Buying second hand, avoid low shutter count.

[rfdesigner]

Being a cheapskate I like to buy second hand.. I got my old 30D second hand and will probably get my next body second hand.

The shutter count is an important feature on any sales comments. However as an engineer I know how you get a usually get a "bathtub" failure rate curve.

Now I found a website showing shutter life:

http://www.olegkikin.com/shutterlife/canon_eos5dmkii.htm

I chose the 5DII as I've been wanting to get one.. just need to find the cash. Anyway what's interesting is that there appear to be several cameras failing at low shutter count. So I worked out from the data what the chance was that a known used working camera would fail in the next 10,000 shots.

(aplogies for the formatting.. the pertinent column is the right most, being the chance of failure in next 10k shots)


Number of shots on camera Recorded working Recorded failed Cameras surviving Failures in next 10k shots Chance of failure in next 10k shots
100 0 1 1526 26.3 0.0172
200 1 1 1525 25.5 0.0167
500 7 2 1523 24.2 0.0159
1000 3 2 1514 23.3 0.0154
2000 3 2 1509 23.8 0.0157
5000 32 2 1504 45.0 0.0299
10000 67 17 1470 13.0 0.0088
20000 114 13 1386 10.7 0.0077
35000 148 16 1259 5.3 0.0042
50000 122 8 1095 3.4 0.0031
100000 296 17 965 1.2 0.0012
150000 262 6 652 1.5 0.0023
250000 233 15 384 0.6 0.0016
500000 93 15 136 0.1 0.0006
1000000 17 4 28 0.0 0.0011
2.00E+006 4 3 7


By the looks of things I need to look above 50k shutter actuations.. and there's nothing to be feared by looking well beyond.

 

[Mt Spokane Photography]

This is a badly flawed study, since you claim to be a engineer, you should recognize that.

There is no testing, just info from random people with failed shutters searching for information about them, and then signing up to post. How did the information get confirmed? Anyone can post anything they want to make up, and a data point of 1 or 2 is less than useful. There is zero reliable data here.

 

[jdramirez]

I forgive the sample size, because it isn't his study... Though obviously more data points are welcome when drawing a conclusion.

Once I tried to lock up the motor of my 60d and the next day I tried shooting and big surprise, the motor was locked up... So I thought I had something wrong with the camera.... until I remembered and all was well.

So it is possible someone,a newb, didn't actually have their shutter broken... if this was a detailed report from Canon, I think the conclusion would be very interesting.

 

[Hesbehindyou]

This is a badly flawed study, since you claim to be a engineer, you should recognize that.

Oooh, that comes across as quite catty, especially when you say "claim to be".

There is no testing, just info from random people...

Random would be great but these people are not random!

...with failed shutters

The vast majority had working shutters.

a data point of 1 or 2 is less than useful.

There were over a thousand data points.

The major, overwhelming, flaw in this study is that it uses a self-selecting sample, can't believe you missed that!

To the OP: Hadn't given it much thought before, always assuming the lower the shutter count the better, but finding where the bathtub curve 'bottoms out' makes great sense. Thanks for the analysis, however flawed the source data. Pity you had Mr Negative try to kill the thread before it started :-/

 

[IMG_0001]

Probably not of great statistical value and certainly not from an independent and identically distributed random sample. Yet, even without referring to this 'study', I would certainly not worry much about buying a camera with a few tens of thousands of shutter actuation. Even more so for pro level equipment such a 5D or 1D series body. Failure of mechanical parts subjected to dynamic loading tends to come with significant scatter. As a result, if Canon is to have a good reliability (which I doubt would be less than 95% survival with 95% confidence) and limited warranty claims, I would not be surprised that the median life would be close to a full order of magnitude more than the warranty.

On the other hand I would be more worried of a camera with a high video usage because of the heat generated in video. Nevertheless, even high video usage would not be a dealbreaker for me if the price was right.

