NASA Technical Standard: Crimping, Cables, Harnesses, and Wiring (2011) [pdf] | Hacker News Reader

NASA Technical Standard: Crimping, Cables, Harnesses, and Wiring (2011) [pdf]

hq.nasa.gov

70 pointsGuiA18 comments

Page 50: Hey, that's how you're supposed to use heatshrink! Wow, I've been doing it wrong for years.

Page 62: Mildly surprised that they don't want crimped connections soldered, but I suppose that compromises flexibility, and shouldn't add all that much strength if they're properly crimped.

Page 76: Wow, had never heard of "connector saver" jumpers before. Sounds bananas, but I suppose if you're going to test everything ten times for every launch, it's mostly reasonable.

> Page 50: Hey, that's how you're supposed to use heatshrink! Wow, I've been doing it wrong for years.

How did you get it to stay if you didn't do the ends first?

> Page 62: Mildly surprised that they don't want crimped connections soldered, but I suppose that compromises flexibility, and shouldn't add all that much strength if they're properly crimped.

Soldering a crimp really only helps if you've done a poor job of crimping; but it's relatively easy to inspect a crimp. If you solder, you could cause a previously excellent crimped join to be not so great, and it may be hard to see if the crimp opened a bit due to heat, or other things.

> Page 76: Wow, had never heard of "connector saver" jumpers before. Sounds bananas, but I suppose if you're going to test everything ten times for every launch, it's mostly reasonable.

I've seen these in stores for ISA/PCI slots; since all connectors have a limited number of mating cycles, it makes sense to use them to reduce the number of cycles used by testing, since you can't easily go and re-seat your connectors once you've launched the thing.

> Soldering a crimp really only helps if you've done a poor job of crimping; but it's relatively easy to inspect a crimp. If you solder, you could cause a previously excellent crimped join to be not so great, and it may be hard to see if the crimp opened a bit due to heat, or other things.

Solder joints are also more likely to crack and fail than crimped joints due to vibration. You solder as little as possible in air and space craft.

And much of the soldering is done with solder ferrules (aka solder zaps) that sit over a wire and release solder when exposed to heat. This is all contained in a neat little package. The solder is ultimately connected to a wire that you can then run to a connector. This is how we usually will connect a braided shield to a component ground.

Tapers (live music recorders) used to do something like this on the old Sony TCD-D8 DAT recorder. The ten-cent 1/8" jack on your $600 device wasn't exactly readily accessible for DIY repair, so people would put the D8 inside a box with panel mount jacks, and run a jumper from the internal to external jacks. I've used the same principle multiple times since then, in other ways - on headphones, computer audio jacks, etc.

That touches a very interesting problem imo: there are some really cheap components in most products that completely compromise their usage, but are prone to failure. Most times, throwing more money at them won't improve the situation significantly, because they are material-constrained (i.e. the material has to be plastic, and it's not viable to make a plastic that wears out 100x slower than the average plastic). The only solution then is to modularize the problem for easy replacement: you can either put it in "maintenance" perspective, or you can spawn a new product out of it.

Yet I feel many major manufacturers fail to see this.

For example, there's the fans in desktop computers -- a $5 fan is about as good as a $50 fan (slight improvements in noise etc), but they're instrumental in keeping your valuable components from overheating, and they fail from time to time, so if one stops working you just get another one and after a few screws you're done.

My Sony phone has a plastic exterior that just snaps out and you can replace it without any tools -- I can get find this part for like $5 -- and it makes so much sense. If they used some much more resistant not-easily-replaceable material the difference in the end will still be breaking from dropping from 1m in concrete to something like 1.5m.

If you're going to insert/remove a USB device several times a day, a 6" extender cable will keep your motherboard jack from going defective. Less than a dollar apiece, too.

There are cheap extender cables for PS/2, audio, VGA, DVI, HDMI; I don't usually recommend them for RJ45 because you can compromise signal quality on gigabit connections, but unmanaged 5 port switches are cheap enough.

One rule of thumb: if the cost of the labor in replacing a component is half or more the cost of the component, see if you can insulate yourself from replacing that component. When pulling cables through buildings, the labor cost far outweighs the cable cost, so pull many spares at the same time.

I currently work as an aircraft telemetry and instrumentation engineer, and have also done similar engineering on rockets. So I spend a lot of time thinking about (and have been trained very thoroughly about) connections and how to do them.

> Hey, that's how you're supposed to use heatshrink! Wow, I've been doing it wrong for years.

