultrapurple has pointed out via a message, there may be other files needed. I am going to take a very close look at the GUI files installed on the computer and see if Ibcan work out what is wrong with the automatic install process. I am thinking there is a bug in the installation routine.

I have been resting today so began trawling through the FLIR Q&A section on their support web site. I found a few interesting “nougats” of information pertaining to the SC4000 and SC6000 cameras. They have likely saved me some head scratching.

Another issue I was experiencing was the thermal display only working when the cameras pixel scale was set to “counts”. In Temperature and Radiance settings I got zero temperature range and a dark red displayed image. Really weird but I thought likely a settings issue. Hence my desire to use the GUI to investigate.
Amongst the various answered on the FLIR Q&A pages I think I found my exact problem. If the camera is set to stream a linear scale of temperature, it will only show an image in Researcher and ResearchIR when they are set to the “counts” mode. Weird, but true !

I need to command my camera into its normal streaming mode and see if that resolves the situation.

It is all a steep learning curve and the user documentation was clearly written by someone who was very knowledgeable about the camera and Research software. There are gaps in explanations, or simple no explanations at all for some topics !

Some specialist scientific equipment can be hard to master in the early stages of ownership. They can have documentation issues and undocumented issues that confuse the user. An example is the “Counts” issue and effect of linear scaling being set on the camera. The windowing feature is another example. If a window of less than 96 pixels width is set on the camera, the data block sent in the stream is reduced in size to the point that it can no longer contain the images calibrated radiometric data ! The image is produced, but the software can not provide accurate temperature data ! This is likely an issue that results from the SC4000 and SC6000 being “First Generation” high speed thermal cameras. To maintain the frame rate, the image data packets are made as small as possible. This includes trimming the headers. Rather than prevent windowing width smaller than 96 pixels, FLIR permit the user to go smaller and enjoy the resultant higher frame rates, but without the calibrated thermal data.

Correction: The header in the data packets contains temperature sensor data for drift compensation. It is not the radiometric data calibration information. The FLIR CS wording is below:

"The smallest window width supported by Camera + Researcher is 96 pixels. Smaller size will leave too little space for the temperature sensor values in the image header, which in turn makes the temperature drift compensation to fail."


I am making progress  :)

Fraser
« Last Edit: March 14, 2019, 06:02:52 pm by Fraser »
If I have helped you please consider a donation : https://gofund.me/c86b0a2c
 

Offline FraserTopic starter

  • Super Contributor
  • ***
  • Posts: 13149
  • Country: gb
Re: High speed thermal imaging cameras - When 60fps is just not enough !
« Reply #53 on: March 15, 2019, 12:19:21 am »
I have made some progress on the Configuration GUI. I discovered that the Bullard PC software uses he ISC_camera.dll as it contains an Indigo core. I transferred all the loose files in the Bullard folder into the SC4000 GUI folder and tested the GUI. My thinking was that any important Indigo files would be in the loose files along with ISC_Camera.dll.

I was correct, the GUI works and offers me a choice of connectivity options now. Sadly the one I need is greyed out as I still do not have the USB driver for the SC4000. It is supposed to be on the GUI disk but there is no sign of it anywhere. The user manual does identify the USB driver but sadly I have yet to find it.

The USB module in the camera came from a company called “HardSoft” and is called the “Cielo Communications Interface”. The USB PID is DEDA and this is indeed HardSoft. The correct inf file is from Silicon Labs but I tried their drivers and no dice. HardSoft were bought by XIMIA and I tried their USB generic drivers as well. That was a long shot and it did not work.

Oh well, small steps. Time for bed :)

Fraser

« Last Edit: March 15, 2019, 12:30:20 am by Fraser »
If I have helped you please consider a donation : https://gofund.me/c86b0a2c
 

Offline TooQik

  • Regular Contributor
  • *
  • Posts: 56
  • Country: au
Re: High speed thermal imaging cameras - When 60fps is just not enough !
« Reply #54 on: March 15, 2019, 12:44:34 am »
Might be a long shot, but have you tried the FLIR drivers found on this page:

https://flir.custhelp.com/app/answers/detail/a_id/47/related/1
 

Offline FraserTopic starter

  • Super Contributor
  • ***
  • Posts: 13149
  • Country: gb
Re: High speed thermal imaging cameras - When 60fps is just not enough !
« Reply #55 on: March 15, 2019, 09:48:41 am »
Hi TooQik,

Many thanks for the link.

