-
Have you designed early life failure into a product?
Posted by
aargee
on 07 Aug, 2014 00:53
-
Or in other words, knowingly or been forced to fit a design to a life cycle much smaller than customers/end users would expect?
I can understand that the race to the bottom in quality causes corners to be cut, etc. but designs that are planned to fail close to the end of warranty (or other milestones in a product's life)
I have never been in the design cycle for new products, I'm a field service engineer, fixing stuff or inventing kludges to fix problems. Over my many years of experience in lots of facets of the electronics industry, it *seems* like some designs I have run into have been purpose built for a lifespan of "X" whereas the customer would expect "Y" (where X<<Y).
I know there's all the evidence out there from the light bulb consortium at the turn of the century through to the printer ink marketing, I'm just curious about the people who knowingly design this stuff.
-
#1 Reply
Posted by
retrolefty
on 07 Aug, 2014 01:30
-
Interesting question. I can't talk much about electronics products other then their real cost (inflation adjusted) to consumers has never been better. A related consumer product I can relate to is automobiles. Growing up in the 50s it seemed that most middle class+ people would look to replace their car around 3+ years or so. The last two cars I've owned have lasted me 13 and 14 years, they just seem to last longer and are more reliable.
-
#2 Reply
Posted by
SirNick
on 07 Aug, 2014 03:46
-
OTOH, I drive a car that was built ten years before I was born.
Not in the industry myself, but I see aspects of products where obvious corners were cut. One can only assume the engineers spec'd parts understanding that they would fail after so many hours of use, and that the part lifetime and warranty period were considered with that in mind. Or that the user would likely abandon the product anyway, if it failed early. (LCD TV and monitor PSUs particularly.)
Interestingly, I wandered into a forum thread once where the (unverified but believable) engineer of a brand-name product discussed the statistical analysis of his component choices. Namely, PSU caps for an amplifier that were operating very close to their rated limits, where you would usually expect to see more safety margin. The engineer spoke about how the calculated lifespan was "long
enough" -- such that they would probably fail about the same time the user would be replacing the unit for other reasons. I'm not sure if that's planned failure so much as matching a low price point to the cultural trend of throw-away products.
-
#3 Reply
Posted by
westfw
on 07 Aug, 2014 04:27
-
A lot of products are built for an "expected lifetime"; it's not so much that they're specifically designed to "fail early", but they don't get the attention they would need to have a good chance of having a long life...
I've certainly written software that rendered older HW obsolete, usually just by virtue of outgrowing the memory hardware limitations of the old HW. It wasn't so much intentional, as just something that happens w software. And I've been on the other side of that as well "how can we pare down this new version so that it will run on the old HW." But at some point it is VERY MUCH an explicit engineering decision: "This new SW version will NOT run on that old HW."
-
#4 Reply
Posted by
miguelvp
on 07 Aug, 2014 04:50
-
I think the production cost reduction run will be your main culprit. For example, Jeri Ellsworth made a video talking about her
C64 FPGA Joystick, and they flew her to China because the thing wasn't working.
Once there she found out they replaced a lot of parts and cut corners, hopefully she was able to save the day and got it working. But what she designed was nothing close to the final product.
But at least it's not like the GM chief engineer that signed off that faulty switch. I still think he was told to sign it or else but it's ok GM named and fired engineers involved because obviously it couldn't be a marketing decision
-
-
aargee, I also spent 25+ years as a service engineer for several manufacturers of medical instrumentation. Most, if not all the disappointments relating to quality have always been "introduced" into the system as a cost saving measure. Most notable was a roller assy for a tubing pump to which a $35 cost reduction was implemented. This was an attractive savings for a system that contained 3 pumps ($100 total savings per instrument in an instrument costing #25,000 USD). Unfortunately the improvement involved hollowing out the individual rollers on the roller cage, allowing the roller to collapse under the strain of the pumping action and eventually compromise the life of the roller.
You guessed it ... either a service call (avg $1000 cost) was needed for mysterious intermittent pumping issues or replacement parts had to shipped over night to the account. Eventually the original design was re-implemented and new rollers were sent to all customers.
The story I tell about my first R&D interview goes like this. First question I was asked was "We have a design choice to make in a transistor used in a power supply assembly", (remember, this is medical instrumentation large enough to require an on site service visit - the unit cannot be shipped back for repair.).