 

[Tinky]

Probably not of great statistical value and certainly not from an independent and identically distributed random sample. Yet, even without referring to this 'study', I would certainly not worry much about buying a camera with a few tens of thousands of shutter actuation. Even more so for pro level equipment such a 5D or 1D series body. Failure of mechanical parts subjected to dynamic loading tends to come with significant scatter. As a result, if Canon is to have a good reliability (which I doubt would be less than 95% survival with 95% confidence) and limited warranty claims, I would not be surprised that the median life would be close to a full order of magnitude more than the warranty.

On the other hand I would be more worried of a camera with a high video usage because of the heat generated in video. Nevertheless, even high video usage would not be a dealbreaker for me if the price was right.

And of course the laugh is that most shutter count programmes do not take account of video use.

When I sold my 600D it was in very clean condition with a 'shutter count' of around 6'000.

Did I explain to the vendor that I shot mainly video with it, and that in there were probably 20'000 actuations on top of that which weren't recorded (still wear on the mechanisms). Of course.

The same way as when I had a big timelapse project on the go, I didn't buy a 550D new, run the shutter until it failed (a week old, but must have been a bit over the 100'000 mark) and then took it back to the shop for a replacement camera "IT'S LESS THAN A WEEK OLD" I didn't say as I didn't stamp my feet.

 

[pwp]

...on the other hand I would be more worried of a camera with a high video usage because of the heat generated in video. Nevertheless, even high video usage would not be a dealbreaker for me if the price was right.

And of course the laugh is that most shutter count programmes do not take account of video use.

Interesting. I'm looking at a pre-owned 5D3 which the seller says has a 5500 shutter count and 8290 minutes of video usage. On the face of it, it looks like a very good deal. What are the genuine implications (if any) of 8290 minutes of video usage?

I always find the shutter count threads perversely fascinating. It's possible I have been absurdly lucky. As a heavy shooter who runs bodies well up into the hundreds of thousands before retiring them, the only two shutters I have ever had to replace are on a film body where I punched my thumbs through a shutter while changing film in a bumpy helicopter flight shooting yachts during the start of an ocean race, and a 5D3 which was DOA. Canon swapped the body without question.

I don't personally know a photographer who has had a shutter fail. I know it does happen on a daily basis, my CPS will verify this, but I get the feeling there is far more significance put on shutter life than is warranted. Nevertheless, why not buy a low shutter count body? Or a low mileage car? The entire mechanicals have simply had less wear and tear.

-pw

 

[TeT]

Certainly not as crisp and detailed a study group as you would wish if you were going to publish. Though there is some merit.

a Bad shutter is certainly likely to fail earlier than later. Just as a worn out shutter will fail later, shutter in range of 20K to 50K is likely safe by that reasoning. Of course a camera approaching 50K (depending upon the user) will start to exhibit other wear and tear.

Obviously the OP doesn't require a certified study group. Opinions and loose info will do:

Welcome to CR, you are in the right place...

just a thought (loose at that)...

 

[agierke]

I had a shutter literally explode during a shoot with my 5dC. It failed at 186,000 actuations.

The cost of replacing a canon shutter is cheap. Mine was replaced for around 200.00 by canon. That's peanuts relative to the added years of use I got out of it. For that reason, I no longer care about shutter counts. If it fails I'll have it repaired quickly and cheaply. No big deal.

Much ado about nothing.

 

[jdramirez]

Here to fore I've sold my two prior bodies after two or so years... But I think I'll keep the mkiii for a good long while. I haven't even come close to wearing my shutter, but when I do, I'll get it fixed for less than the cost of a mediocre point and shoot.

I had a shutter literally explode during a shoot with my 5dC. It failed at 186,000 actuations.

The cost of replacing a canon shutter is cheap. Mine was replaced for around 200.00 by canon. That's peanuts relative to the added years of use I got out of it. For that reason, I no longer care about shutter counts. If it fails I'll have it repaired quickly and cheaply. No big deal.

Much ado about nothing.

 

[Tinky]

...on the other hand I would be more worried of a camera with a high video usage because of the heat generated in video. Nevertheless, even high video usage would not be a dealbreaker for me if the price was right.

