Have you designed early life failure into a product?

[macboy]

In Edison museum there are lamps about 100 years old, shining 12 hours a day, 7 days a week.

Those old lamps run relatively cooler, dramatically improving lifetime at the expense of much lower efficiency. That's why Halogen bulbs were invented; running the filament hotter improves efficiency (less IR/heat output, more visible light output) but would have quickly evaporated the filament, prior to Halogen technology.

[wagon]

In Edison museum there are lamps about 100 years old, shining 12 hours a day, 7 days a week.

Those old lamps run relatively cooler, dramatically improving lifetime at the expense of much lower efficiency. That's why Halogen bulbs were invented; running the filament hotter improves efficiency (less IR/heat output, more visible light output) but would have quickly evaporated the filament, prior to Halogen technology.

But how do the facts fit in around the conspiracy theory?

[Refrigerator]

In Edison museum there are lamps about 100 years old, shining 12 hours a day, 7 days a week.

Those old lamps run relatively cooler, dramatically improving lifetime at the expense of much lower efficiency. That's why Halogen bulbs were invented; running the filament hotter improves efficiency (less IR/heat output, more visible light output) but would have quickly evaporated the filament, prior to Halogen technology.

But how do the facts fit in around the conspiracy theory?

Illuminati ?

[wagon]

Illuminati ?

It's a good thing I'm wearing my tinfoil hat then.

[SeanB]

I actually have a few of those lamps, and they are very long lasting. Only have had one fail in 20 years, and that was because it was dropped and the glass broke.

[ciccio]

I recently had a discussion with a customer about the "longevity" of the equipment he is still building and selling.
I've been responsible for the design of these products for the last 15 years, after he bought the small company that was his supplier, and that was closing because the owner wanted to retire.
The products are audio amplifiers and mixer for PA system, mostly installed in churches, and there are more than 1000 systems installed.
The oldest  are about 40 years in service, and most of the failures come from broken mic cables or lightning strikes during summer thunderstorms (a church audio system, with 100 V speaker lines and mic cables, acts as a big lightning attractor).
The problem is that, with the oldest systems in service , the risk that one of the units burst "naturally" into flames increases every year (I must admit that they are an example of REALLY CONSERVATIVE design, but there is a limit to the life of every component).
Most of them are installed in historical buildings, and this risk of fire propagating to the building structure or furniture is high (Installation standards were not really strict, in these times) .
He asked me to design something that will break after a certain amount of time (let's say 5 years), and will require an exchange of the equipment.
The "breaker" shall be a device that is proprietary and not available to external technicians, so an house-call will be required.
This will allow to check the equipment status, upgrade it to the latest version, and have it ready for other 5 years of service.
As a secondary effect, it will generate some money from maintenance work.

I thought a little about the question, and asked the customer to proceed in another direction (e.g. contact the customers to explain that their system require preventive maintenance), but I'm still puzzled by the problem.

I know that similar things are programmed into industrial equipment, or software programs, but I'm asking if other members think that this is "legal" or "moral".....

Best regards

[jancumps]

Conspiracy part two
Down here, the canonical example given for intentional failure is the TL (fluorescent tube) lamp of Philips.
The gossip mill keeps saying (for 40+ years I think) that the initial design was modified because its life span was too long.
Of course I'm not backed up by any facts.

[wagon]

Conspiracy part two
Down here, the canonical example given for intentional failure is the TL (fluorescent tube) lamp of Philips.
The gossip mill keeps saying (for 40+ years I think) that the initial design was modified because its life span was too long.
Of course I'm not backed up by any facts.

There maybe some merit in this one.  My house has tubes up that are probably as old as the hose, and the house was built in the late 1960's. I changed the tube in my kids room yesterday, and it was an old 40w unit.  The new tube is so much brighter though!

[SeanB]

Simple reason the TL does not last as long as the old T12 tube is that it is likely the old T12 tube was a full dose tube, with probably 100mg of mercury inside the tube so it would not go mercury starved over it's lifetime, and would only degrade due to heater losing emission and the phosphor aging. They typically will last 10 years on a magnetic IS ballast, or around 8 years on a pulse start ballast as there the cathodes get stripped faster. Newer T8 tubes have a lower mercury dose, 12mg or less, and the typical Alto ( what a good word, they stop working fast) has under 4mg, so that you do not have to dispose them as hazardous waste. That the tube typically lasts 2 years or less, degrades fast lumen wise and you need 6 relamps to meet the old T12 is not considered.

This is for tubes made by a reputable manufacturer in a plant with good process control, cheap tubes made in a OHL plant with pretty much QC being " did at least half work in the box? Ship the box" are lot going to reach much past a year.

I find the tubes made in Poland are the best of current, Hungary, then Russia ( strange that), the odd old batch from France and then of the far east Indonesia, Thailand and bottom of the barrel China. Then you get the tubes with no country of origin. I have only had one outlier of the cheapies, a Ya Ming tube that did 8 years, long survivor of the rest of the box of 25 that all lasted less than a year, typically failing in 6 months.

