Mµ Thermal Imager - real or fake?

[l3iggs]

Please forgive me if this is a repost, but at first glance, it seems likely to me that they're using this sensor:
http://www.ulis-ir.com/index.php?infrared-detector=25--m-160x120hv

The specs match up very nicely.

I can't comment too much about the cost of this thing. Can anyone else?

[rem256]

Wow, it's unfortunate to see some of the frustrated supporters of this project.

In March, this article was published about the project: http://reelchicago.com/article/mcgrath-s-innovative-device-goes-twitter-wide-monday130304

If a supporter of this IndieGoGo project connects with the editor of this publication and they choose to run an article about the current status of the project, I suspect the MuOptics team will become more responsive and accountable.  Hopefully then, the supporters will feel a better sense of involvement and overall satisfaction with the project. 

Good luck to all current supporters!

[EEVblog]

Latest update:

Hello All,
Just an update on what we've been working on this week.
As you all know, one of the ways that we are using to reduce the cost of this IR camera is by offloading as much as is possible to the host device (smartphone, tablet, laptop, ect...). This week we've been working on the part of the software that takes the raw thermal data from the sensor and converts it into video. This is anything but simple software, but we are happy to say that so far we have been able to keep the overhead very low. Again our goal with software is to open-source and provide an SDK for nearly everything.
I'll post more this weekend.

So they are only just getting around to working on the bit of the software that produces a video output?
What about the previous updates:

A few more tweaks to the portion of the MCU firmware that deals with the ADC and the sensor itself and the firmware should be good to go.

or from 5 months ago:

Abe is in the other room right now tweaking bias voltages and the like to get the best picture possible in the field.

 

[firewalker]

Raw data of the sensor. Digital or analog data?

Alexander.

[AndyC_772]

Depends how you'd classify 'no data at all'. I guess you could call it either a stream of logic '0's, or 0V with a high impedance 

[EEVblog]

Raw data of the sensor. Digital or analog data?

They have said they are using an ADC to capture the sensor data, which was of course practically finished months ago according to them.

[firewalker]

Yes. I know.

So they shouldn't send raw sensor data to the phone.

Alexander.

[EEVblog]

So they shouldn't send raw sensor data to the phone.

According to them they do:

The data stream going from the camera to the host device is entirely raw data. That data is processed into video in the Mu App. By doing all of the processing in the app, we can push updates related to video without requiring hardware firmware updates.

and

With an image resolution at 160×120, that gives us 19,200 pixels in each frame. At 30 fps, that gives us 576000 data points each second. For the best thermal resolution, we are sampling at 12bits. Giving us about 6.9Mbps of output.

[tom66]

I don't think a SAM3U would have enough power to compress images beyond RLE type, which is no good for noisy sensors like thermal imagers. An STM32F4 could probably kick out ~10fps of 160x120 JPEG, but there's little point in doing so if you've got enough bandwidth.

[zapta]

Fluke has a professionally looking thermal imager for $750 (http://amzn.com/B00APPPL2W). It is made of higher resolution visible spectrum image super imposed with a low resolution IR image. This is a significant price reduction in this market and so far the reviews are very good. Possibly Mu uses similar approach, with some corner cutting and using existing display and computation resources of a smartphone (e.g. via a BT link). Just speculating but it does not sounds to me too good to be true.

[EEVblog]

Fluke has a professionally looking thermal imager for $750 (http://amzn.com/B00APPPL2W). It is made of higher resolution visible spectrum image super imposed with a low resolution IR image. This is a significant price reduction in this market and so far the reviews are very good.

 I don't know why, it's hopeless compared to a Flir i3 which is only a hundred or two more. I've compared them, no contest.

[Fraser]

I don't think Dave owns them, so a tear down would be difficult. Fluke told him that they didn't want a teardown of their thermal cameras, and I suspect FLIR will be even more resistant to such an idea. When you own such an expensive bit of FLIR kit, you are often less enthusiastic about taking it apart ....the micro-bolometer engine is definitely not intended to be dismantled to component level. I have been repairing several industrial TICs recently and thankfully have not had to delve into the micro-bolometer module beyond simple voltage and waveform checks. Damage t this module = written off TIC. Mike did an excellent repair video on a FireFlir TIC that may interest you.



There actually isn't a great deal to see inside these units. There was a commercial costing investigation into an older FLIR i7 version detailed in another thread. The investigation showed the internal parts including the micro-bolometer sensor. As I have found myself when repairing TICs, it basically looks like (but isn't) a CCD chip fitted inside a vacuum chamber that has a germanium crystal window facing the Germanium optical assembly. A mechanical shutter assembly moves a shutter in front of the sensors window for regular calibration and pixel levelling. The output of the micro-bolometer sensor passes to an image processing chip that in turn passes the processed signal data to the video ADC which feeds the SBC microprocessor. An image is then generated on the LCD by the microprocessor image output stages.

The larger industrial grade TICs are far more complex and interesting in terms of build, but the simpler consumer grade units are really just a boring high density SBC and the micro-bolometer engine. They aren't that complex, just expensive on the engine and optics front.

I will see if I can find the link to the FLIR i7 strip down for you.

UPDATE: FLIR tear down added.

The Fluke will use similar technology but with a lesser micro-bolometer in a smaller imaging engine.

[EEVblog]

I don't think Dave owns them, so a tear down would be difficult. Fluke told him that they didn't want a teardown of their thermal cameras

It wasn't Fluke as such, it was Uncle Sam, or Fluke's legal departments fear of Uncle Sam. The camera was export controlled technology, because, you know, the world is full of terrorists who search the internet for technology videos so they can make their own, because it's just so hard to simply pay the money and buy one... 

