calculator longevity



#32

In an earlier thread (archived), Robert Rozee posted the following:

Quote:
a 2011 model 12C/15C is built using a processor with a designed materials lifespan of perhaps 10 years, keyboard (including contacts) and plastic parts that are 'just good enough'. the LCD is attached with a printed flex + glue system that is known to fail over time (several years) - though said failure will in 99% of units be safely outside of the 1-year limited warranty. and the PCB assembly uses a lead-free soldering process that is guaranteed to fail before this decade is out.

I have heard conflicting information with respect to the solder. I have heard only negatives in postings here. But in my real work, I asked about this with one expert who said it is untrue.

What is the story? On what basis is the lead-free approach deemed to be less durable long term?


#33

It's poorly understood at best. Many devices (highest profile is the first generation xbox 360) designed near the date of switchover had reliability problems. The wrong temperature profile was used during reflow which filled joints with microcrystals, rather than full blown cold joints. Even good joints over time are more prone to fatigue failure in my understanding. If anyone tells you they know with absolute certainty how SnAg or SAC solder will perform over time they're crazy because the data do not yet exist.


#34

There are quite a lot of testing results comparing leaded and lead-free solders, particularly from NASA. If you really want to get into it you can read the Executive Summary it's *only* 78 pages long! I guess that NASA executives have a lot of time on their hands these days.

I agree that the results are not all in yet and won't be for many years but there is a lot that is known. The general conclusion from this study is the leaded solders are better for some things than lead-free, and vice-versa.

Edited: 9 Oct 2011, 11:49 a.m.


#35

Quote:
There are quite a lot of testing results comparing leaded and lead-free solders, particularly from NASA. If you really want to get into it you can read the Executive Summary it's *only* 78 pages long! I guess that NASA executives have a lot of time on their hands these days.

Have a gander at the volumes of data produced by the
defense industry. I suppose those folks get a little
excited by the phrase "brittle solder joints". :P

Quote:
I agree that the results are not all in yet and won't be for many years but there is a lot that is known. The general conclusion from this study is the leaded solders are better for some things than lead-free, and vice-versa.

Personally I don't really care for the workability
of the lead free alloys nor the required higher
temperature profiles. Fortunately I'm hardly alone
such that leaded solder still is accessible at reasonable
cost. I expect that will eventually change but by that
time I should have hoarded enough to make it a non issue.

#36

One of the problems that the xBoxes as well as other electronic devices had was problem with Ball Grid Array (BGA) packaging. In a BGA package the interconnect between the chip package and the board it is mounted on is a tiny ball of solder. The problem occurs as the result thermal cycling of components, as the temperature rises and falls, the circuit board and the chip package expand and contract at different rates which puts stress on the joint and can lead to the joint shearing. Devices that are turned on an off frequently are more prone to suffering this kind of failure for obvious reasons. One of the ways some manufacturers got around this problem was to move to an interconnect technology known as solder column, where instead of using a ball of solder, which is very rigid, a column of solder is used, the column is flexible enough to absorb the microscopic movement. I do however know of one instance where the use of solder columns lead to a different problem, a large very dense chip that was mounted on solder columns with a big heat sink on top of it the heat sink was held against the chip by spring tension this spring tension over time, cause the solder columns to buckle with catastrophic results.

I know of devices that suffer from the shearing BGA connection problem that where manufactured before RoHS complaint manufacturing was common.

#37

Also, "designed" lifespan of about 10 years is probably about on par with old HPs. Many of them lasted a lot longer, but even out of those, how many have we (or some previous owners) had to repair? I don't deny that many classic HPs were better constructed than current models (But also some weren't, for instance the solderless Spice models).

I do agree about the glued on LCD ribbon cable, though. I've seen a lot of those fail on older TI graphing calculators (Although I think TI's assembly practices exacerbated this by making the cable longer than necessary and folding it into a 'Z' shape after bolting the LCD and keyboard PCBs into place. If they had waited until bolting the boards into the case, and applied the cable without all the extra length that just gets folded up and pinched by the rear case, I don't think the traces would fail nearly as easily).

