i won't buy chinise parts from aliexpress or ebay

[cmrincon]

Hey! Check my new wear for manipulate mosfets

 

[janoc]

If there's Neopixel's with ICs without input protection you should definitely avoid them, there's no real excuse for that. Half century old CMOS ICs without input protection are curiosities you shouldn't be wasting on a practical circuit either. Junctions even of a laser diode aren't as vulnerable as a MOS gate, they can at least avalanche a little without damage.

I don't know whether or not the neopixels have input protection but they are certainly easy to kill.

Even if there is some protection, it isn't foolproof - it is usually just a pair of diodes and with a sufficiently energetic zap you will kill them or cause a latch-up, potentially destroying the part. It is certainly not a replacement for even basic ESD precautions. ESD mats are cheap, humidifiers and/or plants as well.

[Circlotron]

That's not quite true. They were good enough for his project, but a claim of a part "being within specs"  (as in datasheet specs) is something very different. He didn't really test that - and I am prety sure some would have failed if he has pushed them to the actual published datasheet specs.

Good point. My example is I've bought several thousand "ST" M27C801 1Mx8 eproms from a particular supplier over the years. Upside is they are inexpensive, reliable, and only about 1 in 50 won't program fully. Downside is if you scrape off the paint there is some Chinese writing underneath so not made by ST in Singapore as per the outer layer, they take about 2 hours to fully erase and some won't even do that, and quoted access time is 100nS but in reality is closer to 1uS. But... they are reasonably consistent and quite good enough for my intended usage so I stick with them.

I've bought a similar amount of SPW47N60C3 mosfets from the same supplier (TO-247 package) and early in the piece, when I would bend the leads at right angles they would crack and break off. Some were leaky and no functional, some even pressure sensitive in that they would stop working when they were clamped to a heatsink but come good when they were unclamped. Metal surface was all scratchy looking too, not shiny like a genuine one. Upside with these is when used well within their specs they were just fine. Some have been in daily use for 20 years now. Just been lucky I guess.

[nigelwright7557]

Chinese businesses are like most countries, some good and some bad.
The problem is do you just trust your luck or buy from reputable dealers.
I got caught out a few times with bad ebay components and now just buy from RS Components and Farnell components.
In the last 10 years only had 1 problem with Farnell and they had put wrong resistor in the packet.
The packet had what  I wanted on it but resistor inside was wrong.

[blueskull]

If there's Neopixel's with ICs without input protection you should definitely avoid them, there's no real excuse for that. Half century old CMOS ICs without input protection are curiosities you shouldn't be wasting on a practical circuit either. Junctions even of a laser diode aren't as vulnerable as a MOS gate, they can at least avalanche a little without damage.

And they are heat sensitive, plus the bad ones killed by heat may not die immediately, making reliability horrible.
I'll never touch them, and from my limited experience with them they are worse than those fragile GaAs RF chips.

[vk6zgo]

Meanwhile, half century old  CMOS gates have been working happily for, yes, half a century, in old equipment.
Perhaps the "ancients" knew something you don't!

[Circlotron]

Meanwhile, half century old  CMOS gates have been working happily for, yes, half a century, in old equipment.
Perhaps the "ancients" knew something you don't!

The good ones have certainly been going all that time.
The bad ones conked out and got replaced long ago.

[DBecker]

Meanwhile, half century old  CMOS gates have been working happily for, yes, half a century, in old equipment.
Perhaps the "ancients" knew something you don't!

No, the original MOS chips were as fragile as you have heard *as unmounted components*.  Once installed in the circuit they were much less likely to be destroyed.

Pretty much every part today has ESD diodes or similar protection.  Old CMOS didn't have the die area or process capability to do that.  Plus there was much more process variation between wafers, leading to some being especially vulnerable.  The especially weak ones died early, leaving only the robust ones.

[jfiresto]

... Pretty much every part today has ESD diodes or similar protection.  Old CMOS didn't have the die area or process capability to do that....

What sort of old CMOS do you have in mind? Even the original 4000A-series logic found room for ESD networks. Without them, the chips would have probably been too fragile.

[Psi]

The 2nd and 3rd price grouping is what you want to order. These will be good parts.

99% good is gross overestimation. Depending on what you buy they easily may be the same crap as the cheapest tier.

I'm only going on my experiences using aliexpress.
i've encountered more fake parts (that cause actual problems) from pcbway PCBA runs than from aliexpress.


The danger zone for fakes is parts that are specialized or a more expensive version of something common.
Like an extra stable voltage reg
or a low noise opamp
or a low rds mosfet
The most common fakes i've run across are where they take a similar but much cheaper chip and restamp it to be a more expensive version. These sort of fakes are not caught so quickly because the circuit still works most of the time.

[TheDane]

The problem people run into with aliexpess is ordering by price and buying the cheapest option, because they are so used to doing that on other sites.

