Thermal eye 2620AS not working (solved)

[sam1275]

Hello.

First, here's the manual for Thermal eye 2620AS, because we need it to discuss further:

















Today I try to play with it, it was bought as new and I didn't use it because it have a rare connector, however this time I soldered some cable direct on to it and just want to see if it works.

First I used the low voltage input, and because I don't have a 1.8-3.2v power source, so I use Li-ion battery with 2 IN4007 in series, but the camera did not work at all, despite I verified the voltage is in range. Then I tried the high voltage input, the camera's NUC shutter works, but no video output, I tried both 12v and 24v adapter.

What can be the problem?

Thanks.

[Bill W]

Stating the obvious, but just to be sure....

Have you got pin 8 = ground ?
Otherwise no video out

Have you connected all the power pins as it says ?
Seems rather insistent that you need to connect the lot so that might actually be true, and you have a bit of a problem without the JST 14 way

All I have are the 13 ways from the 2000 so not much help.

regards
Bill

[Fraser]

Sam,

Thank you very much for sharing the L3 core user manual. I have been after such information for some time just in case i get one of these cores.

 

Fraser

[Fraser]

Sam,

If I were testing this camera I would definitely use the high voltage input as that is the pre-regulator for the core and so provides better overvoltage protection etc. The low voltage input is positioned after the pre-regulator and an over voltage event here could be damaging to the cores regulator and even other components.

Note the grounding requirement of pin 7 and, as bill said, correct video output selection via the grounding of the video select pin.

An operating FFC shutter usually indicates a happy core that is running so this does sound like a video output selection issue. The board can also drive a KOPIN LCD and it may be running in that mode.

Fraser

[sam1275]

Stating the obvious, but just to be sure....

Have you got pin 8 = ground ?
Otherwise no video out

Have you connected all the power pins as it says ?
Seems rather insistent that you need to connect the lot so that might actually be true, and you have a bit of a problem without the JST 14 way

All I have are the 13 ways from the 2000 so not much help.

regards
Bill

Hello Bill.
Thank you very much, err... okay I'm so stupid this time, I didn't connect pin 7 and 8 to ground...
I'll report back when I have time to fix that and test again.
Sam

[sam1275]

Hi Fraser.
You are more than welcome and thank you very much for your help.
I will report back when I have any progress.
Sam

[Fraser]

Hi Sam,

Pin 7 floating causes unpredictable behaviour of the core.
Pin 8 floating allows the pin to drift to a positive potential. This causes selection of the direct LCD drive for a KOPIN Cyberdisplay. The NTSC output is switched OFF in that mode. You definitely need to ground pin 8 in order to get an NTSC video output.

Fraser

[sam1275]

Hi everyone!

I'm very glad to announce, with your help, now my Thermal eye 2620AS is officially alive

Screenshot first, I record clips using Philips PMC7230, then take a frame shot on PC using VLC, unedited:





The first one is a normal frame with lens cover on, I heard A-Si is noisier than VoX, but this camera is amazingly clean.
The second picture is a clock on the wall.
The third picture is a heater, note the red arrow, it point to a circle area that's not normal, it appears brighter than normal when looking at a cold object, and appear darker when looking at a hot object, it does not move when lens turning. It was a big dust on the sensor surface, I fixed it easily.
I just added the fourth photo, that's the boot screen

I saw other people's thermal eye 2620AS have a temperature indicator, but mine havn't... Link:
http://www.rtftechnologies.org/general/thermal-imager-L3-thermal-eye-2500AS-2620AS.html

The focus distance is pretty long, it can barely focus closer than about 3 meters, with the adjust of the lens of course.
Edit: confirmed this in the Manual.

I also noticed the high volt line can work on a normal 5v USB, much lower than the supposed 8v, but I have no luck to make it work on a Li-ion battery.

I wonder if I can adjust the minimal span temperature to get more detail, I cannot find that info in manual.

I really appreciate all your help!

Sam

[Bill W]

Hi Sam,

Congratulations on getting it going.   

I believe that these cores were running a fixed temperature - brightness ( = video voltage) relationship.  The reason it does not look too noisy is that the range is kept fairly wide. 

Additionally the AS2620 appears to use a sensor with quite big pixels (46u) compared to the earlier AS2000 which was 30u.
Additionally the AS2620 appears to use a sensor with quite big pixels (46u) like the earlier AS2000.  The 3500 went over to 30u.
[Text above corrected, got the pixel sizes wrong]

regards
Bill

[sam1275]

Hi Sam,

Congratulations on getting it going.   

I believe that these cores were running a fixed temperature - brightness ( = video voltage) relationship.  The reason it does not look too noisy is that the range is kept fairly wide. 

