Based upon the diagrams in the Galbraith article even a slight misalignment of the OBLONG CONCAVE secondary mirror wil have large effects on the image since this is essentially an offset folded telescope with an ocular lens to image onto the AF sensors.
Thermal problems would perhaps tilt the mirror as expansion/contraction mismatches play in to the submirror supports to generate up/down tilt.
Alignment and spherical abberration would play in a lateral or side-to side image problem. This would cause the sensor to not "see" well in terms of contrast both slowing and degrading the AF performance since the processor has to "guess" more about where sharp focus really is... thus leading to the camera focussing "behind" the subject coming toward the camera if time is a problem.
If the problem is S.A. then it all will depend on the contrast ratio that triggers an "in focus" signal to the software. S.A. will "smear" the spot and since the mirror is wider than it is tall, it will likely smear it more laterally than vertically leading to a difference in the distance where the vertical sensor and the horizontal sensor in the "cross sensors" believe the image is at maximum contrast. Canon -- if you are listening -- if your vertical and horizontal sensors are not seeing a diagonal line as in sharp focus at the same point then this is likely the problem. (Note that the horizontal sensors are likely to be the more sensitive ones as the mirror is not circular and the effective aperture in the horizontal is larger than in the vertical.....)
This can be corrected by replacing the image transfer lenses but these are generally TINY and quite difficult to get "right" to correct S.A. generated further up the optical path. In a telescope this would be done by a correction plate (essentially a very accurate aspheric lens) BEFORE the ocular.... There is no place to put that here so either they have to replace the concave submirror or they have to replace the lenses....
If I was designing this baby I definitely would not have done things this way... Making the submirror concave just about requires it to be an asphere or demands a correction plate on top of the relay mirror.... You want to do the correction at a distance where things are still pretty large since it will have to be an aspheric element -- remember that these are sort of expensive even if they are molded out of acrylic (I would prefer glass but....).
Right now I'd hate to be the guy that sucessfully argued for this mirror to be concave over at Canon.....
Sincerely,
Grover Larkins
As I have said before -- it ain't fixed 'til it's fixed and this puppy just ain't fixable without a major modification -- Canon needs to stop producing this abortion, stop selling it, re-design the thing and replace all of the ones out there currently with new product... or face a class action lawsuit for fraud....
Based upon the diagrams in the Galbraith article even a slight misalignment of the OBLONG CONCAVE secondary mirror wil have large effects on the image since this is essentially an offset folded telescope with an ocular lens to image onto the AF sensors.
Thermal problems would perhaps tilt the mirror as expansion/contraction mismatches play in to the submirror supports to generate up/down tilt.
Alignment and spherical abberration would play in a lateral or side-to side image problem. This would cause the sensor to not "see" well in terms of contrast both slowing and degrading the AF performance since the processor has to "guess" more about where sharp focus really is... thus leading to the camera focussing "behind" the subject coming toward the camera if time is a problem.
If the problem is S.A. then it all will depend on the contrast ratio that triggers an "in focus" signal to the software. S.A. will "smear" the spot and since the mirror is wider than it is tall, it will likely smear it more laterally than vertically leading to a difference in the distance where the vertical sensor and the horizontal sensor in the "cross sensors" believe the image is at maximum contrast. Canon -- if you are listening -- if your vertical and horizontal sensors are not seeing a diagonal line as in sharp focus at the same point then this is likely the problem. (Note that the horizontal sensors are likely to be the more sensitive ones as the mirror is not circular and the effective aperture in the horizontal is larger than in the vertical.....)
This can be corrected by replacing the image transfer lenses but these are generally TINY and quite difficult to get "right" to correct S.A. generated further up the optical path. In a telescope this would be done by a correction plate (essentially a very accurate aspheric lens) BEFORE the ocular.... There is no place to put that here so either they have to replace the concave submirror or they have to replace the lenses....
If I was designing this baby I definitely would not have done things this way... Making the submirror concave just about requires it to be an asphere or demands a correction plate on top of the relay mirror.... You want to do the correction at a distance where things are still pretty large since it will have to be an aspheric element -- remember that these are sort of expensive even if they are molded out of acrylic (I would prefer glass but....).
