Electronics design for failure

[edisonobilly]

Hi Everyone, I am taking an electronics design module at university and I've been asked to discuss how an electronics designer can design electronics into "white goods" such as refridgerators, cookers, toasters, etc (these products are increasingly having electronics integrated in them). Taking account of the short life span of the overall machine, whilst maintaining fitness of purpose for the electronics parts involved.
I have some experience in electronics design so I have an idea what this means, but I need someone with a lot of design experience, to give me some tips and ideas to guide my work and research into this, given their experience. Thanks very much, I will really appreciate whatever anyone can throw into this.

Eddie

[VIPR]

I'm not sure that I fully understand what you are asking for. Are you wanting to talk about how to get a product to market quickly and cheaply using known good shortcuts (not safety compromising shortcuts, more like "trade secrets") and thus not over engineering but also not under engineering for the application?

[Sionyn]

indeed i am not sure what your asking either
at the age 25 i have never been asked to purposely (maybe a bit strong lack of a better word) design things to fail.

what do you mean by 'limited life' is it marketing talk ?

when i have more to go on ill try and help

cheers

[edisonobilly]

ok, given that such products have limited life span, what steps should be taken during the design, to not over engineer it and making the system too expensive(considerings its short life span), at the same time it should serve it's purpose properly during its short life.

thanks

[mikemxyzzy]

How long is a "limited life"? I think a major problem with the electronics modules in appliances is that they are *not* designed for a long life, so they usually crap out before the rest of the appliance. If a replacement board is available it is usually very expensive, so it encourages people to junk them. I think they should be designed more like auto electronics. More emphasis on reliability and long life.

[saturation]

White goods, as you call it, have extremely long lifespans.  You go can to any old home and still have refrigerators, range ovens, cookers, toasters still working, they are extremely reliable unless they were made poorly.  My grandfather still had his original circa 1930s appliances still working, and what was replaced since that time were far over 20+ years old, it outlived him and my father.  What tended to kill refrigerators was the gasket would deteriorate or the coolant leak or the compressor blew, but after decades of service.  So many refrigerators live so long, the current movement to upgrade them to modern ones is based on their higher energy efficiency: c1950s refrigerators my brother used cost $1000/year to run, he upgrade his this year, and power consumption show it will cost $50/year to run, that's amazing.

However, my experience with digital controllers has been poor. What takes me one twist to set an old analog timer, or heat setting, now takes me 3 push buttons plus enter to do, and every now and then, the digipad would put an extra digit in, so what was 55 seconds, is now 5:55!  Maybe you can do better.  

In washing machines, small amounts of bleach erodes the protective coating on digipads, causing premature failure; nothing like that happens on old analog controls, old washing machines often have their motors burn out from use or the drive belt snap. This is more than my experience, you'll still find appliances for sale today with analog controls because the digital versions are still not as robust or cost effective to make.  My experience with friends who have gotten chic appliances is that these MCU controllers become a major PITA, that cause more problems that they solve.

However, I have seen no change in reliability in microwave ovens, but its the digipad again, fails from repeated use.

[Simon]

what goal are you trying to acheive ? To design electronics into something that does not need it is not very clever.

[deephaven]

Everything needs electronics 

I can think of a possible example which might help you. Supposing you want a rotary control so that you can adjust in fairly fine steps things like temperature, or even just selecting through a list of options/programs. An expensive but reliable way to do this would be to use an optical incremental encoder. This produces pulses when its spindle is rotated and also tells you which direction it is rotated in due to its quadrature outputs (2 square waves at 90 degress phase). These pots cost £20-£30, but are very reliable because there are no wiping contacts, they just rely on internal contactless opto-sensors. Now, you can get a mechanical version of these pots which have actual contacts to produce the signals, but work with the same principle of having quadrature outputs. The life of these pots is very short compared to the their opto counterpart, but there cost is in the region of £1.

[saturation]

You'd expect that, but the proof is in real time, appliances with analog controllers are still working and are very old.

But I think they may have a fixed cost to make in terms of raw materials which doesn't get cheaper over time, but its possible for all digital controllers to be cheaper long term to manufacture.

