Fluke VT02 thermal camera /visual thermometer teardown

[Fraser]

For thermal camera fans I offer my latest purchase, a VT02 prototype, torn down for your enjoyment and information.

An interesting purchase for me in that I had never seen the original IRISYS visual thermometer in the public domain. I bought a pair    The Fluke VT02 was designed and built by IRISYS in Northampton, UK. It uses one of their in-house fabricated ceramic pyro-electric detector. More on that later.

As some may know, FLUKE recently purchased IRISYS to obtain their knowledge of thermal imaging and other associated specialisms such as detector fabrication.

The units arrived today and I have no history on them so after an initial test (they both work fine) I opened one up for inspection. I already suspected that they might be prototypes and |Mike was of a similar opinion when I mentioned them to him.

As you will see in the pictures that I am providing, the unit is well built and does not look to be an early prototype with many modifications and bodge wires. I did note that the trigger switch has been shifted off of its originally designed PCB pad location so this may be a bit of fine tuning of the design. Also of interest are the laminated labels on the PCB's These ID the camera as 'B16' Beta #16 ? The cameras do not have any serial numbers or consumer type ID labels. I am of the view that these units were never intended for sale to the public and were likely produced for UAT or final design testing.

The units do have evidence of having been in contact with moisture as the memory cell has rust on it ! I would not be at all surprised if these units were recovered from a Skip after FLUKE bought IRISYS (I am told they had a big clear-out)

As can be seen in the pictures of the uncooled pyro-electric detector, there are finger prints all over it ! I am pleased to report that these were easily removed with a microfibre cloth dampened with IPA.

The IR lens is a single element type made of chalcogenide glass and not an expensive Germanium type. Chalcogenide glass lenses are becoming very common in lower priced thermal camera applications. The lens may be adjusted via its large course threaded mount. The visible light camera is a typical 'pill' type as found in mobile phones. The design is clever in that the IR lens and visible light camera are mounted on a single assembly to make alignment easier

A shutter wheel is required when using a pyro-electric detector array as change in the image must be induced to create an image. The shutter wheel in the VT02 is a very simple affair, simpler than that found in BST based thermal cameras. The detector is a staring array and no other 'scanning' is present. Some thought the motor noise that they could hear was a scanning assembly as found in early single element detector thermal cameras such as the Agema Thermovision 880. I can now confirm that this is not the case.

The pyro-electric detector is as yet unidentified but MAY be the 31x31 pixel IRISYS RedEye-6A (it has '6A' on its rear  . The Dewalt DCT416 (15x15 pixel) camera uses the 31x31 pixel Redeye-6A. I have noted that there is a Link with A & B positions on one of its pins. No idea what the link does yet but as the VT04 offers twice the resolution (30x30 vs the VT02 15x15 pixels) I am wondering if the link has any influence on the thermal resolution Knowing my luck, likely not but you never know 

The Redeye-6A may be seen here:

http://www.i-micronews.com/reports/IRISYS-REDEYE-6Pyroelectric-Array-Sensor/19/373/

The report found below also infers that the VT02 contains the Redeye-6A

http://www.i-micronews.com/reports/FLUKE-VT02-Visual-IR-Thermometer/19/418/

This would explain how the VT04 can achieve twice the thermal resolution of the VT02.

As my units are likely prototypes of the series, they MAY be able to operate in the VT04 guise 

So all in all an interesting little pair of units to play with 

[Fraser]

Pictures continued

[Fraser]

Pictures continued

[Fraser]

Some additional information on the unit that I dismantled.....

The plastic case and components are well formed and look to be final versions. The two halves of the case have different moulding dates. One is 01/12 and the other is 06/12. There is the possibility that if these units were being dismantled and rebuilt during testing, that the case halves got mixed up and so caused the date disparity.

Definitely a design from early 2012 though.

[tom66]

I'm guessing with the low processing requirements all the thermal imager compensation is done in the MCU, which is fairly basic/low end (Cortex M3, 120MHz, no hard FPU or DSP capabilities.)

