Sensor Corrosion Analysis and Fix [Merged]

[kristyansen]

Advertisement (gone after registration)

vor 22 Stunden schrieb Einst_Stein:

The distance table I gave in the early post is very pessimistic. It essentially assumes you have infinite enlargement and the lens is perfectly sharp (zero confusion circle). In reality, the "bottleneck" of the focusing beam at the focusing plane is a fuzzy circle. The off-focus blurring is far less than that idealized calculation. 

 So I take my 25mm ZM and M240 to check the more realistic impact of the 0.1mm light path. On my lens, near the infinite, if I turn the lens focusing ring to move the lens about 0.1mm, everything in 30m is blurring. This means the off-focus infinite is far more than the 2.5m. 

I then take the nearest object (0.5m) for the test, I check the blurring if I move the focus to make about 0.1mm focus-off, and check the image through the EVF with focus peaking  at the max enlargement, to my eye, I can't tell the difference. 

These tests indicate: 

1: at infinite distance, the 0.1mm light path difference cam make the object at X distance falls on the image plane supposed to be infinite. The X is far more than 30M. 

2: at nearest distance(0.5M), the 0.1mm light path difference is impossible to judge where is the optimal. 

Note that the above tests are based on the M240 EVF (or live view), which theoretically is applicable to M9. However, with M9 OVF, your focusing judge should be much less accurate. So that 0.1mm light path is even less a concern. 

I would assume the frequently seen OVF infinite off due to the OVF is much worse. Whoever really worry that 0.1mm light path, she/he should maintain the OVF accuracy all the time.  

Thank you for the note.

The mentioned distance table seems to count more for the 16mm lenses than for the higher focus lenses. My question is how much the glass change intervention affects the focus accuracy - not as a function of the EVF, but of the zone focus - i.e. if the area in focus of the lens at 16mm (random pic of the 16mm lens https://www.juzaphoto.com/shared_files/recensioni/leica_tri-elmar_16-18-21.jpg) says that at f16 everything is in focus but the glass change intervention makes everything from 2.5m not to be in focus than this is a big change: or similarly at f4 if everything between 1.5-5m should be in focus that it should be like this.

The follow-up question was: understanding that a 0.6mm glass is applied instead of the original 0.8mm, wouldn't it be more accurate to the original camera to use a 0.8mm glass? As a simple user I would just want to have the same functionality and not think myself of moving the sensor. 

Also, after applying this service, would Leica refuse to accept the camera again for e.g. cleaning, adjusting the focus with lenses, etc.? Just like Apple used not no accept iPhones for repair after "jailbreaking"?

[Einst_Stein]

3 minutes ago, kristyansen said:

Thank you for the note.

The mentioned distance table seems to count more for the 16mm lenses than for the higher focus lenses. My question is how much the glass change intervention affects the focus accuracy - not as a function of the EVF, but of the zone focus - i.e. if the area in focus of the lens at 16mm (random pic of the 16mm lens https://www.juzaphoto.com/shared_files/recensioni/leica_tri-elmar_16-18-21.jpg) says that at f16 everything is in focus but the glass change intervention makes everything from 2.5m not to be in focus than this is a big change: or similarly at f4 if everything between 1.5-5m should be in focus that it should be like this.

The follow-up question was: understanding that a 0.6mm glass is applied instead of the original 0.8mm, wouldn't it be more accurate to the original camera to use a 0.8mm glass? As a simple user I would just want to have the same functionality and not think myself of moving the sensor. 

Also, after applying this service, would Leica refuse to accept the camera again for e.g. cleaning, adjusting the focus with lenses, etc.? Just like Apple used not no accept iPhones for repair after "jailbreaking"?

I have no clue why the 0.6mm glass is used instead of 0.8mm. The only thing I would guess is to reduce the dispersion effect at the edge of the image. But this guess has nothing to back up. I also don't know how Leica service views this modification. Again I would just guess they might refuse to service. 

As far as how much the off-focus distance can affect the image, I think my early assumption and calculation is too pessimistic. You might want to try a far simpler test using the LCD or EVF:

1: Turn the focus ring while look at the lens movement, find out the amount of turn that can make the lens move about 0.1mm (should be very tiny).

2: Shoot your target object according to the normal (best) focus.  This will be picture 1.

3: Shoot the same target again, but this time turn the focusing ring to make that 0.1mm off-focus. Note, the glass thickness business effectively (conceptually) moves the sensor forward (the light path is shorter than the original) 0.1mm. make sure you turn the focusing ring in the correspondent direction. If you are not sure, just try both direction. This will be picture 2 (or 3 and 3).

