jaapv Posted August 2, 2010 Share #161 Posted August 2, 2010 Advertisement (gone after registration) I said before - trust me, the ones reported here are the vast majority of sensors cracked. Link to post Share on other sites More sharing options...
Advertisement Posted August 2, 2010 Posted August 2, 2010 Hi jaapv, Take a look here Crack!. I'm sure you'll find what you were looking for!
elgenper Posted August 2, 2010 Share #162 Posted August 2, 2010 I doubt that about 20 cameras on about (guess!) 20.000 made represents much of a valid statistical sample. Before my retirement, I worked in reliability and statistics. I wouldn´t call those 20 cameras a ´sample´ at all. The 20 000 cameras (for the sake of discussion, let´s accept your numbers) is the entire population, so far, and the total number of crack cases divided by 20 000 then is the observed failure rate for that failure mode (possibly there are different reasons for the cracking, but let´s not complicate things. Had we known the total number of cracks, we would have had the best possible estimate of the failure rate. Only, as discussed above, we don´t know how many of the crack cases that are actually reported here. And, those reported isn´t a random sample at all, since it depends on things like the individual users´ web usage patterns & c. It´s a non-random report filtering (not at all usual in industrial settings), and the ´strength´ of the filtering can only be guessed at; quite different from analysing a true random sample. The number of broken sensors is so small that it does not help much to do a statistical analysis imo. I have a feeling that the number of exposures would be more interesting than the age of the camera, btw. That diagram does look as what one would get over a limited time for a system with low failure rate; not much of a pattern, but some hints. The disturbing thing is that it would (right now) look the same whether Leica did indeed find and remedy the case within the last couple of months or whether new sensors remain just as failure-prone as those before. Time will tell.... One thing is certain: this kind of analysis is performed by Leica, and they have access to the correct numbers that we can only guess at. The light at the end of the tunnel is Stefan Daniel´s statement that Leica will replace sensors free of charge. The rest is silence... The worried old man (to paraphrase a certain other LUF regular) with a yet uncracked 3836*** body, delivered Feb 5 Link to post Share on other sites More sharing options...
250swb Posted August 2, 2010 Share #163 Posted August 2, 2010 I have a feeling that the number of exposures would be more interesting than the age of the camera, btw. FYI 7845 Steve Link to post Share on other sites More sharing options...
AlanG Posted August 2, 2010 Share #164 Posted August 2, 2010 If the number of exposures have something to do with it, what could be the process that causes the cracking? (I am not dismissing this as a possibility, I simply have no idea.) Link to post Share on other sites More sharing options...
250swb Posted August 2, 2010 Share #165 Posted August 2, 2010 If the number of exposures have something to do with it, what could be the process that causes the cracking? (I am not dismissing this as a possibility, I simply have no idea.) Well if you divide the number of my shutter actuations (7845) by the average length of time for each failure (214 days) it comes to 36.6, which is the temperature of the human body........strange indeed if it forms a pattern. But personally I think its sudden changes in air pressure, something I don't think Leica have investigated yet. Steve Link to post Share on other sites More sharing options...
jaapv Posted August 2, 2010 Share #166 Posted August 2, 2010 Alan, the cause being unknown, I have just a theory to offer. The filter is bonded to the microlenses by a thin-layer bonding (that being the type of bonding I know about professionally) Failure of that type of bonding is usually in the bonded material, not the bonding itself. If there is the smallest discrepancy in the thermal coefficients, the thickness of the bonding layer or the homogenity of the bonding layer, thermal cycling can bring on breaking-strength stress. Each exposure on a CCD will cause a thermic cycle. Don't forget the IR filter is 0.8 mm as compared to 0.5 mm on the M8, causing loss of flexibility, and the surface is double the M8, causing four times the stress problems. Link to post Share on other sites More sharing options...
ArtZ Posted August 3, 2010 Share #167 Posted August 3, 2010 Advertisement (gone after registration) Guys, I have two theories about sensor crack. 1.- Some bad quality SD cards can produce hi frequency sound which is not audible for the human ear. This sound could simple break the IR filter like a soprano can break a glass. 2.- It’s just a firmware problem that will be sort out in a future firmware release. So, don't worry about it. Change SD card brand (and avoid buying fake ones) or wait for a new firmware release. PS: Sorry, I couldn't resist an early morning joke Link to post Share on other sites More sharing options...
