5D Dichroic Mirror Etching: Send to Canon Repair?

Discussion in 'Canon EOS' started by danny_klein, Mar 20, 2009.

  1. Images from my 5D (at f/16+) show etching of the dichroic mirror, which is confirmed by test shots (f/22, uniformly lit subject, etc). I see two options:
    1. Send the camera to Canon - it's less than a year old and I have a warranty card, but I don't know if they will warranty this repair.
    2. Forget about it.
    I have a backup Rebel XT, but naturally would rather not part with my 5D for months. Has anyone had good or bad experiences with Canon repair? How long will my camera be in service? How expensive will it be for them to fix the dichroic mirror?
    FYI, test and real world shots here:
  2. I think what you see is residue of some liquid on your sensor. I would start with sensor cleaning.
  3. While there is some residue on the sensor, the problem is certainly etching of the dichroic mirror. The spots are bright (unlike dust), and do not move with a wet cleaning.
  4. If it is etched, how did it get etched? What have you used to clean it?
  5. This camera was purchased used, and I should have been more careful about checking the sensors before buying (duh, ugh, argh!). I have no clue if this is due to a manufacturing defect, or a problem caused by the previous owner. I cleaned the sensor with E2 + SensorSwabs (in hopes of getting rid of the bright spots, but they didn't budge while the dust and everything else did).
  6. So does it effect your pictures? Show in the viewfinder. Is there a real problem? After twenty years or more with Canon single lens reflex cameras( quite a few of them) I hardly clean anything. I just put up with stuff on the focusing screen until I get around to blowing it off. If it doesn't budge I just leave it there. There is probably something I don't know but how do you tell the spots are on the mirror and not on the focusing screen or the sensor? I occasionally use a hand blower on the sensor also but only if I see it in a picture like say a hair. I currently have a 5D and an XTi.
  7. Hi Dick,
    The spots are etched into the IR filter of the sensor (not the reflex mirror). Thus, they do not appear in the viewfinder. They affect pictures with small apetures.
    If the IR filter on the sensor of your camera was etched, would you send it in to be repaired?
  8. It's looks like either residue from a wet cleaning on top of the sensor glass, or under the sensor glass. It might even be that nasty oil they sometimes use around the sensor that smears onto the glass during cleaning.
    Try a wet cleaning first. See if it moves.
  9. Thanks Ed. However, I have tried a wet cleaning, which is effective at moving around the dust and leaving lots of new residue spots, but I do believe that the IR filter is truely etched. SO, assuming for now that there IS a problem with the IR filter (dichroic mirror), how would you proceed?
  10. Now I get you Danny. I think there is usually a few specs of dust on the sensors in both my cameras. I just leave them there because I almost never shoot above f 11 because of diffraction. I can't see them even 18x24 prints I have taken. I occasionally shoot at a white wall at f22 to see what's there but am loath to wet clean sensors unless absolutely necessary because stuff shows in a print. If it were my 5d and I thought it affected my pictures I would have it fixed. If not I would stay with the status quo. I think, if I were you and I was really bothered by it I would send it to Canon and have the sensor cleaned and examined and then follow their recommendations. If it were me I would really have to see something in each print. I probably would have difficulty living with it but it would have to be more than the tiny dust bits that I know resided on the sensors in my cameras.
  11. I would call Canon support, they will let you know if it's covered under warranty. I have had to send in my 30D (error message) and SX10 (sensor problem) and I had my cameras back in 2 weeks TOPS, usually less. yes, it sucks to have to pay to ship it, but I would much rather have a fully functional camera (and have repairs done before the warranty expires). They will send you instructions and then you can track the repairs online, usually once they have received your camera repairs and return shipping are usually completed in 7days, and they usually clean it up and service it for you while it's there. And neither repair cost me anything more than the shipping and insurance.