[Fraser]

I was surprised at how low the chip count was in the unit but I believe the Redeye-6A does a lot of the work on its integrated readout electronics. The total number of pixels is so few even win 31x31 mode that, unlike the E4, a simpler approach may be taken to processing the data from the detector. This is not using an unruly micro-bolometer so an NUC shutter is not required. The rotating shutter wheel just creates the required change that the detector needs, effectively resetting all pixels to nominal levels, ready for their next thermal reading after it has passed in front of them.

Such systems have the advantage of more stable and predictable detector technology, but the disadvantage of a mechanical rotating shutter wheel. The micro-bolometer is an unruly beast that is tamed with an NUC shutter and calibration tables, but it avoids the need for the potentially unreliable motor assembly needed for a shutter wheel.

The VT02/VT04 appears to use a very similar principle to the BST based thermal imaging cameras that preceded micro-bolometers in the marketplace. The most significant difference is the low resolution of the VT series detector array.

It should be noted that IRISYS are past masters of cubic interpolation that can achieve far better thermal imager performance than would be expected from relatively few physical detector pixels. 

[Fraser]

The only data I could find quickly on the Redeye-6A is a reverse engineering report sample....nice pictures though 

It may be downloaded here:

http://www.i-micronews.com/upload/Rapports/Yole_IRISYS_REDEYE-6A_Thermal_Sensor_Sample.pdf

Sadly, no pinout 

A sample reverse costing report on the Dewalt DCT416 contains images that suggest to me that it comes from the same factory as the VT02    The main PCB image is blurred but it certainly appears very similar to the prototypes that I have. A significant difference is the use of a lithium coin cell in place of the supercap that is fitted in the prototypes. The shutter assembly is the same. Just the exterior shell differs significantly. Firmware is also different.

http://www.i-micronews.com/upload/Rapports/Yole_Reverse_Costing_DEWALT_DCT416S1_Imaging_Thermometer_Kit_Sample.pdf

The sample reverse costing report for the VT02 has been redacted so heavily so as to be pretty useless for comparison purposes.

http://www.i-micronews.com/upload/Rapports/Yole_FLUKE_VT02_Visual_IR_Thermometer_Sample.pdf

[Fraser]

I mentioned Irisys experience in interpolation to improve image appearance when using a low resolution detector. Please see the attached picture from Irisys for an example of such interpolation of data from a 16x16 pixel detector.

[mikeselectricstuff]

That sensor package  is different from the ones used in the newer people counting products, which I'm fairly sure are still 16x16 like the older ones.
Packaging is more compact, so probably more specifically designed for thermal imaging.

[Fraser]

I am a little surprised at how little information is available on the Redeye-6A. I would have expected IRISYS to be marketing it as widely as possible to other manufacturers for use in innovative products. The redeye is not subject to IATA and is an affordable alternative to micro-bolometers where low resolution is acceptable. I expected at least a data sheet on the IRISYS web site with detailed tech info available after signing an NDA etc. But no sign of such.

These units may be very low resolution but they will be taking the place of my cheap IR thermometer in my tool kit. The ability to see exactly where you are measuring the temperature and the removal of the measured area error that can occur with IR thermometers being under illuminated is a definite advantage. Very neat little design. It is in a different world to the FLIR E4 though and the two should not be compared in terms of imaging...... a case of apples and pears.

[mikeselectricstuff]

I am a little surprised at how little information is available on the Redeye-6A. I would have expected IRISYS to be marketing it as widely as possible to other manufacturers for use in innovative products. The redeye is not subject to IATA

Would that not depend on how fast you clock it (to keep under 9fps), or are pryoelectric arrays not covered by export regs?

[Fraser]

My understanding of the IRISYS Redeye-6A is that it is not only low resolution, but also low speed with a refresh rate of less than 9fps. It is true that a BST will generate much higher frame rates but it looks like IRISYS intended the Redeye-6A to be an ITAR exempt device. I may be totally wrong though as there is so little data available on IRISYS detectors and their capabilities. I do know that they produced 'real' thermal imaging cameras with decent resolution but I have no data on the detector used in them. I see that they are now offering much more advanced cameras on their web site.

