symmetry

the Russians lost a Proton-M at launch because an angular velocity sensor was installed upside down

whit3rd snipped-for-privacy@gmail.com wrote in news: snipped-for-privacy@googlegroups.com:

A good assembler would learn it in no time and even be able to convey it to other assemblers to never perform the assembly incorrectly. A good engineer could also do that thar convey thing.

Ed Lee snipped-for-privacy@gmail.com wrote in news: snipped-for-privacy@googlegroups.com:

snip

"8 sets of batteries"

Batteries or "cells"?

What is the base storage unit? A battery used to be multiple cells, but they started calling everything batteries back in the '60s or so. Dumb America. Maybe the French are right. Anyway the individual cell declared voltage and topped off fully charged value on each cell would be something to know as it would give insight to the internal resistances of the cells. And there are charging watchdog chips out there that would allow each cell to top off and you would never acquire any kind of memory effect.

Or it could be a bunch of lead acid bateries, where each one actually is a "battery" by the original definition.

ZIF socket? You talking about a hard ribbon cable with a row of contacts across the nose and a little locking device you flip after putting it in the socket?

John Larkin <jlarkin@highland_atwork_technology.com> wrote in news: snipped-for-privacy@4ax.com:

Larkin the NASA hater.

No. That abberation was not an "indexing" issue.

Yours is though. Like that.

John Larkin <jlarkin@highland_atwork_technology.com> wrote in news: snipped-for-privacy@4ax.com:

Large mirrors (what Hubble has) typically have honeycomb back sides to reduce weight and retain stiffness. It only has one 'side' from an optical POV.

Such a lens...

Is that like "Cad" "You're such a cad."

You're such a lens.

Phil Allison snipped-for-privacy@gmail.com wrote in news: snipped-for-privacy@googlegroups.com:

onsdag den 8. september 2021 kl. 23.51.50 UTC+2 skrev LM:

and then you get packages like this

The manuals do not show, but google found out. I'm glad you pointed this out.

Quotes:

Pin 1 is the smaller, concave mark in the lower-left.

Confirm with a continuity tester. For example pins 27,28 and 49,50 are a unique grouping that doesn't occur on other sides.

The other mark is, I believe, an unfortunate artifact of the package encapsulation process, and is not intended to actually be a mark.

But we have customers too. So we want everything to be obvious.

And we don't want assemblers to be working from memory and folklore.

A test fixture was assembled wrong, and nothing was cross-checked. The theory was that the mirror was so good, nothing else was good enough to test it. They never imaged *anything* before launch!

A amateur telescope maker, the ones who grind their own mirrors, could have spotted the error.

Italian semiconductors made in China? That's a win-win for s*it design, assembly and documentation.

I just finished bringing up a board with some 0603 IRLEDs, Osram SFH4050. There is a (tiny) chamfer on one end to indicate polarity, but the data sheet is silent on which pin is which.

Conventionally the mark is on the cathode end, but Osram decided to do this one the other way. They're also very difficult to remove from the board without destroying them. The replacements should be here any day now...

Apart from that issue, they're nice parts. Best output power I could find in such a small package. They're 850 nm, which just-about shows up on a phone camera to check they're working.

snipped-for-privacy@highlandsniptechnology.com wrote in news: snipped-for-privacy@4ax.com:

Ahh... memory and folklore... sure, bub.

Bad stuff. People go on vacation. Or retire. Or forget.

Or A shows B how to assemble something. B tells C. Maybe they don't communicate perfectly.

So we make it easy to understand the drawings and minimize ways to build it wrong.

I've known engineers who created technically perfect ISO-compliant drawings that were incomprehensible without a day of study by another engineer.

The Osram LEDs have beautiful pure colors. I especially like the orange. A fraction of a mA is plenty bright for an on-board test indicator.

I tested a green one for bare visibility, dark adapted in the cabin on a moonless night. I could see 1 nA.

Some LEDs have a T mark on top to indicate polarity, with no obvious meaning.

Or, get sloppy. Or, forget to update/correct the documentation when they encounter an error -- because "we all KNOW that a #6-32 screw is intended for that PEM nut, despite a #4-40 being indicated on the BoM".

Usually, there are connectors and other physical protrusions that refine how components can be fitted together.

That may not be a bad thing -- depending on the intended lifespan of the document and its relative importance in the *system* that it is part of. The last thing I would want is The Next Guy underestimating a design and potentially screwing it up with a poorly conceived revision.

"Gee, here's this (apparently) redundant letter-to-sound rule. I can just elide that and trim the codebase down a bit..."

(Yeah, and now you've caused "thin", "thimble", "theory", etc. all to have a *voiced* 'th' sound, instead of unvoiced. Have a listen to what you've wrought... oh, you're a lazy tester, as well and will rely on the users to tell you of your screwup??)

I have notes on most of my drawings and in my code that effectively tells the next Reader: "There's more here than meets the eye! Read carefully before you go dicking around THINKING that it is 'obvious'"

It's labeled "This side toward object".

snipped-for-privacy@highlandsniptechnology.com wrote in news: snipped-for-privacy@4ax.com:

"A" is a well made set of assembly drawings.

"B" would be an expert at using them if he had any mettle at all.

I do not hire assemblers that cannot connect A and B together when JL is on vacation.

If you have a problem depicting a design incorporating symmetry, then you incorporate an indexing mark. It is really that simple. Oh and have a good, depictive set of drawings. You do not produce product from piless of part you had showing up at the dock without an assembler who is familiar with the first article AND a set of drawings.

another

Then the drawing needed to be broken down into layers to improve discernement of the elements incorporated. Drawings can be "too busy" even though they are "simply showing everything".

Complex mechanical assemblies for example need cutaways, multiple part drawings, yada yada yada...

As do complex IC chip layouts.

Assembly drawing are for assemblers. You know that "design for manufacture" thing. They only need convey the info that particular assembly process step(s) require(s).

One does not have a contract manufacturer make an assembly for one willy nilly and screw up several boards before one catches it. One goes there and makes sure the assemblies get done right and then leaves the contractor to fulfil the contract compliantly.

One step at a time, Johnny.

One well made drawing set at a time if one wants others to interpret what is locked away between one's ears.