"One transistor costs $3USD and is near its design limits. The other choice is a $7 transistor which will last the life of the unit. Which would you choose? Please take some time to think about your answer before you give it".
My reply was "choose the $7 transistor".
WRONG !!!
"May I explain why I made this choice?"
"If that transistor fails for any reason in the next 15 years (life of system), an onsite service visit will be needed. This will involve buying a plane ticket for $2000 to fly across country on a day's notice, grab a $200 rental car, $300 hotel bill, at least $100 in meals and incidentals, and about a $300 dollar two way round trip for FedEx charges to ship the 50 lb part out overnight freight. So I say spend the extra $4 to save $2500 cost to the company in service costs."
R&D response was "you are not looking at the bigger picture. We have a price point to bring this system to market and we must hit that mark".
I mumbled something about who is not seeing the "bigger picture" and told the interviewer I could never work for him under those constraints - not with the "bigger picture" I have come to see from working in the field.
Later that year I was paged out to service one of those systems. A "fix it before the sun comes up or get it out of here" scenarios. I met my wife that night. She was a technician at the hospital and saw me working on the system at 2 in the morning and let out a string of curses I can't repeat here.
We have been married going on 35 years now. That's my story and I'm sticking to it.
-
#6 Reply
Posted by
Bored@Work
on 07 Aug, 2014 05:42
-
I still think he was told to sign it or else
Twenty years ago I would have also thought this. Or that GM management created an atmosphere of fear and obedience where no one dared to step out of line.
But since this was just ten years ago I wouldn't dismiss that the involved engineers just gave a flying fart through a rolling doughnut. The attitude about safety or work ethics in general has massively changed among young engineers.
But back to GM. Even if it was an "I give a fuck" engineer, the management is still responsible. For hiring him, for giving him that responsibility and for not properly supervising him. I.e. for getting everything wrong they are paid for.
-
#7 Reply
Posted by
pickle9000
on 07 Aug, 2014 05:49
-
I'd have to say more about cutting down the BOM bottom line, same effect though.
-
#8 Reply
Posted by
miguelvp
on 07 Aug, 2014 06:12
-
I still think he was told to sign it or else
Twenty years ago I would have also thought this. Or that GM management created an atmosphere of fear and obedience where no one dared to step out of line.
But since this was just ten years ago I wouldn't dismiss that the involved engineers just gave a flying fart through a rolling doughnut. The attitude about safety or work ethics in general has massively changed among young engineers.
But back to GM. Even if it was an "I give a fuck" engineer, the management is still responsible. For hiring him, for giving him that responsibility and for not properly supervising him. I.e. for getting everything wrong they are paid for.
http://www.autonews.com/article/20140605/OEM11/140609881/blame-for-%E2%80%98switch-from-hell%E2%80%99-falls-heavily-on-one-gmIronically, it was DeGiorgio who ultimately fixed the part he referred to in a 2002 e-mail as “the switch from hell.” But by the time anyone else in the company figured out what he had done, GM says the switch had cost 13 people their lives.
If he referred to the switch in an e-mail as "the switch from hell" 2 years prior being implemented doesn't that sound like coercion? 10 years or 20 years doesn't matter, look at Enron for example. Let's just say the 2000 decade is not one of our most glorious times, well this decade is not great either, maybe 20/20.
-
#9 Reply
Posted by
dannyf
on 07 Aug, 2014 12:51
-
Or in other words, knowingly or been forced to fit a design to a life cycle much smaller than customers/end users would expect?
some designs I have run into have been purpose built for a lifespan of "X" whereas the customer would expect "Y" (where X<<Y).
The two are slightly different. The first sounds closer to designing for failure within a set period of time and the 2nd closer to designing for a life of a set period.
ie. "I want the product to fail within a year" vs. "I want the product to last at least a year".
The 2nd approach is wide spread. and it is not common to see the same / similar products at different points that offer different performance (expected life included). Quality is expensive and if a customer expects a life of X, s/he should pay for life of (at least) X.
The 1st approach is rarely seen - it is mean spirited I think.
