Thermal Camera Teardown - NEC AVIO F30 compact camera aka Mikron Mikroshot

[Fraser]

I recently purchased some Mikron Mikroshot-B thermal imaging cameras.

These camera are made by NEC AVIO and hold the NEC designation of F30 and F30S. The Mikroshot-B is an F30S with the upper temperature limit set at 100 Degrees Celcius but twice the sensitivity of the F30 standard model.

The data sheet for this camera is attached.

The cameras arrived safely and I have had some time to test them before embarking on a teardown. They perform very well and offer 160x120 pixels at 8.5fps and a manual focus lens. The lens is a very unusual looking profile but it allows the camera to focus well at close range using the focus ring, yet act as a focus free lens at longer ranges which avoids having to mess around with the focus ring all the time. Minimum focus is 10cm without any additional lenses needed. I have found the camera to be excellent for PCB thermal profiling.

The camera stores its images on a standard SD card and in radiometric .JPG format. The files may be read by most image JPG viewers but a free program is available from NEC to view the radiometric data, modify images and create reports. The software is named NS9500LT. You need to ask NEC to send a password for the download link.

Ok enough about the camera, on with the teardown.

I will add some messages below this one with the images. the originals are high resolution and 3MB each so I can supply better quality on request. I am supplying only a sample of the images as I did not want to bore you. More available of specific areas of the PCB if wanted.

The camera was relatively easy to dismantle and is well built. It was designed to be a light weight camera that may be carried in a pocket so weight savings have meant that the case is plastic instead of the more common alloys that NEC use. There is an aluminium chassis on to which the PCBs attach.
PCB interconnects are via flexible ribbon cables and nice quality PCB connectors. All a bit small and fiddly but definitely a tech friendly design 

It is worth noting in the following pictures that there is one main PCB that carries most of the interesting chips. The small top PCB does contain a PIC but is mainly minor ancillaries and nothing very interesting. It is also worthy of note that the large (18mm across diagonals) microbolometer is deliberately masked to reduce the illuminated area, using a metal shield in front of it.The same metal shield carries the temperature reference IC.
The microblometer may well be larger than 160 x 120 and so masked to reduce the resolution for the F30, but I have no definite information on this.

The main chips used in the F30 design are listed below (as marked, not identified yet) :


TI DAVINCI $NA-8CAILNW TMS320 DM6441AZWT  .................Media processor  Released by TI in Jan 2007
SPANSION GL512P10FFIR1  labelled F30 ROM U102 V1.1   ......Mirror Bit 512Mbit (64Mb) Flash.
M 9CF12 D9KHQ................................................................ Micron MT47H32M16HR-3 IT -  DDR2 SDRAM 512MBIT 3NS
X 3S100E D35869-0837.....................................................  Xilinx SPARTAN 3E
AD780B............................................................................. Analogue Devices Bandgap Voltage Reference 2.5V/3V
AD5551..............................................................................Analogue Devices 14bit serial DAC
LT2245   N29508.................................................................Linear Technology 14bit 10Msps ADC
8054 ARUZ #0831...............................................................Analogue Devices Rail to Rail amplifiers (Quad) 150Mhz 3dB bandwidth
B5301 U0118  0841.............................................................
PIC16LF819.........................................................................Enhanced Flash Microcontroller
Microchip 4FC I823...............................................................EEPROM    This may be where the F30 identifies itself as a standard or S model ?

No visible ID on the Microbolometer but even if there was I doubt a data sheet would be available. NEC used to use Boeing thermal cores but they also made some in Japan.

Pictures follow in the next few postings........

Enjoy 

Aurora

[Fraser]

Picture Set 1 - Case

[Fraser]

Picture Set 2 - PCB's etc

[Fraser]

Picture Set 3 - PCB close-ups

[mikeselectricstuff]

Quite a lot in there- obviously the sensor isn't as well integrated as the Flir ones with the built-in ADCs.
can't see why a mask would be to limit res - probably to mask spurious modes coming through the lens - e.g. off-axis sunlight

[Fraser]

I mentioned the mask because the hole in it is far smaller than the window of the Microbolometer. If NEC only wanted the middle portion of say a 320x240 microbolometer to be 'in-play' they could mask the area around the centre 160x120 pixels ? All supposition and probably wrong but I have not seen such a mask in use on their 320x240 cameras that I also own.

I should state that the mask sits against the microbolometer as can be seen by the witness marks on the painted shield, and purple ceramic of the microbolometer. It could just be noise reduction or off-axis screening though 

From memory this camera was released in 2007 so it is 7 year old technology. Hence the lack of integration. They have done well to fit the camera into a relatively small format though. These days it could be made much smaller as you have proven with your Lepton based 'smallest thermal camera' project 

Original retail price on the F30 camera was nearly $5000 

Aurora

[Fraser]

The Davinci chip is an interesting beast.

Its got its own Wiki Page and was released by TI in January 2007 which correlates with the F30 release year.

http://en.wikipedia.org/wiki/Texas_Instruments_DaVinci

TI Page

http://www.ti.com/product/tms320dm6441

[mikeselectricstuff]

If NEC only wanted the middle portion of say a 320x240 microbolometer to be 'in-play' they could mask the area around the centre 160x120 pixels ?

Why bother - they could select the required region in software.

[amyk]

Why bother - they could select the required region in software.

...and make it easy for people to "upgrade"?

To be certain, put a scope on the clock signal.

On the other hand, I wouldn't be surprised if the microbolometer itself is capable of 60Hz.

[mikeselectricstuff]

Why bother - they could select the required region in software.

...and make it easy for people to "upgrade"?

To be certain, put a scope on the clock signal.

On the other hand, I wouldn't be surprised if the microbolometer itself is capable of 60Hz.

It would probably be quite hard to make one that wasn't, however as soon as you start integrating electronics as Flir have done, it gets a lot easier.
However as we've seen with the E4, it is still an advantage to run at a higher framerate to reduce signal-to noise & allow use of a smaller lens.
One thing I've not determined for sure is whether the E4 sensor actually outputs different data on each 60fps frame - the Lepton seems to send duplicated frames at around 30fps - however it seems unlikely that the E4 sensor would have any onboard memory to  make this possible. It's concviable there could be some analogue filtering to limit the useable framrate though.
Must get around to hooking up that raw datastream to a monitor sometime....

[marshallh]

[Fraser]

Great pictures Marshall 

I have a new Digital camera with better zoom capability than my previous one but, as you can see, it's not great for close up work with the flash as the lens casts a huge shadow on the subject 

I shall re reverting to my trusty Canon iS82 in future.

Aurora

[bat21]

Hi All,

I'm a newbe here and do not know much about electronics, so first of all please be indulgent And I do hope my inquiry is not too far from the topic...

I've come across this post while searching for some more information on NEC Thermal Shot F30 camera with a straight intention to use it hanging down on a gimbal under my UAV (hexacopter) for different kind of thermal inspections. This one has a video output which is great (compatibility with a standard AV transmitter onboard). But I wonder how to release the shutter having no physical servo arm to push the button up in the sky.

Do you have any ideas or suggestions how to release the shutter not pushing the button? Is there a standard remote steering solution? Any additional circut which could be steerable with a PPM signal from my ground transmitter? 

Kind regards,
Martin