11
\$\begingroup\$

I put in an order for 50 Nexperia 74HC173D,652 with JLCPCB for some kits I'm having built and assembled by them. These are going into electronics kits that are sold to high schoolers as part of a digital electronics curriculum, so there are no automotive or military requirements.

The supplier sent me an email this morning stating

C546651 you requested to purchase belongs to a batch that was produced in 2016.

May I know if it is acceptable for you? If yes, we will help to proceed further.

Do folks think this is an issue assuming it is coming from a reputable distributor such as Mouser, Digikey, Verical, etc., or should I run from this?

\$\endgroup\$
10
  • 1
    \$\begingroup\$ Can you edit your question (don't post new info in comments) and add how many you're building, what they're for, safety/reliability requirements and much more. Without knowing that, answers must be guesswork or opinions and the question could get closed. Lots of info about the application, please. \$\endgroup\$
    – TonyM
    Commented Nov 22, 2023 at 14:09
  • \$\begingroup\$ Depends on your application! For general stuff: No, not really.... 7years is not a lot for ICs. Automotie, aerospace, etc.... You really want to buy your parts from a company which gets the "cheapest available at Schenzen market today!"? Different storys for prototypes, but for sale.... run from it. \$\endgroup\$ Commented Nov 22, 2023 at 14:26
  • \$\begingroup\$ Old leads can have solderability problems if they've not been stored properly. The silicon inside ought to be OK, regardless of date. \$\endgroup\$
    – Neil_UK
    Commented Nov 22, 2023 at 14:27
  • 4
    \$\begingroup\$ I'm still using chips I bought from Maplin and Watford Electronics (both UK) in the late 1970s! These are only for one-off home projects but I don't remember finding a dud. \$\endgroup\$
    – Transistor
    Commented Nov 22, 2023 at 14:36
  • 2
    \$\begingroup\$ @ElectronicsStudent, each of your repeated exclamation marks reads as shouting and angry. Your comments would express themselves much better without them. \$\endgroup\$
    – TonyM
    Commented Nov 22, 2023 at 15:46

2 Answers 2

13
\$\begingroup\$

Short answer: This is fine, just let them assemble the PCBs with those parts. There won't be any noticeably elevated defect rate. (You'll still have to expect that there can be one or two defective boards in the batch, for all kinds of reasons.)

Long answer: It depends on the part and its package.

In general, solid-state devices have practically infinite shelf life. They're solid-state after all, which means that there are no moving parts and no liquids that could dry out. They're essentially just "fancy rocks" - the atoms within them are bound in solid crystal lattices and aren't going anywhere as long as they're not in an environment that degrades them chemically (i.e. acidic or basic conditions).

Another aspect to a part's shelf life is its package. Very fine pitch SMD packages, especially BGA, can accumulate moisture in the bonds between their metal leadframe / substrate and polymer encapsulant. When you then heat them up in a soldering oven, the moisture suddenly turns into steam and essentially explodes the package from the inside. This is called "popcorning".

Your SOIC packages, on the other hand, are so large and simple in comparison to modern BGAs that popcorning in them is practically unheard of. You could ask JLCPCB to bake the old parts as a precaution if you're still worried about this. Baking (i.e. 24 hours at 110°C for your SOIC parts) is the industry standard procedure for drying parts before soldering if they might have absorbed moisture during storage.

You have to keep in mind that there are distributors that specialize in stocking old and discontinued parts, like Rochester - if silicon chips rotted in storage, they wouldn't be in business.

Non-solid-state parts are different, of course - I wouldn't want to use electrolytic capacitors that are a decade old for a production run, for example. (Ceramic caps, on the other hand, are fine even when they're quite old.)

\$\endgroup\$
1
  • 1
    \$\begingroup\$ This is a great answer. I would add also that you should be wary of new old stock for electromechanical parts, such as relays. Ask me how I know... \$\endgroup\$ Commented Nov 23, 2023 at 10:23
9
\$\begingroup\$

NXP's standard is 5 years, and they can sit around their customer for another year after that. There is no real shelf life concern with semiconductors like that, just solderability. If properly stored most parts will remain reasonably solderable for 20+ years in my experience.

On the other hand, according to an EEVBLOG posting, Digikey have refused return of electrolytic capacitors that are already 7.5 or 8 years old. Aluminum electrolytic capacitors, even if stored in a good (cool dry) environment do eventually dry out, so that's a chunk taken out of their useful life.

It sounds like JLC (presumably through their sister distribution company LCSC) are being proactive. In mass production a part that's a bit less solderable could result in lower yields and rework. In one-off kit assembly, it's just a bit more time with the soldering iron and maybe the joint doesn't look quite as wonderful.

I would have zero concerns myself on the basis of age (of this general magnitude), and find them proactively broaching the issue refreshing. LCSC seem to be a good company, even if their prices are sometimes higher than others in the China market. Of course the manufacturers they carry sometimes offer parts that may not be what you expect, so always read the actual datasheets for the specific manufacturer. In this case, it's Nexperia so no worries.

\$\endgroup\$
1
  • 2
    \$\begingroup\$ There are many things that can go wrong in the manufacture of a board. If using six-year-old parts increases the likelihood of failures by a ten-parts-per-million, someone who worked hard to reduce failures from all other sources below one part per million would likely view that as unacceptable, but someone whose processes were cruder and expected a 1% failure rate wouldn't even notice something that increased the failure rate by 10ppm. \$\endgroup\$
    – supercat
    Commented Nov 23, 2023 at 17:37

Not the answer you're looking for? Browse other questions tagged or ask your own question.