Yes I did try those drivers that FLIR state cover all their USB connected cameras. Sadly the driver set does not include the one needed for the SC4000 :(

This is one of the challenges of owning a more specialist camera like the SC4000. When you hunt for drivers or other information, you keep meeting the “Contact FLIR to discuss your needs”  statement. There is very little public facing website support available. These cameras seem to reside within a ‘bubble of secrecy’ so direct interaction with FLIR customer support is required.

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

Offline FraserTopic starter

  • Super Contributor
  • ***
  • Posts: 13149
  • Country: gb
Re: High speed thermal imaging cameras - When 60fps is just not enough !
« Reply #56 on: March 15, 2019, 12:32:46 pm »
I have decided to contact FLIR customer support to see if they will assist me with the GUI and it’s USB driver needs. I also need to find out why the camera is only offering a raw pixel data “counts” output data stream in Researcher and ResearchIR. 

The “Counts” situation..... for those unaware, scientific thermal cameras often stream their data in the “counts” format and the host software then creates the temperature and Radiance values using those counts. Counts are the true pixel values either completely raw, or with NUC and dead pixel corrections applied. The camera is normally 12 or 14 bits and that sets the range of counts values available per pixel. It is like the most raw thermal data available that enables whatever processing the users needs. In science, such raw data is important as it is not adulterated by processing that could alter pixel values, unless that is desired.

So “counts” for each pixel are a good thing, but not very user friendly when wanting to know the spot temperature on a target. Researcher and ResearchIR offer the user different pixel data modes. There is the Raw “Counts” mode or you can have the displayed image with Temperature readout or Radiance readout. Temperature and Radiance are calculated by the software from the Counts by using calibration tables for each camera temperature range. These tables are part of the cameras calibration data package that it downloads to the host computer when initially connected. The software recognises the calibration package contents and offers the user the available options or “modes”. My software is offering Counts, Object Temperature, Temperature and Radiance. If I select anything but Counts I get a blank display and no change in levels indicated on the vertical scale. It effectively goes to “Zero Span”. Really weird.

The SC4000 comes in two calibration types from the factory. There is factory calibrated and user calibrated. Factory calibration means the calibration table package is populated with the common temperature ranges plus any additional ranges requested by the customer. My camera has additional ranges for the filters that were supplied. His setup allows the customer to basically switch on the camera and measure temperatures. The user calibrated camera comes with no factory calibration tables populated so the user creates all the calibration tables they desire using their own calibrated black body reference source. In the case of a user calibrated camera, the host software indicates this by placing (User) after the modes of operation instead of (Factory). My camera is isplaying “Temperature (Factory)” yet is behaving like there is no calibration data present. Head scratching time  :-/O

The calibration data in the SC4000 is stored in flash memory. I am wondering whether the flash memory has been erased by accident but I can see no way for the user to do this and for the software to still report “Temperature(Factory)” in the available modes. This is why I think I need to speak with someone at FLIR who is familiar with these cameras. I know that this camera used to be able to generate temperature readout in Researcher 2.8 as a quick user guide created by its original owner contains screen shots of my camera producing such a readout whilst viewing a heating element. This camera was used for a very basic test of heating elements. Total overkill and there was no reason to mess around with the calibration tables.

The good news is that the camera will happily operate as a user calibrated unit as well. If I set up some user calibrated ranges with my black body, I can save them to the cameras flash memory and check that the host software is receiving the calibration package correctly over Ethernet at initialisation.

Sadly other draws on my time may mean I have to park this little adventure for the moment. I may poke FLIR and see if they are feeling helpful  ;D

Fraser
« Last Edit: March 15, 2019, 12:39:47 pm by Fraser »
If I have helped you please consider a donation : https://gofund.me/c86b0a2c
 

Offline TooQik

  • Regular Contributor
  • *
  • Posts: 56
  • Country: au
Re: High speed thermal imaging cameras - When 60fps is just not enough !
« Reply #57 on: March 15, 2019, 12:38:14 pm »
Hi TooQik,

Many thanks for the link.