Edit: Just found the datasheet for one of the IRISYS thermal cameras and I see they state a micro-bolometer is used and not one of their own pyro-electric detectors. I cannot see a frame rate spec on that camera though. I do seem to recall that IRISYS developed UK manufacturing of their ceramic pyro-electric detectors to get around the USA monopoly on such, including ITAR issues for export. Sadly without some data on the Redeye it is hard to know, or even guess at its true capabilities 

[Fraser]

I found an interesting article commenting upon this new generation of deluxe IR thermometers that provide a visual context to measurements. It is worth a read so I have pasted it below.

Credit for the following goes to the Snell Group.


========================================
WHAT DEFINES A THERMAL IMAGER?

There is significant market confusion between infrared temperature measurement and thermal imagery. This is compounded with consumer-based low pixel IR cameras (such as the Mastercraft Maximum thermal imager) which superimpose a very low resolution IR pixel array on a visual image. These new class of instruments, however can produce relatively accurate multi-point (e.g.200) temperature measurements. Real thermal imagers, on the other hand, should produce detailed patterns of infrared radiation leaving a surface. Thermal imagers may, or may not, include the capability to calculate surface temperature based upon values of emissivity and background input by an operator.

Many military, police, firefighting, industrial and building applications do not require temperature measurement. They rely instead on a trained operator’s ability to detect, categorize, recognize, and identify patterns associated with a specific application, whether it be a person, vehicle, moisture, missing insulation, air infiltration, electrical fault, or machine malfunction. Linked directly with the ability to identify a particular pattern is a minimum number of detector pixels across or over the pattern or object of interest. For repeatability purposes, every thermography application should have minimum detectability requirements defining three basic items: minimum Field of View to be observed, the number of pixels required for pattern recognition, and the minimum thermal sensitivity.

At issue currently is a confusion arising from these low pixel instruments labeled as “thermal imagers” that may be capable of detection and measurement but not capable of pattern recognition no matter how close or how large the thermal anomaly is. These instruments are creating confusion because they can in fact measure multiple spots of temperature which are overlaid on a high resolution visual image. They should be called Visual Multipoint Thermometers but not Thermal Imagers. Watch for our upcoming white paper on this topic which discusses fundamental pattern recognition concepts, and the consequent existing and proposed minimum thermal imaging performance requirements for various commercial applications as established by ASTM, ISO and the National Master Specifications of Canada.

Many thanks to the Snell Group for their great article. If you’re looking for a wide selection of real thermal cameras, you can find them on ShopThermography.com
======================================

I totally agree with this articles comments. The FLUKE VT02 and its kin have a place in temperature monitoring but for users needing a true thermal imaging camera with pattern detail, look elsewhere.

[Fraser]

Some Wiki sourced comment on the IRISYS Pyroelectric detector....

====================
The detector array hybrid as manufactured at Irisys comprises the sensing chip, which is created from a thin (typically 200 ?m thick) pyroelectric ceramic wafer; and an ASIC chip which reads the signal from each array element in a regular scan. The array elements are created by a patterned contact array printed onto the underside of the ceramic chip. The circuit is completed by a continuous contact layer on the upper side of the ceramic, which is transparent to the incoming infrared radiation from the scene.

The performance of the detector array depends critically on the geometry, and the ceramic thermal and electrical properties. In particular, the ceramic electrical conductivity is optimized to provide an appropriate shunt resistance at each element, (the required resistance being far too high to provide as on-chip resistors in the ASIC). Thermal design is optimized depending on the thermal targets to be imaged. The pyroelectric responds only to changing thermal targets. If static scenes are to be imaged, (as in hand-held thermal imager applications, for example), the image is chopped with a rotating blade to provide a time-dependent reference. Time constants limit the chopping to 10 Hz, and typical operation is at 8 Hz.