Additionally the AS2620 appears to use a sensor with quite big pixels (46u) compared to the earlier AS2000 which was 30u.

regards
Bill

Thank you very much Bill.
I did a test and fortunately, it does have AGC instead of fixed gain, however the max gain is not very high and thus limit the details when looking at low temperature delta scenes.
Also could you give me any reference of the pixel size? I see the sensor window is much smaller than the Alpha Indigo.
Sam

[sam1275]

Hi everyone.
I just have another question. In the P.6 of the manual, it says the shutter will trigger if the camera temperature changes more than 12 degrees C, so it's not TEC stabilized? Also 12 degree C is pretty large, i remember Flir camera will trigger the shutter when sensor temperature changes 0.1-0.2 C, is that a error of the manual?
Thanks.

[sam1275]

I just found this camera will overload the pixels when looking at hot objects (which are not hot enough to overload my other thermal cameras).
Here's an example, it a 1800 watts heater:

The pixel pattern is interesting, also it seems the core does not use all available brightness level of the video signal when display a normal image, because you can see the overload pixels have a big brightness jumping from others. I hope I can change this behavior but I think it's hard coded.

[Fraser]

Interesting ?

These and similar cores are to be found in fire fighting cameras so I would expect them to be multi range capable of adjusting to very hot targets.

I am wondering whether these cores have a software application that may be used to set them up for different behaviour, temperature displays and colours?

Can the seller of your unit provide any comment on this for you ?

Fraser

[sam1275]

Interesting ?

These and similar cores are to be found in fire fighting cameras so I would expect them to be multi range capable of adjusting to very hot targets.

I am wondering whether these cores have a software application that may be used to set them up for different behaviour, temperature displays and colours?

Can the seller of your unit provide any comment on this for you ?

Fraser

Thank you Fraser.
I don't think I can reach the seller because that's a realy long time ago and I don't even remember that ebay account...
I would be happy if there's anyway to modify the settings.
Here's another guy have a similar core, he didn't mention the software either:
http://www.rtftechnologies.org/general/thermal-imager-L3-thermal-eye-2500AS-2620AS.html
His camera have temperature display while mine don't.
Sam

[Fraser]

I am sad enough to remember that 'RFT' camera when it appeared on eBay and he bought it. It was sun damaged when looking out to the horizon and the Sun passes within its view. It was the first Sun damaged A-Si core I has seen. I avoided it due to the sensor damage.

RFT is asking for the PC software on his front page.

The Raytheon/L3 core is likely configurable for different tasks but what we do not know is whether the configuration software was made available to end users or was it 'factory configuration at time of order' only.

These cores saw service in many different scenarios and equipment embodiments. As such there are different lenses and software configurations, such as the somewhat crude temperature scale. I would suspect a core with the temperature scale had come from a fire fighter camera rather than a surveillance camera application.

I note that the cores are still available as old stock from SPI. You could try dropping them an email enquiringly about userconfiguration options. Sadly I got no response from them when asking about software for another camera so do not be surprised if you also get no response. SPI are selling the 2000AS cores quite. Cheaply on their web site but there is mention of documentation only......no software detailed.

I am soon to receive a camera with a Raytheon/L3 3500AS core and will be keen to explore user configuration options, if any.

Fraser

[sam1275]

Thank you Fraser.

So the bad pixels on RFT's camera is sun damage? That answers parts of my another post .

I believe the camera must have some kind of firmware, and on theory, the settings must can be changed, but I'm afraid it might be impractical for end user, if the factory didn't want us to change them.

Sam

[Fraser]

Yes, sun damage was stated in the auction that sold that camera to RFT. The seller was honest enough to detail the issue with the sensor and its permanent nature.

I have asked a friend about whether these cores have OEM configuration software. If they do, it may be that SPi would provide such to citizens within the USA. They are very funny about discussing thermal cameras and cores with those outside the USA.

The firmware will contain configuration options for certain. What those options are and how they may be accessed is a different matter. The L3 3500AS core brochure states that it is configured via USB 2.0 and an easy to use GUI. Sadly it does not state whether the core contains the GUI or if an external PC running configuration software is needed.

Wth regard to your core hitting the pixel saturation point.....it is quite common for cores intended for general use to have a temperature range of -20C to +150C. Other ranges can be provided but need to be configured in the firmware, with automatic or manual selection, and individual range calibration is often required.

Fraser

[sam1275]

Thank you very much!
So the 3500AS have a USB interface and don't need serial cable? That's great.
My 2620AS didn't mention anything in the manual though, and there are many ports onboard that don't have a word in the manual.
Please kindly post again when you find more information about it.
Sam

[Fraser]

Sam,

I have done some research and confirmed that the 3000 and 4000 series cores do have OEM PC based configuration software. They are configured via a USB interface. The GUI is provided on a Developers CD but L3 also provide access to a more extensive command set for the core via direct commands.

The cores command set is detailed in the ICD document (attached) There is also a nice core integration document that I have but it is too large to attach. An expansion connector guide is also provided and I attach this for interest.

