Leica KAF-18500 CCD used in the M9 and M Monochrom is discontinued, Sensor replacement no longer possible

So you’re talking about rewriting the firmware? That would be amazing! As for the sensor I’m not sure if it’s a failure or possibly a dirty connection, in all the excitement I forgot to clean the contacts before reattaching the cable so I’ll the mainboard and give everything a quick cleaning and see if that helps. How do you want me to send over the test shots?
Not rewriting the firmware: if Leica made it open source, I would have liked to. I'm not up to reverse engineering it...

What I do for my files: I have software that opens all of the DNG files in a directory, reads them in, applies corrections, and writes them back out. The .exe file is small enough to copy onto the SD card once it is inserted into the computer, apply the corrections there. Better yet- copy the DNG files to your computer then run it. The "GAMMA" code reads in M Monochrom DNG files, applies a curve, changes the pixels values to 16 bits (instead of 14-bits) and adjusts the parameters in the DNG header. I'm "thinking out loud" that a similar correction could be applied to the RGB values in your DNG file to bring the values in line with the older color balance. "It's worth looking at" category. Essentially applying the difference in the two spectral responses.

What would be good: If you have a color test chart, take some pictures of it with the sensor before and after changing the cover glass. Use the same lighting, lens, etc- so that the only variable was the transmission of the cover glass. Do this under dim light and under bright light, but do not saturate the test image. The "first thought" would be to apply a correction to bring the new image back to the values of the old image.
 
This is what I get for thinking about code in the Hot Tub.

Take each Bayer Cell, average the two green values, map into an RGB color cube. Compare the two cubes, before and after new filter. Build a multiplier for each RGB value to map each Bayer Cell into the old position that it occupied on the original color cube. Kind of like code written in 1988 to map an RGB image into an automatically generated optimized look up table of an IBM Professional Graphics Controller. That seems like yesterday.
 
BIG news! So I successfully swapped the filter, onto the damaged sensor, took some test images and compared them to my control image. Turns out the color temp is a little warmer and I gained about a half stop of sensitive and a barely noticeable amount of sharpness. I am running into an overheating issue caused by me accidentally removing the thermal epoxy connecting the sensor to the heat sink.

Here’s the before

62026040-72D6-4568-9E94-0165D7B77D2C.jpeg

and after the conversion with AWB
FD31BB0C-FBE2-4DA5-8C9E-86D43218DD45.jpeg
 
Let me know what you all think, keep in mind this process is by no means a controlled test. I the excitement of everything I forgot to take a bunch of test shot before I removed the original filter. The lighting is mostly from my bench and the fluorescent lights above and I shot them both at 800 iso I believe.
 
So I did some indoor shots under tungsten light and boy is it warm. The person I bought the glass from said that using a 1.3mm filter will bring the range down to around 700nm. This will affect the focus so the sensor would have to be shimmed which I’m not sure how that will affect the micro lens array. The current filter has better visible transmission but I get a little bit of IR. Being that I’ll be offering this as a service what are your thoughts? Better transmission with slight IR or Less visible transmission with No IR? It seems there’s no ‘perfect’ option but I can get as close as I can.
here’s the before
86A0DAF4-A158-4BF5-84E3-43EBDE7B3B3A.jpeg

And after the conversion, to some the added warmth is no problem and an upgrade but it could also be a hindrance to some.
66B4B243-5D62-4A50-AF05-50FE3AE9FF3C.jpeg
 
So I did some indoor shots under tungsten light and boy is it warm. The person I bought the glass from said that using a 1.3mm filter will bring the range down to around 700nm. This will affect the focus so the sensor would have to be shimmed which I’m not sure how that will affect the micro lens array. The current filter has better visible transmission but I get a little bit of IR. Being that I’ll be offering this as a service what are your thoughts? Better transmission with slight IR or Less visible transmission with No IR? It seems there’s no ‘perfect’ option but I can get as close as I can.
here’s the before
View attachment 235416
And after the conversion, to some the added warmth is no problem and an upgrade but it could also be a hindrance to some.
View attachment 235417
My "Unpaid Opinion" on the 1.3mm glass- I would keep the filter stack the same as the original and try to keep the original shimming as much as possible. By moving the sensor back 0.4mm you are going to have a large affect of performance of wide-angle lenses. The "Non-Uniformity-Correction" in the firmware for handling corners based on the lens code will be thrown off. The optimization for the offset microlens array is critical in all of this. Better to have a little IR leakage, the S8612 used in the original has some- much more than my Nikon Df. I have not tested the BG55.