Anyhow, those TIs are fixable by carefully attaching wires between the two PCBs, not sure if something similar could be done with an HP LCD ribbon.


#38

Ribbon or flex cable is exactly how the two halves of the HP-71B were connected.


#39

As well as the 28C/28S. Ribbons can work, but you have to be careful how you fold them and if you expect them to put up with stress, make them out of thicker/better material. I've seen old ribbons that were heavy plastic with decently thick metal traces, the aforementioned TI-83+ ribbon was thinner than a piece of bible paper and seemed to have painted/printed carbon traces.

#40

Lead-free is less durable because without the lead, the tin will gradually form "whiskers" which end up shorting signals together. While some vendors have attempted to minimize this with the formulations of RoHS solder and plating, there really isn't any certain method of eliminating it. This is why military equipment and medical equipment is generally exempt from the RoHS regulations. For medical equipment the exemption doesn't do much good because the RoHS regulations have resulted in most things *only* being available in RoHS versions. For military equipment it is not uncommon for the electronic devices to be sent out to have the RoHS plating, solder balls, etc. replaced with non-RoHS, but they can only afford to do that because the military (and by extension, the legislature) doesn't care how much the taxpayers have to spend to get it done.

For example, see the National Electronis Manufacturing Center of Excellence argument for a five-year delay in RoHS requirements for electronics: Tin Whiskers and the Impending RoHS Conversion to Lead-Free. As we know, they did not get their exemption, because the issue is political, not technical. The governments don't care whether they increase the amount of electronics going into the landfill, as long as they can claim progress on lead reduction.


Edited: 9 Oct 2011, 9:22 p.m.


#41

So the whisker thing is part of it. I remember seeing a NASA paper on whiskers once about 8 years ago. It was pretty interesting.

This political thing reminds me of the fluorescent lightbulb law. I am stocking up on 100W bulbs--when I can find them. See my video on the subject:

http://www.youtube.com/watch?v=JR0ws9Gu2kY


Edited: 9 Oct 2011, 9:40 p.m.


#42

Quote:
See my video on the subject:

Bill, that's you in the video?
#43

The audio is unintelligible.


#44

It came in plain for me.

#45

Our son does maintenance in a school and recently got a big box of CFLs that were supposed to last 10,000 hours. About 40% of them didn't even make it to 250 hours. Add to that that it takes them at least two minutes to come up to full brightness, they can't be used as oven lights, etc..


#46

Heh. The maintenance people at my college once tried replacing an oven light with a CFL. Luckily we noticed before using the oven and complained about it.

#47

Not to mention they have a completely different spectral distribution from regular incandescent bulbs. Where incandescents emit light over the entire visible spectrum (they approximate the famous planck distribution), CFLs emit only a few discrete frequencies of light, designed to trick our eyes into seeing white light. You can verify this by looking through a pair of those cardboard diffraction grating glasses for kids. Instead of rainbows, you see several narrow bands.

The problem with CFLs is that the frequency response of the cones in our eyes varies among humans. CFLs look white on average, but I imagine they could be pretty annoying for those on the edges of the bell curve. The 3500 K CFLs don't look that bad to me, but I can't stand the higher color temperature ones (they look too blue). I've talked to others who prefer the 4100 K or 5500 K CFLs and think the 3500 K look too orange.

The other problem with CFLs is color renditioning. Some objects will look really weird under a CFL. While a CFL may look white while looking straight at it or watching it reflect off a white object, it can give colored objects a weird tinge. Many pigments respond to a narrow range of wavelengths, and if the CFL does not emit that wavelength then the color of the object will appear different than it would under sunlight. In fact, some objects which would appear identical in color under sunlight could appear to be different colors under a CFL.

The regulation of light bulbs is insane considering that residential lighting (where most incandescents are used) accounts for less than 1% of the energy usage in the US. Still, our regulators will pat themselves on the back for making the world a safer place and we will be forced to buy overpriced, ugly light bulbs from a cartelized light bulb industry (who will constantly lobby for additional regulation to further restrict competition).


