As a working pro, dedicated amateur, or just plain photographer, the last thing you want is a dead camera, especially when it's through no fault of your own. Failures are part and parcel of the working life of a camera, however, what can the the Soviet T34 tell us about this?
Any mechanical or electronic system will fail at some point. That is the nature of designing and manufacturing products that perform multi-faceted and involved tasks. When the number of parts, and so manufacturing processes, increases — necessitating a multi-stage production line — then the potential for the product to fail also increases.
Tank production during World War 2 is a good proxy to understanding complex mechanical systems. By the end of the war, production looked like this: U.K. (36,000), Germany (47,000), U.S.A. (102,000), and Russia (106,000). What is remarkable is that the Soviets had little money, a moderate labor pool, and relatively little coal and steel. In stark contrast the U.S.A. had a significantly higher GDP, a large labor pool, and lots of coal and steel. In between, Germany was wealthy, with a relatively smaller labor force, but reasonable quantities of coal and steel. How then did Russia miraculously manage to produce so many tanks?
The answer lies in what was a modern approach to manufacturing: targeting a design life and ruthless cost reduction. The Soviets looked at the lifespan of a tank on the eastern front and discovered it was 6 months, which reduced to just 14 hours in combat. In short, any design that extended its utility beyond these limits was an over-design. This became their target lifespan and as a result the engine was designed to last just 1,500 km; to lower costs they opted to select simpler designs, cheaper materials, and innovative production. In fact, the T34 was subsequently reduced to just 5,461 parts, with the cost of production halving from 1941 to 1943. "Money is time" — the Soviets flipped this saying on its head and managed to reduce production to about 40,000 hours.
In contrast, Germany designed and produced a high quality tank. Rather than opting for a rolling production line, it was built in-situ (station based.) Whilst the Soviets' had a total of 3 tank designs, the Nazis had 14 variants for the Panzer 3 alone. Not only that, they liked tinkering, with the Henschel factory introducing 250 modifications to the design over a 2 year period. As a result of this, the Panzer 3 took some 300,000 hours to produce! By way of comparison, the U.S.A. fully embraced production line assembly, with a large and well organized workflow that reduced the time to produce a Sherman to just 10,000 hours!
Mean Time Between Failure
When products get small, utilize a range of different manufacturing materials, have a large number of parts, and incorporate micro-electronics, then the complexity significantly increases with greater potential for failure. Cameras epitomize all of these elements in bucket loads and end up being used in harsh and stressful environmental conditions, yet photographers — fairly or unfairly — expect zero down time. The T34 had a design life of just 6 months and production was oriented around this, although the actual failure rate is unknown. The starting point to understanding camera failure is termed "Mean Time Between Failure" or MTBF.
At its simplest, this describes the amount of time you can expect before a product fails, at which point it needs repairing. Taking a fairly simple example — a bicycle — how many riding hours would it take before there was a fault that needed repairing? The most likely scenario is a puncture, but as time goes on you have the added complexity of other components failing such as brake blocks wearing, cabling snapping, or chains breaking. That is why you have a servicing schedule and the MTBF is designed to reflect, for a serviced system, the expected longevity to failure.
When you design a product you might well have an MTBF you are targeting, but the actual value can only be calculated from known errors that are reported and these are generally not released by the manufacturer. One exception, however, are often vehicle safety tests; in the U.K. all tests are in the public domain which means you can compile a list of which vehicles failed and why, although that doesn't necessarily equate to MTBF for a well serviced vehicle.
This highlights that for complex systems comprised of different components, the product is only as reliable as the weakest component. In cameras, that may well be the shutter as it is both a high precision and high use component. The additional likelihood of failure of other components adds to this, but are generally much lower. In fact, manufacturers do change the shutter design for a target market. Nikon's D6 is "tested to 400,000 actuations", whilst the D3500 is down at around 50,000.
It also highlights that a product return doesn't necessarily mean a product failure for two reasons. Firstly, it you damage the camera then whilst that may be a fault it is externally attributed, rather than an intrinsic failure. Secondly, a component may fail that isn't a requirement for the operation of the camera. If the rubber on the handgrip detaches or LCD screen stops working, whilst the latter is more serious you can still use the camera. It's also worth remembering that where you have a product that isn't repairable it's called the Mean Time to Failure (MTTF), at which point you through it away; memory cards are a good example of this.
What It Means for Cameras
As a working pro you want a camera to be usable whenever you need it, although downtime may well mean that you have windows of opportunity for servicing or repair as necessary. What you don't want is failure on a job, which is obviously the reason you carry a backup camera (and battery, memory card, flash etc). However you largely base your choice of camera on it being "pro spec" in some form, as well as which camera system you shoot. In short, you don't know how reliable a camera or lens is, rather buying it because you shoot Nikon or Canon and need that product.
