☰

By Ked · June 2026

The CCD Problem: Buying a Leica M9, M9-P, or M-E in 2026

June 2026

Most used Leica cameras are simple to value. You grade the cosmetics, count the shutter actuations where you can, check the rangefinder, and the price follows. The M9 generation is the one corner of the digital Leica market where that is not enough, because these bodies carry a known sensor defect that Leica can no longer fix, and the market has quietly reorganized itself around that single fact. It is the strangest segment we track. Two M9s in identical cosmetic condition can sit well over $1,000 apart, and the reason is invisible until you know to look for it. This is the long version of what is going on, what actually fails, how to tell a sound body from a ticking one, what a repair costs now that the factory has stepped away, and how the prices in our own listings reflect all of it.

It applies to four cameras and only four: the Leica M9 (2009), the M9-P (2011), the M-E "Typ 220" (2012), and the M Monochrom "Typ 230" (2012), along with the special editions built on them. All four use the same 18-megapixel Kodak KAF-18500 CCD, and that shared sensor is the whole story.

One clarification up front, because the Monochrom names cause real confusion. The affected camera is the first Monochrom, the "Typ 230," which is simply a black-and-white version of that same KAF-18500 CCD, so it carries the same flaw as the others. The later M Monochrom "Typ 246," launched in 2015, is a different, CMOS camera and is not part of this at all. The two share the Monochrom name and almost nothing else, and getting them mixed up is an expensive mistake in either direction. Throughout this piece, "Monochrom" means the Typ 230.

What actually corrodes

The defect is not in the silicon that captures the image. It is in the thin sheet of cover glass that sits in front of the sensor, a piece of infrared-filter glass of a type made by Schott. Over time, in some cameras, that cover glass corrodes. The corrosion shows up in photographs as small light-colored spots or stains, and the important detail for a buyer is that these are not on the surface where a cleaning swab can reach. They are in the glass itself, so no amount of sensor cleaning removes them. That is the tell that separates corrosion from ordinary dust: dust wipes off, corrosion does not.

The spots hide at wide apertures and reveal themselves when you stop down. Leica's own guidance was to photograph a plain bright surface, a clear sky is ideal, with the lens stopped well down, roughly f/11 to f/16, and look at the result. Wide open, the out-of-focus blur of everything close to the sensor plane smears the marks into invisibility. By f/16 they snap into focus as discrete spots against the even tone. This is why a casual buyer can shoot an affected camera for a while and see nothing: if you live at f/2 you may never notice, and then one landscape frame at f/11 shows a constellation of marks. Where it has been documented, the corrosion tends to spread over time rather than hold steady, so a body with a few faint marks today is not a stable situation.

One point of housekeeping, because the internet is full of confident wrong explanations. A popular story holds that an adhesive behind the glass turns to acid and eats the sensor wiring. That account does not hold up against Leica's own description of the problem, which is corrosion of the cover glass. We mention it only so you can discount it when you read it. The documented, defensible mechanism is the glass.

The four cameras, and why the M-E is not the safe one

It is worth being precise about which cameras are exposed, because the naming invites a costly misreading. The M9 came first in 2009. The M9-P followed in 2011 as the discreet professional version, no front engraving, sapphire-glass rear screen, otherwise the same camera. The M-E arrived in 2012 as a stripped, cheaper entry into the system, and the M Monochrom of the same year was the black-and-white-only variant. The trap is the M-E. It launched three years after the M9 and wears a different name, so some buyers assume it is a newer, later, presumably improved design. It is not. The M-E is a cost-reduced M9 with the same original sensor and exactly the same corrosion risk. A 2012 box does not buy you a 2012 sensor. Treat the M-E as an M9 in this respect, because that is what it is.

Two neighbors in the lineup are clear of all this, and they bracket the affected group on both sides. The older M8 uses a different sensor and is not part of the problem. The newer Leica M "Typ 240," which replaced the M9 in 2013, moved to a CMOS sensor and is unaffected; if the CCD risk is what worries you and you can live with a different rendering, the Typ 240 generation is the clean alternative. The corrosion belongs to the CCD bodies alone.