Guest malland Posted August 3, 2010 Share #168 Posted August 3, 2010 ...I have a feeling that the number of exposures would be more interesting than the age of the camera, btw.In my case, the crack was there when I got the camera from the dealer and appeared already in the first exposure, although I didn't know until a week later because I picked up the camera from the dealer, made a few exposures, then left the following day from Paris on a flight to Bangkok and didn't have time to examine the files until a week later. —Mitch/Bangkok Shophouse Demolition Link to post Share on other sites More sharing options...
hoppyman Posted August 3, 2010 Share #169 Posted August 3, 2010 Mitch that is understood to be the case with some of the reported cracks but the others are reported to have appeared later and spontaneously or rather without known cause. So I guess some folks are interested to see if there is any visible trend on the latter. The numbers that we are seeing here thus far are so small against the estimated production that I'm not convinced that the reporting has any statistical validity. I'm not sure what people can actually do with the figures at this point except maybe have increased stress levels . At this point every single M9 is of course well within its international warranty period. Link to post Share on other sites More sharing options...
andy_mclean Posted August 4, 2010 Share #170 Posted August 4, 2010 Alan, the cause being unknown, I have just a theory to offer. The filter is bonded to the microlenses by a thin-layer bonding (that being the type of bonding I know about professionally) Failure of that type of bonding is usually in the bonded material, not the bonding itself. If there is the smallest discrepancy in the thermal coefficients, the thickness of the bonding layer or the homogenity of the bonding layer, thermal cycling can bring on breaking-strength stress. Each exposure on a CCD will cause a thermic cycle. Don't forget the IR filter is 0.8 mm as compared to 0.5 mm on the M8, causing loss of flexibility, and the surface is double the M8, causing four times the stress problems. Hi... perhaps there is something being overlooked; the sensor cracks all appear in the upper left and there is asymmetric vignetting with the W/A lenses that has a hint of diffraction/birefringence about in so much as the left side shows greater vignetting and there is L/R colour imbalance too. Michael's bayer filter notwithstanding... it seems to me that perhaps the two effects are interrelated. Selectively stressing one area of the sensor could affect refractive indices/light scattering leading to the sorts of L/R colour effects and vignetting... and selective mechanical failures that are cropping up. It could be due to sensor delamination, or excessive local mechanical stress, all sorts of things... but repeated mechanical failures in the same sensor area together with uneven vignetting and colour balance may very well be inter-related... Link to post Share on other sites More sharing options...
lct Posted August 4, 2010 Share #171 Posted August 4, 2010 M8.2 crack (source) Link to post Share on other sites More sharing options...
Ario Arioldi Posted August 5, 2010 Share #172 Posted August 5, 2010 @ andy_mclean:"the sensor cracks all appear in the upper left" This is simply not true. Cheers, Ario Link to post Share on other sites More sharing options...
ho_co Posted August 5, 2010 Share #173 Posted August 5, 2010 Andy, good reasoning from your side IMO. But so far there seems to be no relationship between the cameras that show the "red edge" and those that crack. Several folks who had no "red edge" problem were the envy of the rest of us--until one day their sensors cracked. There have been a lot of theories advanced, adjusted, and advanced again. The day when we find the reason is coming closer. Till then, it's not worth worrying "will my sensor crack?" because so few do; and the more theories we can get, the better. Link to post Share on other sites More sharing options...
jaapv Posted August 5, 2010 Share #174 Posted August 5, 2010 Hi... perhaps there is something being overlooked; the sensor cracks all appear in the upper left and there is asymmetric vignetting with the W/A lenses that has a hint of diffraction/birefringence about in so much as the left side shows greater vignetting and there is L/R colour imbalance too. Michael's bayer filter notwithstanding... it seems to me that perhaps the two effects are interrelated. Selectively stressing one area of the sensor could affect refractive indices/light scattering leading to the sorts of L/R colour effects and vignetting... and selective mechanical failures that are cropping up. It could be due to sensor delamination, or excessive local mechanical stress, all sorts of things... but repeated mechanical failures in the same sensor area together with uneven vignetting and colour balance may very well be inter-related... Mounting stress causing a crack would of course be the very first thing investigated by Leica. I can assure you that that is not the cause. Link to post Share on other sites More sharing options...