Now compare the two or three pictures to see how much blurring it contributes.

Note again: use tripod if possible to minimize the error due to human movement.  

 

 

[dllewellyn]

Another point.  You certainly wouldn't want to change the sensor coverglass unless you already have a problem.  If you do have sensor coverglass oxidization, if it is bad enough, the camera will be unusable.  Seems to me at that point, your options are:

• Use it as a trade in towards a new camera
• Not to do anything
• Sell the camera
• Get the camera repaired and run the risk that Leica won't do other repairs if something else breaks. 
• Get the camera repaired and turn it into a monochrome camera and run the same Leica repair risk

Here is how a sensor looked with really bad oxidization.  The brown residue is from the person trying repeatedly to clean with a solvent of some sort.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

Edited September 15, 2020 by dllewellyn

[dllewellyn]

Here is the same sensor cleaned up and turned into monochrome.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[Einst_Stein]

I should point out the precision requirement is not the same matter for the camera engineer and camera users. Here I am taking my personal view as the user.

If you want the engineer's precision, it could be very different. See this:   https://www.photo.net/discuss/threads/film-flatness-practical-considerations.23078/

A brief summary of the above discussion says, from the "image plane curvature" point of view, the required engineer precision was in the range of 0.15mm (old lens design) to 0.015mm (the modern lens design). I assume this can be a reference to the glass thickness discussion. But the curvature error should be more visible than the focus plane shift since it is easier to attribute to the camera error vs. the user error.

So, for people who care the engineering precision, the 0.1mm focus plane shift could be unacceptable, but I doubt it matters that much practically. 

Of course, YMMV.

Edited September 15, 2020 by Einst_Stein

[jaapv]

1 hour ago, BrianS said:

plan to use has too much IR leakage for the M9, and will give results closer to the M8.

Which would not be a problem - The user would need to use IR cut filters in front of the lens, which is no hardship.  The only drawback would be the need to use flat field correction for certain WA lenses.

[BrianS]

Advertisement (gone after registration)

49 minutes ago, jaapv said:

Which would not be a problem - The user would need to use IR cut filters in front of the lens, which is no hardship.  The only drawback would be the need to use flat field correction for certain WA lenses.

We've used Hot Mirror filters for years, with our M8 and with some lenses on the M9. I have about 12 of them. The M8 drew much criticism over the need for external filters and the IR problem. Leica/ONSEMI avoided this completely by using BG55 glass, which is much more stable environmentally than S8612.

The protective coating of the S8612 as used in the M9 did not stand up to the job of preventing corrosion. Dan claims that the Data Sheets published for years, revised a few times- are all wrong- that the M9 sensor's cover glass has no protective coating at all. If that had been the case, as per the S8612 Data Sheet- the corrosion problem would have been much worse than it was. Schott states that unprotected S8612 glass will corrode after a few months. My M9 went 7 years.  Dan's article on MaxMax labeling Kodak and Leica as "Stupid, Stupid, Stupid" does not identify the glass, which is specified in the Datasheet. The Data Sheet also states that the cover glass is provided by Schott, and the performance of the cover glass in the Data Sheet is from Schott. I believe the Data Sheets from Schott and Kodak/TrueSense/ONSEMI. 

 

I work with groups that make custom glass, and make custom sensors. To state that silicon sensors  are all the same was just way too much. A couple of years ago one of my coworkers needed to make a low-cost VNIR sensor, I went through the data sheets for a number of CMOS and CCD COTS sensors. The IR had been bred out of the CMOS sensors of the time, Hamamatsu CCD's had 10dB higher sensitivity in the IR. All were Silicon sensors. Same with ONSEMI CCD's without IR absorbing glass, better IR- but the emphasis had been placed on Visible. The spectral response of my DCS200ir is about the same in IR as Visible. Kodak started using Indium Tin Oxide to extend blue response. Once again- follow a technology for 40 years, you pick up some things.

 I hack lenses, known for converting Sonnars to Leica mount and adjusting Jupiters. That's lens hacking for relaxation.From working in an Optical Research Lab for over 40 years. I'm a Lens Hacker, but had two optical engineers work for me to design the real stuff. Kodak came to us for IR technology back in the 80s. Seemed fair that I give them a call in 2009 and ask for a Monochrome M9. I actually asked for a full-spectrum M9, told them I'd buy 6 of them. It would have been under $50K for all of them. Way cheaper than Sensors Unlimited stuff. The Kodak Engineer told me that had a meeting with Leica coming up, and they would discuss it. Kodak had wanted to do a Monochrome M8 to pick up where the DCS760m had failed- due to a banding problem. Preaching to the choir.

