Excellent UV camera outside my price range

[dabateman]

Seems like experimental sensors are hitting the market faster than I ever would have thought.

Rising camera, a Chinese microscopy and Astrophotography camera manufacturer got back to my questions about a different camera, with this suggestion.

 

https://www.ebay.com/itm/high-sensitivity-USB3-0-4-2mp-1-2-Monochrome-Digital-UV-ultraviolet-camera/124065525412?_trkparms=aid%3D111001%26algo%3DREC.SEED%26ao%3D1%26asc%3D20160727114228%26meid%3D9b108a431eb04309ada18661c4122853%26pid%3D100290%26rk%3D1%26rkt%3D4%26sd%3D124065525412%26itm%3D124065525412%26pmt%3D0%26noa%3D1%26pg%3D2060778&_trksid=p2060778.c100290.m3507

 

Only $5800, but uses the GSENSE2020BSI sensor I saw in the senor blog announced in last December 2019.

 

This is the sensor specifications:

https://gpixel.com/products/area-scan-en/gsense2020bsi/

 

Thats right 60% QE at 250nm. Its only 4Mega pixels though.

[Andy Perrin]

4 Mpixels is much better than my TriWave's 0.3 Mpixels...Honestly, though, with the fluorescent imager, I feel like I have a reasonable way to take UVB/UVC photos already, if I ever get around to working out a safe way to deal with the light source.

[dabateman]

Andy,

Two interesting things though. Our imager QE maximum at 254nm is 25%, this is 60%. That is more than a stop more sensitive using a silicon chip. This camera does have a quartz 1mm thick coverglass without microlenses or color filter array. So its obviously monochrome.

 

The second thing is the timming. This was announced in December. And QHY and Rising Camera seem to be selling cameras. The finger print sample images in the ebay announcement look quite good. But those SWIR cameras I posted about were also November/December. I will have to see if Rising or QHY have a SWIR camera coming.

 

[dabateman]

This is the QE graph from the sensor manufacturer. Just over 3 stops more UV light than my full spectrum converted Em1mk1 in the UVB range. Then that amazing UVC sensitivity spike.

[colinbm]

Dave, so this graph should be the same as Bayer CFA with the cover glass, microlenses & the CFA removed....ie any good mono camera ?

[JMC]

Colin, not quite. The one David has posted about has a BSI sensor. Not all sensors will behave like this when mono converted. My two for instance (one Canon, one Nikon) have a little bit of sensitivity at 300nm but essentially nothing below that, and they have conventional sensors. I do however have a d850 on the way which has been mono converted, and with a fused silica window. I hope that will give sensitivity down below 300nm.

[colinbm]

Thanks Jonathan, enjoy you world at 300nm...:-)

[dabateman]

Dave, so this graph should be the same as Bayer CFA with the cover glass, microlenses & the CFA removed....ie any good mono camera ?

 

This sensor was designed from the ground up to be sensitive to UV. I never thought it would be commercially available. Like most of the dreamy sensors I read about.

 

No idea if any other sensor would have this kind of response. Its also crazy thin, which might help for the UV detection.

[colinbm]

Thanks Dave, it is pretty impressive. I guess the price keeps the crazies away......;-)

[nfoto]

A camera that could be useful combined with a UV-Nikkor. It is tempting ... must try to resist ...

[dabateman]

Yes an odd camera. To add perspective, its said to be 1.2" sensor. But thats not really the whole story.

Its 13.3mm x 13.3mm square. So actually larger than a micro four thirds cameras. If you owned an Olympus camera and only photographed in 1:1 square, this sensor is actually larger. This leads to 1.8x crop vertical and 2.7x crop horizontal, from a 135 format sensor.

 

Its only 2048x2048 resolution, so 4Mega pixels. But that allows for what looks like low noise and great UVC signal. The wells are 6.4um, so larger than current cameras, but smaller than the 7.3um pixels of my Nikon Df.

 

The QE is amazing, only bested by the in research phase Sony 3 layer (organic green top, blue, red) sensor. But that sensor would still only have just 1 stop more UVB response and less UVC response. Advantage of the Sony is global shutter and scaleable to 24Mpixels. But still just at research stage.

 

So this really would be the best UV camera for sensitivity you can buy.

 

But what do you get for your $5800?

A sensor with T-mount, that is at a C-mount back focus distance, so a T to C adapter is included and software to run it. So you would need a Windows 10 tablet, with USB3 port to power and control it. It does have 8xM5 screw threads on the sides. I am not up to speed on my metric since living in the US for 12 years. But should be able to screw an Acra plate to the side.

A T-mount to Nikon mount adapter is available, so you can mount your Nikon 105mm quartz lens.

But no screen, no IBIS, no EVF, No battery.

[JMC]

Interesting ones here too (one of the graphs looks the same as the one you shared David); https://www.photometrics.com/wp-content/uploads/2019/10/Prime-95B-Datasheet-June-2019.pdf

[Andy Perrin]

This is also pretty neat:

https://www.horiba.c...hd-system-1839/

 

The SceneScope RUVIS UHD uses intensified UV reflectance instead of fluorescence as in Forensic Light Sources.