Out of curiosity, how have you been doing it? The way shown in this manual is the standard way to do it in my world.

> Mildly surprised that they don't want crimped connections soldered

I'm confused as to why you would think about soldering a crimped connection. Properly crimped connections will stand up to a good deal more vibration (and are pretty much gas-tight, staving off corrosion) than soldered connections. Plus, crimping is quick and easy with the right tools. Maybe there are some niche applications where you'd do both. I've never seen it.

> Wow, had never heard of "connector saver" jumpers before. Sounds bananas

Totally not bananas when you look at the spec sheet for something like a D38999 series connector. Connector savers are a normal thing in the aerospace world. Most connectors are only rated for a few hundred mate/demate cycles (Usually 250 or 500). Every time you mate or de-mate a connector you run the risk of damaging a pin or socket. So the connector savers are sacrificial for when you test. They get mated to the real connections once at the beginning of your tests, and demated at the end. Then you give your real connections a thorough check at the physical level and hook them up.

I work with 1000-1200C resistance furnaces, and my understanding is that the internal connection between the heating element and the power leads is both crimped and welded. It is probably a niche application, and welding and soldering aren't quite the same, but it does happen.

And this is one reason why I love HN, and why I threw in that caveat, because I knew someone would come up with a scenario where it's done. Thanks.

Like I said, I come from the aerospace world, where as another user here said, we solder as little as possible. Sure, components on PCBs are soldered (and even then, they are often covered with an inert, non conductive coating or RTV to fix them in place), but when it comes to wiring and connections, solder joints are asking for trouble when they get vibrated.

I'm confused as to why you would think about soldering a crimped connection

Because most crimped connections (at least in the hobbyist/DIY world) are crap. So after crimping, the only way they hold together is with solder.

There are no applications where this is better. Every connector manufacturer I know recommends against doing it. You can get away with doing it if you support the soldered end against vibration, but you should do that anyway.

It's really a training issue. People haven't been taught better and there is a lot of bad advice floating around hobbyist forums. I have only been using good crimping tools for the last few years since I started making a product with a 50-conductor harness. Until I spent $200 on a crimping tool and took the time to research how to make good crimps (Molex has an excellent document), I never realized how they were supposed to look. Now most of my tools are used, purchased at auctions of dead companies, but I have $400 crimping tools I paid pennies on the dollar for. Even so, I normally farm out crimping to a company that does it with automated machinery, better and faster than I can do by hand.

Most of the crimpers I use are $1000+. I was totally shocked at the expense when I first switched from home hobby to professional work, but the quality of the crimps is simply incomparable. Very much worth it for reliability.

I don't think I've seen a manual tool that was so expensive. Are you using some kind of pneumatic crimper? I thought of getting even better tools, but I found an outfit that will cut and strip wire to my spec. and crimp on any contacts I need, and they do it cheaper and faster than I can.

No, that's about how much a good manual crimper costs. But man, the results you can get out of them are fantastic.

Yeah, this is the difference between the hobbyist world and the aerospace world. As I've said before, that's where I spend my time...But I can understand why people don't have crimp tools and dies at home because just the die for a certain type of crimp connection can be several hundred dollars. That's just not doable for the vast majority of hobbyists.

Thanks for pointing this other side of the equation out.

Technical standards would be vastly improved by adding a because-clause to all of the do-this and don't-do-that requirements.

Section 9.5: "NOTE: Do not use spiral wrap sleeving on mission hardware including launch vehicles."

If you knew why, then you could make intelligent decisions for cases to specifically covered by the standard. Possible reasons: Poor strength to weight ratio, unreliable under high-g load, makes visual inspection difficult, there is a better but more expensive substitute, etc.

From a point of view of a person not employed at NASA manufacturing plant who would like to learn something from that document, I agree. But my guess is that this standard is created for people who are paid to obey, not to think, because there are decisions made higher up the chain that depend on the components being deterministic.

This reminds me of an anecdote I saw on HN once, about Apple hardware team sending boards back to Chinese manufacturers who replaced a capacitor (or some other part) with a cheaper one without realizing it was vital for the product to work for reasons they were not aware of.

Years ago I trained as an (industrial) electrician with a particular focus on hazardous environments (flour mills etc, lots of suspended flammable dust), lots of the stuff in here is familiar :).

Working with armoured cable all day was enough to make me decide to go do something else as a career.

Reply on news.ycombinator.com