Yes I did try those drivers that FLIR state cover all their USB connected cameras. Sadly the driver set does not include the one needed for the SC4000 :(

This is one of the challenges of owning a more specialist camera like the SC4000. When you hunt for drivers or other information, you keep meeting the “Contact FLIR to discuss your needs”  statement. There is very little public facing website support available. These cameras seem to reside within a ‘bubble of secrecy’ so direct interaction with FLIR customer support is required.

Fraser

No worries. I figured you might have already tried, hence the "long shot" in my post.

Hopefully FLIR come to the rescue and provide you with the necessary drivers. *crossed fingers*
 

Offline FraserTopic starter

  • Super Contributor
  • ***
  • Posts: 13149
  • Country: gb
Re: High speed thermal imaging cameras - When 60fps is just not enough !
« Reply #58 on: March 15, 2019, 12:41:47 pm »
Contacting FLIR could go two ways......

They will either be helpful and agree to assist me with the above challenges, or they will go into headless chicken  :scared: mode when they hear that I now own one of their heavily controlled science cameras and I am not the original purchaser ! We shall see.

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

Offline TooQik

  • Regular Contributor
  • *
  • Posts: 56
  • Country: au
Re: High speed thermal imaging cameras - When 60fps is just not enough !
« Reply #59 on: March 15, 2019, 12:44:51 pm »
Do you know of any other users of the SC series cameras that may be able to provide some assistance?
 

Offline FraserTopic starter

  • Super Contributor
  • ***
  • Posts: 13149
  • Country: gb
Re: High speed thermal imaging cameras - When 60fps is just not enough !
« Reply #60 on: March 15, 2019, 01:08:17 pm »
Sadly these are pretty rare cameras. Other users would also be rightly wary of helping someone they do not know with a heavily controlled thermal camera.

I will gently prod FLIR and see if I awaken a monster or someone willing to help.

Fraser
If I have helped you please consider a donation : https://gofund.me/c86b0a2c
 
The following users thanked this post: Durrfred

Offline FraserTopic starter

  • Super Contributor
  • ***
  • Posts: 13149
  • Country: gb
Re: High speed thermal imaging cameras - When 60fps is just not enough !
« Reply #61 on: March 15, 2019, 02:47:43 pm »
Well I have done it..... I have raised a support call on the SC4000 with FLIR CS.

I have asked for the full SC4000 GUI and comment on the "Counts" to "Temperature" issue.

Now to see whether they are willing to help me. I have been registered with them for many years so at least they can see my history with their products.

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

Offline FraserTopic starter

  • Super Contributor
  • ***
  • Posts: 13149
  • Country: gb
Re: High speed thermal imaging cameras - When 60fps is just not enough !
« Reply #62 on: March 15, 2019, 04:01:01 pm »
I am currently in communications with a 2nd level Science camera support tech at FLIR. He has already made some suggestions and we are working the "counts" problem. Fingers crossed that he can help me.

FLIR's support response time is simply superb  :-+

Fraser
« Last Edit: March 15, 2019, 05:05:37 pm by Fraser »
If I have helped you please consider a donation : https://gofund.me/c86b0a2c
 

Offline Ultrapurple

  • Super Contributor
  • ***
  • Posts: 1027
  • Country: gb
  • Just zis guy, you know?
    • Therm-App Users on Flickr
Re: High speed thermal imaging cameras - When 60fps is just not enough !
« Reply #63 on: March 16, 2019, 11:16:35 am »

FLIR's support response time is simply superb  :-+

Fraser

Funny how companies that pare prices to the bone don't offer much in the way of support yet those that sell better kit with higher margins are able to do so...

I have written elsewhere about how I see the future bringing inexpensive thermal imaging to the mass market. Hopefully there will still be room in the market for top-tier equipment with support to match. Try getting support for that budget no-name smartphone you bought last week and compare it to the spares situation for, say, a two or three year old iPhone or Samsung device.

Adage 1: you get what you pay for.
Adage 2: you pays your money and takes your choice.
Adage 3: the poor man always pays twice.
Rubber bands bridge the gap between WD40 and duct tape.
 

