Jump to content
UltravioletPhotography

Multispectral monochrome Nikon d850 - testing and shakedown


JMC

Recommended Posts

I recently got a monochrome converted Nikon d850 with a fused silica window, and this thread will include some of my observations on my testing of it.

 

The reason behind getting this was because of the sensor. As a BSI sensor I thought it may be very useful for UVB imaging for my research. So I pestered Dan at MaxMax for a while to see if he could make me one with a fused silica window, so that it would be useful at 300nm and below. I'd got into my mind from looking at BSI sensor data that they should be useful at 300nm and below, hence my interest. However, as is often the case, it's not as simple and straightforward as that....

 

I'll include a summary of the work here, and then add in more details on the tests later in the rest of the thread.

 

Summary

 

The monochrome converted d850 seems to be more sensitive than the monochrome d800 and Eos 5DSR conversions I have.

 

The monochrome converted d850 is offering more sensitivity at 280nm than the other monochrome converted cameras, although a relatively small improvement (about 1/2 stop).

 

Under relatively extreme circumstances (high ISO), the mono d850 shows evidence of IR fogging from the internal shutter monitoring LED. The degree of fogging increases with increasing ISO. The fogging seems to occur during the first 1/30s of exposure, and does not increase for longer exposure times.

Link to comment

A quick example image with the mono d850, using the Rayfact 105mm lens, and a Baader U filter (ISO200, f11, 1.3s exposure, sunlight at about 10AM). Image taken as RAW and JPEG in the camera. I used mirror lockup and a 3s exposure delay.

 

The images were taken from the JPEG. For the first image, it was just reduced in resolution for sharing here. For the second photo below, the centre of the original image was cropped and shows the original file resolution.

 

The picture is of the end cut oak log.

 

Full image, but reduced in resolution

post-148-0-50030800-1581945204.jpg

 

Cropped portion of original resolution image

post-148-0-41817100-1581945184.jpg

Link to comment

Sensor sensitivity for monochrome d850

 

Measurement of sensor sensitivity between 280nm and 800nm using the method I developed (discussed more here - https://www.ultravioletphotography.com/content/index.php/topic/2580-build-thread-at-home-measurement-of-camera-uv-spectral-response/page__fromsearch__1) was carried out for the mono converted d850, and compared with a monochrome d800 and monochrome Eos 5DSR. The comparison graph is shown below.

 

post-148-0-60509800-1581945842.jpg

 

The converted d850 looks to be more sensitive than the mono d800 and mono Eos 5DSR for a given wavelength.

 

Note, the converted d850 has a fused silica window, while the d800 and 5DSR have WG280 windows, effectively limited their short wavelength response, irrespective of whether or not the sensor is sensitive.

 

Comparison photos will be done with different bandpass filters and the three cameras when I get chance.

Link to comment

IR fog

 

I'm calling this IR fog, as that is what I am assuming it is (IR from the shutter monitor LED). One o the first tests I do with a modified camera is images at different ISOs with the body cap on to look for fogging.

 

Results of a few tests here. Taken with the body cap on, and the eyepiece shutter closed, and in a darkened (but not pitch black room). Images taken as RAW and JPEG. JPEGs used here and shrunk for the table.

 

First test, 1s exposure, different ISOs

post-148-0-15175700-1581955565.jpg

 

Second test, 1-30s exposure, different ISOs

post-148-0-09744700-1581955553.jpg

 

Third test, ISO H1.0, 1/30s down to 1/1000s

post-148-0-86194700-1581955542.jpg

 

At high ISOs I'm getting fogging as a band at the right hand side of the image, which gets worse with increasing ISO. Increasing the exposure from 1s did not make the problem worse. As the shutter speed increases, the degree of fogging remains the same down to about 1/30s, and then decreases as the the shutter speed gets faster.

 

I'm assuming this is IR shutter monitor related. However does it makes sense that it occurs during the first 1/30s after the shutter opens? Originally I assumed this was happening at the beginning of the shutter opening, but is that the case? Could it be occurring at the beginning and the end of the shutter being opened? I know nothing about the LED shutter monitors other than they are there and can be a problem with converted cameras.

 

Don't get me wrong, at normal ISOs that most people will be using, this isn't an issue, however if you're like me (no comments on that please) then at very high ISOs it does become a problem.

Link to comment

Congratulations of a worthy addition to your deep-UV tool box !!

 

As to the "IR fogging" issue, I believe there is a similar problem with the full-spectrum converted Z6. Just to add to the already long list of troubles with that model. The fogging is on the right-hand side and becomes noticeable from ISO 3200 and upwards.