Does every original CCD eventually corrode? Leica was careful to say the problem affects "some" cameras, not all, and no public serial-number or production-date cutoff has ever been published that would let you sort safe bodies from at-risk ones by number. There is no quantified failure rate either. So you cannot reason your way to safety from a serial; you have to test the individual camera in front of you. Assume any original-sensor body is capable of corroding, and verify rather than trust.

How Leica handled it, year by year

The reason the program history matters is that almost everything written about it online is true for some date and false for now. Here is the sequence, with the dates attached, because the dates are the point.

Leica acknowledged the problem in 2014, with a full public statement around that December. The initial response was generous: a free sensor replacement offered as a goodwill arrangement, regardless of the camera's age, covering even sensors that had already been replaced once, and with refunds for owners who had previously paid for the work out of pocket. For a stretch, this was about as good as a known defect gets handled.

In June 2015 Leica announced the permanent fix: a new generation of the sensor built with a different, corrosion-resistant cover glass. Two things about that replacement are worth holding onto. First, it is still a CCD, the same KAF-18500 imaging design, not a switch to CMOS, so a camera with the new sensor keeps the CCD character that draws people to these bodies in the first place. Second, Leica had outside users test it specifically to confirm the imaging characteristics were unchanged, which is the basis for treating the common claim that a replaced sensor "looks different" as rumor rather than fact. Replacement bodies began going out around September 2015.

The terms then tightened. On 15 and 16 August 2017 the open-ended free program ended. After that date, free replacement continued only for cameras that had been bought new within the previous five years, a rolling window that quietly excludes most bodies on the used market. Owners of older cameras now paid a share of the cost, set at 982 euros (825 euros plus VAT), which at least included a general overhaul of the camera and a one-year warranty. In December 2018 that price was raised again, reported at around $1,595. The trend line is the story: each step made the fix harder to get and more expensive.

Then it stopped. In August 2020 Leica announced that production of the KAF-18500 sensor had ended, and that as a direct consequence defective or corroded sensors in these cameras could no longer be replaced. That is the line that changed the market. Since 2020 Leica has offered only a trade-in "upgrade" program: you surrender the affected CCD body and Leica applies a credit toward a new camera purchased from them. The list of cameras you can move up to has widened over the years and now spans the digital M line (the M10 and M11 families), the Q line (through the Q3 and Q3 43), the SL system, and, as of 2025, even new analog M film bodies. What the program has never included is the only thing an owner with a corroding sensor actually wants: a working sensor for the camera they already own. An upgrade is not a repair. As of 2026, there is no way to have Leica put a working sensor back into your M9. The factory door is closed.

What a fix costs now, and who still does it

With Leica out of the picture, one documented route to an actual repair remains: the US firm Kolari Vision. They offer two different services, and the distinction matters to a buyer pricing risk.

The first is a cover-glass repair, priced around $999, for a camera whose sensor still works but whose glass has started to corrode. They replace only the cover glass, with a corrosion-resistant Schott BG61 glass of their own, and leave the working sensor in place. The second is a full sensor replacement, around $3,500, which swaps in a functioning Leica CCD with the same resistant glass and fixes the harder failures: a fully dead sensor, the "half screen" fault, the "purple curtain." Special editions cost about $1,000 more, optional calibration adds a little, and turnaround runs four to five weeks. Kolari warrants their glass against corrosion and gives a one-year warranty on the sensor against failure.

There is one hard limit to know before you buy a particular body on the strength of "well, I can always get it fixed": Kolari cannot replace the sensor in the M Monochrom (Typ 230). The black-and-white CCD is not something they can source and install. So of the four affected cameras, the Monochrom is the one with the thinnest safety net, and that belongs in your risk calculation if you are shopping for one.