Offline FraserTopic starter

  • Super Contributor
  • ***
  • Posts: 13149
  • Country: gb
Re: High speed thermal imaging cameras - When 60fps is just not enough !
« Reply #64 on: March 16, 2019, 12:25:58 pm »
Ultraviolet,

I totally agree.

With the SC series cameras costing between $85K and $145K in 2008 I suspect a healthy chunk of the price covers the after sales support costs associated with such equipment. After sales support staff can be an expensive investment after all. I have normally had great support experiences with FLIR. The only time they let me down was with a faulty FLIR One G2. I returned it as requested and heard no more. I complained to the VP (Bill) covering that area of FLIRs business and he stepped in and sorted the situation immediately. Along with a profuse apology he refunded the full cost of the camera as well as providing the replacement ! Excellent customer focus  :-+

I normally dealt with FLIR’s professional thermography division and with cameras in that sector being more specialist and costly, you tended to deal with very customer focussed and knowledgeable support staff. Just like I am currently with the SC4000 support call. There is an argument that such support is due to a customer who paid a lot of money for a product ...... just think of the car industry though...... after sales support can be very patchy, even with brands like BMW and Audi (I speak from experience!). I personally find FLIR to be very helpful in time of need so long as you are honest with them and can comply with the basic requirements for support. They will normally want a serial number in order to check an equipments ownership and history. If you are not the original purchaser of the equipment you can experience difficulties and they can decline to help. This often relates to the Dual Use technology status of the equipment and their lack of knowledge of who they are being asked to help. The Dual Use technology regulations cover the equipment, technical documentation and support.

If is only right to balance the above praise with some facts of life when dealing with large companies and/or specialist equipment though.......

I used to maintain Racal HF transmitters that cost a small fortune. Spare parts were eye wateringly expensive. An older engineer and my mentor explained to me that if a single 5 pence screw were needed for one of the transmitters, Racal would likely charge £10 for it. The reason is overheads..... the screw cost 5p and is nothing special, but it has been bought in by Racals purchasing department, placed in a spares box on a spare parts shelf in the spares holding facility on a site that has building management and staff to administer it. Added to that the part I order has to be picked by someone, packaged, shipped and billed to me. All of the above have costs associated with them and someone has to cover those costs. The screw I ordered would likely cost me 10p with 100% markup applied but the £9.90 would be the overheads for holding and processing that screw as a spare part. Someone has to pay, and in the case of a spare part, it is the requisitioner. Granted, there is likely a “management” fee added to the order as well and that often equates to pure profit, but such is life.

The other story I read recently that amused me was as follows.....


A women joins two friends for a chat in a coffee shop. The two friends order coffee but the lady orders just tap water with ice and a slice of lemon. When the bill arrives there is a charge of £2.50 for the water (London prices !). The lady is outraged at being charged for “just tap water” as it is “free” and demands to speak to the manager. He arrives and the lady demands that the £2.50 is removed from the bill. The manager declines and explains why.....

“You are my customer, you have enjoyed the fasciities of my premises whilst meeting your friends, you have ordered tap water with ice and lemon. That order did not come at no cost to my business. The staff who served you have to be paid for their time, a part of which they spent with you.The water is free but the ice comes from an ice maker that has to be paid for and maintained, the lemon was freshly cut by the staff which takes time and was placed in a glass which has to be cleaned by staff after you have finished with it. The water is free. The coffee shop premises, staff’s time, ice and lemon are not. “

The lady sheepishly paid the bill.

Now you know why you pay more to “eat in” rather than take away..... overheads can be significant.

Very little in life is free so if a customer wants the cheapest possible prices, some things can suffer. Aftersales support is one such possible loss as it does not always influence the original sale and is a hidden cost that customers often do not consider. In a gambling world the seller may offer reduced costs in order to gain sales and hope that little, if any support is requested. The customer only finds out about the effect of cost cutting later when support is requested and found to be either absent or of poor quality. As Ultraviolet says, you often get what you pay for.

Now back to FLIR and my SC4000..... FLIR have every right to decline my request for “support”. They have received no payment from me in the form of original purchase or a support contract. They are effectively spending their time helping me out of goodwill and a wish to maintain their good reputation as a quality provider of advanced thermal imaging equipment. They could still have declined to help me though and I would truthfully have little reason to criticise them.... remember the coffee shop scenario.... nothing is truly free where service or support is involved.