 

Like your observations, all evidence points towards an initial problem at long shutter speeds. It's time for me to take the camera to the Nikon repair facility and learn if they can make sense of the irregularities.

 

I do wish we UV photographers could step out of the current role as alpha testers :sad:

Link to comment

That result is very surprising. You see 1 stop advantage as I would expect, but only to about 350nm. Then the loss of silicon sensitivity seems to kick in. Looks like all sensors loose out at 280nm. That might be the limit for our non uv optimized sensors.

 

The fog you see is equally odd. At first it looked like amp glow to me. But that doesn't behave that way. You maybe correct, and it is IR shutter leakage. At least its minor, maybe easy to subtract it out using these references. Also being limited to 6400 ISO, is not much of a limit and still should work.

 

You have an amazing camera there to push the resolution limits of our lenses.

 

With all the different ones you have it would be interesting to know whom the sharpest is at various wavelengths. Or maybe I just hope that the UAT beats out the Nikon 105mm, because I own it. I know the KSS 60mm would fail quickly. But does look good through low resolution film.

 

I still need to test it out on a Nikon for you. Hopefully tomorrow.

I did see direct c-mount to nikon F-mount adapters on ebay. Why I don't know. But the macro capabilities of the kss help for using it way beyond its intended purpose.

Link to comment

As a fellow chemist I can ask silly questions. Compared to your third test, which way does the shutter run, left to right or the other?

 

This looks a bit like the problems Mark had with his D750 at high ISO which I think I have managed to also reproduce at high ISO but which I think can be avoided by using shutter lock-up and delaying shutter release for a few seconds. If that fails you can always go back to the Witches in Macbeth :rolleyes:

 

Dave

Link to comment

I think this illustrates a shutter monitor problem with the D850 UV/IR in general maybe, not just the monochrome version. I don't know?

I know Bob's MaxMax monochrome has a BG38 installed in it to reduce the IR because it causes a lack of sharpness. Perhaps the BG38 removes enough of the shutter monitor cloud to render it unnoticeable?

Of course, you can't use BG38 if you want to make the sensitivity below 300nm.

This is the first D850 full spectrum or monochrome I have seen so far. I think Bob's is a D800, anyone know how the shutter monitor problem might differ between the D800 and the D850?

https://maxmax.com/maincamerapage/monochrome-cameras/nikon-monochrome

Link to comment
Can the IR shutter monitor LED be replaced by (mad idea) a SWIR LED and the photodiode with a gallium-based one? Is this a problem that all cameras have?
Link to comment

Thanks all, and I'm sure there will be plenty of questions as I look more at it.

 

Stefano, Yes I'll be looking at the 308nm filter at some point, but I see that as more of a filter test than specific to this camera, so will probably post that elsewhere. I also want to do that with sunlight, but will need to wait a while - 308nm is thin on the ground in the UK at the moment. In theory yes the LEDs and photodiodes can be changed, and it has been reported before. Not sure whether it can be done or not on the d850 though.

 

David, it surprised me too. I plan on doing some images with band pass filters and different light sources to check these results with a second approach. Lens testing would be interesting, but I need a more solid tripod (or camera mounting method) for that. My Benbo has too much flex with it.

 

Birna, yes we do end being tester for these things. For most people the issues I have with the fogging wouldn't be an issue. Because some of my work is with cross polarised UV, I need every bit of sensitivity I can get at times.

 

DaveO, not sure about the orientation. I have tried various combination of mirror lock up, and shutter delays, and so far nothing has an effect on it.

 

Steve, yes, this would impact a multispectral d850 (with the Bayer filter intact). Diglloyd has a monochrome d850 (https://diglloyd.com...D850+monochrome). I doubt he's using it at very high ISOs though, as his work looks to be more about ultra high resolution imaging. Unfortunately a BG38 would kill the short wavelength UV usefulness.

Link to comment

Jonathan,

Next test would be to compare the Sony A7m3 with the Nikon. Both being BSI, I would be curious as to the gain if any that the Nikon has at 280, 300nm for being monochrome.

Does the Sony CFA still kill 2 stops of signal at the lower wavelengths?

Link to comment

Can the IR shutter monitor LED be replaced by (mad idea) a SWIR LED and the photodiode with a gallium-based one? Is this a problem that all cameras have?

Not YOUR mad idea, I'll note. Nick Spiker did that on one of his. And no, it's not a problem all cameras have, only certain recent models.

Link to comment

Jonathan I tested the KSS 60mmMacro on my DF using a Nikon to M42 thin adapter and a C-mount to M42 adapter screwed in.

With lens at F4, the working distance from front of lens is 7.5 inches to 3 inches.

So will work, but will be for mostly macro.

 

On to your fog test. If you boost the 6400 ISO shot 2 stops in software does it look like the fogged 25600 shot?