  12. Hi Shannon,
    Thanks for the encouraging information, I am glad to hear that Canon serviced your camera quickly and for free. My case is a bit less clear because the etching could have been caused by a manufacturing defect (their fault), or by improper cleaning (by the previous owner). But I am now leaning towards sending the camera for repair, even if I have to pay (ugh).
  13. Assuming it is etched, I agree with Shannon. Call Canon. If it isn't under warranty though, you'll be looking at a new sensor module I think. It's probably going to be pricey.
  14. How does Canon tell if my camera is under warranty? Is this something they have in a database, or do I need to obtain the date of purchase from the previous owner?
  15. If you're sure it was like this when you bought it, any chance you can return it to whoever you bought it from, and get your money back?
    I think it unlikely it left Canon with those marks. More likely the previous owner fouled it up while trying to clean it.
  16. Hi Alan,
    I have been in contact with the previous owner today to try to find an original receipt, but think it a but much to ask him to take the camera back. In an ideal world, I would have realized the problem before purchasing the camera. I do have the option to sell the camera as-is to someone who is less particular than me (or does not frequently use settings for which the problem appears).
  17. You have made your point clear that you found a problem with your camera and that are not happy about, so send it in! It sounds like this problem would not affect most people, but if it bothers you then my all means try to get it repaired.
  18. Phew, thanks Kevin. I think I've been thinking that all along, but just worried about sending it to Canon. I have a daughter due in two months, and am worried that Canon will still have my camera at that time. But from what everyone has said here, Canon is fast and I should just send it in and get it fixed. Thanks!
  19. Another possibility - there's a company that adapts cameras for infrared work, and they do 5Ds - basically it involves removing the sensor filter that is (presumably) the thing damaged in your camera. Since they remove these as their primary job, one supposes that they must have a whole stack of them laying around their shop. I checked their website and it includes the promising text
    "Scratched your stock camera sensor filter? We can fix it!"
    They quote $425 for the 5D. Not cheap, but beats buying a new camera. They're at http://www.lifepixel.com/index.html . They even have do-it-yourself tutorials if you're brave enough.
  20. Alan, that's an interesting though. But I think if Canon will not fix the camera under warranty that I would prefer to sell the camera as-is, letting the buyer know of the issue. Good to know that Life Pixel is available though, sounds like a cool service.
  21. Danny,
    Are you near a Canon Service Center? I've taken equipment to Canon in Irvine, Calif and they repaired it pretty quickly. If the service rep is really sensitive to your plight they can ask for a faster service.
    As for Canon having your camera when your daughter is born, if you think about it, there won't be a time in the next few years that you can do without your 5D. At what point in her life will you have a couple months (although it shouldn't take that long) to go without the camera. You have a good backup and if the 5D doesn't come back in time, you will still get the shots. This is why it is extremely important to get this done ASAP. Congrats on your daughter.
  22. Thanks John, I live near Santa Barbara, CA - so Irvine is likely the closest service center. Did you physically drop the camera off at the service center? I didn't know that was an option! Irvine is a serious drive for me (about 5hrs), but good to know they are sensitive to their customer's needs.
  23. I dropped it off there. You might call them up since the service time depends on how many cameras they have in for servicing (and if they have the parts in stock)