When I first heard that FLIR was entering the consumer thermal camera market I had concerns for them as a company. Other companies that made the same move have quickly regretted it as there is often a massive difference between selling products to professional institutions, Industry and the general public (consumers). Most notably, the public tend to want cheap prices, something for nothing and still get gold standard after sales support. That is just how life is in the consumer age. The companies that venture into the consumer marketplace often experience an exponential increase in after sales support calls, duration of support calls and criticism ! Not great for business. DATONG were a company I worked with who eventually decided the overheands of supporting consumer products were so excessive compared to their sales to industry that they pulled the plug on all of their consumer products in order to concentrate on just the Industrial products and sales. I feared that FLIR could experience the same issues...... and they did !

FLIR tried to be smart in how they entered the consumer market place. They created a whole new division of their company that was based in Europe. That division was responsible for consumer product sales and support. This would leave the Industrial divisions of FLIR to continue to focus on their established customer base that FLIR traditionally support.  The new Consumer division of FLIR began selling their new products that included the FLIR One G2 and other Lepton based equipment. There were teething troubles with some products as is to be expected. It was now that FLIR’s brand new Consumer Support team were tested....... and there were some problems with the  quality of support provided. Again this was to be expected as the Consumer support team in Europe were new to the job, having been recruited from outside FLIR. It is very easy for a companies support reputation to become tarnished, especially in this modern age of social media. I feared this might happen to FLIR but they are a very robust company and appear to have weathered the storm well. Having spoken to the VP of the European operations I know that, as a long term employee of FLIR, he intended to install the high operating and support standards into the new European branch of operations. He was horrified by my experience with the F1G2 failure and resulting poor customer support. Because of this, and other negative reports coming onto his desk, he flew from the USA to the European branch for some serious discussions about their operation and performance. During those meetings he tasked the support team with sending me a brand new FLIR One G2 immediately and charge it to FLIR’s internal account. He advised me of this action whilst still at the European Division.
As stated He also refunded me the full cost of the F1G2.

We are a ways down the line since FLIR entered the Consumer market and I still feel it was a risky move on their part. They appear to be coping but I still have some serious doubts about some of their Lepton based products. They are built down to a price in order to meet consumer price expectations, and IMHO, it shows  :(

Fraser


« Last Edit: March 16, 2019, 05:10:37 pm by Fraser »
If I have helped you please consider a donation : https://gofund.me/c86b0a2c
 

Offline eKretz

  • Frequent Contributor
  • **
  • Posts: 870
Re: High speed thermal imaging cameras - When 60fps is just not enough !
« Reply #65 on: March 16, 2019, 03:22:52 pm »
We are a ways down the line since FLIR entered the Consumer market and I still feel it was a risky move on their part. They appear to be coping but I still have some serious doubts about some of their Lepton based products. They are built down to a price in order to meet consumer price expectations, and IMHO, it shows  :(

Fraser

While that latter statement is true, I think most consumers at that price point aren't as technically savvy as many of the folks you'd find around here. As Arthur C. Clarke's commonly quoted line so aptly demonstrates, a very large percentage of the average consumers consider highly technical products "magic." It is in my opinion unlikely that these types of consumer even notice most of the deficits you or I might consider major issues.

The biggest problem with that is that the large volume of these clueless purchasers make them a major driving force for product development - and often they drive it to ever lower cost, reduced features and lower quality. The best we can hope for is that they will maintain at least a few different price/quality tiers.
« Last Edit: March 16, 2019, 03:26:21 pm by eKretz »
 

Offline mahony

  • Regular Contributor
  • *
  • Posts: 156
  • Country: de
Re: High speed thermal imaging cameras - When 60fps is just not enough !
« Reply #66 on: March 18, 2019, 11:47:35 am »
Hi Fraser,
a was able to work a little bit with data from some SC6xxx and SC7xxx cameras on ResearchIR some time ago and your symptoms sound to me like a missing calibration file or a file containing invalid data.
Typically what I have seen so far a those cameras stream AND store only raw(ish) 16bit counts or digital level (DL) data. All the calibration to radiance or temperature is done in software via calibration file of various flavor. Some formats include basic calibration in their header (i.e. the old FLIR *.ptw file and new *.sfmov may do so) or the header refers to some external calibration file. You habe a lot of possibilities to create your own calibrations in ResearchIR however ...

