Old or fake chips? See pics.

[rankot]

Hi fellows, what do you think about those:



Fake or not? I don't know how to test anyway, so I must ask gurus!

[ElectricDruid]

What makes you think they might be fakes? Did you get them for some unbelievably low price or something?

Frankly, neither seems likely candidates for fakes: LM13700 is still in current production, and so is the LM4562, although National aren't called that any more, I don't think - bought out by Texas Instruments.

As to testing, I guess you'd run up the datasheet circuits on the breadboard and see if it works.

HTH,
Tom

[rankot]

Thanks Tom! I've actually bought them through Aliexpress. I did the same in many previous occasions, with various results (and decided not to buy JFETs from there anymore, cause all their sellers sell some rejects).

LM13700 chips that I bought before were all working fine, op amps too, so I decided to have another try. But I usually bought some not so fancy op amps - OP275, LM308, LM308H and NE5532AP.

Those are low noise and I bought them intending to build a mic pre (except LM13700N which are used for Engineer's Thumb and other various pedal projects). Prices were:

- LM4562NA -> $1.10
- LME49720NA -> $1.16
- LM13700N -> $0.56

So if any of you have some experience of how those original ICs look like, please take a look at those and compare.

I wanted to buy them through Mouser, but some of them are not available and they charge me $30 for shipping to Serbia, which is huge for such a small order 

[MrStab]

I've actually bought them through Al-

fake!

is the LM13600 significantly cheaper than the 13700? could be relabels? no clue about the op-amps, they could be literally anything, from what i've heard.

[pinkjimiphoton]

far out!
i gotta bunch of these in a bag of parts an older friend gave me when he shut down his recording studio, had no idea what they were. now i'll have to have a play with 'em!!

cool!

[MrStab]

what's the deal with some of those chips having both dots and notches for orientation? just noticed that and never really thought about it. that's a bit weird and redundant.

[Rob Strand]

what's the deal with some of those chips having both dots and notches for orientation? just noticed that and never really thought about it. that's a bit weird and redundant.

Not sure why they have both.  The dots are used for identifying pin 1, you see the dots on flat-packs where top and bottom isn't so clear.   Some automatic part loading machines, and checking machines, use cameras.  It might help different software versions identify chip orientation - but I'm only guessing here.

[anotherjim]

Does the JR prefix indicate they were actually produced by or for Japan Radio?

[rankot]

I've just thought that some of you used any of those original ICs before, so you can do a visual comparison.

I'm pretty sure LM13600 are unavailable cause they stopped producing at least 5 years ago, so I don't think LM13700 are relabelled. LM13700 ICs I bought before (working fine) have a little bit larger font, but have the same letter color (a little bit yellowish) and also have a dot and a notch.

Jim, which ones do you have of those? Can you compare them visually?

[Rob Strand]

I've just thought that some of you used any of those original ICs before, so you can do a visual comparison.

Not by sight.  I've seen explanations in the past showing small clues.  It's a lot of small details.
- printing: type of print, too thick or thin, has blobs or runs, color wrong
- surface of plastic smooth vs textured.
- Re-marked ones show evidence of being sanded or modification and the printing is of poor quality
- inconsistent markings and letters for the manufacturer
- in some cases the logo is replicated poorly
- there's examples which are hard to tell visually but they don't work at all or don't meet spec.

Things like the N logo in the dot might be mentioned in the datasheet.

When you see real and fake side by side it is often easy to see but if you are given one sample and have to judge it you need to have studied up on all the details - unfortunately that's not me.

Try web searching for fakes from the sellers you have bought from.

You will find a lot of examples like this on the web,
https://www.maximintegrated.com/en/app-notes/index.mvp/id/5458

[rankot]

Plastic doesn't look sanded (I've seen that before). But prices are really low. Maybe some kind of rejects???

Letters look quite fine and sharp.

[duck_arse]

PRR once pointed us to this doc, very informative.
http://cti-us.com/pdf/CCAP-101InspectExamplesA6.pdf

moulding marks are moulding marks - they just depend on the moulds. they might have a notch, and a dot at each end, or any combination thereof. one thing to look for is consistency across a batch #. if you've got a half dozen parts w/ same part # and same batch #, and they don't all look to be from identical mould, print, plastic colour and lead dress, you might start to suspect.

also, get the "manufacturers" datasheet, see if you can measure basic parameters within quoted specs, like current draw.

[Rob Strand]

if you've got a half dozen parts w/ same part # and same batch #, and they don't all look to be from identical mould, print, plastic colour and lead dress, you might start to suspect.

That's an excellent way to check for rebadgings of miscellaneous chips.

[MrStab]

just had a wee look through some of my own chips, from (hopefully) reputable sources, and a standard TI TL072 had both the notch and the dot. interesting to know it might be for the manufacturing process, and not the copyright-infringement process.

[rankot]

So I've finally build a test rig and tested my opamps, as described here http://www.ee.buffalo.edu/faculty/paololiu/353/opa1.pdf. Here are the results with 10kHz / 1V input signal:



It is interesting that my locally bought NE5532P seems to be working worse than NE5532P I've bought from Aliexpress, and LM4562NA from the same source seems to be working fine, also. However, LME49720NA from that source is very suspicious, 'cause it outputs only half the voltage of the rest and it has some kind of spikes at square edges.

I didn't make audio listening tests.

And there is another batch of double opamps I have:



What's your opinion, guys?

[ElectricDruid]

I'd be looking for the mistake in the third circuit. Probe stuck on "x2" or similar? "Produces exactly half the expected output level" is just too damn suspicious to be a genuine fault.

Otherwise...jeez, your local store has some shite parts, doesn't it?!? Neither the 5532s or the 1458's are any good!

T.

[Rob Strand]

NE5532P

The NE5532 usually produces nice clean waveforms.

I believe the NE5532 has input protection diodes.  You should be able to measure those.
No diodes =>  doubtful it is a true NE5532.
(There was a recent post on the free st bx forum about identifying sanded opamps.)

[rankot]

Well, this could be a problem with my breadboard, I've always had problems with various BB circuits. I will make a PCB tester with proper sockets and run those tests again. Maybe this circuit (see PDF above) is not good? I will try to make something like this, with unity gain:

[Rob Strand]

Maybe this circuit (see PDF above) is not good?

If you want change the ground connection like that you might be better off using smaller R4, R5  (say 1k), make C2 100uF,  add another 100uF across R5.   Even better would be use an opamp to create the mid rail voltage so you have a low impedance output.

[Kevin Mitchell]

Hey gang I've got another one for ya. These are suspicious because I haven't seen any quite like em' yet (stamp color, blank marker on the bottom). I'm going to test them tonight.

I haven't experienced the dolor of dealing with fake ICs yet - only one pruchace of fake jfets from ebay.
Hoping the tinker gods are looking over my shoulder 



-KM