A simple example is your cellphone, look at the keypads after one year of use and what kind of environmental abuse can it take?  Imagine the same after 5 or 10 years.  It can be made better, but will it be cheaper to make for harsh working conditions? 

In appliances, you have electrically noisy environments that generate lethal voltages for digital electronics, so it requires suitable protection.  You have very high heat from ovens, mostly above 100C, at or above the operating temperature of most ICs, that now requires insulation or spacing to protect against.  In washers you have very high working humidity not to mention the exposure to water, and cleaning chemicals.  So the challenge is to make digital controls within equal or better capability without more cost, and that includes not just the electronics but the input-output: keypads and LCD or LED readouts.  On analog controllers, you can just paint the dial with ink.

Digital controls work well in microwave ovens as besides high line voltages and the magnetron inductive voltages, temperature and humidity is limited to the food item being cooked, and yes, you can hear the fan whirl in there too.  But it others, its a mix bag, in 30 years, I've not seen an appliance with digital controllers work better, or cheaper, than analog, it usually sold as a chic premium item. 

For adjustment, appliances are crude in terms of 'steps'.  For heat, such as for ovens, typically 25-50F resolution, microwaves about 10+ seconds, washers about 1 minute etc.,

Everything needs electronics 

I can think of a possible example which might help you. Supposing you want a rotary control so that you can adjust in fairly fine steps things like temperature, or even just selecting through a list of options/programs. An expensive but reliable way to do this would be to use an optical incremental encoder. This produces pulses when its spindle is rotated and also tells you which direction it is rotated in due to its quadrature outputs (2 square waves at 90 degress phase). These pots cost £20-£30, but are very reliable because there are no wiping contacts, they just rely on internal contactless opto-sensors. Now, you can get a mechanical version of these pots which have actual contacts to produce the signals, but work with the same principle of having quadrature outputs. The life of these pots is very short compared to the their opto counterpart, but there cost is in the region of £1.

[edisonobilly]

The reason for the initial question was because, goods are now being designed to have "limited life" and "white" goods are in this category, and are increasingly having electronics functions. Now how can a designer design electronics for these kinds of products, taking into account the short life of the machine, whist maintaining "fitness for purpose". What considerations will the designer have to make in this situation as not to make the design and the  finished product unnecessarily expensive, since it will be replaced in a very short time with a newer, better and probably, more efficient product. 

thanks

[EEVblog]

This question doesn't really make sense.
Whitegoods are typically designed to have a LONG life, not shorter.
High tech consumer electronics are the ones with short life spans (mobile phones, computers etc)

Even if your fridge has an internet web terminal in it, customers will expect it to have a 5 year warranty and last in the order of decades, not years.

The key with whitegoods design is good serviceability.

It's actually hard to "design in" short life spans whilst optimising cost, as it seems you are asking. So not only does your premise seem incorrect, it does not have an easy answer.

How to design whitegoods for LONG service life on the other hand has very definite engineering answers.

Dave.

[the_raptor]

Do you mean "white goods" like toasters/microwaves or white goods like fridge/stove/washing machine?

There is no way I or anyone I know would touch a fridge etc that wouldn't last 20+ years. There might be a market for digital bullshit at the top of the market but people who pay those prices expect decent quality.

[Simon]

put it ths way. In computing we all prefer buttons to keypads and displays. but we have to use a keypad. companies have gone to lengths to design digital mixers and similar kit with "old fashioned" controls to keep the prefered interface but still have the revolutionary technology underneith. Camera's, serious amateurs and pros want DSLR camera's with more buttons (like a mechanical camera) because they hate scrolling through menus and so the more primitive approach is prefered. If there is no call for something there is no point in doing it.

The only thing I know of that has had marginal success despite being so badly made and vauely fit for purpose is windoz$

[saturation]

Not to discourage the OP, but it should be a good exercise for a class, particularly once made equivalent of an analog controller, how cost effective it will be. Who knows, you may succeed, but you won't know if you don't try.

As for the real world, take a look at any catalog of white goods, and check the number of digital vs analog controller on them and the cost of the appliance.  Since any basic refrigerator works well, the only chic item are gadgets on them, and AFAIK, they don't sell well.

