Author Topic: Identify real OPA627BP (Read 3895 times)

ROSELJR · « **on:** July 02, 2021, 07:16:18 am »

Hey, guys. I just bought an OPA627BP last week. I want to know whether it is real one? And how to identify the real and false one, thanks in advance.

If you need, here's the detail information of OPA627BP:https://www.utmel.com/components/opa627bp-operational-amplifier-pinout-feature-and-datasheet?id=97

magic · « **Reply #1 on:** July 02, 2021, 08:12:05 am »

I was under impression that BB DIP packages have dots near pin 1.

This looks more like old Philips perhaps?

Is the bottom of the package perfectly flat?

At any rate, test them. Plug into a breadboard. Ground IN+, short IN- and OUT. Apply power, check quiescent current. Check offset voltage. Put 1MΩ between IN- and OUT and see if the output changes - this will "out" a bipolar opamp like NE5534 masquerading as a JFET one.

BTW, opamps may oscillate and do weird stuff when loaded with capacitance of DMM probes. Connect a 100Ω resistor to the output and test output voltage through that resistor, not directly on the pin.

eplpwr · « **Reply #2 on:** July 03, 2021, 12:58:22 pm »

Power the OPA627 with +/- 15V to 18V, connect it as a voltage follower and apply a 100 kHz square wave 20Vpp with short rise and fall times to the input. Compare the output to the datasheet "Large signal response". An original '627 is a very fast opamp; all fakes I have encountered are much slower.

magic · « **Reply #3 on:** July 03, 2021, 01:35:54 pm »

That's a good idea too. Basically, look at the datasheet and test any specs you can.

And test with DMM in diode mode between IN+ and IN-. If there is a diode there, it's not OPA627. If there is 12kΩ (or anything else for that matter) between pin 8 and pin 7 (VCC) or any other, it's not OPA627 either. If there is no resistance between pins 1 and 5 and pin 7, bad news again.

If you have a caliper, measure the height of the package. Grinding original markings from the top leaves the chip thinner than it's supposed to be (see the datasheet for dimensions).

My guess: it's Philips NE5534

Kjelt · « **Reply #4 on:** July 03, 2021, 07:37:17 pm »

Look at the underside it should be made in Phillipines or IIRC Korea but in no way China

PA4TIM · « **Reply #5 on:** July 03, 2021, 08:13:54 pm »

Most simple way is to buy them from a company like Farnell, digikey, Mouser etc. And if the real part is 10 dollar and you bought it in China for 10 ct then it probably is fake.

Kjelt · « **Reply #6 on:** July 03, 2021, 08:24:18 pm »

The real parts are around $35 , They were even in the $60 start if the century.

Kleinstein · « **Reply #7 on:** July 03, 2021, 08:48:01 pm »

The OPA627 has low bias (much lower than a BJT based) and also quite good LF noise. So most low cost OPs would give away there lower quality quite easy. Things would get a bit tricky to if they have send you a OPA827 - but usually the fakes are more like the really cheap ones (5534 or tl071 or similar), a reasonable substitute is rare, unless there is a cheap one.

magic · « **Reply #8 on:** July 03, 2021, 08:56:11 pm »

There absolutely is a chance of scoring authentic recycled chips from sites like AliBay at very attractive prices, but given the popularity of OPA627 among audiophools who couldn't tell it apart from µA741 and the resulting huge volume of counterfeits, such chance with this IC is very slim

floobydust · « **Reply #9 on:** July 04, 2021, 01:02:22 am »

Ones I got from Digi-key, look like this copied pic. There is a notch, a smooth shiny dimple, no Ⓡ symbol, and Thailand on the backside dot.
2014 chinese fakes are usually AD744 based on die shots of the SOIC-8's, and are unstable as a voltage follower/buffer and a bit noisier.

I have other SOIC-8 that are fake OPA627's but have no idea what they are. They have the speed.

Jwillis · « **Reply #10 on:** July 04, 2021, 05:21:42 am »

Not sure if the dot at the pin one is conclusive evidence of authenticity .
I brought both these OP445AP at Digikey but about 3 years apart . I do have a few left I bought on EBay and they work the same as the ones from Digikey and look like the ones on the left . All work well above 50V between rails .
All I know is TI has several facilities around the world since Bur Brown is a part of TI . Not sue where exactly they are produced .

I had to adjust the colour and brightness to get the labels to show . Still trying to figure out this camera . Looks better with HDMI .

PA4TIM · « **Reply #11 on:** July 04, 2021, 07:05:18 am »

There is a way, decap the chip and read the type-number direct from the die. If you do not have the fuming acid stuff, you can burn the housing until it crumbles. Then read the die. (the 10 dollar I mentioned was just in general, I had no clue what this opamp costs.

magic · « **Reply #12 on:** July 05, 2021, 07:13:10 am »

Somebody tell me what this metal can Burr-Brown chip really is

https://zeptobars.com/en/read/OPA627-AD744-real-vs-fake-china-ebay

BB logo, some weird drawing (self-portrait of the designer perhaps

) and unidentifiable numbers. JFET input with laser trim and a quasi-complementary output with two arrays of 18 small NPNs each. What the hell. Is it supposed to be high-speed something? Clearly an old or low cost part because it doesn't use their complementary bipolar / low leakage JFET process, unlike true OPA627 and many others.

Offset adjustment on pins 1,5 to ground. "Upgrades µA741 sockets with improved performance", or whatever.

Kleinstein · « **Reply #13 on:** July 05, 2021, 08:36:03 am »

The OPA627 is made in a ratgher special SOI (DIfet) process. This makes it rather expensive but also low leakage.

