devices of unecessary complexity

That started in the 1980's with surface-mount electronics, which are substantially more difficult to repair by hand than thru-hole, and not reliable unless the tech who solders on the new parts is more than usually skilled and experienced. I got the experience on lab prototypes where a solder failure was only a brief inconvenience instead of costing a field service call.

Compared to thru-hole SMT is very cheap to manufacture, costing little more to make than the Bill of Materials, and the ICs themselves are cheaper to make due to the smaller lead frame with much less metal. I first encountered the no-repair policy on computer add-in cards for Winchester drives, when the vendor didn't want us to return defective ones.

The Army taught us troubleshooting to the transistor level. They had so much difficulty finding recruits who could learn it that they changed to training LRU (Line-Replaceable Unit) board-swappers. The four (of ~80) of us who graduated all had science degrees. The washouts had a choice of other repair schools so they weren't wasted.

-jsw

"Jim Wilkins" fired this volley in news:lvpki2$j5c $ snipped-for-privacy@dont-email.me:

Jim, are you familiar with "Chip Quick" alloy?

It's a 'solder' with a large hysteresis between melting point and re- solidification point. It allows you to 'dope' all the leads of an SMT component, then simply run a hot iron 'round it once, and lift it off the board as if it were not even soldered down.

Re-soldering is just as easy, as the stuff has amazingly high surface tension (automatically centering the chip on the leads), and a very low tendency to oxidize.

Lloyd

I've substantially modified densely packed Segway control boards for factory testing and then ridden the machine, with padding and a helmet. There's no way I can match the reliability of the original manufacturing process. Afterwards the engineer who designed it asked me what changes would have made working on it easier, but I couldn't give him any that wouldn't unacceptably increase the size of the board. On surface mount prototypes I extend the pads out half a mm to give a spot to heat them with an iron.

I did the layout for the 2nd generation Balance Sensor Assembly circuit board ("gyroscope"), crammed with much effort into the same footprint as the first one, and am very glad I didn't have to hand-solder one.

I've tinkered with the Mass Air Flow and Oxygen sensors on my 1991 truck, and thoroughly scoped out the signals to and from the ignition module. I found some help on line but most of the instructions and operational parameter data came from my complete set of factory repair manuals.

Amazon just loads quickly over dialup. I bought it at NAPA.

-jsw

I've retired and hope to never solder another SMT component.

It was a big parson's-type table with a good laminate top. And I had a *very* tolerant mother.

You would have appreciated the board-level repairs I had to make, in a shirt and tie, at the Rocky Flats bomb plant. We had sold a Sodick EDM to them and we had to fly out to do a repait -- two young Japanese guys from our staff, one an engineer and the other a technician, and me (Marketing Manager).

But they make (or made) nuclear-bomb triggers there and the Japanese couldn't get past the lobby. So they sent me in with some test equipment and got on the phone with me. We had a box of discrete parts and the boards with us that we thought were the problem, but that wasn't it. So I had to solder a couple of components right on the shop floor.

Fortunately, they were through-hole, two-side boards, not multi-layer.

It's not just a mindset. I would love to be able to reliably design simple solutions to simple problems. I can't, easily. I can COPY someone's simple solution to a problem, I can, eventually, figure out simplifications to some complex solution that I (or someone else) has come up with, but a dirt-simple solution that actually works often evades me.

Fortunately, there are plenty of Really Complex problems out there just crying out to be solved, and that I can do.

Tim Wescott fired this volley in news: snipped-for-privacy@giganews.com:

Dirt-simple solutions are the epitomy of engineering. It's hard to get to that point.

I just spent two years of my life designing (and building) a prototype machine for an explosives manufacturer. Some of the most difficult aspects of its operation were solve by those dirt-simple mechanisms invented in the 1920s and 1930s. Some others required complex mechanisms I'm not totally pleased with, but must endure, because there seemed no other way to accomplish them.

To be sure, it is a complex machine, full of potential failure points. Mitigating them required "over-engineering" to make those points robust enough to stand the duty. When complex overcomes simple, that's the cost.

I'm not sure any complex machine (like a mechanical camera with 47 functions!) can be made simply.

Lloyd

At Mitre my machine shop was secured behind a cypher lock yet the Ph.Ds still got in to try to cut lawnmower blades on the bandsaw, at the wood speed.