Similarly does the 3200 ISO shot boosted 3 EV in software look like the 25600 shot?

I ask as it might not. You maybe able to get your higher ISOs by under exposing 3 stops and using software you get you back up.

Link to comment

David, I'll add the A7III vs d850 test to the list.

 

EDIT: Done a comparison at 310nm and 360nm using bandpass filters, and imaging a 60% diffuse reflectance standard using a 302nm fluorescent light source. I don't trust the blocking on my 300nm, hence the 310nm one. I calculated average channel response from the raw files using Rawdigger. At 310nm the monochrome d850 is about 5 stops faster than the multispectral A7III (which obviously has the Bayer filter still present). At 360nm, the mono d850 is about 3 stops faster than the multispectral A7III. This is similar behaviour to what I am seeing if I compare my monochrome 5DSR with my multispectral 5DSR. Obviously I'm comparing a Sony with a Nikon here (although a Nikon using a Sony sensor from what I have read online), but the Bayer filter/microlens setup is behaving similarly to the Canon ones.

 

With regards to the fog test, if I take the ISO 6400 raw file and add 3 stops exposure in Darktable, it looks like the ISO H1.0 one. So it looks like underexposing and then boosting using software afterwards wont work.

Link to comment

Thank you Jonathon. So the color filter array and microlenses rob us of 5 stops at 310nm. That is very significant.

The BSI improvement seems to only kick in at and after 330nm with 1/3 to 2/3 stop improvement.

Your canon 50:50 tests imply that the microlenses don't impact the lower UV. So maybe all that is due just to the CFA. You might just be convincing me to get a ZWO ASI1600MM or Rising Camera ATK3-1600 to test. They have the same Panasonic MN34230 sensor but without the CFA, but with microlenses. I wonder if I could see past the magic 280nm mark with either of them.

 

Link to comment
This result for the Nikon D850 has finally convinced me to go for that Panasonic S1R instead. I am going to push that button tonight.
Link to comment

Andrea,

That seems like an odd conclusion. Why the S1R? Is it because it doesn't have a BSI sensor?

Yes Jonathan test seems to indicate little gain for visibly optimized sensor that are BSI. Looks like 2/3rds stop gain for BSI over FSI sensors, but only after 400nm. So oddly for UV with off the shelf tech we don't get the benefits of BSI for UV.

 

Actually, off the self BSI sensor might be worse for UV photography. As I see 1 stop gain from his plot in the IR but only about 1/3 stop gain in UVA range. So you better have stronger IR blocking filters with a BSI sensor.

Link to comment

Andrea if you do get the Panasonic S1R, could you add the Canon 40mm f2.8 STM lens and the Sigma EF to L-mount lens adapter to your cart?

I am curious how the Canon auto focuses in UV on the Panasonic. The Canon 40mm f2.8 STM lens, is my best UV AF lens right now and is fun to do focus stacks with on an Olympus camera.

Its also really good in IR too. So you may need a Baader to test it out with.

Link to comment

You might just be convincing me to get a ZWO ASI1600MM or Rising Camera ATK3-1600 to test. They have the same Panasonic MN34230 sensor but without the CFA, but with microlenses. I wonder if I could see past the magic 280nm mark with either of them.

The only thing I add there David, is whatever you go with make sure you are clear on any cover glasses they are using. If the coverglass material is blocking the light then it doesn't matter how good the sensor is.

 

I am hoping to get a 254nm light at some point, just to see if I can see anything with my BSI cameras.

Link to comment

Jonathan,

The cover glass will be an unknown. All I do know is that Astrophotographers complain that it is not AR coated. Which is a good sign.

 

If you do get a 254nm bulb or filtered light, just make sure it says ozone free. These new ones are much easier to handle. They seem to have titanium doped glass to cut out all wavelengths below 230nm.

Link to comment
David, the first thing I'm going to try for 254nm are a pair of 8W fluorescent tubes to fit into my UVP lamp. The tubes are relatively cheap, and while the filter in my lamp wont be any good for 254nm, the lamp should be fine for running them without the built in filter. I'm going to use the 254nm filter from the Sirchie lens. I'll try a 'with and without' WG295 to see whether it is actually the short wavelength light I am seeing, or whether it is leakage of longer wavelengths. It'll be very interesting to see whether the A7III can see anything down there, and if so what colour it comes out as.
Link to comment
  • 2 years later...
lukaszgryglicki

My first tests of 4 bulbs 254nm "non-ozone free" are not promising :-(

It is 80W+ INPUT power (output unknown :-()

I can probably registr them, but I lack filters to block any longer weavelengths.

 

Link to comment

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...