Imagine fixing a damaged lens by just replacing the bad element with a new one of the same focal length and almost the same diamater- the lens works, but the performance gets a little wild at the edges. I've done this with a Nikkor 10.5cm F2.5 and Canon 50/1.5...That's how I see changing the position of the offset microlens array...

Repeat of data sheets here-

Which filter did you settle on, BG40?



ONSEMI/LEICA used BG-55, but changed the color dye and firmware so it would better match the original. You can't change those. But- the curve of the BG40 is close to the S8612, but the IR leakage of the BG40 is more than the BG55 and S8612. It is no where near as bad as the M8 cover glass.

Is there a reason to not use BG55?


This is where I turn on the Light Bar of the Wii at home to test IR leakage- superbright IR LED's. A rough indicator.
 
Last edited:
My "Unpaid Opinion" on the 1.3mm glass- I would keep the filter stack the same as the original and try to keep the original shimming as much as possible. By moving the sensor back 0.4mm you are going to have a large affect of performance of wide-angle lenses. The "Non-Uniformity-Correction" in the firmware for handling corners based on the lens code will be thrown off. The optimization for the offset microlens array is critical in all of this. Better to have a little IR leakage, the S8612 used in the original has some- much more than my Nikon Df. I have not tested the BG55.

Imagine fixing a damaged lens by just replacing the bad element with a new one of the same focal length and almost the same diamater- the lens works, but the performance gets a little wild at the edges. I've done this with a Nikkor 10.5cm F2.5 and Canon 50/1.5...That's how I see changing the position of the offset microlens array...

Repeat of data sheets here-

Which filter did you settle on, BG40?



ONSEMI/LEICA used BG-55, but changed the color dye and firmware so it would better match the original. You can't change those. But- the curve of the BG40 is close to the S8612, but the IR leakage of the BG40 is more than the BG55 and S8612. It is no where near as bad as the M8 cover glass.

Is there a reason to not use BG55?


This is where I turn on the Light Bar of the Wii at home to test IR leakage- superbright IR LED's. A rough indicator.
I’m there with you on keeping the same thickness. Plus I got the data sheet and the 1.3mm would have terrible transmission around 80%. The BG40 and the S8612 are at 98% transmission. The main reason I don’t want to use the BG55 is the loss of transmission and the color shift. The BG55 Has 93-94% transmission. I started looking around at other brands with no luck. Unfortunately the glass used was some really high quality stuff so there’s really no perfect replacement. The BG40 seems to be the closest I can get
 
I have never owned or used - and likely never will - a Leica camera, but as someone who has absolutely no engineering skill, I find this fascinating. Well done.
Thank you! Since highschool I’ve been obsessed with Leica and wanted get into repairs, I’ve been working on the film Ms for about 2 years now doing repairs that included CLAs, blocking out unwanted frame lines, and even re-cementing and resilvering damaged rangefinder optics. This is my first digital camera as well the first digital I’ve taken apart. Up until recently there hasn’t been a lot of info about replacing these filters so this is mostly just trial and error
 
Jadon- I'm glad that Leica users have you on their team. Film shooters and Digital both. It is getting more and more difficult to find shops that do repairs, and now- that do conversions.
 
I’m surprised everyone has been as open to this conversion, I know some people are very particular about what goes in their camera whether it works or not. I should add that if anyone needs work on their Ms feel free to reach out to me, I go strictly by factory spec. None of that “within tolerance” nonsense.
 
This morning I pulled the Bg40 filter off the sensor so now it’s shooting full spectrum. I have a few other IR cut filters and just put them in front of the lens, I know this will affect the focus and it’s not the most accurate way of testing but it gives me a rough idea of what works without needing data sheets. I used the original filter as a control, then a filter off a phase one 645 back, a d700 and then the BG40. The phase one filter is almost perfect. The problem is I don’t know what the glass is so I can get more

this is the Mamiya/phase one
55C4176C-6046-43C1-9A6B-550DB939C11E.jpeg

This is the schott
 

Attachments

  • E3C102D0-8256-4196-9A1D-3E922F4BBA25.jpeg
    E3C102D0-8256-4196-9A1D-3E922F4BBA25.jpeg
    199.6 KB · Views: 183
This morning I pulled the Bg40 filter off the sensor so now it’s shooting full spectrum. I have a few other IR cut filters and just put them in front of the lens, I know this will affect the focus and it’s not the most accurate way of testing but it gives me a rough idea of what works without needing data sheets. I used the original filter as a control, then a filter off a phase one 645 back, a d700 and then the BG40. The phase one filter is almost perfect. The problem is I don’t know what the glass is so I can get more

this is the Mamiya/phase one
View attachment 235447
This is the schott
Sorry THIS is the schott s8612, as you can see they look very Similar. At least similar enough to mix up haha
9863C539-DBFF-4983-96DA-8A58DB0E62D9.jpeg
 