    I think this is their number: (949) 753-4200
  24. I'd say it's very unlikley that Canon would do a warranty repair. My guess is that it was most likely damaged by over-zealous cleaning or using the wrong solvent. I've never heard of a manufacturing defect resulting in anything like what you're seeing.
    I don't think the cover filters are cemented onto the sensor in the 5D (they are in the 1D series bodies), so they can be replaced. The high cost is mostly due to the amount of labor involved in getting at the filter assembly. My guess would be at least a few hundred dollars.
  25. Thanks Bob, I was a bit afraid of that. Good to know that the entire sensor will not need to be replaced. As far as I know, the sensor has been well cared for, but there is no way for me to be absolutely sure that the problem is due to a manufacturing defect. If I decide to send it in, I will do so with my fingers crossed!
  26. Do most people send in the camera with a lens attached, or just the body? I imagine that Canon provides instuctrions for this sort of thing...
  27. They prefer just the body and its cap - no strap, no batteries, no lens - the less you send, the less they are liable for keeping track of.
  28. zml


    Call Canon, if you are in the US 1-800-OKCANON, and set up your repair. Then send it in as instructed. It's all in the manual, by the way...
  29. Danny: I find it unlikely that the filter is etched. Glass is generally fairly resistant to etching (I learned it at UCSB!). I find it more likely that it is the residue of a cleaner that is not soluble in Eclipse 2. First off, you should ask the previous owner what he used to clean the sensor and use the same cleaner. If you have no success there I would try to remove this with Smear Away from Visible Dust--which, in fact can leave a residue not unlike what you have. But the Smear Away residue goes away with Sensor Clean, also from Visible Dust (Smear Away may not get this residue if the residue is old, thus use Smear Away first). Get the swabs that are recommended for these cleaners, to avoid making things worse.
  30. Danny: I find it unlikely that the filter is etched. Glass is generally fairly resistant to etching (I learned it at UCSB!). I find it more likely that it is the residue of a cleaner that is not soluble in Eclipse 2.​
    Same here. Etching requires some really strong chemicals like hydrofluoric acid. Who has access to chemicals like that knows that they would damage the filter (and the camera) and would not use them. Besides, the maps really do not look like etching to me. I can be obviously wrong, but I am confident that these are some cleaning residues and that professional Canon service might be able to clean them.
  31. I know you got a lot of answers on this but there were some coatings in earlier 5ds that could be sensitive to certain liquid cleaners. I had heard if your sensor has a yellowish/orange hue as you look at it that these were the sensor filter coatings vulnerable. I've only read of this but can't verify from experience. I would at least have a qualified person inspect it and get a quote for repair. If it is the filter it could be fairly expensive. I do wish you well on this. Let us know what you discover, it could be helpful to us 5D owners.
  32. Hi Everyone - thanks for the great responses. This is a newer 5D (serial # begins in 3), so I don't think I have the "thin coating" problem experienced by first generation 5D owners. Peter E and Mirek Elsner raise a very good point - that there may be residue from cleaning fluid that is not soluble in E2. I had assumed the Eclipse/E2 would disolve just about anything, but in hindsight that may not be the case. Also, I agree that etching a film does seem difficult. It does, however, remain possible that some residue lies under the front filter, thereby explaining why the Sensor Swabs were unable to touch it. I have decided to send to camera to Canon, and will update this page as more information becomes available.
    Thanks again to all who contributed!
  33. I would guess that some liquid made it between the sensor and the filter. Cleaning it would require to remove the anti aliasing filter first which you cannot do at home.
  34. I find it unlikely that the filter is etched. Glass is generally fairly resistant to etching (I learned it at UCSB!).
    The dichroic coating on the filter is not glass, it is usually organic thin film (in some cameras this layer also absorbs IR) which can be easily removed/etched/damaged, hence manufacturers do not recommend any cleaning procedure that involves physical contact with the filter, otherwise Canon, Nikon etc. would be selling their own magic cleaning solutions etc.
    Anyone who has doubt about this can try the following (knowing that it will destroy your camera!) try some high purity methanol and swab your sensor, then take a photo to see what has happened ;)
  35. Methanol is used in many of the sensor cleaning products. What are you talking about, Arash?
  36. > Anyone who has doubt about this can try the following (knowing that it will destroy your camera!) try some high purity methanol and swab your sensor, then take a photo to see what has happened ;)
    I have to say... I've swabbed my 5D sensor with high purity methanol several times, over three years, without any problems.
  37. Mark,
    You are wrong, methanol is only 1-2% of the E2 solution, the rest is IPA. 100% Pure methanol will dissolve the coating.
    http://www.photosol.com/documents/MSDS - E2.pdf
    Alec, where did you get 100% pure methanol? I have used JT Baker semiconductor grade methanol on a Nikon D80 and the sensor coating was fully dissolved. If you really want to make sure you destroy the coating you can use Acetone ;)
  38. Arash, I'm not wrong, thank you. Eclipse and e-Wipe are 100% methanol based cleaners. I've quoted a portion of the MSDS sheet to make it convenient for you...
    From the MSDS sheet:

    "Product Use :​

    Eclipse is intended for cleaning optical lenses such as those used in photography, binoculars, telescopes, etc. It
    is applied by dropping a few drops onto a clean PEC*PAD®; and wiping across the surface of the glass.
    e-Wipe is intended for cleaning critical optic CCD's, mirrors and lenses such as those used in digital camera and
    photographic equipment, binoculars, telescopes, etc. It is applied by opening the sealed packet and gently
    wiping in a single direction across the surface.​

    Section II Composition and Ingredient Information​

    Chemical Identity : 100% Methanol - CAS # 67-56-1"
    Now, if you have some data to back up your claims that 100% methanol will dissolve the sensor coatings, I'd sure like to be pointed in the right direction.
  39. Mark,
    Your are confusing two different products, The MSDS data sheet you attached is NOT for the eclipse E2 solution but for a different product which is an optical lens cleaner. These are two different products one for cleaning surface of glass optics and CCD optics (such as CCD cameras in optical scopes not DSLR sensors)
    This Eclipse E2, i.e. this product
    is meant for cleaning DSLR sensors with the following data sheet which you obviously did not read http://www.photosol.com/documents/MSDS%20-%20E2.pdf and it is not made of 100% Methanol, any ways I am not going to debate this, you can try pure ethanol on your sensor and see what happens.
  40. Sorry, I meant Methanol above. Here is spec from Eclipse E2 MSDS section, this is primary Ethanol solution which does not dissolve ITO or organic coatings.
    here is a link to manifacturer website
    9. Physical and Chemical Properties