"CCD ID 16," "neuer Sensor," and what those listings actually mean

Once you start reading M9 listings closely, you notice the sellers talking to each other in a private shorthand. Bodies are advertised as "CCD ID 16," "neuer Sensor," "Tauschsensor," "Werksumbau" (German for factory conversion), "new sensor ID 16," or plainly "REPLACED SENSOR." All of it is pointing at the same thing: the seller is claiming the camera has the post-2015 corrosion-resistant sensor rather than the original at-risk one. In a market where the factory fix no longer exists, that claim is the most important thing a seller can say about the camera, which is exactly why you see it splashed across titles. Whether the market actually pays extra for the claim is a separate question, and the prices further down answer it in a way that may surprise you.

Here is the honest caveat. The "CCD ID" number lives in the camera's service menu and the community reads the higher numbers as the replaced generation, but Leica never published an official table mapping those IDs to the original-versus-resistant glass. It is strong, widely held lore, not a manufacturer-confirmed spec, and a title is not a receipt. So do not pay the replacement premium on the word "new sensor" alone. Ask for the dated Leica service invoice that documents the replacement. If the camera is in front of you, you can read the CCD ID yourself, and you should still run the f/16 sky test regardless of what any paperwork says, because the test is the one piece of evidence that cannot be faked. Paperwork plus the ID plus a clean test is a replaced body. Any one of them alone is a claim.

What our market data shows

This is where the abstract risk becomes a number. We pulled every M9, M9-P and M-E listing in our index from the last five months, active and already-ended alike, which leaves a few hundred bodies once bundles and collectible editions are stripped out. About a third of them, 107 of 314, say something about the sensor directly in the listing, either advertising a replacement or, to their credit, disclosing an original sensor and visible corrosion with a "please read" note. A third of a market volunteering information about a part you cannot see in a photograph is remarkable. You do not get that behavior around shutter counts or rangefinder calibration. It happens here because the sensor can make or break the value of the whole camera.

Stripped down to clean body-only listings, with bundles and collectible editions removed, the asking prices across that set run roughly as follows. The M9 sits at a median around $3,750. The M9-P, scarcer and more sought after, runs higher, near $4,900. The M-E, the budget entry, lands around $4,350. These are advertised asking prices over the January-to-June 2026 window, active and recently-ended, not confirmed sale prices, so read them as the shape of the market rather than an appraisal.

The more revealing cut is to take each model on its own and sort it by what the seller says about the sensor:

M9, M9-P and M-E listings in our index over January to June 2026, active and already-ended, body-only, collectible editions removed. Advertised asking prices, not confirmed sales. Rows marked with an asterisk come from fewer than ten listings and are indicative only.

One pattern survives being split by model, and it is the important one. In every body, a claimed replacement and a listing that says nothing sit at essentially the same median: $3,850 against $3,790 on the M9, $4,920 against $4,900 on the M9-P, $4,880 against $4,310 on the M-E. Three times over, independently, the market declines to pay extra for an unverified "new sensor" claim. A finding that repeats on its own in three separate models is one you can trust.

The discount for an admitted problem is real but smaller and patchier than a pooled view suggests, and it is only cleanly visible on the M9, the one model with enough disclosed examples to trust. There, the fourteen bodies that own up to an original or corroding sensor sit about $750 below the rest, in a tight band entirely beneath the others. On the M9-P and the M-E the disclosed samples are too thin to lean on, eight bodies and two, which is why those rows carry an asterisk. Lumping all three models together would suggest a wider gap, closer to $1,400, but that is partly a trick of the mix: disclosed-corrosion bodies skew toward the cheaper M9, so pooling the models exaggerates the discount. Model by model, the honest figure is narrower, on the order of the glass-repair bill rather than a full sensor replacement.

For a buyer the lesson is unchanged in direction and sharper in degree. The words "new sensor" in a title are worth nothing extra by themselves, because the market has already learned not to pay for a claim it cannot verify. What is worth real money is proof, a dated Leica service invoice. What is genuinely cheap is an honest confession of corrosion you can negotiate against, priced with the repair already subtracted, bought by someone who has run the f/16 test and knows exactly what they are taking on.

