DIY Nikon D70 Full Spectrum Modification and Fun With Filters

[toups]

I really didn't need more projects, but decided to modify a D50 or D70 for full spectrum.  Decided to do the D70 because I wouldn't need to pull out the soldering iron to remove the shield unlike the D50.  Followed the instructions in the tutorial at lifepixel.com and used a window from tangsinuo (more later).  At first everything seemed to go fine but then the display started acting strange.  Long story short, it turns out that one of the tiny ribbon cable catches broke on the lower board.  That's the bad news, the other bad news is that the other day my second D70 abruptly stopped working (probably a bad power board.)  So it turns out that was also the good news, I swapped the board from the "dead" camera into the one with the damaged connector and now have one full spectrum modified D70.

Regarding the hot mirror, I replaced it with a "JGS1 UV Grade Fused Silica Quartz Glass Window disk deep UV Grade" sized at 29*25*1.0mm and it was a perfect fit for length and width.  These are sold in batches of 5, so I can do my D50 and have some spares.  Plus I need one as a window for an astrophotography project I've been slowly working on.

Of course now, I need to investigate the focus, it looks like I may need to re-calibrate.  I found some information at http://www.leongoodman.com/d70focus.html

Regarding IR/UV filters, I ordered a batch to experiment with from tangsinuo's e-bay store.

Took the camera outside prior to sunset to see how things worked.  A few observations regarding the UV filters I ordered.  These are just a set of quick observations (qualitative, a quantitative set will have to wait.)

"IR Cut Filter TSN575 used for stacking with ZWB2 for UV photography"
As will be mentioned later, I found it to at least qualitatively be quite effective and needed for the 3 UV filters tried.  At least this was true for natural light outdoor UV images.

"TSN340 52mm UV Pass Filter for pure UV Photography" 
This is an 8mm thick ZWB1 filter.  In spite of the claim there is no IR leakage, I found that it really could use additional IR filtering via a TSN575. (See third post, the need for an IR filter might be lessened with a lens with better UV transmission.)

"Various sizes ZWB2 UG1 UV Pass Filter Glass for 365nm light source flashlight" and
"ZWB1 UG11 UV Pass Filter Glass lens for 302nm 312nm light source" both appeared to play nicely with a TSN575.

"Various sizes HB630 630nm IR Infrared Long Pass Filter Visible light Absorbing" used for testing for IR leakage (and has other uses.)

The technique I used to determine if there was IR leakage was to stack the HB630 with the candidate filter(s).  All three of the UV filters showed significant light leakage when stacked by themselves with the HB630.  However adding a TSN575 resulted in a significant reduction in the transmitted light.  With a stack consisting of a UV filter (passes UV and IR), the TSN575 (blocks the IR) and the HB630 (blocks the UV and passes the IR) most everything should be blocked.  Without the TSN575 but with the UV and HB630, only the IR leakage should be displayed.

The next step would be to redo this a bit more systematically and measure how well the non-UV wavelengths are blocked.  Should also probably look at the visible region below red.  It will be interesting to see if the 8mm filter ends up performing better or worse than the thinner filters in UV purity and image quality aspects.

I know there are related discussions over at https://www.ultravioletphotography.com/content/

[toups]

Interesting, loaded the images on my computer and I found the first issue with using a D70 is that as far as I know the only histogram available on the camera is for the green channel.(?)  It looks like the exposures between the different color channels are wildly different (no surprise there.) So I will need to understand what the relative values are in the color channels are for a pure UV (or pure IR for that matter) exposure are so that I can estimate exposure quality based on the available histogram.  In some cases images were saturated when from the histogram I thought there was plenty of headroom available.

I am thinking for a future camera, I might want to acquire and modify a D610 since it can show histograms for each channel and I am now more familiar with its operation.  Besides the obvious advantages of a D610 over a D70, there is a lot to be said for only having a single interface to remember.  For example, I have trouble remembering which buttons pull up which display options when going back to the D70 and D50 especially when these two have different interfaces themselves. 

A modified D610 is a future concept.  Right now, I have a lot to do to understand the full spectrum modified  D70 and the filters I now have.

For future considerations, does anyone know what the hot mirror filter size (length x width x thickness) is in a D610?

[toups]

Of Course I need to take into account the UV transmission of the lens used.  The previous observations were using a Nikkor pre-AI 50mm f/1.4 and an AI 28mm f/2.8, neither of which are specialized UV lenses.  So if there is little UV transmission then maybe the need for a TSN575 with the TSN340 might be lessened if one used a lens with more UV transmission so the UV to IR ratio is increased in the final image.  I will have to do some more quantitative measurements in the future.

I have some diffraction gratings so a real nerd project would be to set up measurements of the system transmission as a function of wavelength.  Too many projects, too little time.

[toups]

P.S. I also have in my collection a MidOpt (Midwest Optical) BP365 UV Bandpass filter that I acquired some time ago.  It nominally covers 335 to 400 nm.  I need to add this to the set for UV testing.  It's specs are at https://midopt.com/filters/bp365/ 

Also over the years, I have collected a number of color bandpass filters and IR filters to add to the mix including tangsinuo HWB780 in 52mm (and 40.5mm for use with my Nikon 600mm f/5.6 with a modified filter holder (yet another project)), and a variety of MidOpt filters in sizes larger than 52mm for use with my lenses with step up adapters.  These include the MidOpt bandpass filters BP365 (already mentioned), BP470, BP525, BP635, BN660, BP660, and BN850.  I also have some broad filters (UV/IR Cut) such as the BP550 and SP705.  So all in all, I have filters that should allow me cover a wide range of wavelengths.

[Bernard Delley]

I look forward to read what you did with the full spectrum camera.
With my background on spectral properties, I would be curious to see the recorded spectrum of a halogen lamp. The general arrangement could be similar to this simple setup:
post 2 at https://nikongear.net/revival/index.php?topic=10662.0 . All you need in addition to you full spectrum camera is a slit with illumination and a cheap grating (2$). -- Maybe yet another project for you.

[toups]

I look forward to read what you did with the full spectrum camera.
With my background on spectral properties, I would be curious to see the recorded spectrum of a halogen lamp. The general arrangement could be similar to this simple setup:
post 2 at https://nikongear.net/revival/index.php?topic=10662.0 . All you need in addition to you full spectrum camera is a slit with illumination and a cheap grating (2$). -- Maybe yet another project for you.

I actually have some good gratings and will be working on putting together a spectroheliograph for solar work with my telescope.  And yes I do want to measure the response of the camera and the Bayer colors of the sensor.  While for the spectroheliograph I intend to use a monochrome industrial fire wire camera, I still want to measure the spectral response of modified (and unmodified) D70 and D610 cameras as well as some of the lenses I have.  In particular I am very curious as to the UV response of some of the lenses I have.  I am also very interested in the false color properties of the Bayer layers.

So I have converted a D70 and now have a second D610 that I intend to modify.  More to come, but be patient, as always, too many projects, too little time.