 

Once I got the M Monochrom, I had to get the shim for an F1.5 J-3 and Sonnar within 0.01mm. I have about 20 5cm F1.5 Sonnars and J-3's in my collection. Past 100 lenses in Leica mount now. For the M Monochrom: the shims are optimized for different color filters. Deep color filters shift the focus towards infinity, enough to have to adjust the shim. 0.1mm is far off for a lens. I no longer do lens work- Skyllaney in the UK offers the service professionally. Gone are the days that I'd convert a Sonnar to Leica mount for $25 and left-over parts. Some of those Sonnars that I converted sell for a lot these days. All a clever plan to shoot with and take apart hundreds of lenses.

Edited September 16, 2020 by BrianS

[dllewellyn]

The Schott BG types of glass are IR absorbing not IR reflecting (hot mirror) like the vapor deposition coated filters.  Most decent consumer cameras have both types as part of the ICF.  Some of the inexpensive point and shoot cameras use only a hot mirror.

I have personally measured the spectral response of many consumer digital CCD and CMOS cameras.  I don't have experience working with sensors out of something like a Hamamatsu device.  What I can tell you is that for consumer cameras, assuming no coverglass acting like an ICF, IR response is pretty much the same.  

Another point is that Brian is assuming he knows the specification of the glass I use for a visible light monochrome conversion.  I am not sure how he knows that since I haven't published that specification.  The glass I use is a combination of ionically colored glass plus vapor deposition coatings to further control the ICF. 

[jdlaing]

What’s the warranty on the conversion?

[dllewellyn]

1 year warranty on the sensor itself.  Sensors rarely go bad though.  If it happens, failure usually happens during the conversion process, and then I have to source another sensor.  That's part of the reason the monochrome conversions are expensive.  Oh, and you need $100,000+ of semiconductor fabrication equipment, know how to use it, maintain it, have a building to hold it, have the 110V single phase and 220V three phase power, and a lot of other stuff.  I have, literally, millions of dollars of equipment in the building, but I am sucker for buying tools and equipment. 

With the M9 cameras, I really, really don't want to get a failed sensor - they are too hard to find.  If I could find out where Leica put all those failed M9 sensors, I would buy a bunch in heartbeat.  Anybody know where those bad sensors went?  Any Leica people sneak some out from the trash?

BTW, teaser alert, I am planning on attempting an S2 monochrome conversion fairly soon.  Keeping my fingers crossed!  I picked up one mint body for $1,000 that has a few tiny specs of sensor spots for $1,000, and a 2nd for $1,600.  I wish their lenses were that cheap!

Edited September 16, 2020 by dllewellyn

[kristyansen]

One question on the KolariVision conversion and the explanation of the sensitivity difference of the BG60 06.mm glass: https://kolarivision.com/product/leica_m9_repair/. Must the increase in exposure of 1/4 to be done in camera by the photographer, or is it something physical done during conversion? I am asking, since the Leica M9 needs all the light it can get due to its lower ISO capabilities and everything to be done through software will just increase noise even faster. Also, there is no AA filter being added, correct? 

On the MaxMax conversion _keeping the color capabilities_ of the Leica M9, how do the sensitivity and color profile of the sensor change? 

[carbon_dragon]

I'll just say that the reason I keep the M8 is to use it as an IR camera. So that IR sensitivity is a plus for me. And as a landscape camera I never needed an IR cut/hot mirror filter (though I owned them). I guess I'm saying I regard that as a PLUS, not a MINUS.

[DandA]

Can someone clarify something (in simply terms) please.  What was the original thickness or the original non coated (corrosion prone) cover glass used in the M9's.  0.8mm?  Secondly, once Leica replaced the corroded cover glass on a M9 with a coated no corroding one, what was the thickness of this newly replaced coverglass?  I assume it was 0.8mm too, like the original?  If so, isn't it possible for these third parties now having the ability to replace M9 cover glass, to spec a 0.8mm one instead of a 0.6mm cover glass, so focus stack remains the same and no refocusing or changes to the focusing of the M9 takes place in any measurable way, even using ultra wide angle lenses?  Thanks!

 

Dave (D&A)

[jaapv]

Simple answer: yes.

[jaapv]

3 hours ago, kristyansen said:

One question on the KolariVision conversion and the explanation of the sensitivity difference of the BG60 06.mm glass: https://kolarivision.com/product/leica_m9_repair/. Must the increase in exposure of 1/4 to be done in camera by the photographer, or is it something physical done during conversion? I am asking, since the Leica M9 needs all the light it can get due to its lower ISO capabilities and everything to be done through software will just increase noise even faster. Also, there is no AA filter being added, correct? 