[SeanB]

And if Columbian marching powder can come in in ton lots then the same way works for taking stuff out...............

[EEVblog]

So it's now been 6 months since the campaign ended and not even a real thermal image yet, and no update in 17 days.
Is it time to confidently right this project off as simply never going to happen?

[grenert]

Three of the four members of the Mu "development team" make up three of the four members of the development team of a forgettable me-too restaurant app called Sqipp (scroll down the screen):
http://www.sqipptheline.com/

What's to stop them from taking the Mu money and putting it toward this other venture?  Is there any accountability regarding where the money goes?
I assume probably none.  It amazes me that anyone would throw their money away toward stuff like this.  A sucker born every minute, I suppose...

[grenert]

Fluke has a professionally looking thermal imager for $750 (http://amzn.com/B00APPPL2W). It is made of higher resolution visible spectrum image super imposed with a low resolution IR image. This is a significant price reduction in this market and so far the reviews are very good.

 I don't know why, it's hopeless compared to a Flir i3 which is only a hundred or two more. I've compared them, no contest.

Flir has just released the E4, which has 80x60 resolution (versus 60x60 in the now-obsoleted I3):
http://www.tequipment.net/FLIRE4.html?idevd=f1d73e45a4554c07b60e7fe457a66e0b&idevm=5796ba9b89e94fbca82b49e0fe97b42c&idevmid=363816
Under a thousand bucks!   

[EEVblog]

Flir has just released the E4, which has 80x60 resolution (versus 60x60 in the now-obsoleted I3):
http://www.tequipment.net/FLIRE4.html?idevd=f1d73e45a4554c07b60e7fe457a66e0b&idevm=5796ba9b89e94fbca82b49e0fe97b42c&idevmid=363816
Under a thousand bucks!   

Nice. I wonder what the focus distance is?
I believe a problem with the i series was that you couldn't focus close to see temp spread across a chip for example.
EDIT: Ah, focus is 0.5m minimum. That would almost certainly mean no good for close up PCB work.

[Corporate666]

Fluke has a professionally looking thermal imager for $750 (http://amzn.com/B00APPPL2W). It is made of higher resolution visible spectrum image super imposed with a low resolution IR image. This is a significant price reduction in this market and so far the reviews are very good.

 I don't know why, it's hopeless compared to a Flir i3 which is only a hundred or two more. I've compared them, no contest.

Flir has just released the E4, which has 80x60 resolution (versus 60x60 in the now-obsoleted I3):
http://www.tequipment.net/FLIRE4.html?idevd=f1d73e45a4554c07b60e7fe457a66e0b&idevm=5796ba9b89e94fbca82b49e0fe97b42c&idevmid=363816
Under a thousand bucks!   

That illustrates another problem with this crowdfunding stuff.

Technology moves so fast, and crowdfunded campaigns almost universally move slowly (delivering very late, if at all), that by the time they are completed, they are already obsolete.

Look at Lifx, Philips Hue has already been on the market for almost a year and is now sold through the Apple store.  They lost any earlybird advantage they would have had. 

Look at the Elevation Dock.  By the time they shipped them all, the phone had been replaced with a newer version.

Many many more too - and based on this post, looks like Mu may be in the same boat.  I doubt they ever deliver anything, but if they do, it will be behind the current latest/greatest before it ever ships.

[mikeselectricstuff]

I am interested to see FLIR's take on visual and Thermal image merging. There is normally an image registration error but FLIR appear to have sorted that issue.

I suspect its only solved over certain distance ranges.

The down side of the stated new E series camera is the low resolution of the base unit. I surmise that the images in the brochures are from the higher resolution models and not the base unit.

Would be very interesting if all the E series used the same sensor, like the I series apparently did. 
Also may be some scope for tweaking lens distance for close-up  use.

[tom66]

I'm sure Flir could market a cheap camera with just a CVBS or HDMI output and the hobbyist market would lap it up. No LCD, case or anything, just a dev board breakout for one of their sensors with an optional digitisation chipset. $200 list price. I doubt the sensor costs that much -- you're paying for R&D for most of it.

[MFX]

BTW any UK members see the British wildlife program on BBC1 tonight? Looks like they borrowed a British army night surveillance vehicle, the resolution of it's thermal camera was pretty incredible. Wondering what cost more the vehicle or the camera.

[Fraser]

It was a mobile platform offering from Selex Galileo

http://www.armedforces.co.uk/companies/raq4315b2a67e1d3

Suppliers to the worlds Western friendly Military. Good company, excellent products 

Selex TIC Data sheets attached. Nothing very secret about their specs.

[mikeselectricstuff]

I'm sure Flir could market a cheap camera with just a CVBS or HDMI output and the hobbyist market would lap it up. No LCD, case or anything, just a dev board breakout for one of their sensors with an optional digitisation chipset. $200 list price. I doubt the sensor costs that much -- you're paying for R&D for most of it.

They maybe could, but why would they? There wouldn't be any profit in  it for them. The sensor quite probably does cost that - vacuum packaging, germanium window & optics, calibration etc.

[AndyC_772]

Fluke have a new product out, the VT-04 not-quite-a-thermal-camera:

http://www.farnell.com/datasheets/1751644.pdf

"Four times sharper" than the VT-02, apparently, though still no actual resolution figures quoted.