Having had only a few cases of issues vs. loads of great experiences I can add this to your excellent list:
Also:

Doing some research - take a VERY close look at the pictures posted and compare to what is in the datasheet or google pictures - and you can avoid wasting your money.

Buying from a seller which offers tons of other (non-related/electronics) stuff is also a potentially bad sign.
If they offer a wide variety of related goods, imho they also tend to be more serious and has 'better' parts in general.

Be kind, establish a long term relationship by re-ordering + don't bargain down the price - then ask the seller to look for component xyz at a trusted store/place.
Be a , dispute any little  whilst only grapping the cheapest offerings - and you might make it onto a  list (Many store names, fewer shops, couple of shippers, one receiver)

You get what you pay for, and you get treated as you treat others/'them'.
Be a sceptic and do your own pre-research. Sure it takes a bit longer to order and be kind - but it takes a lot longer for the order to arrive, even longer to re-order.

[wraper]

Doing some research - take a VERY close look at the pictures posted and compare to what is in the datasheet or google pictures - and you can avoid wasting your money.

in 90%+ of cases pictures do not show actual product you will receive. Picture may contain genuine item and you will receive counterfeit or vice versa.

if they offer a wide variety of related goods, imho they also tend to be more serious and has 'better' parts in general.

Nope they still usually sell a lot of fake crap.

[Ivan7enych]

I've had opposite experience recently.

5v regulator from ST-microelectronics 78L05 officially bought in russian supplyer has failed my project - by documentation it is rated up to 30v input, but on my board it fails at 28v input, burned itself and several other ICs.

On the opposite, 78L05 from Ali with unknown marking "WS 78L05" works fine up to 35v input.

[wraper]

I've had opposite experience recently.

5v regulator from ST-microelectronics 78L05 officially bought in russian supplyer has failed my project

Officially bought from crappy supplier which applies to 100% of Russian suppliers to some extent. Whenever you'll buy parts, all have fake parts. Some are crappier than others though. For example you can buy something  from Чип и Дип at ridiculously high price and get a fake crap.

[janoc]

I've had opposite experience recently.

5v regulator from ST-microelectronics 78L05 officially bought in russian supplyer has failed my project - by documentation it is rated up to 30v input, but on my board it fails at 28v input, burned itself and several other ICs.

On the opposite, 78L05 from Ali with unknown marking "WS 78L05" works fine up to 35v input.

That is most likely due to the poor design of your circuit and e.g. heat caused parameter drift/manufacturing tolerances than any poor quality parts. If you are using a part rated for 30V in a circuit where it can get 28V and then you are wondering that it has eventually failed (and blew up other things), you have only yourself to blame, IMO. Never heard about component derating? Operating any component right up to its absolute maximum ratings is a recipe for disaster.

And why, for heaven's sake, are you running an 78L05 from 30V in the first place?! Even at the 5mA quiescent current you are going to dissipate around 125mW in that plastic package. Add any load and it gets much worse pretty quickly. That thing is going to be pretty toasty! Fortunately it has internal thermal protection, but it still makes one wonder what are you trying to achieve there with such setup.

[jmelson]

Yesterday i was learning about mosfets amps and i decided to use a few 2n7000 N-channel mosfet which i got from aliexpress time ago. I did the maths, even i simulated the circuit. When i mounted the circuit in my breadboard i could see that the mosfet was in ohm region and with a volt above VTh the drain current was more than it could handle, no matter if i slightly drop down vgs, the circuit still didn't work because my mosfet didn't behaves as datasheet and it was useless for amplification ( maybe for switching purposes).

Few weeks ago i get capacitors from another chinise vendor on ali, and i used them to repair a sat receiver. Yesterday i got to repair it again, was hard because i couldn't see at first glance what was wrong with the receiver, but after a few test i repaired it. Guess what was wrong? the chinise capacitors!


I can't belive that all chinise parts are like this ones, and i gave my last chance to lcsc. I bought some MMBTH10, ceramic caps and few rf npn bjts from not chinise manufacturer.
In a month i will know if i will loose the money or it is a cheap way to buy parts.

These Chinese junk merchants are my absolute LAST place to get anything from.  I only buy from franchised distributors except for obsolete parts where that's the only people who have them.  Sometimes I need to repair an old product of mine, or make just one more board of an old design that uses an old part.  Other than that, I just won't do it because of this sort of problem.
You wanted a 2N7000, but got some kind of FET device, but who know what?

Jon

[Psi]

Buying from a seller which offers tons of other (non-related/electronics) stuff is also a potentially bad sign.
If they offer a wide variety of related goods, imho they also tend to be more serious and has 'better' parts in general.

Agreed

If the aliexpress seller has listings for 16bit DAC chips, door bells, handbags and bed sheets it's a good sign they don't really know what a 16bit DAC chip actually is.  This means there is a extra high chance of mistakes in the technical listings/descriptions.
They may have added keywords that are incorrect and make you think you're getting something you're not.
And if they're that unfamiliar with the product they're more likely to buy fakes themselves without knowing.
This is why it's usually better to order from a seller that is only selling electronic items.
At least they're more likely to know something about what they are selling.

It's always good to look for inconsistencies in a listing, If the photo implies one thing but the description says something else do not buy from that seller. Find someone selling the item that seems to know what they are talking about.

When i'm looking for a seller to buy a particular item from i'm always looking for any reason to skip to the next, more expensive, seller.
It usually goes something like...

 -Way to cheap. Next.
- WTF. Why are they saying this mosfet has a Rds on of 10V. Next.
- No specs listed, just a photo.. Hell no.
- 91% feedback rating, Nope. Next.....
- This looks ok. (Clicks buy)

[Ivan7enych]

I've had opposite experience recently.

5v regulator from ST-microelectronics 78L05 officially bought in russian supplyer has failed my project - by documentation it is rated up to 30v input, but on my board it fails at 28v input, burned itself and several other ICs.

On the opposite, 78L05 from Ali with unknown marking "WS 78L05" works fine up to 35v input.

That is most likely due to the poor design of your circuit and e.g. heat caused parameter drift/manufacturing tolerances than any poor quality parts. If you are using a part rated for 30V in a circuit where it can get 28V and then you are wondering that it has eventually failed (and blew up other things), you have only yourself to blame, IMO. Never heard about component derating? Operating any component right up to its absolute maximum ratings is a recipe for disaster.

And why, for heaven's sake, are you running an 78L05 from 30V in the first place?! Even at the 5mA quiescent current you are going to dissipate around 125mW in that plastic package. Add any load and it gets much worse pretty quickly. That thing is going to be pretty toasty! Fortunately it has internal thermal protection, but it still makes one wonder what are you trying to achieve there with such setup.

My circuit was designed for 24v input, current is 1mA (3 74hc chips), so my choice of 78L05 looked just fine. Derating was done, no overheating problems. But when input voltage was accidentialy rised above 28v (below 30v, still whithin specs of this regulator), this ST78L05 at first outputs 7V, then blows itself and killed all 74HC chips...

[worsthorse]

If you can buy a part for eighty cents that currently is sold by digikey for twelve dollars, there's only a few reasons for that price differential: it is a degraded copy, remarked fake, reject, or stolen.  There is just no other way to account for such price differences, at least for distributor stocked parts.  The same goes, generally, for NOS parts, especially NOS parts that weren't heavily stocked in the day outside of the United States or Europe.  Just ask yourself, why would there be a big stock of forty year old NVRAMs sitting in some warehouse in China or Vietnam?  Can you find NOS parts? Yes, and I often do, but invariably they are from individual sellers getting rid of personal stock or well established vendors that sell little else. 

I guess it depends on how you want to spend your time and money. Personally, I hate trying to track down circuit problems caused by marginal components, even when building throwaway circuits. And I don't want to have test every part that I buy to figure out if it is what it says it is on the label before I can use it.

So I've never found the money saved buying really cheap parts to be cost or time effective, whatever vetting process one uses to find a "reliable" vendor.

[janoc]

My circuit was designed for 24v input, current is 1mA (3 74hc chips), so my choice of 78L05 looked just fine. Derating was done, no overheating problems. But when input voltage was accidentialy rised above 28v (below 30v, still whithin specs of this regulator), this ST78L05 at first outputs 7V, then blows itself and killed all 74HC chips...

That looks like you got unlucky with a dud part. However, with such high input voltage I would rather use a switching regulator instead of having such a large drop over a tiny linear regulator. If you add the quiescent current of the regulator to your load current, that's quite a bit of power being dissipated.

At least I would pick something with a higher input voltage rating, e.g. LM2936M (40V) or TPS79850QDGNRQ1 (50V). 30V rated part for a 24V input is a bit close for comfort, IMO. 

[wraper]

That looks like you got unlucky with a dud part. However, with such high input voltage I would rather use a switching regulator instead of having such a large drop over a tiny linear regulator.

There is little point going with switcher at 1mA load. Quite likely you will have higher quiescent current than 78L05.

[KE5FX]

Usual cause of 78xx regulator failures is having too much capacitance on the output side.  If the input rail falls faster than the output rail, perhaps because the input is feeding other loads, the capacitance at the output will backfeed the part and cook it.  This happens at turn-off time, of course, so the next time you turn it on, your "regulator" will send the full input voltage downstream.

Designing those parts without an internal diode was nothing short of irresponsible, but it is what it is.  Most 3-terminal regulators have the same hazard, and most of the time it's discussed in the data sheet, but not always.

[nigelwright7557]

I can't belive that all chinise parts are like this ones, and i gave my last chance to lcsc. I bought some MMBTH10, ceramic caps and few rf npn bjts from not chinise manufacturer.

There are good and bad suppliers in most countries not just China.
The trick is finding reputable dealers.
I only use Farnell and RS Components and have very few problems.