The current chic item in appliances is 'restaurant' chic, all stainless steel.  Looks nice, but its too easy to get fingerprints on them, constant cleaning, PITA.  Good old cheap white enamel is hard to beat.

What I think has made a difference in modern appliances is reduced energy consumption, as Dave suggest.  I don't know how they do it, but for refrigerators, going from $1000/y to $50/y operating costs is worth an upgrade.  Induction ovens?  Maybe if you cook 3x a day, toasters?  Water heaters?  Anything to insulate hot water and prevent leakage, maybe an cyclical on/off timer so it can turn off when no one is home.  Can't think of more.

[Alex]

A washing machine designer/manufacturer had designed a very long interval timer into the motor contol software routine that caused the motor to vibrate and make weird noises giving the impression that there was a malfunction.

[scrat]

A washing machine designer/manufacturer had designed a very long interval timer into the motor contol software routine that caused the motor to vibrate and make weird noises giving the impression that there was a malfunction.

Unbelievable! Is it for sure? It seems an engineer's joke become true!

[grenert]

That sounds like a tall tale.  I've never seen a washer with any kind of long-term clock/memory, and bit would be crazy to put in such a thing just for this purpose.

[Alex]

You can be sure it did happen, but not for long. Can't say more, sorry.

My point is that designing the system for a given lifetime makes sense. It is excellent input for the big mean marketing machine.

[Mechatrommer]

on mechanical engineering side, this "white good" or limited lifespan design or whatever you want to call it... is a common thing and easy to do. i've seen most products failed mechanically but electrically still good, but for regular consumers, that product is rendered useless. from electronic point of view, i've seen most failure due to usage of cheapo no-brand part that can fail variably from the first day of usage to several weeks, to months to years, it depend on luck and how hard do u use it, and the fun part is that, it can fail if prolonge non usage, than suddenly fail when u use it. i got products where ants built a colony inside the electronic "city" for months in drawer, then u know what when we switch it on. still its mechanical (casing) design failure, from my point of view.

[rf-loop]

I think a major problem with the electronics modules in appliances is that they are *not* designed for a long life,

maybe it just this way but other side it is art of science to design (and make) just for "controlled" short life time. It is not so easy. First must keep care that it work reliable in warranty period becouse warranty time fails is extremely expensive and not only direct money for repair/change. Then need design so that it die fast after this reliable lifetime. Designing for short lifetime is not easy if you still keep warranty period reliability good. There are methods but...

[Simon]

really ? just ask half the printer manufacturers how they do it. Ever heard of lexmark printers with a serial fail at 24 pages syndrome ? clearly someone forgot a couple of pages in the counter. After 3 machines from the same batch the guy gave up

[FreeThinker]

I think the real question is cost. Most white goods fail due to mechanics ie bearings etc on a washing machine or compressors on a fridge. Cheaper items (toasters microwaves etc) are essentially throwaway products. So the real question is can the electronic part do it's job cheaper,better than a mechanical equivalent. In the case of a programmer for a washing machine the answer is probably yes, if it is internet connectivity for your fridge then the answer is more complicated. Predicting the time and point of terminal failure is the first requirement. If the electronics can outlive the critical mechanical bits and are cost effective then you are are half way there. I remember one of the design engineers at a F1 meeting saying that the perfect situation is for his car to perform faultlessly for the duration of the race and fall to pieces after it crossed the finish line, that is it performs it job but is not over engineered, a neat trick.

[XynxNet]

My personal experience is, that todays white goods seem indeed designed for limited life.
All the high quality white goods, which had to be replaced in my family (after 25+ years of service) curiously seem to fail now after 5-10 years of service due to electronic errors. I definitely think the old analog controls are much more reliable than modern electronics.

[Wim_L]

Those Lexmark printers simply use a lousy page feed mechanism which, after a little bit of wear, draws the paper through the machine at an angle instead of straight. If it goes to the side far enough, it jams.

Another common problem are low quality plastic knobs that break off their shaft after some use. Fortunately those shafts are usually of the same type as standard potentiometer shafts, so that's easy to fix with a button from the local electronics supplier that's both considerably stronger and cheaper then the official replacement part, even if it might not follow the style of the original.