[Fsck]

I once had to use 85°C Electrolytic capacitors and not 105°C parts because "management" thought they would last longer then the warranty period and that is long enough.

The best way to deal with this is to trap management into demanding something too cheap and creating a lot of in-warranty failures. Once they have been burned and maybe one of them made into a scapegoat they will listen to you next time.

They will just argue that it is your fault for not being persuasive enough.

You can't win against management ar$eholes, salesmen & accountants.

if you create enough of a papertrail, you can win. avoid in-person and phone interactions at all costs.

[Refrigerator]

Simple reason the TL does not last as long as the old T12 tube is that it is likely the old T12 tube was a full dose tube, with probably 100mg of mercury inside the tube so it would not go mercury starved over it's lifetime, and would only degrade due to heater losing emission and the phosphor aging. They typically will last 10 years on a magnetic IS ballast, or around 8 years on a pulse start ballast as there the cathodes get stripped faster. Newer T8 tubes have a lower mercury dose, 12mg or less, and the typical Alto ( what a good word, they stop working fast) has under 4mg, so that you do not have to dispose them as hazardous waste. That the tube typically lasts 2 years or less, degrades fast lumen wise and you need 6 relamps to meet the old T12 is not considered.

This is for tubes made by a reputable manufacturer in a plant with good process control, cheap tubes made in a OHL plant with pretty much QC being " did at least half work in the box? Ship the box" are lot going to reach much past a year.

I find the tubes made in Poland are the best of current, Hungary, then Russia ( strange that), the odd old batch from France and then of the far east Indonesia, Thailand and bottom of the barrel China. Then you get the tubes with no country of origin. I have only had one outlier of the cheapies, a Ya Ming tube that did 8 years, long survivor of the rest of the box of 25 that all lasted less than a year, typically failing in 6 months.

I have a 40-year old russian radio that uses tubes, a couple years back one of the tubes failed but i happened to find a bag of the same kind of tubes thrown away, found one that fits, popped it in and it worked flawlessly, i can listen to  retro music again.
I was lucky to find those tubes because they cost a pretty penny over here where i live ( 40 bucks a pop! ).

[SirNick]

The problem is that, with the oldest systems in service , the risk that one of the units burst "naturally" into flames increases every year ....  Most of them are installed in historical buildings, and this risk of fire propagating to the building structure or furniture is high (Installation standards were not really strict, in these times) .
He asked me to design something that will break after a certain amount of time (let's say 5 years), and will require an exchange of the equipment.
The "breaker" shall be a device that is proprietary and not available to external technicians, so an house-call will be required.
This will allow to check the equipment status, upgrade it to the latest version, and have it ready for other 5 years of service.
As a secondary effect, it will generate some money from maintenance work.

I thought a little about the question, and asked the customer to proceed in another direction (e.g. contact the customers to explain that their system require preventive maintenance), but I'm still puzzled by the problem.

That IS a tough one.  From my experience working with / in / around that kind of organization (be it a church, school, or non-profit org), there is very little attention given to maintenance and preventative replacement.  It's often a shoe-string budget, with no one really filling the role of lifecycle analyst, which pretty much guarantees that anything not strictly necessary will be put off until becomes strictly necessary.  And then, if it can be bodged, it will be.

Still, 5 years seems a little short -- short enough that it would still be memorable to anyone there.  "It's broke again?  Didn't we JUST get this thing fixed a few years ago?"  It also depends on how expensive that service call is.  Labor isn't cheap, and if you make an amp that takes $300 to service every five years, the owner is better off replacing it with an off-the-shelf Class D model that he can source at Guitar Center within half an hour of the failure.

Up-front service contracts might be an option -- that way, it's pre-paid, so there's no struggle over wrangling the funds when maintenance is due.  OTOH, if it isn't mandatory, it won't be purchased most of the time.  If it IS mandatory, many of your clients will go elsewhere.

I'm sure I'm repeating discussions you've already had with yourself and your boss at this point.

[David Hess]

The problem is that, with the oldest systems in service , the risk that one of the units burst "naturally" into flames increases every year (I must admit that they are an example of REALLY CONSERVATIVE design, but there is a limit to the life of every component).
Most of them are installed in historical buildings, and this risk of fire propagating to the building structure or furniture is high (Installation standards were not really strict, in these times) .
He asked me to design something that will break after a certain amount of time (let's say 5 years), and will require an exchange of the equipment.
The "breaker" shall be a device that is proprietary and not available to external technicians, so an house-call will be required.
This will allow to check the equipment status, upgrade it to the latest version, and have it ready for other 5 years of service.
As a secondary effect, it will generate some money from maintenance work.

I think it is a monumentally stupid idea.  What if the "breaker" is the first thing to fail?  What if a customer figures out how to bypass it thereby creating a larger safety issue?

If the device has an increasing risk over time of suffering catastrophic failure which could spread, then design in safety in depth.  Typically this involves metal shielding, fusible links, and fire resistant materials but design counts as well if you consider fold-back current limiting and crowbar circuits.

