Several manufacturers for the same part number

[RDSCANNING]

I'm trying to put together a doc for University students that will give some advice on buying parts and finding datasheets. I basically want to know what differences there could be for the same part number, different manufactuers.

For example a search for LM317 on https://www.datasheetarchive.com/ returns 17 different manufacturers.

Are there any manufacturers to avoid. For example, I've read that you should only use NTE parts as a last resort. That old Motorola parts are very good.

If a student buys the cheapest LM317 they can find, would there be any performance or reliability issues?

Any advice appreciated.

 

[Dean Huster]

I'm trying to put together a doc for University students that will give some advice on buying parts and finding datasheets. I basically want to know what differences there could be for the same part number, different manufactuers.

For the most part, things are identical. However, there are some cases where in critical designs that depend upon a specification that's pushing the envelope for the original part or for an alternate manmufacturer that was originally used, another manufacturer's part may not work. But that's more poor design when that happens.

For example a search for LM317 on https://www.datasheetarchive.com/ returns 17 different manufacturers.

For the most part, any (for example) LM317 from a reliable, well-known manufacturer should work in any properly-designed circuit.

Are there any manufacturers to avoid. For example, I've read that you should only use NTE parts as a last resort.

I agree that NTE and other "replacement type" parts should never be specified in new designs or for published projects. That's like a manufacturer designing, building and selling a car made strictly from third-party parts bought at Auto Zone! Especially in the area of discrete parts like diodes and transistors, consider that NTE may use one of their parts to cover for the substitution of 850 different part numbers. Who's to say that NTE is going to use the same generic transistor for all of their subs for that NTE number?

That old Motorola parts are very good.

Oh, yeah? Then why did I have to replace every Motorola Nixie driver in a DMM due to failure while the drivers from NSC held up fine? I think that you can only make broad and rash statements like that if you say something like, "I've heard that Fluke's current crop of handheld DMMs are 'bulletproof'." Too many component failures are due to poor circuit design, abuse by students or the user and shouldn't necessarily be chalked up to the part or manufacturer.

If a student buys the cheapest LM317 they can find, would there be any performance or reliability issues?

Danged right! Too many "surplus" outfits sell "floor sweepings". In the old days, Poly-Paks was good at that. You end up buying a bag of what have been generically marked "LM317" or "317" -- no date code, no manufacturer's logo, none of that -- and you end up with a part that works similar to an LM317, but it doesn't meet specifications at all. For the most part, as long as the device has been purchased from a reliable source, you're going to be OK.

Nuthin' to quote here.

Do be aware that not all similar parts will work exactly the same. For instance, I made the drastic mistake of assuming that a 7400, 74LS00, 74L00, 74H00, 74S00, 74C00, 74HC00, 74ALS00 -- well, you get the picture -- all logically worked the same and only varied in specs with things like power consumption, speed, driving capabilities, etc. For the most part, that's true. But then I tried substituting a 74LS123 dual one-shot in a circuit where I'd been using a 74123 to find that it wasn't working right. After careful studying of the data sheets, I found out that the reset function is different with the two versions. So, be careful in that area of subsititution.

Dean

 

[RDSCANNING]

Thanks for the reply.

I'll suggest sticking to the well known manufacturers. I'll also suggest keeping away from generic replacements. Digi-key, mouser and alike all seem expensive but I guess they are reliable sources.

I'm surprised about your Motorola parts. I had been told to always buy Motorola or On Semiconductor if they manufacturer the part. I guess this potentially bogus information I've been relying on is a good example of why i need to get it right. Information like this sticks.

I missed a bit off my original post. I was also wondering about the shelf life of parts. My presumption is that they should last forever.

For example, I bought a large batch of TO-220 parts some of which have been obsolete for many many years. They are unused and not pulls, however they are really grubby looking. Lots of dark dirt buildup over the years. Can the age of a part raise any issues if they are unused.

Also, the suffix denotes the package for many parts. Lets say their is no suffix, is it possible that there can be varying packages for a part number when no suffix is mentioned.

Sorry to fire back with more questions. I'm just trying to get as much info as possible.

Thank you.

 

[audioguru]

A Japanese semiconductor manufacturer copies American ICs. They decided to copy the Texas Instruments TL07x opamp and make it better.
They made it always oscillate then had to re-design it!

 

[JimB]

I missed a bit off my original post. I was also wondering about the shelf life of parts. My presumption is that they should last forever...

... Can the age of a part raise any issues if they are unused.

Generally age is not a problem, however some components do die of old age:

Electrolytic capacitors

Carbon composition resistors (although you would be hard pushed to find those these days).

Some older germanium transistors develop internal short circuits due to "tin whiskers", but again you would be hard pushed to find any and would not want to use them anyway.

The biggest problem with dirty components is that the leads can be difficult to solder or make bad contact in a socket or breadboard until they have been cleaned.

JimB

 

[JimB]

Deleted duplicate post
JimB

 

[RDSCANNING]

Interesting stuff,

Thanks for the replies. I've read a few articles on tin wiskers and it seems like the problem is coming back due to RoHS. ie no more lead parts.

Military and aerospace components are exempt from RoHS in some cases. They dont want to take the risk with tin components.

 

[Dean Huster]

Absolutely DO NOT store components, especially DIP ICs in "leaded" conductive foam (the black, matte-finish stuff). It's harder to find these days, but you may come across some. It doesn't hold up well over time and will disintegrate. Worst of all, it will suck the tin plating right off IC pins after a couple of years. The newer black foam is OK and pink poly is always nice to parts.

DO be sure and store static-sensitive components properly to insure that they maintain their electronic integrity. Just throwing sensitive parts into styrene drawers or boxes is asking for trouble. The original "A" CMOS ICs would zap if you just looked at them funny. The "B" CMOS is a lot more robust because of built-in protection on the pins.

The suffixes can get pretty complicated these days, indicating not only differing packaging but different device specifications as well. Generally, no suffix would indicate that a device has no alternate forms, but in these days of converting over from thru-hole to SMT, suffixes are almost mandatory.

Dean