    Clear, colorless liquid.​


    Mild pleasant whiskey-like odor.​


    Miscible in water.​


    0.79 @ 20C/4C​


    No information found.​

    % Volatiles by volume @ 21C (70F):​


    Boiling Point:​

    78C (172F) (ethanol)​

    Melting Point:​

    -114C (-173F) (ethanol)​

    Vapor Density (Air=1):​

    1.6 (ethanol)​

    Vapor Pressure (mm Hg):​

    40 @ 19C (66F) (ethanol)​

    Evaporation Rate (BuAc=1):​

    ca. 1.4 (CCl4=1) (ethanol)
  41. Au contraire, Arash...I did read the MSDS you referenced. I did not see where it states the composition is, as quoted directly from your post, "methanol is only 1-2% of the E2 solution, the rest is IPA".
    I am not going to debate this, you can try pure ethanol on your sensor and see what happens.​
    First, we were talking about methanol, not ethanol, and second, I could really give a rat's behind what you consider debatable. All I asked was that if you have information that shows that methanol will dissolve sensor coatings, then let's see it.
  42. Mark,
    Many DSLR sensors are coated by ITO (Indium Tin Oxide) as a dichroic and conductive film, if you are in this business you will immediately recognize that Methanol will dissolve ITO (and other organics such as photo resist and other refractive polymers), if not here is a link http://www.earthboundlight.com/phototips/indium-tin-oxide-sensor-cleaning.html
    you can also google search ITO to learn more about its properties.
    Using Methanol solution for ITO coated sensors will dissolve the coating and thus destroy the camera, I just wanted to point this out because I have already killed a Nikon D80 by using Methanol as I did not know there was an ITO coating at that time.
  43. In fact the damage caused to OPs sensor coating might have been caused by the previous user using a Methanol solution as it has partially dissolved the coating-so be careful!
  44. Arash, this is my last post about this. It's obvious you are making claims that you can't support. Your link only directs to an article written by a photographer who gives no causal fact that ITO is soluble in methanol. It offers no support for your claims and gives absolutely no proof of what you state. So beyond your anecdotal experience, I can't find anything that definitively states that ITO is soluble in methanol. Thanks for your very feeble attempt to enlighten those of us too dense to understand your brilliance. I'm done.
  45. Mark,
    I am really sorry I was not able to convince you! Not a problem, I understand as this was probably the first time you heard the word ITO. If you have a chemistry major in your family that you can trust maybe they can tell you more about ITO and effects of Methanol on it. Also try to familiarize yourself with Google Scholar search, it makes life so much easier :)
    I am sure you know more than than company who makes and recommends this solution, I suggest you email them and tell them that they are wrong in their product composition and recommendations and that they should use Methanol instead, they might pay you something in return ;) good luck!
  46. Mark and Arash, you are referring to two distinct products, I think, which are Eclipse and Eclipse E2. One (the original Eclipse) is apparently pure methanol, while the other is a mix of methanol, ethanol and isopropylalcohol. To my knowledge, E2 was developed because Sony wanted something milder for their ITO coated sensors (even though they did not find any damage caused by the original formula)
    Anyone who has doubt about this can try the following (knowing that it will destroy your camera!) try some high purity methanol and swab your sensor, then take a photo to see what has happened ;)
    I worked in chemical research for about 5 years, using various solvents on a daily basis and find it hard to believe that the sensor coating would be so easily soluble in methanol so that you can wipe it out - if at the same time it is insoluble in all of the following: water, acetone and isopropylalcohol mixed with methanol.
    I am sure you know more than than company who makes and recommends this solution, I suggest you email them and tell them that they are wrong in their product composition and recommendations and that they should use Methanol instead, they might pay you something in return ;) good luck!​
    An article on the manufacturers site indicates that methanol does not cause any problems with ITO coatings and in fact the E2 caused some issues with earlier 5D sensors, while the Eclipse (methanol only) did not cause any. They say E2 is gentler, that is all. Please correct me if I'm wrong.
    Many DSLR sensors are coated by ITO (Indium Tin Oxide) as a dichroic and conductive film, if you are in this business you will immediately recognize that Methanol will dissolve ITO (and other organics such as photo resist and other refractive polymers)​
    Indium tin oxide is not an organic compound.
  47. Mirek, thanks for providing the link and some sensible information.
  48. Mark Elsner,
    I pointed out that Eclipse E2 was different from original Eclipse in that it is a Ethanol solution instead of Methanol, please read the posts.
    Secondly I do research in semiconductor fabrication facility on a daily basis, we use Acetone and Methanol to lift off (i.e. dissolve) photoresist on Si wafers, Acetone will surely cause damage not only to moat sensor caoting but also to the interior of the camera as it dissolves any plastic.rubber part. Both Acetone and methanol are perfect organic solvents for most refractive polymers.
    By the whcih chemical research facility do you work at? What are your experties?
    I know ITO is not organic, Organic coatings that are used in camera sensors include infra-red coatings and are made of polymers of 3612 family.
    This is really beyon the scope of this forum but if you have the background and access to technical journals you should read this to learn more about effects of methanol adsorption on ITO.
    Study of methanol adsorption on mica, graphite and ITO glass by using tapping mode atomic force microscopy