Most likely the calibartion data is somehow invalid as I remember missing files cause the software to not even offer you a radiance or temperature view.

What file format does your camera produce when recording? Maybe some inspection of the associated files gives some insight.
 

Offline FraserTopic starter

  • Super Contributor
  • ***
  • Posts: 13149
  • Country: gb
Re: High speed thermal imaging cameras - When 60fps is just not enough !
« Reply #67 on: March 18, 2019, 12:21:45 pm »
Dear Mahony,

I agree, this does sound like the package of calibration files (that the camera uploads to the host when first connected) is in some way corrupted or not valid. As you also state, the actual calibration file must be present as Researcher and ResearchIR only offer scale options that are present in the calibration file.
This camera is very different to the likes of the Ex or Exx series. As you say, the conversion of Pixel level “Counts” to Temperature or Radiance unit’s is carried out on the Host computer. This offers the greatest versatility to the user. The loss of the Factory calibration is not as serious as if it happened to say a FLIR Ex series camera as these SC4000/6000 cameras are actually available from FLIR with no factory calibration installed. This camera is intended to be operated with either a Factory or User provided Calibration. The menu’s for creating User Calibration are comprehensive and include four levels levels of complexity, depending upon the users experience and needs. The calibration process just requires a variable temperature black body reference. The process is detailed via on screen user instructions as a calibration is carried out. Very well thought through. If I have to, I will create my own set of 2 Point Calibration tables. I wanted to ask FLIRabout the issue first in case it was something that I could rectify to enable the Factory Calibration.

The update on the FLIR support call came in this morning. The 2nd Tier Scientific Support chap has contacted the specific FLIR engineering team who know about this camera. As a Legacy model he needs their assistance to diagnose what has happened. I have asked him whether I can access the contents of the Factory calibration files, either on the camera or the Host PC. From what I can tell, the calibration file package is downloaded to the Host PC when the camera connects but I cannot find it in the computers file system. I considered trying to capture the Ethernet traffic between the camera and PC during camera initialisation and seeing what the calibration package contained. That is getting a little time consuming for me at the moment though as I am in the middle of a bathroom upgrade ! Hopefully FLIR will have some suggestions. They may even hold copies of these cameras original calibration files. Such is not unheard of with specialist scientific cameras.

Thank you very much for your comments. I will keep this thread updated with FLIR’s comments on the matter. If all else fails I will set up my Black Body and try creating a nice fresh two point calibration using the User calibration routine. The NUC and Dead Pixel replacement routines appear to be working fine as those are regularly updated upon user request.

Fraser
« Last Edit: March 18, 2019, 01:06:00 pm by Fraser »
If I have helped you please consider a donation : https://gofund.me/c86b0a2c
 

Offline FraserTopic starter

  • Super Contributor
  • ***
  • Posts: 13149
  • Country: gb
Re: High speed thermal imaging cameras - When 60fps is just not enough !
« Reply #68 on: March 18, 2019, 06:28:19 pm »
Having read some documents on refreshing the cached calibration file in ResearchIR, I now know that with the SC4000/6000 cameras, the calibration file is persistent after prog.ram closure and reboot. It is an XML file held somewhere on the PC. In order to refresh the data held in the calibration cache, the user has to select “clear cache”  in ResearchIR and the program clears it out and reloads a fresh copy from the Camera. I have used this function upon guidance from FLIR but sadly it did not solve the problem. I will go on the hunt for the caged calibration file in an effort to investigate its contents.