For very low-cost detectors, such mechanical chopping is not preferred, and in this case, (for people counting and monitoring applications for example), the natural movement of the objects in the scene is relied upon to generate the pyroelectric response. This has the substantial advantage that background clutter is automatically removed from the scene to be analyzed, (in contrast to a conventional video scene), and only the moving (people) targets are observed.



====================

[Fraser]

Chip list for the VT02: Markings are shown and identity given where known.

1. Redeye-6A                  31x31 Pixel Pyro-electric detector array with integral ASIC readout chip.
2, STM32F207                 ARM based MCU - 32bit Cortex M3 CPU - 120MHz
3. AAY 942 RM                Maxim 810ZXR Microprocessor Reset circuit. Threshold = 2.32V
4. LTGF                          Linear Tech LT1761ES5-3.3V LDO Regulator
5. 3534 N43200 LT137    Linear Tech LT3524 Buck-Boost DC-DC converter
6. AAT 944 M4                Maxim 810LXR Microprocessor Reset circuit. Threshold = 4.63V
7. OAVQ                         TI OA365 Op-Amp
8. 9801M 111B2K           Microchip MCP9801-M High accuracy 2 wire temperature sensor
9. 24LC64                      Microchip 64K I2C eeprom


PCB Numbers:
 
PPT1001501
PPT1004001
PPT1006103
PPX1006102

'PP' = pre-production ?

Also worthy of note is the fact that the plastic casing has the additional cut-out in the left side that is only used by the VT04 and is for the charging socket. So the VT04 was part of the original design phase and likely not just a development from the VT02 ? I have noted that the field of view on the VT04 (28x28 Deg) is wider than that of the VT02 (20x20 Deg). Could this suggest that the VT02 is only using the centre 15x15 array of pixels on the Redeye-6A ? The visual camera also has a really non standard pixel format which suggests that only the centre of that imaging chip is being used as well.

I am starting to think that the VT02 is just a hobbled VT04 with software determining the resolution and facilities rather than hardware. Where have we seen that before I wonder 

The 24LC64 eeprom will certainly be worth investigating to see what is loaded within its memory.

The only firmware updates that I can find are suitable only for the later version of the VT04 and so will be of little use on this early camera unit.

[Fraser]

I have just dismantled the second unit to compare it with the first.

The second unit does differ to the first in minor details and is a +1 revision of PCB's. It has one bodge wire fitted.

The shutter wheel has been made from a thinner piece of metal to reduce weight and likely the issues of an unbalanced load. This unit is noticeably quieter than the other when running. The Chip set is unchanged. Minor PCB layout tweaks but nothing of great interest. The pyro-electric detector is a Redeye-6A and, just like the previous one, it has a dirty great thumb print across it  No idea why, maybe no one cared as they were for testing only.

The second unit has laminated labels applied to the PCB's with the ID 'C6' on them, so there goes my theory about 'B' indicating Beta. The trigger switch position has been corrected on the new revision of PCB. 

I am of the opinion that these two units are pre-production pieces, used for final testing and tweaking of the VT02 design. They both work very well, so the changes may be minor in nature and effect. The noise reduction by the use of a thinner, lighter shutter wheel is noticeable. I was surprised to see that neither unit has any form of counter weight on the shutter wheel to help balance it properly.

I do not know how these units came to be in the public domain but suspect that IRISYS binned them and they were salvaged. Great condition though.

[Fraser]

Having just revisited my IRISYS datasheet archives I can see how the VT02 came to exist. IRISYS originally produced the IRI-1001 low resolution thermal camera. It used a 15x15 pixel pyro-electric detector. In front of the detector a chopper wheel 'modulates' the incoming image from the lens. The detector output was processed and passed to an iPaq PDA for display. Does all this sound familiar ? It should. The VT02 is basically the same design with the iPaq PDA replaced by a dedicated ARM MCU, the addition of a visible light camera and an integrated LCD display. IRISYS took their original design and condensed it into a smaller package. The natural progression of a novel product.

I attach the IRI-1001 and later IRI-1011 datasheets for interest.