It is worth being blunt about that silent majority, because the charitable reading is not the safe one. On most used cameras, a seller who does not mention the sensor is simply describing condition the normal way. On these four, the sensor is the single fact that decides whether you are buying a camera or a future repair bill, and anyone selling an M9 in 2026 either knows that or ought to. A listing that carefully notes the shutter count, the rangefinder calibration, and the original box and papers, yet says nothing about the one component that cannot be fixed, is not making a neutral omission. At best the seller has skipped the homework that unloading a $4,000 camera calls for. At worst the silence is deliberate, a quiet bet that you will not think to ask. On this camera you should treat it the second way until the seller proves otherwise. Silence is not the absence of a problem, it is the absence of an answer, and here those are not the same thing.

So do not hand a silent listing the benefit of the doubt the market hands it. Put the question to the seller in writing, plainly: is the sensor original or has it been replaced, and if it was replaced, may I see the Leica service invoice. A straight answer backed by paperwork moves the camera into the documented column and earns the higher price. A vague reply, a sudden pivot to how clean the body looks, or no reply at all moves it into the risk column, where you price it as an unverified original and not a cent more. The honest sellers in this market disclose the sensor without being asked. The rest you have to ask, and how they answer tells you most of what you need to know.

The risk profile, and how to buy

Put it all together and the M9 generation sorts into three buying situations.

A documented replaced-sensor body, invoice in hand and clean on the f/16 test, is the low-risk buy and the one to prefer. The encouraging part, from the prices above, is that you do not have to pay a steep premium for it. Verified-replacement bodies sell for about what unmarked ones do, so the cost of certainty here is mostly the effort of finding a seller who kept the paperwork, not a large surcharge. That work cannot be bought from the factory at any price today, which makes a documented body the closest thing to a worry-free M9.

An unmarked body that tests clean today is the middle option, and the data holds a quiet warning for it. Because the market assumes an unmarked body is probably fine, these listings are not discounted; they sell for roughly the same as documented-replacement ones. So you are paying close to full price for a sensor with no proof behind it and no factory backstop, self-insuring a possible $999 to $3,500 repair at some unknown future date. It can still be a sound buy, the camera works and may never corrode, but only if your own f/16 test comes back genuinely clean and you have made peace with the standing risk. The thing not to do is pay a documented-body price for an undocumented body, because that is exactly what the market invites you to do.

An original-sensor body that already shows corrosion is, paradoxically, where the real bargains in this market sit. On the M9, where there are enough of them to judge, these disclosed bodies trade roughly $750 below the rest, about the cost of the glass repair, and for a buyer who has done the f/16 test and knows what they are taking on, that can be a fair trade. The conditions are firm, though: the price has to have the repair fully subtracted, and you must first confirm the specific camera can actually be serviced. Remember the Monochrom exception here. A corroded Monochrom is a harder problem than a corroded M9, because the usual third-party fix is not available to it.

Whichever you are looking at, the physical inspection is the same and it is not optional. Mount any lens, stop down to about f/16, photograph a clear sky or an evenly lit blank wall, and view the file at full magnification. Spots that stay put through a sensor cleaning are corrosion, not dust. Pair that test with the written question about the sensor's history covered earlier, and you have the two pieces of evidence that cannot be faked: what the camera shows you and what the paperwork proves. On a camera where the most important component is the one you cannot see, the five-minute sky test is the cheapest insurance you will ever buy.

The M9 generation remains a genuinely special set of cameras, the last CCD M bodies Leica made and, to many, the source of a rendering nothing since has quite repeated. None of the above is a reason to stay away. It is a reason to buy the sensor, not just the camera. Check live asking prices on our M9, M9-P and M-E guides, and read the listings the way the market now writes them, with one eye always on the glass in front of the silicon.