On the MaxMax conversion _keeping the color capabilities_ of the Leica M9, how do the sensitivity and color profile of the sensor change? 

The sensitivity is influenced bu the light transmission through the glass. I may well be that the 0.6 mm glass has more dense filtering which would mean a lower transmission. I would need to see the transmission curves of the specific IR filters. However, since the camera meters from the sensor, the exposure metering should be correct in all cases. It may make a difference using an external exposure meter.

[dllewellyn]

I would tend to go with the factory  0.80mm thick glass.  One can second guess the factory and put in 0.60mm glass and change the shims, but there is probably a reason the factory chose 0.80mm.  The factory could have put in 0.50mm glass if they wanted.  I also have doubts about how much improvement your are going to see with 0.60mm glass.  If enough people want it, I can easily make up 0.50mm glass and change the shims accordingly.

RE changing the color profile.  To get a certain answer, I would want to measure camera spectral response of a stock and converted camera because the color response is a combination of the filter and the camera RGB response.  For example, the glass in the M9 I measured goes down to 300nm but the sensor has very little sensitivity down there.  But I think you would notice very little difference and if you set a custom White Balance, then probably no difference at all.

Here is the M9 ICF that I measured versus a typical ICF.

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[dllewellyn]

BTW, you see those ripples on the top of the red line and the sharp drop at 420nm?  That's from the *coatings*.

[Herr Barnack]

Saw this article the other day:  https://petapixel.com/2020/09/12/leica-m9-sensor-corrosion-due-to-dumb-design-decision-report-claims/

Leica M9 Sensor ‘Corrosion’ Due to Dumb Design Decision, Report Claims

Quote

"...Manufactured between 2009 and 2014, an estimated 180,000 Leica M9 cameras were produced and sold with a price tag of $8,000.

Inside the camera was an 18-megapixel Kodak full-frame CCD sensor. Covering the image sensor itself is an IR Cut Filter (ICF) which also serves as the coverglass.

“Typically, the ICF is a separate piece of glass held in front of the clear sensor coverglass,” writes Dan Llewellyn, president of LDP LLC. “I don’t know if Kodak did that to lower costs or perhaps increase optical performance since every piece of glass light has to go through degrades the image somewhat.”..."

And -

Quote

"...What I found was a filter that extended down to 300nm,” Llewellyn writes. “That filter doesn’t have coatings! […] Leica / Kodak used an uncoated Schott BG type glass! The coatings seal the glass. Somebody really screwed up.

“All those BG and UG types of glasses are susceptible to oxidization. Variable include time, temperature, humidity, ozone, and even the particular glass melt, but you would never put uncoated BG or UG glass in a camera and not think that one day you are going to have a problem. That Leica / Kodak would do that boggles the mind..."

I would agree that "somebody" really screwed up.  The question is this:  Was it Leica or was it Kodak?

Quote

"...Kodak sold its image sensor business to the Platinum Equity-owned Truesense in 2011, and Platinum Equity sold the business to On Semi in 2014..."

Given that Kodak/Truesense/Semi were playing hot potato with Kodak's image sensor business (and the  liabilities attached to it) tells us something IMHO.

Edited September 16, 2020 by Herr Barnack

[dllewellyn]

Keep in mind that around the time the M9 was being made by Kodak, they had laid off about 50,000 of their 160,000 employee company.   Their business was floundering, and by 2011 they had spun the sensor division to a company called Truesense which was owned by a private venture capital firm Platinum Equity who three years later flipped it to On Semi.  I am speculating here, but, the specifications called for a coated coverglass, but someone at Kodak dropped the ball and forgot.  The company was in turmoil then.  Have you ever heard of any other camera have this sort of sensor 'corrosion'?  When I took out the sensor, the glass looked exactly like what uncoated BG glass is prone to do.  Measuring the glass shows no coatings.  Looking at the glass eroding away shows no coatings.  I know I poked a stick in a hornet's nest suggesting that someone made an error, but that's the only explanation that makes sense to me.

Here is what the glass looks like after polishing off the oxidization.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[Advertisement]

Hallo,

Look here: Leica M

[Herr Barnack]

Quote

"...  I am speculating here, but, the specifications called for a coated coverglass, but someone at Kodak dropped the ball and forgot..."

That was my point in my post #118. 

If that is indeed what happened, I wonder why Leica did not pursue legal action against Kodak to recover the costs incurred by them in resolving Kodak's, uh,  "forgetfulness?"  That seems like a reasonable course of action.

Edited September 16, 2020 by Herr Barnack