And of course the laugh is that most shutter count programmes do not take account of video use.

Interesting. I'm looking at a pre-owned 5D3 which the seller says has a 5500 shutter count and 8290 minutes of video usage. On the face of it, it looks like a very good deal. What are the genuine implications (if any) of 8290 minutes of video usage?

If they shot entirely in29minute chunks then that is a mere 286 shutter and mirror actuations.
If they shot in 1 minute chunks then thats 8290 actuations.
If they shot in 6 second chunks thats 82,900 actuations.

Of course most folk shoot a mix of clips from long interviews to brief cutaways. It's impossible to tell the amount of use in terms of shutter life by minutes.

150 hours? Sounds like it's had a professional working life, treated as a tool of the trade, rather than say, kept on a velvet cushion, and only used on sundays by a doting amateur. Your call. The 5d3 is still current, the question to my mind would be... why are they selling it?

Not very helpful. In professional tape camcorders you could display 'head hours' which was handy as a meter of use / servicing, as tape had mechanical contacts and wear,oxide dirt etc. For flash storage cameras like a 5d3 it's less helpful to know

 

[sagittariansrock]

Being a cheapskate I like to buy second hand.. I got my old 30D second hand and will probably get my next body second hand.

The shutter count is an important feature on any sales comments. However as an engineer I know how you get a usually get a "bathtub" failure rate curve.

Now I found a website showing shutter life:

http://www.olegkikin.com/shutterlife/canon_eos5dmkii.htm

I chose the 5DII as I've been wanting to get one.. just need to find the cash. Anyway what's interesting is that there appear to be several cameras failing at low shutter count. So I worked out from the data what the chance was that a known used working camera would fail in the next 10,000 shots.

(aplogies for the formatting.. the pertinent column is the right most, being the chance of failure in next 10k shots)


Number of shots on camera Recorded working Recorded failed Cameras surviving Failures in next 10k shots Chance of failure in next 10k shots
100 0 1 1526 26.3 0.0172
200 1 1 1525 25.5 0.0167
500 7 2 1523 24.2 0.0159
1000 3 2 1514 23.3 0.0154
2000 3 2 1509 23.8 0.0157
5000 32 2 1504 45.0 0.0299
10000 67 17 1470 13.0 0.0088
20000 114 13 1386 10.7 0.0077
35000 148 16 1259 5.3 0.0042
50000 122 8 1095 3.4 0.0031
100000 296 17 965 1.2 0.0012
150000 262 6 652 1.5 0.0023
250000 233 15 384 0.6 0.0016
500000 93 15 136 0.1 0.0006
1000000 17 4 28 0.0 0.0011
2.00E+006 4 3 7


By the looks of things I need to look above 50k shutter actuations.. and there's nothing to be feared by looking well beyond.

There are a couple of things to consider when using population statistics:
1. How do the odds favor you? As you interpret, if it was a faulty (or suboptimal, in any case) shutter assembly to begin with it would fail below 50K, and if it survives that number there is a good chance it would serve you for quite sometime. Here, you are correct to look for a 50K+ camera.
2. On the other hand, yours might be the one to break. Question is, how much did you save by buying a used (or well-used, in this case) camera to dole out the cost of shutter repair? Here too, it makes sense to go for a high-mileage camera as the cost of a potential repair is offset by the low price. You buy a "5K only" camera for $ 1300 and then end up paying $ 200 for repairs- that would suck.
So I support your conclusion. However, I wonder- considering the low price differential between the 5DII and a new 6D, the much better features of the latter, and the fact that a new camera will come with a couple of years of warranty (the second year through your credit card), why are you considering a 5DII?

 

[danski0224]

At some point it won't matter when the camera is obsoleted by Canon or parts are no longer available.