Manufacturers obviously do have an understanding of the failure rate of their products and these sometimes lead to a product recall. Nikon had oil spotting on the D600, as well a faulty shutter mechanism on the D750, whilst Apple was infamous for antennagate. Perhaps the biggest potential source of advice is from large independent repair centers. For example, LensRentals has reported on "weeks to repair" for their sample of 12,000 lenses representing 2,000 repairs, however bear in mind this is for heavy and hard use reflecting expected failure as well as damage inflicted. End users would expect a much longer serviceable life. You can also get an idea of shutter life at the Shutter Life Expectancy Database where end users enter the number of actuations for their current camera; for popular cameras this can give you a really good idea of failure rates, although it also shows most photographers don't shoot many photos!
This highlights the importance of servicing for heavily used equipment. For cameras, the shutter count can give you an indication of use, but it won't tell you how many flights the camera took, how many times it was thrown in to the back of a car, or how many shoots were on a beach. All it is is a proxy. Perhaps a better marker is who you buy it off, although if it's from a used store you won't know that.
All of which brings us back to the T34. It had a design life of some 6 months or 1,500 km, which was intended to maximize production for the intended use (time is money!). In the commercial world, the decision is to build to a specific budget and, like the Soviets, a manufacturer will look to simplify, select cheaper components, and lower production costs. If you can do that, whilst maintaining MTBF, then you have a winning design.
Body image by kteague and used under Creative Commons. Lead image by Pavlofox and used under Creative Commons.
My 88-year-old Rolleiflex still works like it came out of the factory last week. Will it fail one day? Sure, but I'm not holding my breath :)
Ok but how many Tiger tanks has it had to fend off?
Wouldn't a Rolleiflex be fighting Shermans and T34s?
I realize the T-34 reference was mostly click bait to pull in “World of Tanks” players, but it still seems like a labored comparison. After all, most T-34s' service life was not limited by component failure but by enemy action, right?... whereas it's uncommon to have to replace a camera because it was hit by a Stuka. I would guess that even professions seldom replace cameras because a key component wears out... it's more likely because of some technology shift (“I could get more salable shots if my camera had eye-detect autofocus” or “A lot of my competitors are offering 4K video now, so I need to match them”), right?
It was a stretch...
So the MTBF for cameras should be adapted by manufacturers based on the expected economic life of the camera -- when it's expected that its users will replace it by newer models with such upgrades that they can't be ignored.
The cameras enemy that will knock it out of the field, is the future camera model that will replace it.
In that sense the comparison can be stretched to fit. ;)
Oh the merits of soviet production... Thing is, the T34 was pretty much the Zenit of the tank world. Sure, people love to claim how their Zenit is "built like a tank" and so solid. But frankly, look at what your average surviving Zenit looks like these days... most of them have de-silvered prisms, rotten shutter curtains and stuck mechanics. Stalin just decided he wouldn't care about how many soldiers he'd lose along the way. Not that the Nazis had more respect for human life, they just didn't have them in such quantity to flood their enemies over.
As for that shutter life database, well, I would be surprised if it's not very strongly biased towards gearheads that spend much more time fondling the newest and fanciest they just bought rather than being out in the big wild world photographing something else than test charts.
That shutter life database, what does that actually say about shutter life?
Are you supposed to only enter there the number of clicks of your camera after the shutter has failed? If yes, and it's verifiable, then it means something.
Otherwise, all it means is that "this camera has survived at least X clicks without failure". Doesn't say anything about when it'll fail.
I found the comparison of Russian, German and US approaches to WWII tank engineering and production quite interesting even if it doesn't perfectly correlate to camera design/production.
The T-34 was brilliant in that it used sloped armor in the front. The Sherman was fast and could turn on a dime, but that gasoline engine....oy! And it was under gunned. The Germans had the firepower advantage through most of the war. The evolution of the main battle tank over the past 25 years has been remarkable, just like camera evolution. See how I did that? :-)
Apparently the Germans called them Tommy Cookers due to their propensity to "brew up" when hit. A Tiger, with its vastly better armour and gun could take on multiple Shermans and win. Unfortunately for the Germans the US had multiple x1000.
And those magnificent German tanks had a rough go when the mud and snow hit them on the Eastern Front. All of that terrific armor and firepower was very heavy and with that, the tanks got bogged down.
In the end, it was Hitler's worst strategic error attacking Russia, besides starting the war in the first place.
So is it your contention Nikon are over engineering their cameras because their business will fail?
Why do you think that their imaging division will fail? Just clear data, please.
Go and read his other articles, and kindly don't talk to me ever again.
Your article prompted me to find out how many times the shutter has been actuated on my Nikon D600, which is 69,000 thus less than half of its expected lifetime.
I bought my camera almost eight years ago, it looks like heck now and I'd have a hard time selling it for more than $300.
On the other hand I'm happy to see it might last for another eight years…
This seems a rather laboured comparison, and the main flaw in comparing tanks to cameras is that the former is designed to literally destroy other models, not compete with them for sales or do a standalone job. The failure rate of a tank is also dependent on the quality of the competing tank. This is a major difference compared to camera failure rates. It is better to compare cameras to sports cars, which are designed simply to be the fastest both in relation to each other and in relation to a mean standard.