Right now I'd hate to be the guy that sucessfully argued for this mirror to be concave over at Canon.....
As I have said before -- it ain't fixed 'til it's fixed and this puppy just ain't fixable without a major HARDWARE modification -- Canon needs to stop producing this abortion, stop selling it, re-design the thing and replace all of the ones out there currently with new product... or face a class action lawsuit for fraud....
DavidP wrote:
Ummm, you CAN take the camera for a test drive. Buy it from a place that has a return policy, then TEST it.
Nowhere in Australia is there a store that will allow me to take a NIB camera out for a few hundred test frames then bring it back for a cash refund.
They will give me a store credit though.
If they had a full cash refund policy do you think I would own any equipment? everytime I needed something I'd go buy it use it for the weekend then ask for a refund monday morning. So would thousands of others. That way I'd always use the newest gear suffer no depreciation and not have to worry too much about theft and insurance.
If you ever open a store like this please let me know.
Simple fact is you buy this camera at your own risk knowing the issues it has. I myself for my needs would not chance it. Even for $500, I'd pass on it.
Thats not true, it is not a demo days it is a return policy. The M3 I bought I could return in 2 weeks if there was a problem. (and of course I did not trash it.) When I brought the M3 back they did not have one to exchange with but told me I could when they got one in. I did and it also had a problem so they gave me the option of cash back or try again. Well the blue dot came in and I was given that to try and it works much better. Some places have up to a 30 money back if not satisfied.
larkinsg wrote:
(Note that the horizontal sensors are likely to be the more sensitive ones as the mirror is not circular and the effective aperture in the horizontal is larger than in the vertical.....
Grover - could this explain a variance in focusing ability/consistency when puting the camera in portrait mode from landscape? Mounted on a tripod, I did the focus adjustment point on a lens. But then to test, I turned the lens to portrait, and it required a different adjustment to be accurate. I repeated 3 times, and each time, I had to adjust differently depending on the orientation of the camera. It was quite confusing....Show more →
darryn patch wrote:
Once again, I am reading about buying decisions based solely on reading comments from users that are having problems rather than trying it for themselves. There are plenty of us that are not, but don't necessarily post about it every day.
****************************************************************************************
This type of comment I find intriguing.
If your looking at a car that you've heard has major flawes and it was going to cost you plenty would you buy it? given that you can not get a refund and you can not take it for a test drive, It just sits there in the window when you buy it your stuck with it moneys gone no refunds.
Do you relly expect me to beleive you when you say you would buy it after hearing all the issues from other owners of that car who are stuck with a car that wont go? Sure some of the cars have no probs but many do would you buy that car??
I have no problems returning defective merchandise (including cars). Mine, however, has proven not to be defective. Not sure I understand about the "Major Flaws that would cost you plenty" reference... To each their own, and I respect your decision not to purchase. I am just stating that I am more than happy with my Mark III thus far...(had it since July). I know of many that feel the same way but are not constantly posting about it.. (and I AM NOT suggesting that those that do have problems, not voice them....) It just seems to me this is like most thing on the Internet. Those that have problems post about them passionately and those that do not, post much less frequently and with less urgency.
My post originally was placed due to comments by those that have never owned the camera or used one. Solely basing decisions by Internet forum comments doesn't seem prudent to me nor worth posting negative comments about. i.e., (in another car reference), I wouldn't go out to a forum about Ferrari's and say I'd never buy a specific one because of x, y and z if I had never owned or driven one.
Enough from me on the subject. Canon definitely has a problem and hopefully it'll be resolved fairly. I'll be sending mine in as soon as I have some downtime...Not because I am having problems, but because I do want the extended warranty and the new assembly in case I decide to sell someday.
Numfar wrote: larkinsg wrote:
(Note that the horizontal sensors are likely to be the more sensitive ones as the mirror is not circular and the effective aperture in the horizontal is larger than in the vertical.....
If you are focussed on the same exact point on a fixed target, YES!
I also noted this but was not going to mention it until AFTER someone else did 1st....
??