-
#10 Reply
Posted by
dave_k
on 07 Aug, 2014 13:18
-
I have said this of most contractors/tradies:
"There is no money to be made in fixing the problem, but there is plenty to be made in prolonging it"
-
#11 Reply
Posted by
wagon
on 07 Aug, 2014 13:34
-
I spend my days fixing electronic stuff. I repair a lot of electric fence energisers : my workshop is in a small city that is a bit of a 'hub' for a lot of agriculture. Now, I'll get to my point!
I fixed an old unit yesterday, that was over a year older than me. It was built in 1975. The main capacitor was still within spec, and the unit really only needed a new 47nF 'X2' cap and a new mains cord/plug. It could well see the farmer out. I could have sold him a new one, too. I don't expect a new one will still be going nearly 40 years from now, since it is a lot more complicated and its little case with its thin walls is likely to suffer in the sun. Is the new unit built with failure in mind? I don't think so, but it does have to be sold in a market where the alternative products are, so manufacturing costs must be kept down. Price is as important as specifications, quite often.
With a lot of consumer electronics, like TV's, I feel that the 'market' demanded cheaper stuff, and the suppliers worked out that they just weren't selling the better stuff, so they just stopped making it! Look where that has gotten us, now that we can only get cheap & nasty.
-
#12 Reply
Posted by
digsys
on 07 Aug, 2014 14:02
-
Sadly, even going back 30-35 yrs, a LOT of companies I contracted R+D for, often paid big bucks for designs that just made it past
their warranty. We got bonus amounts on how close we could get it. In some service departments (who we worked with), they even
ran competitions ! The one big difference to these days was that - There was NO compromise on quality / (design) reliability !
Some of the "fixes" were ingenious :-) I occasional still get asked, but stopped doing that cr@p a long time ago. It still happens a lot though.
Edit: OK I have to confess.
It was the mid 70s (IIRC), someone had just come up with the idea to connect one of our Monroe calculators (nixie tubes) to a cash drawer !!
It was a gold mine, but they needed to ensure that it had a limited life. I made up the interface card to tie the systems together, which gave the
cashier - items, totals and date - and controlled the cash drawer. I then made up a separate diode logic pcb (100's of diodes from the nixie array)
and set up links to create a "date code". At this point the AND array would latch and turn off the drawer control.
The service techs simply swapped out with the next date code ! I still remember my bestowed hero status ... ahhh memories
-
#13 Reply
Posted by
JuKu
on 07 Aug, 2014 16:06
-
Most manufacturers give a MTBF for their parts. Designers simply pick parts that have a MTBF slightly longer than the warranty period. The EU is looking at requiring appliances to have the lowest MTBF of all components printed on the efficiency rating label.
No, that would mean close to half of the products fail within warranty period! But you are right, mtbf is an important spec, products are not designed to last forever anymore.
When lightbulbs were used, it costed a little more to make them fail. Controlling the impurity of the inert gas costed something, when putting in pure inert gas would make the bulbs last much longer. In Edison museum there are lamps about 100 years old, shining 12 hours a day, 7 days a week.
-
#14 Reply
Posted by
Marco
on 07 Aug, 2014 16:48
-
Red glowing space heaters.
-
#15 Reply
Posted by
David Hess
on 07 Aug, 2014 17:12
-
Comparing the designs for old switching power supplies to new ones reveals a difference even over the past 10 years.
In the past, the filter capacitors were commonly not specified or poorly specified for ripple current rating so they were derated more than strictly necessary to make sure they would have an acceptably long operating life. Newer switching power supplies use much more precisely rated capacitors so they can be selected for exactly the expected operating life with high accuracy.
-
#16 Reply
Posted by
SirNick
on 07 Aug, 2014 18:46
-
That's just ugly. Power supplies especially
should be designed as the most robust, reliable, and safe module in a system. At least for consumer electronics, it's usually the one piece most capable of personal or property damage if it fails. Dirty game.
The story I tell about my first R&D interview goes like this. First question I was asked was "We have a design choice to make in a transistor used in a power supply assembly
One transistor costs $3USD and is near its design limits. The other choice is a $7 transistor which will last the life of the unit. Which would you choose? Please take some time to think about your answer before you give it".
My reply was "choose the $7 transistor".
WRONG !!!
R&D response was "you are not looking at the bigger picture. We have a price point to bring this system to market and we must hit that mark".