[SirNick]

To be fair, you're kidding yourself if you think any given piece of electronics can be totally safe.  In just my experience as a consumer, I've seen several reputable power supplies fail with smoke and projectiles.  While unlikely (they were designed to contain a certain amount of this), they could've started a fire.

For high-current devices (like an amplifier), it takes a beefy fuse to support the maximum rated load.  However, typical use will be well below that limit .. and as such typical users will do dumb things like plug them into undersized extension cords and power strips, and get away with it during non-fault conditions.  E.g., I had a car amp short out internally and melt the insulation off its power cables.  The upstream fuse blew, but not before the remaining cable adhered itself to my car's upholstery.  That could've easily been a fire.  It was my own fault for using the minimum rated power cable -- a mistake I won't make again.  But it only takes once.

[T3sl4co1l]

To be fair, you're kidding yourself if you think any given piece of electronics can be totally safe.  In just my experience as a consumer, I've seen several reputable power supplies fail with smoke and projectiles.  While unlikely (they were designed to contain a certain amount of this), they could've started a fire.

Once saw an industrial MeanWell (480VAC input, 24V 5A output, DIN rail mounted -- good compact size and efficiency for its ratings, and excellent price at that) go pear-shaped.  It had gotten moisture inside.  There was severe blackening around the traces near the switching transistors (obviously, some droplets were up to no good there), and the fuse...there was no trace of.  In this model, they used a 3AG size fuse, which I'm not sure if it was glass or ceramic (I hope it was ceramic!), but yeah, it got destroyed with prejudice.

The chassis (which is perforated on most sides for ventilation) did its job containing the arc flash though.

Depending on just where it failed, it could probably be better or worse, but that's the main point of safety procedures in that sort of domain (480V+ industrial): keep away from it, because it is (literally) explosive when it goes!

Tim

[ciccio]

I think it is a monumentally stupid idea.  What if the "breaker" is the first thing to fail?  What if a customer figures out how to bypass it thereby creating a larger safety issue?

If the device has an increasing risk over time of suffering catastrophic failure which could spread, then design in safety in depth.  Typically this involves metal shielding, fusible links, and fire resistant materials but design counts as well if you consider fold-back current limiting and crowbar circuits.

In my post, I said that I did not agree with my custome's idea, but I actually I do not believe that it is a "monumentally stupid idea".
I'm talking about 30 to 40 years old power amplifiers. The simple fact that they are still working (maybe with some problems, a dryed-up capacitor or two) confirms that the original design (the designer is retired, and well alive at the age of 92 years) was sound and safe and "state of the art".

My new car forces me to change oil at some intervals, and if I do not oblige,  after about 1000 km from the first signal the engine will not start.
One of my CAD programs says that the licence has expired (don't remember of a time limit, but cannot find the original papers) and does not work anymore.
I this really different from some "part" with a limited life that will break without other damages before any other part, and will require some technician to open the rack, check cables for undamaged insulation and rodent's bytes, and some unexpected work from an unknown and incompetent technician?

I know the end users: as was pointed out by SirNick some posts above, it is practically impossible to program any preventive maintenance with the "standard" customer: they simply refuse the idea.

Another example: I have a friend who was in TV service, for a BIG manufacturer. He showed me a board that was populated with 1/8 W resistors, that were overheating by design.
The standard failure was after 3 years of service. He replaced with 1/4 W parts, and the sets are still working after many years.  Was this  BAD design, or planned obsolescence?

Following my advice, my customer will write a letter to the owners of the oldest systems, and disclaim any liability for any damage. His lawyer is working on it.

Best regards

[David Hess]

My new car forces me to change oil at some intervals, and if I do not oblige,  after about 1000 km from the first signal the engine will not start.
One of my CAD programs says that the licence has expired (don't remember of a time limit, but cannot find the original papers) and does not work anymore.
I this really different from some "part" with a limited life that will break without other damages before any other part, and will require some technician to open the rack, check cables for undamaged insulation and rodent's bytes, and some unexpected work from an unknown and incompetent technician?

It is different if it was not made clear to the customer that the design life is enforced making the product a rental instead of a sale.  I am dubious of examples like your car because all too often manufacturers take advantage of what could be a genuine safety feature to extract rents.

Maybe I was a excessively harsh.  It is "unusually stupid" if the artificial restriction is in lieu of proper safe design.

Another example: I have a friend who was in TV service, for a BIG manufacturer. He showed me a board that was populated with 1/8 W resistors, that were overheating by design.
The standard failure was after 3 years of service. He replaced with 1/4 W parts, and the sets are still working after many years.  Was this  BAD design, or planned obsolescence?

They were probably just being cheap.

Occasionally I run across designs like this where the under specified part fulfills some other function.  Carbon composition resistors for instance make pretty good fusible links because they tend to fail open and a low wattage part might be selected deliberately in that case.

Sometimes it is because no other option is economically available.  Precision metal film resistors for example are not generally available with high power ratings.