 

[Andrew Davies Photography]

I have a 5d3 with 50k on it , a 5d2 with 170k on it which works perfect and I have just sold for a good price and a 5d2 with over 100k on it which i still use as second camera. As said already its not that expensive to get the shutter replaced so it should really not be that big a deal when choosing second hand. Just make sure if you get a high shutter count that its priced well enough to allow for a new shutter if it goes against one that has a low shutter count.

www.andrew-davies.com

 

[rfdesigner]

Thanks for the overall feedback.

I agree this is a long way from an ideal study. If I had wanted to publish in a journal I would of course have raised the bar significantly... but this is a forum.

What I'm seeing on the second hand market is cameras with say 50k shutter count being marked down by £100~£200 compared to a sub 10k count example, even if in pristine condition.

Given that the probability of failure seems to fall off after a few 10ks of activations (not unexpected really), and a replacement shutter is not too expensive, a higher shutter count is not a worry and if it leaves enough money for a new shutter should it fail then thats better than a guarentee as there's no time limit.

The question in my mind now is "what video counter?".. does the 5DII have one?.. as that would be a item to check.

 

[tinman0]

In my opintion, it's too difficult to just take actuations as a measure of wear on the shutter. The problem (again in my opinion) is that the shutter is an incredible piece of presicion engineering that is probably prone to knocks from the camera.

So whilst 50k may sound like some sort of sweet spot, if it's spent it's life being bashed around, then that shutter (again in my opinion) is closer to throwing an error, for instance the blades are very slightly out of alignment - maybe they have a bit more drag and create extre wear, the actuator components are a little bit out of alignment, not to mention the mirror box components and so forth.

For instance, I bought a completely beaten up 50D with 130k on the clock. It died at 140k (two days into a job). It was repaired for £150, and it then put on 430k images before the actuator died.

The replacement 7D did 290k in 7 weeks (iirc) without a single problem. And somewhere round here, I have a Rebel XTi with 350k (although it sounds a little sick).

Thing about the 50D and 7D - they live in a bag, they are never put down on hard surfaces (the 7D has been put on an unprotected hard surface 4 times in 8 months). And the Rebel doesn't exactly get thrown around. I am convinced it's the little knocks etc that wear hard on the camera.

So my advice, buy a low shutter count as the chance of it being given a hard life is minimal. And pay extra to anyone offering it's original box - as that's where the camera has spent most of it's life.

 

[Mt Spokane Photography]

Thanks for the overall feedback.

I agree this is a long way from an ideal study. If I had wanted to publish in a journal I would of course have raised the bar significantly... but this is a forum.

Since you made it a point to state that you had engineering expertise, you raised that bar by implying it would withstand engineering scrutiny, but, IMHO, it doesn't. As a licensed registered professional engineer, I found that comment to reflect poorly on the engineering profession.

There are many reasons for camera failures, and shutters are only one of them. I suspect, but have no actual numbers that the average consumer camera has well under 15,000 shutter cycles, I've bought two dozen and none of them had nearly that many when I later checked them, and none failed. Without accurate numbers and information about the camera population, its impossible to put the data into perspective. If, as I expect, 90+ % of the cameras of a given model have under 15000 actuations, then the numbers do not make sense. We don't know.

Reliability does indeed involve infant mortality where failures happen more frequently after a new product is put into service. Usually, electronics is the most affected, but mechanical items suffer from the same issue.

Heat is the big reason driving failures, but cold can be bad as well. The thermal expansion / Contraction can literally tear things apart in short order. Product wearout usually happens after a long period of use. Obviously, vibration and shock can also destroy a item, but heat is a killer. A camera that is left in a hot car can have its life shortened considerably.

Then, there is the sensor. They tend to get more hot pixels over time, so a older one will suffer. There is no information about newer sensor designs, so hopefully, that issue is being reduced.

When buying used, I would have no concern about a camera 6 months old with a few hundred cycles, because the electronics has been burned in, and any mechanical issues should have been found. The one area where there would be no information is exposure to high temperatures as found in a car on a hot summer day.

 

[canonvoir]

So many statistics experts.

Maybe not peer reviewed worthy but a good observation.

Thumbs up.

 

[rfdesigner]

Thanks for the overall feedback.