I’m surprised everyone has been as open to this conversion, I know some people are very particular about what goes in their camera whether it works or not. I should add that if anyone needs work on their Ms feel free to reach out to me, I go strictly by factory spec. None of that “within tolerance” nonsense.
I think giving new life to the Leica cameras that develop sensor corrosion is worth an open conversation! Getting me to shutup about a camera or lens is like getting a Teen-age kid to shutup about tricking out their car. Just let me know when you want to talk about 1930s Sonnars converted to Leica mount.

"way back in the 80s" I wrote the software for digital sensors, data capture, calibration, display, and image processing. We had the optical engineers and material engineers all in the same group to design and implement custom IR sensors and have windows made to spec. "Zinc-Selenide" windows for the P3 Orion, cost $40K or so almost 40 years ago. Fun times. That group is still going strong, I went more in the computer engineering direction.

How thick is the Phase I filter? Is it made with a thick sandwich?

Taking a break from tracking down voltage issues on a new computer board...
 

Amazing what you can find on the Internet.
The first two images are from my Fortran codes. One of the scientists in the group submitted them. These were done in the mid 1980s. I used an Intel Hypercube to generate the synthetic cloud imagery.
 
I think giving new life to the Leica cameras that develop sensor corrosion is worth an open conversation! Getting me to shutup about a camera or lens is like getting a Teen-age kid to shutup about tricking out their car.

"way back in the 80s" I wrote the software for digital sensors, data capture, calibration, display, and image processing. We had the optical engineers and material engineers all in the same group to design and implement custom IR sensors and have windows made to spec. "Zinc-Selenide" windows for the P3 Orion, cost $40K or so almost 40 years ago. Fun times. That group is still going strong, I went more in the computer engineering direction.

How thick is the Phase I filter? Is it made with a thick sandwich?

Taking a break from tracking down voltage issues on a new computer board...

yeah Im about the same with Leicas and cars haha I used to have a 68’ beetle but my mechanic left out a Cotter pin and I lost wheel on the Highway! They ended up paying me for the car and i picked up an 83’ Porsche 944

The programming stuff sounds interesting I’ve never had any experience doing, I’m not really great with computers except CAD for 3d printing

the Filter is 1mm same as the s8612, so would change the focus at all once installed. I just need to find more glass to cut
 
yeah Im about the same with Leicas and cars haha I used to have a 68’ beetle but my mechanic left out a Cotter pin and I lost wheel on the Highway!

the Filter is 1mm same as the s8612, so would change the focus at all once installed. I just need to find more glass to cut

I was wondering if the filter could be S8612 with BG38 used as a sandwich to protect it- you can buy S8612 screw in filters on Ebay.


This Ebay seller provides some useful information. BG39 is the same class 3 glass, prone to corrosion.
 
Looks like the glass is similar to the BG39. The data sheet shows that it has 97% transmission vs the BG40s 98%. Beside that the curve is very close to the s8612. A 1% loss seems much better than a 5% loss on the BG55. Im waiting to hear back from the seller on whether or not the coating is susceptible to corrosion
 
I was wondering if the filter could be S8612 with BG38 used as a sandwich to protect it- you can buy S8612 screw in filters on Ebay.


This Ebay seller provides some useful information. BG39 is the same class 3 glass, prone to corrosion.
Ah I see, I thought about maybe putting like a clear filter on the back of the s8612 so it wouldn’t affect the focal plane but I’m not sure how that would work with the micro lenses. The data sheet for BG39 said it was class 2 though? And the s8612 was considered a class 1 even though it corroded. I’m thinking it ultimately comes down to The coatings but I could be wrong
 
So the seller mentioned that every phosphate glass is prone to corrosion, he said that the newer AR coatings protect them well enough. With all this in mind I wonder why schott never updates the coating on the s8612 like they did with the BG55
 
Back
Top