    Li Wang, Yonghai Song, Aiguo Wu, Zhuang Li, Bailin Zhang1 and Erkang Wang[​IMG], [​IMG]

    Journal of applied Surface Science
    Received 13 March 2002; accepted 22 April 2002. Available online 12 June 2002.
  49. Sorry Mirek, I got your name wrong.
  50. Hi Arash,
    I worked in polymer chemistry. What is important though is that the original Eclipse is pure methanol and the manufacturer on their site mentions that it won't damage the ITO coating and even Alex - a poster in this thread - used methanol to clean his sensor with no damage. So there seems to be evidence that methanol works safely on EOS5D. If the manufacturer does not have any reports of damaged 5D sensors from methanol, I think it is extremely unlikely that Danny's sensor was damaged that way.
    I would be curious to know what the problem was and hope that Danny will report back after his camera returns from Canon.
  51. Mirek,
    There is one subtle difference between bulk material and thin films (the coatings are usually sub wavelength), stress and surface tension could be quite different in a thin film versus bulk, exposure to solvents can make various thin films to simply peel off, for example if surface interaction between the solvent and thin film is stronger than Van der Waals bonding between the film and the substrate.
    I don't know what exact coatings canon uses in their 5D and in what order but if there are some polymer layers there is always a risk unless the package is hermetically sealed in full. Just imagine if there is a layer of resist or polymer beneath the ITO coating, even if methanol is completely safe on ITO, the solvent may find a way to the polymer layer through microscopic openings on the side walls and "lift off" the layer.
    Here I am attaching a SEM micrograph for you, this is Silicon surface covered with a polymer (PMMA) that has had a thin layer of Pd (Palladium) deposited on it, very rigid metal very fine grain structure. The sample has been sitting in 100% Methanol for 10 minutes @24C, notice how the film has simply peeled off. If you gently scrub the sample these peelings will turn into large spots very similar to what the OP showed, of course I am not sure and unless the filter is inspected in SEM not much can be said but this is certainly a possibility.
    I just wanted to caution users to be aware, of course it is their sensor and their money, after all Canon does not approve this method of cleaning and the risk is on users.

  52. Arash, I think we digress from the original topic and there is no much value for other users. I will close it from my end by saying that
    1. The manufacturer of Eclipse apparently does not know about any 5D user who would have problems with Eclipse, i.e. pure methanol, one user in this forum used methanol on 5D and did not have any problems and there is not evidence that would prove any damage. So I, personally, would not be afraid to use pure methanol to clean my 5D/5D2 sensors
    2. If you repeated your test with your silicon sample and used Eclipse E2 instead of methanol, you would most likely get the same results
  53. I have sent my camera to Canon, and it seems as though they are going to do a warranty repair. However, they have not provided any details as to what was wrong with the camera and what they will fix. I'll keep everyone posted as things progress, but looks good for now!
  54. Canon returned my camera last night! In total, they had it about 10-15 days. They have replaced the sensor, under warranty! Here are the service comments:
    We have examined the product according to your request, and, it was found that the internal component was inoperative the image could not be displayed properly. The internal component (image sensor) was replaced. Other inspection and cleaning and parts replacements were carried out.
    I am very happy with Canon service and my "new" 5D. Thanks to all who contributed to this post.

Share This Page