#48

..AND it is one massive transfer of jobs to China. The profits are higher on the mercury laced CFLs. Regular lightbulbs are a commodity (they really aren't but act like it in the marketplace...)

All the CFLs I've sen are Chinese. And places such as Ikea push them and push them


There was an attempt by the republicans to get this repealed a few months ago but it died in the Senate. Dammit.

We need a grass-roots movement. Unfortunately the rich Hollywood "green" types (green as in greenback if you asked me) have been lending their support to the Big Corporate and have put out little blurbs such as National Geographic's "This Bulb" campaign.

#49

Quote:
Our son does maintenance in a school and recently got a big box of CFLs that were supposed to last 10,000 hours. About 40% of them didn't even make it to 250 hours. Add to that that it takes them at least two minutes to come up to full brightness, they can't be used as oven lights, etc..

In my experience we get the rated lifetime if not more in any CFL we've used in our home. I replaced some 8-10 year old CFLs a few weeks ago and the newer designs seem to come up to full brightness within 15-30 seconds, some newer GE units we bought last year seem full brightness within 5-10 seconds. CFLs don't like lots of cycling (like in a hallway situation) but in our home we've put them in places were they stay on for hours and we have always gotten rated lifetimes if not better. I also like the fact that I'm not climbing on ladders to replace bulbs in hateful locations once a year anymore.

And I don't find anything odd nor objectionable about the light they give off. I've had guests in our home that would gripe about CFL color rendition. When I asked them what they thought about the lights in the room, they always something like 'Why do you ask? They're regular bulbs'. They weren't, they were plastic-covered CFL bulbs and looked like regular incandescents.

I have to wonder if the spiral shape and early problems with CFL light quality 'colors' judgement about newer designs. My limited experience with people who gripe about CFLs and the reaction I get when they're informed the light in the room is from CFLs (but disguised as incandescents) lead me to that conclusion.


#50

The biggest improvements are the electronic ballasts. Older fluorescents have a very noticable flicker. For me, it made reading impossible after a few minutes. I had to disable the ceiling fluorescent fixtures when I worked in an office and I couldn't read in the library at college. I would get triple vision and a headache.

#51

Quote:
In my experience we get the rated lifetime if not more in any CFL we've used in our home.

I've been replacing the incandescents in my office with CFL's for several years. In my experience, they last about 2x, not the 7x or more usually claimed.

#52

The solution is simple: use LED light bulbs. The Philips 60 watt equivalent bulbs:

http://www.homedepot.com/h_d1/N-5yc1v/R-202530170/h_d2/ProductDisplay?langId=-1&storeId=10051&catalogId=10053

...introduced last fall provide illumination that is (to me) indistinguishable from that of a good incandescent. And they output a full 850 lumens.

Yeah, they're $25 each (but they were $50 this time last year), but hey, we're saving the planet! Or something.

I bought several last year and my office has two of them that are generally on at least 10 hours per day. No failures yet.


#53

I call it the Candle Effect.

There was no legislation passed to ban candles. We chose the incandescent naturally.

Why are people not switching to CFL "fast enough"? Because they suck.

What is the need to convert to a transitional technology (CFL) with enormous pblic health risks (mercury) when there is a better technology already in existence and catching up?

$%#$#ing goddam politicians make me soooooo angry. They *all* suck. Suck suck suck.

LEDs have been improving in a way that is similar to the magnetic memory advances. It is called "Haitz's Law".

http://ledlightinggroupllc.com/wp-content/uploads/2010/02/Haitz-email.pdf

#54

I concur on LEDs. I have one in my desk lamp, 40 W equivalent, that probably has at least 5000 hours on it. I just installed a 65 W equivalent ceiling flood light in the kitchen - it is indistinguishable from an incandescent, at 1/5 the energy.