[Fraser]

I have received an email that appears to explain the existence of these units but not how they came into the public domain.

The red coloured units are IRISYS 'in-house' testing and evaluation samples that were used to gain approval for production. FLUKE bought IRISYS and 'Borged' the units into their product range with the FLUKE Yellow and Grey livery.

Fluke thought that the VT02 was a novel bridge between single pixel non-contact thermometers and high pixel count thermal cameras. What FLUKE had not expected was the appearance of the FLIR E4 at its very competitive price. With superior imaging (even at 80x60) the larger screen and FLIR branding, the FLUKE VT02 was left in the E4's shadow.

It IS a novel design that does what it says on the tin. It is argued that it is not a true thermal camera. This is technically incorrect as it is a camera by virtue of its physical construction and capability, just a very low resolution version 

My units will be sat in my tool bag for those occasions when I do not have the E4 with me or do not need the E4's capabilities. They replace a cheap Chinese IR thermometer and an expensive Raytek IR thermometer. I paid GBP125 for mine, which IMHO is about the max I would pay for such a unit.

Aurora

[ratmole]

So, after a 1 year of setting aside my badly corroded Fluke VT02 Thermal Camera, today i decided to fix it.
Thankfully i had a firmware dump taken before the corrosion!

First i cleaned  and repaired all the broken traces.
The camera was powering now but with a white screen. I had missed a broken trace under one of the display filters near the lcd connector! 

Camera powers on now, but hangs on the FW logo!
Had a stm32f207 laying around so i replaced the original stm32f207 and flashed my backup FW...

How? Here are the instructions for windows:

Code: [Select]

1.You will need a stlink v2 programmer/debugger and the SW from here:
    [url]https://www.st.com/en/development-tools/stsw-link004.html[/url]

2. Connect the SWCLK pin from the stlink to the  PA14 pin on the stm32f,
3. Connect the SWDIO pin from the stlink to the  PA13  on the stm32f
4. Connect the 3.3V from the stlink to the VDD pin 100  on the stm32f
5. Don't forget the GND!

6. Install the STM32 ST-LINK utility
7. Run the setup(mine was installed on  C:\Program Files (x86)\STMicroelectronics\STM32 ST-LINK Utility\ST-LINK Utility\STM32 ST-LINK Utility.exe)
8. Click connect and then load the attached file. Click on the program and veriy icon.
9. Cross your fingers and wait....
Linux instructions:
1. Install st-flash following the instructions frme here: https://github.com/stlink-org/stlink/blob/develop/doc/compiling.md
2. Then on a terminal

Code: [Select]

st-flash write fluke_vt02_1.9.000_FW.bin 0x080000003. Cross your fingers and wait....

The Camera boots, but hangs on the FW logo 

Took me a while, but after reviewing the pictures Fraser uploaded (thanks Fraser!!!), i spotted that i had connected the power to the wrong connector!!! (always reassemble your projects before setting them aside...) 
It was connected to the calibration motor/sensor/rotating shutter wheel connector!

Connected the power to the right connector and pressed the power key...
The calibration/motor sensor started to turn a bit slower than before!, finally progress!

FW logo appears, 10 seconds later the camera was working again!!

Note, i have reinstalled the original stm32f207 on the camera, re-flashed the FW and the camera works with the original chip!
Note2, the camera hangs on the FW screen if the calibration/sensor/rotating shutter wheel  motor is not connected. Also the calibration/sensor/rotating shutter wheel has to be on his original place, between the 2 optical sensors, otherwise you get the FW logo hang.

[Fraser]

Excellent news 

No need for my spares donor now then 

Fraser

[beanflying]

Cleanup on Aisle 6   

Thanks for the teardown Fraser made finding the two stickered over screws easy 

Might be a bargain after I cleanup the mess and likely replace the wires from the battery holder and repair the pcb and ..... 

The easy solution might be so simply cut away the tray and fit a two cell (no dc to dc convertor) PP3 battery in it providing it will cope with the higher voltage. Anyone have a circuit diagram for this or the VT04 LiPo versions let me know.