The metal can fake looks like a rather complicated part with quite some resistors. So it is likely some older parts - not necessary low cost when new / still in production. The die is quite large. The actual properties like supply current , GBW could give some more hints on the part.

magic · « **Reply #14 on:** July 05, 2021, 01:19:59 pm »

By Burr-Brown standards, "low cost" is something like their audio grade opamps such as OPA604 or OPA134. Not exactly jellybean, but far behind the DiFET chips

This puppy looks quite scary due to a few common centroid pairs with all connections routed on one layer

but the signal path is actually pretty simple. Main differences from TL072 are input degeneration and cascode and an all-NPN output stage.

Still no idea what it is, but none of the audio parts mentioned above for sure.

edit
I'm inclined to side with the "very old and very good for its time and very expensive" hypothesis.
It's built on a noncomplementary process.
It doesn't use the "noise-free, die area effecient cascode circuit" patented in 1985 but expends a lot of area on cascode JFETs

Probably a fast part, PNPs are eliminated from signal path. Even the PNP making I5 is cascoded with a JFET (the largest of them all, btw).
Laser trimmed input stage resistors, probably a precision part.
Compensation capacitance consists of two metal-oxide-semiconductor caps in anti-series. WTF. Perhaps for distortion...
Common mode input range starts maybe 2~3V above ground and reaches almost VCC.
Output goes close to ground but not up to VCC.

Still no idea what it is.

Terry Bites · « **Reply #15 on:** July 06, 2021, 04:12:16 pm »

As mentioned by others the test is in the in curcuit performance.
See https://www.aeri.com/counterfeit-electronic-component-detection/

magic · « **Reply #16 on:** July 06, 2021, 09:12:43 pm »

Of course

OP should audition the new chip critically.
If it sounds good, it's genuine.
If it sounds ordinary, it's NE5534.
If it sounds like 1960s transistor amp, it's LM358.

magic · « **Reply #17 on:** July 06, 2021, 09:30:39 pm »

Winner winner chicken dinner

The fake OPA627 from zeptobars is OPA445 - a fairly slow, but high voltage (±45) JFET opamp from 1987.
The doubled capacitor has to be simply for increased breakdown voltage

floobydust · « **Reply #18 on:** July 06, 2021, 10:51:09 pm »

How do you get OPA445? Zeptobars die shot is AD744 and to make it worse the Hack-a-day article calls it an AD774

I'd like to send my fakes to see what they really are. For the $30-40 TI is charging, it should have some anti-counterfeit marking technology. As we all know, there is only one laser beam in china doing the work. OP might be able to improve the effort.

$113 for OPA627SM TO-99 can, take that LTZ1000A only $88 USD.

magic · « **Reply #19 on:** July 07, 2021, 06:15:31 am »

Both are fake. One was immediately identified as AD744, the other was initially assumed authentic ("OMG so many lasered resistors and BB logo, it gotta be the real deal") but later it came out that OPA627 looks a little different

See the update at the bottom:

Quote

Metal "OPA627" appeared to be some unidentified BB part remarked as OPA627.

Related:
https://zeptobars.com/en/read/BB-Ti-OPA134PA-FET-opamp-trimmed-low-distortion
https://zeptobars.com/en/read/Ti-JFET-TL051CDRA-TL051
I don't understand the "extremely similar" comment, they are obviously identical and we all know what it means

Quote from: floobydust on July 06, 2021, 10:51:09 pm

I'd like to send my fakes to see what they really are.

I believe the zeptobars guy moved places and currently doesn't have a lab to decap plastic packages, but he accepts metal cans or cerdips. Contact info at the website.

I'm sure that Noopy would be happy to take them too.

Kjelt · « **Reply #20 on:** July 07, 2021, 07:42:10 am »

Quote from: magic on July 07, 2021, 06:15:31 am

I don't understand the "extremely similar" comment, they are obviously identical and we all know what it means

Well the text differs and perhaps the laser trimmed resistor accuracy might differ ?

magic · « **Reply #21 on:** July 07, 2021, 08:39:19 am »

But they are not two grades of the same IC

Specs are quite different, one has phase reversal but not the other, one is supposed to have input cascode, offset trimming is done with a pot to different supply rails and on different pins, ...

And what is 1995 TI doing on an IC supposedly made by Burr-Brown? TI only bought them in 2000 or so

edit
Good point about laser trimming. OPA134 is supposed to use it, but here there is only a bunch of fuses or zener zaps on the left. Anyone cound notice that

magic · « **Reply #22 on:** July 07, 2021, 12:19:41 pm »

Quote from: Kjelt on July 07, 2021, 07:42:10 am

Well the text differs and perhaps the laser trimmed resistor accuracy might differ ?

You aren't wrong, though

In a stroke of genius it occurred to me that if EX051C is TL051, then EX087C...

could possibly be...

drumroll...

Yes, it really exists: TL087

0.5mV max offset voltage

floobydust · « **Reply #23 on:** July 18, 2021, 06:12:47 pm »

A quick test for OPA627 fakes is to measure the resistance between offset trim pins 1 and 5.
Many say it should be around 50kΩ and my genuine OPA627AP's read around 60kΩ. DMM polarity should not matter. Pin 8 is N/C.
A second check is the resistance to +Vs pin 7 of either pin 1 or pin 5 will be half that value.
AD744 reads open between pins 1,5,8 on low power ohms.
Some chinese OPA627's I bought a long time ago read 50kΩ so I guess they are contenders.

magic · « **Reply #24 on:** July 19, 2021, 07:51:22 am »

Also look for parasitic and/or protective diodes between I/O pins and voltage rails. These may be unusual (absent) on OPA627 because it's built on a SOI process, unlike most others.


EEVblog Main Site	EEVblog on Youtube	EEVblog on Twitter	EEVblog on Facebook	EEVblog on Odysee

Author Topic: Identify real OPA627BP (Read 3895 times)

Share me