sclamb Posted August 5, 2010 Share #175 Posted August 5, 2010 Has any correlation between the sensors being cleaned and the cleaning method used been investigated. Could using sticky pads be a common thread here? Simon Link to post Share on other sites More sharing options...
jaapv Posted August 5, 2010 Share #176 Posted August 5, 2010 Maybe - sticky pads pulled on my IR filter and I felt some movement - and it caused a new sensor - but it did not crack. Link to post Share on other sites More sharing options...
sclamb Posted August 5, 2010 Share #177 Posted August 5, 2010 I was thinking back to the video showing M9s being assembled and the way the young lady was jamming the sticky thing onto the surface and pulling it off again. She seemed quite rough and I wonder if some sensors were weakened in the process, or have been weakened by owners emplying the same cleaning method. Simon Link to post Share on other sites More sharing options...
andy_mclean Posted August 5, 2010 Share #178 Posted August 5, 2010 @ andy_mclean:"the sensor cracks all appear in the upper left"This is simply not true. Cheers, Ario Argh... a basic error because i saw a photo with the line running across the top left corner. It would be the opposite on the sensor. Apologies. Although my point was that perhaps the two issues were related. Link to post Share on other sites More sharing options...
andy_mclean Posted August 5, 2010 Share #179 Posted August 5, 2010 Mounting stress causing a crack would of course be the very first thing investigated by Leica. I can assure you that that is not the cause. Yes, but you can't break things without causing some form of mechanical stress. As you mentioned earlier there are several ways in which you could induce stress; heat recycling and associated differences between expansion coeffs etc of the various materials in a composite and so on. What i'm driving at is that the asymmetric vignetting may be a symptom of the underlying problem which in the extreme case also lead to the mechanical failure of the sensor. One side of the sensor shows greater blue vignetting than the other. Some have suggested that this asymmetry is due to the sensor mounting, or the lenses - i'm suggesting that if the sensor itself has an asymmetry in it's light path from left to right, then you may get similar effects. The question then becomes what could be causing an asymmetry in the sensor's light handling from left to right - and it is possible that one area of the sensor is always subject to more stress than other areas, which in extremes results in the sensor cracking, but meantime may also be responsible for the asymmetric vignetting. The stress may be due to mounting of the sensor, or its manufacture, either way localised stress may cause both the vignetting and ultimately, the cracking. For example, if you take two strips of glass and clamp them together very tightly at one end, leaving the other end with a significant air gap between them... and then shine blue light through the two layers of glass... will equal intensities of blue light be transmitted along the full length of the sandwich? They won't - and if you were to over tighten the clamp eventually you will break the glass. Also, if you were to tighten the clamp to just below the breaking point, then tapping the glass with a pencil may be enough to shatter it. Perhaps not with the first tap, or the second.... but after a thousand, two thousand? So depending on whether the clamp is there or not and to the degree to which it is tightened, you may have a different degrees of blue light transmission at one end of the strip sandwich compared to the other. This will be the case whether or not the glass breaks/cracks due to excessive stress. In fact, you could try to use the difference in light transmittance between the two ends of the sandwich to measure the degree of stress applied at the clamped end. The effect would be wavelength dependent too.. red light would scatter less than blue and green somewhere in between. It would also depend on the angle of incidence of the light...depending on colour you would see variations in transmitted light intensity along the glass sandwich as you varied the angle of incident light. Of course I'm not saying that this is what is happening with the M9's sensor - it may be totally unreasonable to suggest that something akin to this is happening. Nonetheless, light transmittance and stress can be related and nobody discussed this... so i thought i'd pop it in. Link to post Share on other sites More sharing options...
bobrudolph Posted August 20, 2010 Share #180 Posted August 20, 2010 Hello again, My M-9 was sent to Leica NJ on August 1, 2010, and returned through my dealer, Camera West today August 20th. The cracked sensor and glass were replaced, as well as the CPU and the camera's leather-like covering. I am grateful for the help from Sean at Camera West, and for the rather quick repair time from Leica NJ. Bob Link to post Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.