Sadly this information may not help you with your 2000 series core but I suspect that a similar, if not identical core configuration process is used, but via RS232 ?

Once my L3 3500 core based camera arrives I shall do some investigation of the configuration process. I suspect I will just need to load an appropriate USB-UART bridge driver and then access the core via its command line interface. Sadly I do not have the GUI software 

Fraser

[Fraser]

Sam,

With regard to the user manual providing no detail of configuration. This is not uncommon. The Core manufacturer can produce a 'plug & play' core for an OEM to install or the OEM can purchase the Development kit that gives them the tools to configure the core as they wish. The Preconfigured core comes with a simple user manual. The Developer CD contains the configuration utility, integration manual, expansion port manual and ICD manual that details the Communications port and instruction set for the core.

You have a preconfigured core with its associated user manual and no configuration detail. The other manuals may be on the internet somewhere but I was looking for 3000 series core manuals and did not find any 2000 series stuff of note. Sorry.

Note that the 3000 and 4000 series cores have some hardware pins to select options such as Zoom and image polarity. Your core likely has the same as these are common modes of operation and hard wired into the boards rather than Communications link controlled. This means an OEM can place the core in a shell, add power and LCD, plus some push buttons, and they have a pretty much fully functional thermal camera for fire fighting or security. Plug & Play thermal imaging really

Fraser

[Bill W]

To see 'hot' things on these cores you need 'EI' mode (electronic iris - or low gain) active.  All I see in the ICD for the 3600 posted is the ability to show the symbol.

Could be that some of the part variations were whether EI was available ?

The grid shown is pretty typical of a bolometer hitting the end stops.  Bright white where you get some overflow or capping and some still register OK.  This is caused by gain and offset variation pixel to pixel.

These are not peltier stabilised sensors, they have multi-point calibrations and interpolate between them.  The separate calibrations are probably the 12 degrees of sensor temperature mentioned.  Can't recall if there were 'long' and 'short' recalibrations but that is how they got around the sensor changing with temperature.  Lots of ovens with lots of charts and a semi- automated system to move trays of cores between them while doing the calibrations.  This is a surprisingly common approach.


regards
Bill

[sam1275]

Thank you very much Fraser.
The documents are incredibly detailed, but sadly they seems not apply to my core. However I'm looking forward to read something about your camera's experiment 
Sam

[sam1275]

Thank you very much Bill.
That's exactly what I guessed, it might have a shutterless FFC table within every 12 degrees range, however the startup calibration is stupid I think, it doesn't even have to heat/cool the sensor on startup, why force to calibrate N times anyway?
My Indigo Alpha actually have a high-temp mode, when looking to/away from hot things, it will trigger the shutter and change the temperature mode, I think this behavior is not even documented.
In compare, Flir i series seems able to adjust temperature range without triggering the shutter. But I remember someone said they are single span design, and not actually very good because they are not using DSP's full range in normal scenes.
Sam

[RO]

Any chance someone can reupload the manual for Thermal eye 2620AS, because I picked one up on ebay recently

(p.s. the thermal-eye Infrared Thermal Security Camera houses this core)

[RO]

Edit for important note: This is definitely a completely different 2620AS board and probably manual; check video

I found the youtube video and other post:
https://www.eevblog.com/forum/thermal-imaging/i-make-a-thermal-imager-with-thermal-eye-2620as/

From the video the listed J1 14-pin connector seems to have the following connection diagram:
1 Low Voltage in
2 Low Voltage in
3 Voltage return
4 Video Out
5 Voltage return
6 -
7 Must be connected to ground
8 Toggle video mode
9 -
10 -
11 -
12 Polarity
13 Touch-up / Manual shutter
14 High Voltage

[RO]

I also found a new problem, J1 has 13 pins 

But it's identical to the one from RTF
http://www.rtftechnologies.org/general/thermal-imager-L3-thermal-eye-2500AS-2620AS.html

with similar pins
(SAM_2033a.jpg)

 

[Bill W]

The 2000AS has a 13 pin J1
See http://www.fire-tics.co.uk/project1/index.htm

I have the 13 way pinout there, but no guarantee it is the same as a 2620AS.

Bill

[RO]

Oh perfect! Thanks Bill! that's exactly what I needed.

I assumed I had a 2620AS as it looks identical to Andrews camera from RTF Technologies
linked at:
https://www.eevblog.com/forum/thermal-imaging/why-some-thermal-cameras-can-stand-direct-sun-light-while-others-cant/msg2367696/#msg2367696
http://www.rtftechnologies.org/general/thermal-imager-L3-thermal-eye-2500AS-2620AS.html

But no sticker is present on the optical block, and now I think it might be 2000AS, most details are conform your documentation, 13pin J1 and the DSP chips. (same goes for the RTF unit)