[dabateman]

Jonathan,

I missed that one. Maybe because of its very low resolution. Interesting how it says the design of the pixels allows for UV capture without microlenses. So either marketing BS or these chips are truly designed differently.

 

Andy,

That 16Mpixels one is getting there. But no information on sensitivity or how it works. Also might be an order of magnitude more expensive based on its target audience.

 

Imaging again today with the imager is a pain, focusing on a focusing screen is not ideal and the blur is hard to fix. Doesn't have a crisp 3D image as your just photographing off a flat screen.

But the QE graph I found for our imagers seems real, the 500s curve. Using a Halogen lamp I was able to get images with #47, #58 and #29 filters. So can get true RGB with my UV filters if I want. Also IR can be detected.

[JMC]

This is the QE graph from the sensor manufacturer. Just over 3 stops more UV light than my full spectrum converted Em1mk1 in the UVB range. Then that amazing UVC sensitivity spike.

David, what's your thoughts on that spike at around 250nm? Is it due to coatings used on the sensor to optimise UV transmission, or do you think it's something fundamental to the BSI sensor design?

[Andy Perrin]

Jonathan, how could a coating add light back at 250nm? Only something like a upconverting phosphor would do that, but they aren’t strong enough to have such a big effect. And if it’s not a phosphor, then it can only absorb light, so it would have to be decreasing the QE between 400 and 250nm. Which would be a weird thing to do.

[JMC]

Not sure Andy. Initial thoughts was anti-reflective coating optimised for UV, but could it make that big a difference. Just looks very sharp to me, so not sure what causes it.

[dabateman]

Jonathan that is a good question to question that curve.

Thinking about it, I now don't think they really are reporting quantum efficiency. As that shouldn't happen.

The best explanation I could think of would be intensity read out from selected light source. That looks like the output from a mercury lamp.

 

So instead of reporting true QE, where yoi know that there are fewer photons at 280nm. But that you are still capturing 60% of them, To me it looks like intensity capture, dropping after 260nm, due to light source effect.

Looking at the y-axis it says QExFF. I wonder what FF is?

 

So the real QE probably isn't as exciting as I thought. Good catch.

[Andy Perrin]

FF is probably Fill Factor.

https://en.m.wikipedia.org/wiki/Fill_factor_(image_sensor)

[dabateman]

This tis the best explanation I have found for that curve and FF stands for fill factor of the wells:

 

"Dave Gilblom

April 27, 2015 at 6:02 PM

Below about 277 nm, the quantum yield for detected photons in silicon exceeds 1 due to ionization. In the range shown, it will be nominally 2. This is not strictly an increase in quantum efficiency because while the additional electrons generated from ionization increase the signal, they do not improve the shot noise. In fact, the statistical variance in the quantum yield will increase the noise where ever the yield exceeds 1. See "Fano factor".

 

There is nothing strange about the curve but keep in mind that it documents two phenomena."

 

Later:

"The minimum absorption depth for silicon is around 254 nm so as this wavelength is approached from above the QE drops because of recombination near the surface as the e-h pairs are created at shallower and shallower levels. Before 254 nm is reached, ionization e-h pairs begin to be generated so the yield begins to rise. At shorter wavelengths still, the e-h pairs from the bandgap become more collectable. Thus the peak.

 

In UV photodiodes, as in backside illuminated CCDs, the surface is prepared in a way that minimizes recombination so the QE curve below 277 nm is flatter and the yield simply rises as the ionization pairs start to be produced."

 

Excerpt taken from first prototype article here:

http://image-sensors-world.blogspot.com/2015/04/towerjazz-and-gpixel-announce-bsi-cmos.html?m=1

 

Looks like TowerJazz, the old Panasonic sensor company actually makes the sensor.

[dabateman]

So the lay person take home is that the detected photons do double at 254nm as reported. But the signal to noise does not. So you will see a brighter image, but will be similar to boosting the ISO at that local wavelength.

 

[Andy Perrin]

Yeah. So we can probably calm down about this one. Although it does look like a nice sensor just generally.

[dabateman]

The first version, the Gense400 BSI sensor is available in a commercial camera. The QHY42. Its just $15000, sadly 17500 pounds.

But I do see some Astro sites have discount on it to $9000, maybe do to this future release of this newer sensor. Or a worse sensor, class 2.

 

No way, I still don't think I would calm down from it. If I could get a used one or discount one in a couple years, I would jump to it. The sensitivity is still excellent and capture time would be fast.

I might want one that is cooled though as Gpixel has optimized the H sensor, which needs cooling for highest resolution.

[dabateman]

Actually this sensor is the real one to go crazy for:

https://gpixel.com/products/area-scan-en/gsense6060bsi/

 

Similar UV sensitivity, 37Mpixels, 61.4x61.4 mm.

Its a true 6x6, just like my Pentacon six camera.

That thing will probably cost more than a house.

But it does allow for movie capture, at 26fps with 12bit!

[Andy Perrin]

Haha, that is like something that would go on a space telescope.