In the meantime, I find CFLs quite acceptable - and I've lowered my electric bill by around $10 / month for the last year. I've seen maybe one infant failure out of more than a dozen.

#55

Quote:
and the newer designs seem to come up to full brightness within 15-30 seconds, some newer GE units we bought last year seem full brightness within 5-10 seconds. CFLs don't like lots of cycling (like in a hallway situation) but in our home we've put them in places were they stay on for hours and we have always gotten rated lifetimes if not better.
I took my light meter to a newish CFL we have at home, and it started up at less than half final brightness, and it took over two minutes to get there. I agree that CFLs last plenty long if they're left on for long periods. That's not the case in the hall or the hall closet or the bathroom at home like you say.

On the mercury issue, we've found that the fluorescent tubes that have slightly less mercury--supposedly to be better for the environment--have extremely short lives compared to the ones with a little more, so they're actually putting more mercury into the environment because you have to replace them so often. Also, our son has found a high DOA rate on those. It just doesn't pay.

#56

Does anyone have at least a rough date of when HP switched over to RoHS solder on their calculators?
Just curious since I have a pair of 12C calculators made in China. One was made in 2002 and one in 2004. This was I think before RoHS was mandatory, but I don't know if HP switched sooner or not. You would think they would have been working on the switch over for quite some time, and well before the deadline to make sure there would be no production delays or stoppages.


#57

in theory the PCB inside should be marked with a clear "RoHS" or "Pb free" logo to meet statutory requirements. certainly, all products i've worked on were marked thus.

#58

10 years for an IC is an industry standard lifespan going back at least since the mid 80s. I started my career as an IC reliability engineer. We used 10 year lifespans, assuming continuous usage, as our long-term reliability target. ICs often last way longer than that since they are rarely used continuously.

Nothing new there, even the older 'classic' HP calcs used ICs that were likely rated for 10 years of practical life.


#59

That is right. However because engineering is incremental, we improve our ability to accurately predict life. So the specification is the same, but the actual outcome is vastly different because the conservatism is replaced with more accurate and therefore less conservative predictions.

Some products seem to reach a maximum quality point at somewherea around 10 years into the development of the thing. Older electric toasters last forever. New improved ones last 6 months if you are lucky. Voyagers from 1982 last forever. Spice (older) and 49g+ (newer) die. I am sure we can find tons of examples of this.

#60

Quote:
What is the story? On what basis is the lead-free approach deemed to be less durable long term?

The other consideration to keep in mind is the unique service
conditions to which an average pocket calculator is subjected.
Unless the design provides a substantially rigid frame
the pcb will likely be subjected to bending and torsion stress
in excess of that seen by other consumer devices. The wide,
flat envelope and economical construction doesn't leave
much room for functional beams and gussets in the design.
I'm not certain whether this gives the
advantage to the improved ductility of lead bearing solders or
rather the increased joint stress effectively dwarfs the marginal
difference between lead bearing and lead free alloys.

Sort of a timely question as I'd been puzzling why HP chose such
a seemingly bizarre packaging for the 42s cpu. But it dawned
on me that was probably the only way to reliably reflow a
170 pin 0.5mm pitch package
onto a pcb exposed to substantial distortion in
normal usage. Case in point an sram housed in a rigid
epoxy flat pack adjacent to a saturn cpu yielded to the pcb
stresses as can be seen
here.
In this case that intermittent existed before the unit was
disassembled.

On a related note the conventional "disassembly procedure"
of peeling apart a 42s until four pops have occurred I
believe just exacerbates the problem.


#61

this is very interesting. I am glad I asked the question!

#62

Now that is a very impressive investigation that sheds new light on the design decisions for the 42s. Thanks!


Possibly Related Threads...
Thread Author Replies Views Last Post
  BATTERY LONGEVITY Howard Lazerson 7 689 07-30-2009, 09:56 PM
Last Post: John B. Smitherman
  SD card longevity Thomas Okken 10 854 05-18-2006, 07:25 PM
Last Post: Crawl

Forum Jump: