Silly resistor values

They wouldn't be *my* assembly guys very long, if they did.

-- "Electricity is of two kinds, positive and negative. The difference is, I presume, that one comes a little more expensive, but is more durable; the other is a cheaper thing, but the moths get into it." (Stephen Leacock)

Does anyone's "assembly guy" stuff 1K 1% resistors, rather than P/N 1234567? Why does the "assembly guy" even care if it's a resistor?

Quote from my link "0.25 Watt (1/4 Watt)"

What in that quote is it that you don't understand?

I have given you another example because it seems you haven't looked around. From Stackpole, 10% resistors that I have designed in.

The Stackpole parts aren't everyday parts? And yeah, those 10% resistors are purpose-specific like most everything else in life. Their purpose is to shave off cost.

Then you have probably never designed very cost-critical stuff. I have, a lot.

I am designing electronics for a living, for over 20 years. Oh, and during a summer job between my time at the military and start of university I actually built resistors that stop trains. Brazilian freight trains, to be exact.

Carbon resistors are cheap, in China. That plus Malaysia and a few other countries is where all low-cost electroncis get produced. The trick is this: You produced resistors with the least amount of process control and preferably no trimming.

You cannot obtain much info about that from US sources. Datasheets like the one I showed you, yes, but not Chinese pricing.

I have, starting around 1994 or so. That was the first time a client came back to me with this request after I finished a design: "Our Chinese contract manufacturer is asking which of those 5% resistors we can ease off to 10% or more. Can you please identify those? Datasheet comes via fax." (This was the day where not all companies had scanners). My billed time for this effort was 1/4 hour. The cost savings were IIRC around 5c per unit, times 50000/year, times 15 years, and counting (it's still in production). That kind of ROI is stunning. Yet none of this seems to be taught at universities or anywhere else and most people don't bother to educate themselves about it, probably because it's perceived as boring.

Someone who has ever designed with a team in Asia (I have) knows the drill. They had reams of computer paper hung on hallway walls, updated about weekly, with fresh commodity pricing of electronic parts. And yes, all the way down to the lowly resistors and diodes. This wall is what drove the design process, not elegance or whatever.

Same for most of my design. But not for all of them.

It's the only way that works. One client's contract mfg violated this, once, and promptly hit a brick wall. An expensive one, had to trash thousands of units. A "more economical" electrolytic had x times the leakage and they didn't notice that it was in a timing circuit ...

They didn't have PCBs before 1976? Wow. Phenolic was quite common then and well known. For hi-rel stuff I really liked alumina. Much better than perspex posts and so on. You could seal the whole thing, and we usually did.

E96 is standard here. But only for prototyping and for gear where cost doesn't matter, or where engineers think it doesn't matter. It always amazes me when I see a circuit from engineers who aren't great cost thinkers and then, after re-designing it, cost drops hugely.

What a bunch of retards.

Sorry, I do not build cheap chinese crap, so I do not need to deal with a customer that wants to dictate stupid part purchases to pare a penny off his rip-off priced piece of shit product.

Contract manufacturers are supposed to build to the print (ours did). If the retards were asking me if they could switch to other values, the only other value I'd be switching to is a different contract manufacturer.

Not all makers are turnkey that way. Some MUST and do make what was designed.

Good for them.

In the HV power supply industry, the maker builds what we design and ONLY what we design. We do not need a batch made one way one month, and then the next batch having entirely different operating characteristics because some dope decided to make some substitutions.

If we want 104 C caps, that's what we got. If we wanted a particular brand due to ESR or 'whatever', that is what we want. We do not need some chump f*ck trying to cut HIS costs compromising the quality of OUR design. So NO, there will be NO substitution, including resistors and diodes.

Elegance or whatever has nothing to do with it.

It does help if assembly folks know what part is what kind. For temperature profiles and such. The guys rigging the machines don't need to know as long as they follow proper ESD procedure.

Our "assembly guys" (mostly women, actually) know what process to use on which products and how to profile them. They have no need to know what's on the boards, other than inventory numbers. The "assembly guy" doesn't decide which profile to use. The people doing rework (sometimes the same people) need to know a little more, perhaps.

Well, I meant the personnel at the assembler, not just the workers on the floor. The foremen and similar folks do need to know IMHO. I would not advice a client to use a shop where that wasn't the case.

Ah, you're including the "manufacturing engineering" people in the "assembly guys" bucket. I was restricting the group to just those who push the buttons and load the pick-n-place machines.

Not always real engineering. But it does help if the line lead or at least the foreman knows a thing or two about this stuff. Older ones can be more knowledgeable than engineers assume. I was once called to a serious production problem, nearly all boards failed final test. Asked their engineers whether they had a sit-down with production people. "Uhm, no, why?". So I had that sit-down. Problem looked like board warpage. The foreman and I quickly figured out what could be done, for lunch we headed to a place where there was a large hardware store. Bought some lumber there, made a few "pre-stretch" frames, sent some boards through -> all passed.

Warpage? Pre-stretch? Must be some warped boards to cause failures. We have some fairly large boards (half-panel) with a lot of heavy stuff (audio transformers) on them and haven't had too much trouble with warpage. Certainly no big fallout in assembly.

This was a design with tons of RF challenges, there was no way to keep the copper areas even remotely symmetrical to center. Buried impedance-controlled traces, umpteen channels, dense as heck, size of an oven pan, early days of SMT manufacturing and this was all SMT. Oh, and double-sided.

If the cheap Chinese "crap" is just as good as what you prefer to build (e.g., the Chiense stuff has various 10% resistors whereas yours are all 5%), why not?

Would it change your mind if you were a stockholder and could garner the company couple a couple million bucks extra each year by going through the designs and making sure nothing was specified any more tightly than it actually had to be?

You can bet that Apple, who has everything built in China these days, to very high specs, is keeping a close watch on exactly how much it costs to build an iPhone, an iMac, etc... they want to keep their profit margins up! :-)

I would welcome a contract manufacturer who mentioned that they had sources for cheaper parts if looser tolerances were OK, and if I wanted to they'd be happy to hook me up.

---Joel

I'd give a guy at least one or two strikes on that before showing him the door, as I'm willing to believe he really was trying to help and didn't realize the potential pitfall (hence, "inadvertent").

While there are people out there who just don't ever seem to "get it" -- and their efforts to "help" unfortunately more often than not create problems -- in general my experience is that people who'll at least take a bit of initiative and *try* to fix a problem (even if *on occasion* they unwittingly create a bigger problem in the process) are more valuable than those who are almost robotic and just immediately give up and wait for someone else to come and tell them what to do the moment they encounter anything that deviates even the slightest from their prescribed procedure.

You do need people who are circumspect, of course: If a guy is trying to get some old VW Bug running, by all means, using balling wire and chewing gum and anything else that might help... but if a guy is trying to get the GE90-115B jet engine on a Boeing 777 running, it's an entirely different matter.

As engineers we have a responsibility to help assembly people make as few mistakes as possible. This includes obvious stuff like clear (and correct) labeling/assembly documentation, but also things like using polarized connectors (unless you just don't have the budget for it), not putting two mechanically-identical-but-electrically-vastly-different connectors side by side (particularly if mis-plugging said connectors blows something up), etc.

---Joel

No, but I remember some stuff like that, like: "See the USA, in your Chevrolet" "A silly millimeter longer" - Chesterfield 101's "The seven-minute cigarette" - Pall Mall Gold 100's

And who could forget Burma Shave? ;-)

Cheers! Rich

Yeah, the attitude of some folks that stuff from China is always crap has brought down whole companies. I will never understand it. After we did a major cost reduction effort at an ultrasound company (without compromising quality at all) one of our two main competitors threw in the towel for our market. That was Hewlett-Packard, no less.

This reminds me of major ROFL episodes among the big three in Detroit when the VW Beetle was presented. Nobody is laughing there anymore ...

Oh yes! There is no purpose in having 1% or 5% pull-up resistors for push-button inputs or digital lines. It's like wearing sun screen under a swim suit. I have spec'd in as loose as 20%. Because it didn't matter.

The most coarse was -10% to -35% in hybrid production, meaning that we simply skipped the laser trim step for many resistors that were deposited on the alumina substrate. That shaved off some pennies. At tens of thousands per year that kind of savings really matters and is worth the several extra engineering hours that had to be put in (which we did).