I agree this is a long way from an ideal study. If I had wanted to publish in a journal I would of course have raised the bar significantly... but this is a forum.

Since you made it a point to state that you had engineering expertise, you raised that bar by implying it would withstand engineering scrutiny, but, IMHO, it doesn't. As a licensed registered professional engineer, I found that comment to reflect poorly on the engineering profession.

There are many reasons for camera failures, and shutters are only one of them. I suspect, but have no actual numbers that the average consumer camera has well under 15,000 shutter cycles, I've bought two dozen and none of them had nearly that many when I later checked them, and none failed. Without accurate numbers and information about the camera population, its impossible to put the data into perspective. If, as I expect, 90+ % of the cameras of a given model have under 15000 actuations, then the numbers do not make sense. We don't know.

Reliability does indeed involve infant mortality where failures happen more frequently after a new product is put into service. Usually, electronics is the most affected, but mechanical items suffer from the same issue.

Heat is the big reason driving failures, but cold can be bad as well. The thermal expansion / Contraction can literally tear things apart in short order. Product wearout usually happens after a long period of use. Obviously, vibration and shock can also destroy a item, but heat is a killer. A camera that is left in a hot car can have its life shortened considerably.

Then, there is the sensor. They tend to get more hot pixels over time, so a older one will suffer. There is no information about newer sensor designs, so hopefully, that issue is being reduced.

When buying used, I would have no concern about a camera 6 months old with a few hundred cycles, because the electronics has been burned in, and any mechanical issues should have been found. The one area where there would be no information is exposure to high temperatures as found in a car on a hot summer day.

In my case RF = radio frequency, and I've spent the last 15 years playing silicon development games... design, test, compliance test, debug and so on (in addition to a lot bit of PCB level electronics research, design and development experience). I'm currently a principle RF engineer in another R&D role, as I work in the UK there has never been a requirement to get chartered status.. that's just the way things work here in my industry.

Early life failure.

My experience is that the "electronic" failures are really mostly software/firmware, unless you're talking electrical (higher power stuff). I'm not aware of any DSLR needing an electronics hardware related recall. Lensrentals have the D800 down as failing most for a battery door and the 5DIII failing due to the memory card pins.

Silicon in hot cars:

Silicon itself will be fine. It all comes down to the question of being used or not. A TV hung above a mantlepiece with a woodburning stove will kill it quite quickly as it will be operating. There can be an 80 or even a 100C temperature gradient between die temperature in product and ambient air temperature... industrial electronics needing to operate to 85C will often have fans etc and much much better cooling. At the end of the day hot silicon will age more quickly.

I'm not saying a camera left in a car in Texas won't die, and I agree thermal cycling will break things, but any camera left in a car will be off.. so cooked electronics really isn't something to worry about. Most silicon devices storage temperatures are rated -55 to +155C even if their operating temp is only 0~70C

If you want to know more about it look up JEDEC HTOL testing (something I've had to drive on a number of chips)

Bad pixels.

CMOS sensors and CCD sensors are different, CCD sensors will usually aquire bad pixels over time due to cosmic ray impacts on the delicate structure. CMOS sensors are far more robust, my 8 year old CMOS sensored 30D has far far fewer hot pixels than my much newer CCD camera.. and old astro shots I took with the 30D still show most of the same hot pixels. Any camera might get a devestatingly bad pixel that kills a column (which is what will always be noticeable) at any time, just one bad hit. The sensor itself doesn't get more sensitive to cosmic ray hits over time, so if you're picking up an old camera, one test shot will show you if there's a problem, if there isn't then you as good as if it were new.

Ulitimately what I'm looking at is cameras that appear almost unused, in pristine condition vs cameras that look well cared for with a few miles on the clock and are therefore ~£150 cheaper.

I'll take the cheaper run in camera thanks.

 

[IMG_0001]

Engineers, we are fussy nitpicking little critters aren't we? Looking back, I feel like most great things we accomplished were made to prove someone else wrong...

*EDIT: please take this post with a pinch of salt.