Later that year I was paged out to service one of those systems. A "fix it before the sun comes up or get it out of here" scenarios. I met my further wife to be that night. She was a technician at the hospital and saw me working on the system at 2 in the morning and let out a string of curses I can't repeat here.
We have been married going on 35 years now. That's my story and I'm sticking to it.
Aww... they weren't trying to hire you, they were match-making. <3
-
#17 Reply
Posted by
SeanB
on 07 Aug, 2014 18:53
-
I bet you were thinking of sending a copy of the total bill to that interviewer along with a line of "Aren't you glad you saved $4 here". Though that might have been better as a message to the shareholders after they got the earnings report.
-
#18 Reply
Posted by
David Hess
on 07 Aug, 2014 19:03
-
The situation with capacitor ripple current specifications over the years reminds me of bridges.
There are Roman bridges still standing and in use after 1500+ hundred years because with a few simple rules of thumb, masonry structures may be easily scaled up. But those same rules lead to a large derating that is not present in modern structures which are more "efficient" and have a correspondingly short expected operating life.
-
#19 Reply
Posted by
SeanB
on 07 Aug, 2014 19:12
-
Most buildings are conservatively designed ( Tongaat mall excepted from the preceding generalisation) such that they have a very large safety factor built in. This is true for passenger lifts, as the typical spec calls for it to be able to operate, run and level within limits with a load equal to twice the rated load.
-
-
Different industry but it just happened:
A friend of mine is in the business to produce flexible exhaust pipes/hoses for a big name truck manufacturer in Germany.
They had to last a minimum of 100,000 km but not more than 150,000 km
His first production batch lasted more than 200,000 km and he lost the contract.
This happened in 2014!
-
#21 Reply
Posted by
David Hess
on 07 Aug, 2014 19:45
-
Most buildings are conservatively designed ( Tongaat mall excepted from the preceding generalisation) such that they have a very large safety factor built in. This is true for passenger lifts, as the typical spec calls for it to be able to operate, run and level within limits with a load equal to twice the rated load.
Sure. But the Romans designed in safety factors an order of magnitude (or more) larger than a modern structure would use and their static loads were correspondingly large compared to their dynamic loads. They did not do it deliberately; they just did not know any better and were using a design methodology which they knew from experience to work.
I suspect the same thing happened with aluminum electrolytic capacitors in power supplies. When they had poorly specified ripple current ratings, large safety factors were used but as those specifications got better, they were not derated as much and operating life was more accurately bounded.
The above may be a justification for over specifying a PC power supply or planning on replacing the capacitors. I am getting annoyed by how many "high quality" PC power supplies fail just out of warranty because of capacitor failure.
-
#22 Reply
Posted by
pickle9000
on 07 Aug, 2014 19:57
-
Most manufacturers give a MTBF for their parts. Designers simply pick parts that have a MTBF slightly longer than the warranty period. The EU is looking at requiring appliances to have the lowest MTBF of all components printed on the efficiency rating label.
This sounds fair for consumer gear, would be a selling point.
-
#23 Reply
Posted by
zapta
on 07 Aug, 2014 20:28
-
Or in other words, knowingly or been forced to fit a design to a life cycle much smaller than customers/end users would expect?
I did it recently for an internal software, a hard coded expiration time, because I am sure that by then it will have a upgrade. It will stop functioning at that date but it is not mission critical.
-
#24 Reply
Posted by
SirNick
on 07 Aug, 2014 21:10
-
I was the victim of that, once. I bought a 24-channel rack-mount audio interface with a PCI interface card. The drivers had a hard-coded failure date in them. Worked one day, then the next week, the computer wouldn't boot. Finally narrowed it down to the PCI card -- when it was removed, the PC booted fine. Logically, I expected the card had gone tango-uniform and threw it away.
Then, while looking for a replacement, I found a forum post where a BUNCH of other people just happened to be having the exact same problem. The company released new drivers that "fixed the problem" shortly (at most a couple days) after. Luckily, I had just tossed the card into a junk box to get rid of later, so I was able to upgrade the drivers, reinstall the card, and get on with life. But I can imagine, if I were running a small commercial studio and that was my primary rig...
Bomb dates are NOT COOL, ever, but especially if you don't know they're there.