However I could leave my soldering station out and the lab door unlocked, certain that none of them would dare risk using it. Instead they'd leave the job on the bench with a little note.

Like tying climbing knots and 7018 rod it becomes very easy with practice.

-jsw

Sometimes it helps to jump the wall and look elsewhere. It may not be original, but I ran a simple A/D converter from an LPT port by sending "print" data to an attached 8-bit DAC and reading back a comparator connected to the DAC output and the signal. The 8 data lines were also connected to other stuff that I ignored during a conversion, and vice versa.

-jsw

I realize you need some spiffy mechanisms but they went way overboard with little things, like how some springs were attached. It's just not necessary, unless you're trying to have the longest parts list.

While not as advanced, my grandmother gave me some sort of mechanical device for summing numbers. It consisted of sliding pieces of formed sheet metal, a reset handled and a probe. She used it at the grocery store or something like that.

I took apart a modern leica rangefinder lense last night and noticed the inside looked really crude and had all sorts of scratch marks and symbols on it like it was hand fitted or something. I was surprised by that.

If I had a spare, sure, why not. While I don't fix anything related to cars, they do seem obnoxously designed with all sorts of pegs and covers that have to be broken apart to get to anything.

Hell, even headlamp changes require a tear down and small hands.

It's the fast way to see what's holding something together, and if you don't need it put back, it's fine.

In fact, taking a hammer to saw to stuff is a good way to learn where the strong and weak points are in something.

In the software world, yes. People get really carried away with stupid, overly complex ideas that were just bad to start with.

Actual requirements are usually really hard to come by.

Well, in the case of the original F camera of early 1970s revision, every damn part it connected. There's no sign of any modules of grouped functionality or subassemblies that are not interconnected in 3 dimensions with 15 other parts. That's why I wonder if labor was free when thing thing came off the assembly line. Even assembling it would have taken ages.

I recall some VCRs that were never designed to be serviced. Replacing one tire involved actually cutting a hole the stamped metal made up the transport. Sanyo eventually woke up and redesigned it enlarged openings where people were previously cutting holes. Stupid design.

Cydrome Leader fired this volley in news:lvq5ev $d3e$ snipped-for-privacy@reader1.panix.com:

Only stupid in retrospect. They did not anticipate the need when they designed it.

You're a mook, CL. Smashing an F-body camera just to complain about the complexity of it is ridiculous. Even lacking the prism, it was valuable to someone other than you.

They were complex because of how much they could do -- mechanically, only. Despite their complexity, they were marvelously reliable. Even many that got dropped and/or banged around in service continued to work just fine. I have a 1948 Meteor SP (1/2-frame 35) rotary focal plane shutter camera that's a lot simpler than a Nikon-F, if you really want 'simple'.

My old Asai Pentax is every bit as complex as a Nikon. Despite its 1980s origins, it's as reliable today as the day it was built. Film... that's another problem. Only one maker still out there...

LLoyd

If they don't know what they want then you have a chance to tell them. We've acquired some really great airplanes that way.

"Kelsey recalled in 1977 that he and Saville drew up the specification using the word interceptor as a way to bypass the inflexible Army Air Corps requirement for pursuit aircraft to carry no more than 500 lb (227 kg) of armament including ammunition, as well as the restriction of single-seat aircraft to one engine. Kelsey was looking for a minimum of 1,000 lb (454 kg) of armament. Kelsey and Saville aimed to get a more capable fighter, better at dog-fighting and at high-altitude combat."

-jsw

Thrashing the heck out of it - "drive it like you stole it" is a better way of finding weak points that matter.

Is that the mantra for the engineers who need to justify their ghastly salaries do the idiots upstairs?

Gunner Asch fired this volley in news: snipped-for-privacy@4ax.com:

'had a whole -good- darkroom at one time. Simmon-Omega D2 enlarger with the condensor color head, polycontrast filter set, the whole thing.

Gunner, stuff comes, and stuff goes. To be honest, a high-end pro- digital can do anything and all things film could, except for manipulation in the darkroom. And that was "once or nothing", at least with the film, itself... like 'pushing' a roll, or cold-processing for higher contrast. At least with digital, if you goof, you can try it again. Screw up a roll of film, and you went out and shot it again.

Lloyd