Fraser

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

Offline CatalinaWOW

  • Super Contributor
  • ***
  • Posts: 5173
  • Country: us
Re: High speed thermal imaging cameras - When 60fps is just not enough !
« Reply #69 on: March 18, 2019, 06:42:32 pm »
We are a ways down the line since FLIR entered the Consumer market and I still feel it was a risky move on their part. They appear to be coping but I still have some serious doubts about some of their Lepton based products. They are built down to a price in order to meet consumer price expectations, and IMHO, it shows  :(

Fraser

While that latter statement is true, I think most consumers at that price point aren't as technically savvy as many of the folks you'd find around here. As Arthur C. Clarke's commonly quoted line so aptly demonstrates, a very large percentage of the average consumers consider highly technical products "magic." It is in my opinion unlikely that these types of consumer even notice most of the deficits you or I might consider major issues.

The biggest problem with that is that the large volume of these clueless purchasers make them a major driving force for product development - and often they drive it to ever lower cost, reduced features and lower quality. The best we can hope for is that they will maintain at least a few different price/quality tiers.

The low end addresses two wildly different market segments.  Obviously the unsophisticated group which will not notice the defects and the compromises.  And also the price sensitive, but very knowledgeable group that recognizes the deficiencies and won't be fooled by them, and can even in many cases find ways to overcome them.
 

Offline FraserTopic starter

  • Super Contributor
  • ***
  • Posts: 13149
  • Country: gb
Re: High speed thermal imaging cameras - When 60fps is just not enough !
« Reply #70 on: March 18, 2019, 07:55:33 pm »
I have found the SC4000 cameras calibration file that is held in flash memory on the camera and downloaded to the Host PC upon request. The file is in XML format so is easily read in a plain text editor  :)

I have had a quick look through the calibration file and can see no glaring issues. I have sent the file to FLIR support as part of my open case. Maybe they can spot why Researcher and ResearchIR both consider the actual calibration data invalid but the 'envelope', detailing what calibration files are present, valid  :-//

The file was easily found once I knew that I was looking for a .XML file format  :) Here is its location in case anyone else ever needs it:

 C:\Users\Fraser\AppData\Roaming\ResearchIR\Cameras\SC4000 (Serial Number)

I have attached the Calibration file in PDF and plain text formats. I have only edited the Serial Number and replaced it with (Serial Number).

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

Offline Max Planck

  • Regular Contributor
  • *
  • Posts: 97
  • Country: pl
Re: High speed thermal imaging cameras - When 60fps is just not enough !
« Reply #71 on: March 18, 2019, 09:37:54 pm »
Hi Frasier,
if this is the camera I saw on ebay, you are a lucky guy.
I have never used that model but a few things from experience.
The FPA is an InSb, which means it is sensitive, roughly, in the 1-5 um band, even if cooled MWIR systems are usually advertised as 3-5 um. To avoid sunlight reflections and other problems, the lens (or sometimes the FPA window) has a filter layer reducing that band. If you can see through the window there is probably no filter here - good news. You wrote you have a wideband lens. If this is what FLIR is calling wideband, it should operate approx. in the 1.5-5 um band (an you are then not only lucky, but very lucky).
If your FPA window is not filtered, basically, it means your camera can see in the SWIR band and you could experiment with old fashion cameras lenses. The old ones coming from 38 mm film systems are the best solution as usually there should be no IR blocking coatings.
The biggest strain on the Stirling cooler takes place during devar cooling down, so it is usually better to leave the camera operating, even for 1-2 h, than to switch it on-off-on.

If the problem with USB drivers is not solved, you could always try with the Camera Link interface.
« Last Edit: March 18, 2019, 09:42:11 pm by Max Planck »
 

Offline FraserTopic starter

  • Super Contributor
  • ***
  • Posts: 13149
  • Country: gb
Re: High speed thermal imaging cameras - When 60fps is just not enough !
« Reply #72 on: March 18, 2019, 11:29:47 pm »
@Max Planck,

The camera is Pt No. 420-0044-04-00BB but I have been unable to determine which cold filter passband is fitted.

The lens is a very nice wideband 1.5um to 5um Janos NYCTEA 25mm optic. I attach the datasheet. As you likely know, wideband lenses tend to be fitted to wideband 1.5um to 5um cameras. The requirements of a 3um to 5um lens being covered by the Janos ASIO range of lenses. There seems little reason to fit the NYCTEA wide band lens to a standard 3um to 5um Camera.

Whilst testing the SC4000 I could clearly see my wife moving around in the adjacent room through the glass panes in the dividing doors. This camera appears to be happy to image through toughened glass  :)

The original use of this camera involved imaging heat sources through high temperature glass and ceramics. The camera came with two FLIR specialist filters. One is the the GHT (Glass High Temperature) and the other is the TGL (Through Glass). The cameras factory generated calibration tables cater for these two filters and I have the FLIR user manual that came with them. Sadly I cannot find their pass and details  :(

As this camera was used to image though Borosilicate glass etc, I believe it might well be the wideband 1.5um (SWIR) to 5um (MWIR) model. I attach the transmission graphs for Borosilicate and fused silica glass. She is also capable of radiometric data collection up to 2000 Degrees C  8)

I did say this was a very special camera :) She is a beauty.

Fraser

« Last Edit: March 19, 2019, 12:32:23 am by Fraser »
If I have helped you please consider a donation : https://gofund.me/c86b0a2c
 

Offline FraserTopic starter

  • Super Contributor
  • ***
  • Posts: 13149
  • Country: gb
Re: High speed thermal imaging cameras - When 60fps is just not enough !
« Reply #73 on: March 19, 2019, 01:10:26 am »
I have been taking a look at the calibration file I found earlier today. I think I now understand why I am having a problem obtaining Temperature and Radiance conversions from the Counts output of the camera....

The camera appears to have been calibrated in only two use scenarios.... when fitted with the GHT filter and when fitted with the TGL- ND2 filter.

Each Range Calibration entry consists of the following:

Calibcase
Calibconst
FPAconst
Stdconst
Tcpconst


These are present for all ranges but the GHT and TGL-ND2 have additional entries after the Tcpconst. These additional entries are the factory Black Body calibration points that enable the host software to establish the correlation of Reference source temperature to camera FPA count reading. My camera was used in a scenario that involved the GHT and TGL-ND2 filters only. As such it would appear to have been calibrated for only those two scenarios.

I can test this easily enough and even do some cutting and pasting within the Calibration file to test this theory. I think I may now have my answer to the weird Counts only situation in the modes menu. It may also explain the “SP” suffix in the calibration Ilene part number for my camera .... “SP” for “Special” ? As in Special calibration for a customer using the GHT and TGL-ND2 filters and not the standard calibration. The standard calibration range file entries were left in place so the Host Software displays it but has no data on which to calculate the Counts to Temperature, or Counts to Radiance conversion. Hence my ‘issue’

To solve this I can create User calibration entries with my Black Body reference sources. I could even use the data collected to add the missing calibration points to the cameras factory calibration files.... if I can find out how to safely over-write the XML file in the cameras flash memory. Alternatively I can just edit the factory XML file in the Host Software or just work with User calibration entries in the range selection menu.

Time for bed now. I think I may have solved this little mystery now  :)

Fraser
« Last Edit: March 19, 2019, 01:40:14 am by Fraser »
If I have helped you please consider a donation : https://gofund.me/c86b0a2c
 

Offline Max Planck

  • Regular Contributor
  • *
  • Posts: 97
  • Country: pl
Re: High speed thermal imaging cameras - When 60fps is just not enough !
« Reply #74 on: March 19, 2019, 10:09:54 am »
From the data you have posted, it looks like your camera was configured for a very specific application, i.e. glass manufacturing process control. As the manufacturer charges for every single calibration file, only what was needed was choosen. Idem for the lens. In fact you have now a 1-3 um SWIR configuration, not a wideband one. Given the prices asked for a 1-5 um wideband lens I can understand that.
Now, an educated guess, GHT and TGL-ND2 stands for glass high temperature and neutral density 2, respectively. The first filter is probably a narrow-bandpass interference filter to measure glass surface temperature during the manufacturing process. The second one is for reducing the amount of energey reaching the FPA from high temperature sources, typical for glass processes. Both filters are so called warm filters, source of noise emission and reflections, thus, each one requiring a dedicated NUC and calibration.

@Max Planck,

Whilst testing the SC4000 I could clearly see my wife moving around in the adjacent room through the glass panes in the dividing doors. This camera appears to be happy to image through toughened glass  :)
It all depends on the glass composition, coatings, thickness and source temperature, but a standard glass should give you an acceptable transmission up to 2.X um for room temperature sources.

Max

 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf