#311

Rumors say that a lot or even most of the active parts like transistors on AliExpress are fake. This would have severe consequences for our purchasing. So, let’s check my randomly purchased transistors to see if this is true and also establish a simple procedure to prevent obvious fakes in the future.
In my last video, we discussed transistors, and I mentioned that parts on AliExpress might be faked. This sparked quite some comments, sometimes quite harsh. Today I will show you how you can test your transistors to make sure they fulfill the most important claims of the datasheets. This video is intended for Makers, not for manufacturers of products.
To save us money and hassle, we will:
- Understand how we can test MOSFETS. They are the cornerstone of our “survival kit.”
- Build a simple setup to do the tests
- Test my transistors purchased on AliExpress to see how many are faked products
- I will propose a simple and fast “incoming inspection” for your AliExpress purchases to avoid you are cheated
I am a proud Patreon of GreatScott!, Electroboom, Electronoobs, EEVblog, and others.
Links:
N-Channel:
IRF3708 (TO-220) s.click.aliexpr...
Si2302DS (SMD) s.click.aliexpr...
P-Channel:
NDP6020P (TO-220) s.click.aliexpr...
Si2301 (SMD) s.click.aliexpr...
Transistor Tester: s.click.aliexpr...
Transistor Curve Tracer incl power supply: amzn.to/2O7FY6t or ebay.to/2t7YX9Z
without power supply: amzn.to/3aQWthi
W2AEW Channel: / @w2aew
CommissionsEarned: The links above usually are affiliate links which support the channel (no additional cost for you).
Supporting Material and Blog Page: www.sensorsiot.org
Github: www.github.com...
My Patreon Page: / andreasspiess
My Bitcoin address: 19FSmqbBzb5zsYB1d8Bq4KbxVmezToDNTV
If you want to support the channel, please use the links below to start your shopping. No additional charges for you, but I get a commission (of your purchases the next 24 hours) to buy new stuff for the channel
For Banggood bit.ly/2jAQEf4
For AliExpress: bit.ly/2B0yTLL
For ebay.com: ebay.to/2DuYXBp
www.facebook.c...
/ spiessa
www.instructab...
Please do not try to Email me or invite me on LinkedIn. These communication channels are reserved for my primary job

[@arcadeuk]

Almost every IC or transistor I have tried to buy from Ali Express has been fake. The thing you did not test here is current limit. With the "working" fakes, they usually use much smaller silicon than the real parts to save money, so you end up with 20amp T03 transistors that smoke at 2amps. If you dremel them open you will often be able to see the tiny silicon

[@svenpetersen1965]

I have my own theory about the "faked" parts purchased from "dubious" Chinese sources: They are actually original parts, that did not pass the quality check. It might not always be sure, that those parts are destroyed, but maybe get picked up, labeled etc. by "evil" sellers and then just be sold to "private" people. The Pi1541 is a floppy disk emulator (for a retro computer), that actually works, but some people experience problems with "Chinese" 7406 drivers. I don't think, that somebody do all the effort and create an almost properly working "fake" 7406, but they might take parts, that do not meet the maximum propagation delay or another paraemter and sorted out as defective.

[@AndreasSpiess]

I agree for special and more expensive parts. I only tested very cheap standard transistors where I even cannot believe there would be enough bad production runs. It could be that they are relabeled. In this case i think I would have discovered it.

[@JanCiger]

@@AndreasSpiess Common way how fake parts are created is by relabeling a lower spec part as a more expensive higher spec one. Or a Chinese clone with slightly different specs gets relabeled/sold as the genuine part. Your test won't really catch that, only the gross fakes (a totally wrong part relabeled). This is common for both opamps, power transistors and even electrolytic capacitors. On the other hand, if your project isn't critical (most hobbyist stuff), it likely won't matter much as long as the part does what it is supposed to and you don't need the limit parameters (i.e. low noise precision/audio opamps, large current mosfets or power transistors for an amplifier/switching supply). For professional application (= something you are going to sell) you would be a fool to buy components from AliExpress or eBay, though. That said, even in Asia it is not a problem to buy genuine parts - e.g. LCSC or other reputable sellers will not sell you relabeled fakes.

[@SianaGearz]

Your theory is easily refuted. There is a large relabelling business in China. Often they just freshen up genuine salvaged parts to make them more presentable and easier to use in manufacturing, they straighten and refinish the legs, and they grind down the label side and repaint and relabel it. Obviously nobody will produce a new IC specifically for a fake, but for many categories of semiconductor products, you will find a lot of possible substitutes. 7406 is a generic part. Soviet Union started making one in the 70s designated K155LN3, and everyone else has one too, if Soviet Union had one, means China had one too going all the way back. There are literally dozens of different die designs (unlicensed) for the whole 40 and 74 series chips with same pinout and function, but different electrical spec. As far as MOSFETs go, you can do a quick check for gate capacitance, for fake ones it's usually not even in the same order of magnitude as the genuine part, so they have a completely different die design. Often enclosure differences are seen, and when they are decapped, a die of different (usually smaller) size is found inside. You will also find that whenever there is a REALLY popular part, whether Western or Japanese, you will find that someone spins up a part in China as a substitute to allow device manufacturers to reduce costs. I could give you a million example but i'll start with just one, the STM32F103 processor has been "cloned" by two different manufacturers, GigaDevice as GD32 series and CKS as CS32 series. They are not direct clones - they are compatible processors with same pinout, similar and largely compatible peripherals, and ARM licensed identical core, but they show behavioural differences and manufacturing process differences. So far so good, this is completely legal and valid. But then you'll find processors in the wild that are marked STM32 and vaguely made to look genuine, but are definitely not, and are instead actually these compatible replacements, see Alex Kenis for proof, and i think i might have one of those too in my pile of parts. And mind you, these CS32 and GD32 chips definitely cost a bloody FORTUNE to spin up, because they've got everything: high-density processor, flash, mixed-signal for comparators and ADCs, power semiconductors for the internal linear regulator, RC oscillators, they have just about every engineering complication you can think of, and yet it evidently made sense to build them. For some parts, it's certainly the case that there exists no fake die and every device is genuine, but often cosmetically manipulated. Some single devices are extremely likely to be fakes. The ratio of fakes varies depending on viability, and many reputable Aliexpress sellers are pretty good at dodging fakes for the most part.

[@juststeve5542]

I can explain how some of the genuine parts can be so cheap. Water absorption. SMD packages need to be dry for the soldering ovens or they can crack, so components have to be dry. They arrive sealed, but will start taking in moisture the moment they are opened. Depending on the component packaging (and climate) you could have from a day to a couple of weeks to use them. After that they're just no good for industrial soldering ovens. However, they're perfectly ok for hand soldering. They can be dried out, but it's usually just not worth it for the companies, so they go in the scrap bin (and get "recycled"). Skip diving and part recycling is an industry in itself in China. I would salute their green credentials if it wasn't done purely for money and their other industries weren't so polluting!

[@EgonSorensen]

I would say inappropriate ESD + MSL precautions during handling, shipping and experimentation/installation is what creates a lot of issues. Want to pay for proper handling and care? - you do pay a higher initial price, and (hopefully) avoid a lot of issues and problems down the line if you do your part right. Want to gamble - buy from an unknown source, some are legit and honest - you (hopefully) win. Others aren't, and you (certainly) loose. This is my experience, not a theory.

[@cls9474]

I buy my semiconductors mostly locally from reliable sources (digikey, reichelt, distrelec). The little extra cost - especially for the small actives like transistors etc - is easily worth the time I would otherwise spend debugging bad parts. I can focus on my own design flaws :) Additionally I find it somehow inefficient and not ecologically viable sending single transistors in small quantities around the globe. I buy stuff from AliExpress but do usually think twice if it makes sense. I think it's a very good discussion you sparked here! Thanks Andreas

[@TheBodgybrothers]

It is also much faster with digikey (or similar). Even in Australia if I order a half reel of transistors it takes 4 days. My last order was made on a tuesday and delivered on the next monday.

[@excitedbox5705]

Why would you buy single transistors? Buy a set of 25-50 of each for the $36-50 it costs and never worry about not having a fitting part. I only order from 4.8+ rating suppliers and have never had a serious problem. Any time something has been broken they refunded my money and let me keep the item. I paid 3.50 for 20 transistors from a store here and could have gotten 180 on aliexpress for the same price. For many products you can buy directly from the manufacturer on Aliexpress and save a lot of money. Voltcraft for instance sells rebaged Hantek stuff for crazy prices. My MSO5202 would have cost over 800 Euros and it sells for ~480 from Hantek. I paid 285 because I had ordered the DSO4102C but the seller sent me the more expensive one by mistake :D Banggood on the other hand is crap. The same Aliexpress products for more money and crap service. They cancelled my first payment and refused to give me the new member discount when I ran it again. Let me wait for 3 months to get part of my order and my multimeter had obviously been tossed around and was dirty. There were multiple items in the box and that was the only one covered in dirt and broken. Whoever packaged my order saw that it was dirty and put it in the box anyway and that is just unacceptable. Also the Microscope I ordered was listed as 3mp for $27 but was really a 1.2mp microscope that sells for $12.

[@skeptical_bystander]

Let's check what is "the little extra cost" :-D. I live in Canada, so I'm using CAD. Aliexpress: 100 of SI2301CDS cost 1.85 CAD + 1.39 CAD shipping. Digikey: 21.36 CAD + shipping. I didn't even bother to check shipping cost. So "the little extra cost" is >18.22 CAD. Maybe this is "little extra cost" for you. But for me electronics is a hobby and paying extra 600% (shipping not included) is way too much. The same story with Mouser.

[@cls9474]

If just one of the 0k1 transistors suffers from borderline behavior and forces you to spend two hours debugging on your hobby wage of as low as CAD $ 9 you already made up for that "reasonable extra". And don't forget the extra time disputing the seller. Additionally the professional suppliers handle their parts correctly. ESD-damaged parts mostly work but have a much lower life expectancy. When it comes to power transistors I'd consider the risk of fire too. DigiKey is expensive but not the only supplier. Here in Europe we have other distributors which are more economical at lower quantities. But yes, if all the factors add up "in phase" it can be worth the risk of course. Simultaneously I try not to forget that someone works for me a a wage which I don't like to work at. The prices that we pay in the western world better match the costs that we cause with our acquisitions - socially and ecologically: shipping just 100 transistors via airmail from the other end of the world isn't probably the most ecological mode of shopping. And counterfeiting isn't something that goes along with my ethical standards as well. But that's my personal opinion. We see: buying transistors isn't that easy :)

[@excitedbox5705]

@@skeptical_bystander Exactly. If you buy quantity the prices do come down to the same levels though. Just small quantities make no sense. I have been trying to organize a giant group buy. If I could find 100 people to buy a full set of all resistors we could have 100 of each value for $5-6. The other option would be to make loot boxes. Each month you get a bunch of passives and a few transistors and connectors, some ICs, some consumables and 1 bigger item like a stepper or an ESP32 dev board. That would only make sense though with at least 1k orders a month guaranteed.

[@skeptical_bystander]

Thanks, Andreas! I like your approach - "fit for the purpose". This is exactly what hobbyist needs.

[@AndreasSpiess]

It is a good concept also for other areas...

[@InssiAjaton]

We got a batch of "Ixys" FETs from a reputable distributor. When they misbehaved, we started a communication with the distributor and eventually with Ixys. Ixys checked a few of our suspect parts and immediately reported that the FET date code was illegal. They also checked the chips (yes, plural like 2) chips in the package. Real Ixys part had just one chip. After that, I myself confirmed the finding. The fake parts indeed contained two chips inside, while a genuine Ixys part had just one, larger chip. The distributor admitted having received the lot in question from a "wholesale rep" of Ixys, not directly from Ixys. Well, so much for reputable distributors...

[@AndreasSpiess]

I still think the chance is smaller if you buy from a source with supplier check.

[@Rob2]

You have to be careful when using a bench power supply as a current source, and even more so when you want it to transition between constant-voltage and constant-current with a varying load. Many bench power supplies have way too much capacitance over the output terminals in an attempt to filter noise from the (switchmode) supply. When you set the supply to 30V limited at 1A you will be charging the output capacitor to 30V. Looking carefully during your video you can even see the charging current indicated on the current meter of the supply! Now, when you suddenly short the output by turning on the FET, the current will rise to much more than the set current limit on the supply, because that current is delivered from the discharge of the output capacitor, not from the current-limited supply. This can lead to current spikes that could easily destroy your device. I would advise to always incorporate a series resistor (as indicated in the first schematic you have shown) that has a value such that the maximum current you want to go through the circuit can never be exceeded even with the supply at the set voltage. E.g. a 1 ohm series resistor to limit the current to 30A at a voltage of 30V. Of course, a larger resistor for smaller parts that would be destroyed by the 30A spike.

[@AndreasSpiess]

You are right. For bigger transistors I thought my crappy connectors and cables were enough. For smaller ones it would have been a good idea.

[@heyarno]

That is why I set the current first and raise the voltage after connecting the part. I did lose some LEDs until I noticed what is going on.

[@Rob2]

@@heyarno That is good enough for static parts like LEDs, where the voltage-current curve is constant and you only need to watch out when connecting the device. However for a FET it is different: you can connect it and then increase the voltage, however when you switch it on it will still draw much more current than the value set in the supply. (of course you can try to turn it on slowly so the peak will be less than when you suddenly switch it on)

[@jimb032]

@@heyarno this is always a good idea as well, because many of those cheaper power supplies with output filtering do tend to overshoot initially. So your solution is again a good idea for this test method.

[@Rob2]

@@Ormaaj Of course the preferable device for such measurements is an SMU (Source measuring unit). That is a power supply that is specified for usage as a voltage or current source in 4 quadrants, and also measurement of the resulting values. For "testing with what you happen to have available" a bench power supply and one or two DMMs is OK, but don't trust the capability of a bench powersupply as a current source... use a resistor to help it with that.

[@muppetpaster]

There is also a lot of production surplus dumped on Aliexpress and Banggood and that type of sites. Also a lot of re-sourced parts (recycled) but then it is stated most of the times.

[@AndreasSpiess]

Production surplus sounds like a good deal because it should be original parts.

[@Moonbase59]

Overall, this is kind of a sensitive issue. I found that most "everyday" parts for maker/hobbyist use are quite okay from Aliexpress, maybe because cheap everyday parts don’t justify the effort of cloning/relabeling. I did have (very rare) issues with more expensive parts like better-quality opamps and microcontrollers, though. So I still order everyday common parts through Aliexpress just for the price, but for "real" prototypes that might turn out into (small series) production runs or more critical devices I tend to use LCSC, Digikey and Mouser, even German suppliers like Conrad at times. Thanks for the video anyway-because I’ve had so few problems in the past, I’ve started getting a little lazy on "incoming quality assurance". Have to check if my transistor tester is still fit to the task, and I’m wondering how the newly-ordered IRLZ44S, IRF3708S and FQP30N06L will turn out. :-)

[@Leif_YT]

When i order from Conrad or Reichelt i get a headache from their search engine :D You search for a specific thing and get thouseands of completely different parts as a result. TME is also a decent option for people that live in Europe or near Poland. Most parts i order from AliExpress or LCSC, if it's a more expensive thing or i need it quick i look at TME and the last option is Mouser where i always have a shopping cart that slowly fills up until i hit the limit of free shipping.

[@cr6925]

Not a "sensitive issue" merely facts. You get what you pay for.

[@nlo114]

Put a low-value series resistor in the gate, plus a capacitor to damp overshoot and oscillation, plus prevents over-voltage failure from switching overshoot. Spec at 4:03 shows Rg as 25 ohms.

[@AndreasSpiess]

Such a resistor would have helped, you are right. I cannot follwy your argument concerning 4:03

[@nlo114]

@@AndreasSpiess I'm not an expert, but you can extrapolate the other characteristics from the rise-time for those particular conditions, then change those times by altering the resistance and capacitance elements of the circuit to prevent ringing in your particular application. Changing local in-circuit values may shift the switching speed into an operational area where there is not a problem. (probably not a very good explanation, but I'm sure you get my drift)

[@_adamalfath]

Last year my coworker buy 500pcs IR mosfet from aliexpress and almost 50% have defect (shorted, abnormal temperature), sure it's cheaper but for medium production qty it just not worthed for test it one-by-one or unfortunately found dead one after soldered to the PCB. Change the sourcing to mouser and no problem ever since. Maybe I still trust sourcing passive component from aliexpress, but for IC/FET, I think I'll pass.

[@Basement-Science]

yep, defective parts seem to be very common when buying on aliexpress. It's always extra work to check everything you get from them. For me it's still usually worth it though.

[@AndreasSpiess]

As mentioned in the video: I would not use these parts for production runs. As you mention: Too much work and risk.

[@efimovv]

Extremely useful and looks like mythbusters series! Thanks!

[@AndreasSpiess]

You are welcome!

[@uwezimmermann5427]

Put a series resistance of at least 1 kohm between the potentiometer and the gate and you will not kill the potentiometer next time! It will not affect the gate voltage but limit the failure current.

[@AndreasSpiess]

You are right. I was too lazy :-(

[@uwezimmermann5427]

@@AndreasSpiess I am lazy all the time - the difference is that I do not show my lazyness on youtube, because I am too lazy 😎

[@TheRcEngineer]

I bought hexfets on Ebay for a spot welder, the rds on was 10 times as high as specified. Had I used them the results would not have been pretty .

[@AndreasSpiess]

I assume you used the proper method to compare the two. It is not easy to measure those small resistances.

[@santopino2546]

Andreas Spiess at the milliohm range I suppose reading resistance with the four wire method is impossible due to active circuit. What would the correct method be? Maybe a low ohm shunt in series to measure the mV drop across it to calculate current, and we can measure the mV of Vdg. Impossible to have a precise current reading and it's not easy to find a high precision shunt, so I suppose a comparison between official and Suspicious parts could be the solution.

[@AndreasSpiess]

The shunt will not help you too much because you always need the voltage across the transistor if you want it’s resistance The comparison is always best.

[@santopino2546]

Andreas Spiess but for a MOSFET the voltage can be measured, if we have the current reading through the shunt we know it's the same current through the MOSFET Drain to Source. So we have the I and V across to calculate Ron.

[@TheRcEngineer]

@@AndreasSpiess i was sceptic from the beginning and compared the device against one bought from digikey . i used a calibrated transistor tester. when i showed the evidence the seller he immeadetly returned the money.

[@Scrogan]

Perhaps you could make a video where you peruse AliExpress pointing out reputable sellers (Fantasy Electronics, Shenzhen Semiconductor, etc.), and pointing out what are likely to be fakes or scams. Personally I always sort by number of orders, and have installed the Aliprice extension in order to see whether prices have recently been hiked and whether the seller has good approval ratings. I don’t think I’ve ever ended up with fake parts, but I’ve definitely had one or two sellers who say “sorry dear, product was out of stock, we refund you now please understand”, and products that just never arrive. Basically, don’t trust Yanwen or Cainiao shipping companies. China Post Registered Air Mail is the only one for me. Epacket was apparently pretty alright also. If a seller tells you to cancel a dispute, either to refund you via PayPal to avoid aliexpress’ deranking, or to send you a new one, don’t. Once the time for a dispute is over (15 days since confirming arrival or after purchase protection runs out), you’re never able to dispute again, so they have no incentive to actually refund you or send a replacement after the 15 days.

[@Sergeant898]

^^^^ All very true. Exactly as I have discovered. You can get good products at great prices, but you have to do your homework and be a smart consumer. The Chinese aren't known for their honesty unfortunately.

[@AndreasSpiess]

Thank you for sharing your experience. I do not buy enough transistors to get a clear picture of who sells good quality and who not.

[@elioalcala4371]

Thky hoy guys

[@pfsmith007]

I really enjoy your narration, you have a great way of verbalizing your subject. Thanks for helping us all learn.

[@malou8484]

Thank you a lot Andreas Spiess for that video, at least you dare to open the sensitive subject. From my rather long experience (more than 10 years now) from the chinese parts, I observed that, the more the chips are "complicated" and originally expensive, the bigger are the risks to find fake parts. A lot of integrated circuits are in this case, the most known are the voltage regulators which are often unable to provide the nominal current with a good regulation. The perfect example was (it begins to be obsolete now) the LM338K, I bought a lot of them from various chinese suppliers and never found only one abble to respect the original specifications. At the opposite, basic transistors are, most of the time, perfectly acceptable as you show in your video. Many other bad examples can easily be found in high end op-amps, unable to reach the specified gain-bandwith ratio.

[@AndreasSpiess]

I would also assume it is more profitable to fake an expensive opamp than a cheap transistor. On the other hand I had never a faked MCU, which is also quite complex. But maybe easier to test.

[@DeadKoby]

Good to know...... As a repair technician, I sometimes have to order parts that are discontinued in order to fix older equipment. I've been fortunate with Ebay sellers, and I always look over the reputation carefully. If the parts worked, I say so in my feedback.

[@AndreasSpiess]

Some commenters mentioned that it is not a good idea to buy rare transistors on Aliexpress because often, they are faked. Your way is much better. Especially giving feedback helps others!

[@BlueyMcPhluey]

I like that you frame results in terms of whether or not they matter because as someone relatively new to electronics it can often be difficult to tell

[@AndreasSpiess]

Glad you like my style. I think we should always concentrate on what matters;-)

[@davidellenberger3852]

I never received a non working transistor but I had the feeling that some repairs didn’t hold long. One inverter for instance where double diodes and mosfets were damaged died again on the next day in similar manner. I did perform a stress tests... When I fixed again from a different source it’s today 3 years later still working every day... I rather wonder if there are lower quality components/impurities or stuff from rejected batches randomly in the market. Actually I wouldn’t be surprised if the seller was in the unknown and tend to not blame them. I guess paying 5 times less has it’s price too!

[@AndreasSpiess]

I would probably not buy products I depend on from unknown suppliers.

[@Basement-Science]

When repairing stuff, the question is always, why did it fail in the first place? So you can never be sure if your replacement parts are to blame or not.

[@Conservator.]

5:30 To save the potentiometers you can add a resistor (about 500 Ohm?) between the potentiometer (middle pin) and the gate of the mosfet. Or use a 10 Ohm 1/4 resister and it will act as a fuse.

[@AndreasSpiess]

Of course. But I did not plan to destroy the transistors ;-)

[@Conservator.]

Andreas Spiess Haha, obviously not. My remark was aimed at those who haven’t yet tried it. 😉

[@streefland]

Last november, I ordered some TO-92 MCP1700 3v3 LDO regulators from Aliexpress. When I used one of them it didn't work and got pretty hot. They turned out to be relabeled NPN transistors! Of course the dispute period was over by the time I found out. In the future, I will do at least a quick test when new parts come in.

[@AndreasSpiess]

Bad luck. That is why I propose to do a quick test on arrival... Also because we would help AliWxpress to sort out the "cheaters"

[@yakacm]

The guys with the Swiss accent? I mean ignore the elephant in the rooms why don't we, it's not the Swiss accent that jumps out it's the mad flippin hat and goggle, lol.

[@mortenbrendefur7878]

The term: "Fit for purpose" is great. However. I did buy 50xIRF3205, 50xIRF3702 and 50x50N06, all of which I was going to test in powerful Buck-converter-setups and Relay-setups for Solar Cells and batteries. And so it happened that I was given an old UPS that I was going to repair. The original blown parts was IRF3205, which I replaced, and it just said POFF, both MosFets blown open with sparks and everything. After a lot more checking, and more Mosfets blown, I cracked open one of the blown original ones, and saw that the die-size was about 20 mm2. The Die Size of my new mosfets was only about 1 mm2.. I also built a test-setup for the On-resistance and used a 100 Watt car bulb as load on a 12 volt power supply. I then found, using Ohms Law that the internal resistance was way higher than specified in the datasheets as well. They all can turn on and off 100 watt car bulb, so fit for many purposes, but when driving an inverter or a Buck-converter for charging my batteries from my solar panels, they simply shortcut or blow open with a bang. I plucked some powerful 90N06 from a square wave inverter I am never going to be using, replaced the bad ones in the UPS, and now it works. I must admit that I did go for the Cheapest Offer of mosfets when I sourced them on Lazada (similar to Ali Express), but I have 150 fake power mosfets now. And that is only those I have tested. I fear that ALL the mosfets I have bough cheap are of same "mediocre" quality. Unfortunately, I did not test on reception, otherwise, I could have returned them all. :p I live and learn. I will build myself a testbench, Arduino-controlled, for quickly testing some parts upon reception. A testbench that can give loads of 50 Amps or so :)

[@AndreasSpiess]

Testing is always good. And sometimes even needed as your example shows. Other comments suggest that the more special and expensive a part is the bigger the chance for a fake.

[@chrisg6597]

One word of warning with some of the cheaper bench PSU's in current limit mode and especially ones without Output On/Off switches. Some of these PSU's have quite large value capacitors across the output terminals to reduce noise, so even though the current limit may be set, several amps may flow on initial connection (You are effectively connecting a charged capacitor to your circuit). If your PSU is of this kind then at 5:00, it would still be wise to include Rd to limit the current.

[@Rob2]

Ah I wrote the same thing before finding your comment! That is so true!

[@AnalogueGround]

"I try hard never to go to the limit with my parts" - that's good advice Andreas and works at many levels!!

[@AndreasSpiess]

:-) Thank you!

[@Slartibartfas042]

You are presenting your sight as a maker and that's absolutely the right point of view. I think Product designers will probably not rely on youtube videos for chosing their suppliers, so this is the only valid point of view here ON KZbin. :) On the other hand I do see your measurement procedures with some pain. Withstanding some voltage without immediate breakthru will not mean that the DUT will withstand exact this voltage under every environmental situation (temperature?) nor will it mean that it will withstand under other external drive situations. (DC voltage vs. voltage spike!) As there were already videos out where transistors got decapped to see what is the "engine" driving the whole thing I can say that there are a lot of rebadged cheap transistors out pretending to be some genuine brand ones that could not hold up against the originals. There are for example TO220 transistors out that use a much smaller die, which then means they could by no means transport that much of heat to the cooling system and by thus meaning they can't do that high power dissipation. One transistor that nowadays almost only exists as fake type is the famous 2N3055 im TO3 form factor.

[@AndreasSpiess]

I agree, there are a lot of fakes around. Usually expensive or rare products are more prone to be faked. The standard Chinese cheap stuff is already cheap enough, so faking is not very profitable, I think.

[@uni-byte]

Thanks for the refreshing outlook. I'm squarely in your camp on this. I have not often bought a component that did not do it's job, and the few times I did the vendor gave my money back without issue.

[@AndreasSpiess]

Thank you for sharing your experience. Some other commenters were cheated. Mainly with rare or expensive parts (at a cheap price)...

[@doost6233]

I purchased a lot of transistors and capacitors from aliexpress do not work as specification get hot and burn out capacitors may burst too, I tried to return it the return site not working is blocked.

[@AndreasSpiess]

Thank you for sharing your experience!

[@stevenbliss989]

I have bought thousands of mosfets from AliExpress. They fall into three catagories. 1) nearly in specs, but lacking some important characteristic, like too thin bond wires (common in high power mosfets), or too high a capacitance on the gate and way out of spec switching characterstics - usually factory "out of spec" rejects. 2) Voltage is in spec, but Rds On is about twice what it should be (plus 1 issue as well), and 3) not often, but sometimes they are utter junk, nothing like the part number says, and if you but 10, you will get 10 different specs! The only good news is for non critical functions, they are great value, and the risk of junk is about 1/5 for most parts.

[@AndreasSpiess]

Thank you for your detailed analysis. I agree, even if they are out of specs, they are usually usable for me.

[@foxabilo]

Just checked 2 of the AO3401's mine start to conduct at -29.7v +/-0.4vand fail at -32.6v +/-0.3v

[@foxabilo]

@Tiago Ferreira Sounds legit to me , mine are about 4 years old from RS Components.

[@janseiffert7799]

@Tiago Ferreira I bought some Hooyi MOSFET (HYzzz), waiting for their arrival, hope what i get matches the datasheet. I also looked at some NCEzzz part from Wuxi, they have some nice ones. How have your NCE hold up? Personally i think that is the "trick" to by on Aliexpress. Yes, some parts are plain fake (duds, esp. for long EOL parts, or a parts uplabeled, say 4A diode becomes 15A, normal OPAmp becomes high precision low noise), but thats IMHO a case of "you get what you pay for". A 35A Thyristor for under 10¢, sure... That doesn't even cover the cost for the silicon needed for 35A but sure... Where there is a margin for counterfeiting, there will be counterfeiting. But other parts are just second source/compatible. That used to be normal in the semi-industry, how many companies made a 555. Or the basic National OP-Amps LM339/LM358/LM393. And the list goes on and on with such "standard components". That has always been an entry point for small non-brand shops. First in Korea, then Taiwan. And now China. For Germans reading this: you guys clearly didn't order from Reichelt in the 80s and 90s ;-) Back then you would get mostly second source chips from them. And often a different every time. Remains of that can still be found sometimes in their order-numbers from when they slowly cleaned up their act. You can now order a "MJ 15016 ONS" (ON Semi/Motorola) or a "MJ 15016 ISC" (Inchange Semi). Back then you could just order a "MJ 15016", surprise what you would get (but never brands like TI/Mot/Philips/etc.) One can shout his anger in the wind getting a "fake" LM2575 (or other from the same family) or IRF. Or one can embrace it and buy the "copy"/similar from XLSemi, Wuxi or other never heard of companies in the first place. "Fit for purpose" At the moment i try to find these hidden gems of Chinese second source with the help of the JLCPCB parts list.

[@shanesgettinghandy]

I've been hoping for a video like this! Thank you!

[@AndreasSpiess]

You are welcome!

[@robertheinemann838]

They have many fake MOSFETs . If you crack the new MOSFET you have bought, you will find a very small silicon die that is much much smaller -almost 5 times- compared to the silicon die size of an original MOSFET. The mosfets will work well at low loads, but fail when the full load is applied.

[@AndreasSpiess]

You are right. It seems that the more expensive, special, or rare the parts, the more fraud.

[@ramoncomellas4245]

Hi Andreas, The transistors tester you are using is the LCR-T4 12864 ? I'd like to buy one, but the are so many variants that i am a bit confused...

[@AndreasSpiess]

I made a video about the testers, but I like this one most

[@D4no00]

dave from eevblog reviewed it, these products are based on an open-source design where all the logic is located on a microcontroller, the only differences between these products are just some additional features.

[@ManfredBrauchle]

take a look at the original project: ogy.de/l46k

[@GRBtutorials]

Don't worry about the current ratings. These are derived from the Rds(on), and the junction-case thermal resistance, and are hugely exaggerated for practical purposes, because they assume you have a perfect heatsink that keeps the case at exactly 25 ºC at all times. The important parameters to determine the maximum practical current are Rds(on), and the thermal resistance: junction-ambient (equivalent to junction-case + case-ambient) if there's no heatsink or fan, or junction-case if you use a heatsink (in which case you also need the thermal resistance of the heatsink). Just use the Rds(on) to determine the power dissipated (with P = I^2 * R), then multiply by the sum of all thermal resistances to get the increase of temperature. So you can measure the Rds(on), and as for the thermal resistances, you can calculate the case-ambient one by measuring the increase in temperature for a given power dissipation, and while you can't measure the junction-case easily, you can increase the temperature of the case until the transistor no longer works (or the magic smoke escapes, whatever happens sooner)

[@AndreasSpiess]

Most of the power FETs are anyway too strong for my purposes and they never get even warm... So you are right.

[@tinayoga8844]

I am a "Maker" and also a designer at a small electronic manufacturer. (20 employees) I have used AliExpress parts in very few products, and only once had a real problem. The problem part was a INA219 current/voltage sensor chip. I used two AliExpress orders on prototypes and everything was fine. Initial production was OK as well. But one batch of chips later proved to be defective. They "worked" but the returned measured values were way off. Some had bad current measurements, some bad voltage measurements. Some both were bad. We then used parts from normal distributors. I had designed one product with a 100 Amp automotive relay. I was skeptical about the rating so I tore it open. Lots of copper. Heavy contacts. It certainly looked capable. We never had any issues with it.

[@AndreasSpiess]

I would not be brave enough to use Aliexpress parts for a manufacturing run. Because of the warranty cost and the risk for a bad customer satisfaction...

[@SidneyCritic]

13:12 The red pin between the MOSFET and resistor means they are not connected in the program. It's best to link parts with lines. Click on a pin and everything electrically connected will turn yellow. To put the resistor junctions in the right place on the horiz line you grab the line, move it and move it back. This creates a bendpoint where you drag a line from the part pin to the bendpoint. You can also right-click, create BP. To go from a bendpoint to a pin you hold ALT when clicking the BP, and drag the line to the pin.

[@AndreasSpiess]

Thank you for your explanations. I was not careful enough, I think.

[@cr6925]

Even if it costs extra I personally always buy from RS, Farnell, Mouser or other well known reputable companies. Peace of mind....

[@AndreasSpiess]

I think this is a good strategy. A good sleep is very important ;-)

[@picwiz2]

My transistor tester does not appear like yours (less infos). Probably you already published the link where to buy it... Can you remind it to us, please? TNX in advance.

[@AndreasSpiess]

I made a video about them. There you find the links.

[@picwiz2]

TNX.

[@vylbird8014]

I built my own super-low-cost curve tracer for this purpose. Perfect for measuring Vfd vs I on diodes as well. It's a simple linear supply with some instrumentation, with the notable ability to set output voltage with milivolt precision all the way down to zero.

[@AndreasSpiess]

The principle behind it is quite simple. So it is probably cheaper to use equipment you already own.

[@Klokopf52]

I also never really had problems with smaller standard parts, but somehow this doesnt seem to apply to bigger ones. About half of my high power stuff bought on aliexpress doesnt even come close to its rating (and i do avoid the "way to good to be true" offers), same is true for very specific not widely used components, like GHz level transistors. They turn out to be relabeled lower frequency ones most of the time :(

[@AndreasSpiess]

Your comment point in the same direction as many others: The more expensive and the rarer a part, the higher the chance for a fake.

[@uK8cvPAq]

The variable pot on the MOSFET gate whilst current is flowing drain to source can exceed the safe operating area of the fet if you're not careful, these devices are mostly designed to act like a switch with as little time spent in the linear region as possible. I found this out when selecting series pass elements for a constant current load project, fets kept dying for no apparent reason.

[@AndreasSpiess]

You are right. I think I mentioned it in the video.

[@Sovvyy]

Great follow-up Andreas! I think if I were developing a product, I would not be using AliExpress as my go-to, so as you concluded, for us makers a "fit for purpose" approach is satisfactory. While I can accept there are fakes out there, it's far to easy to be cynical and make generalisations. Also, you've now made me very curious about running tests on future orders and collecting statistics....

[@AndreasSpiess]

As engineers, we believe what we saw with our own eyes...

[@p_mouse8676]

What's actually also important is how big the die is and/or heat slug is. I have had transistors that measured just fine on all parameters but kept blowing up. After uncapping them, I saw that the metal part for distributing the heat was only half the size compared to the original ones.

[@AndreasSpiess]

Good point. Thank you.

[@W1RMD]

Great informative video! As a side note, some parts may not include everything in them such as cd4028's that won't blank the digits because they left out the inverting nand gates, but you can still drive nixie tubes with them. Also, you save your vr's if you add a resistor in series so the are never a "short" circuit.

[@AndreasSpiess]

Good to know. So far, I never had such parts where they removed functions!

[@W1RMD]

@@AndreasSpiess You can drive the digits, but can't blank the zero's with all four inputs high.

[@addysoftware]

About the transistor tester wich shows a high Rds-on : it may be because the gate voltage applied to MOSFET is quite low and the transistor may not become complete ON. It happen to me once, I declared a lot of IRF840 fake, WHEN i SAW RESISTANCES 1.5....4.5 ohms, especially because I saw a very low input capacitance of only some 750pF instead of typical 2nF. But after, I got a lightning idea: what if? And I made a setup with my Kelvin clamps and a 12V source (for the gate) and a 3.3 ohms resistor hooked to a lithium battery for the drain-source circuit. So far so good: all transistors presented 0.7 OHMS when gate was charged with 12V. So they were not fake, actually the tester was not fit for purpose, the transistors have quite a thicker oxide layer at gate, or a lower gate surface; different technology than the original Infineon. Anyway, they were almost as good at 0.7 ohms (originals have 0,55 typical) and they have smaller gate capacitance. Still I have to test them for real, in power supllies, to see how performs.

[@AndreasSpiess]

You are right. Vgs influences Rds on.

[@Delali]

I love your approach..."Fit for Purpose"..... You were also honest with your thoughts on the last component. I also try not to go close to the max specs of my parts...so there isnt a need to be worried about whether parts meet their max.

[@AndreasSpiess]

That is always a safe way. Sometimes it does not work if you need the last percentages of a part...

[@Delali]

@@AndreasSpiess Indeed. Plus if the parts get damaged, we need to wait another 2 weeks for new ones from China

[@TheBodgybrothers]

A lot of these parts come from the overflow from pick and place. When doing large batches of boards you start with fresh reels and always throw out the leftover parts on the reel when finished the run. This is normal for electronic manufacturing and a good design will have part number counts close to full reel utilisation. A manufacturer will never put half a reel onto a machine, they will sell it as parts. Most companies like mine do not ask for these leftover parts back, so the pcb guys sell it to hobby stores.

[@AndreasSpiess]

That would explain that, despite the cheap prices, the products are genuine

[@Heisenberg2097]

Just a few things I want to mention about the 'common' AliExpress sellers. I was in Shenzhen walking up the eigth floor without elevator. These people make a living out of selling single transistors and shipping them for free. Who of you can imagine living in a factory on less than 10 square metres with their co-worker to support a family? And watching military parades regularly telling the people speak up and get shot? And that is where u get your hobby fun from for prices noone could give you at european rates.

[@AndreasSpiess]

I am never sure if it would be good for these people if I would stop to buy their work. Would they get better-paid work instead?

[@Heisenberg2097]

@@AndreasSpiess U wouldn't make the world a better place not buying stuff from China. But our industries+governments can make a change in China or India or Africa. And we are the ones that have at least some influence on what they do. It is an interesting fact that after 2000 when China opened a lil bit more to capitalism the amount of people that tried to escape to Japan decreased dramatically (in a good sense). For people interested I am not a good person to speak to but can forward you to lovely maker queen Naomi Wu. Please find her channel and let her know again that my gene-stealer agents are still out there and will get that little peace of DNA I need... OOOOPS got a lil bit off-topic now. Just me. And always listen to her talking mandarin... will give you a much broader + complete view of her than just her english version. It doesn't matter if you understand mandarin or not... so much information about her is transmitted in the way she is speaking her mother tongue. She is a true rebel and a punk and HOT beyond believe + always properly dressed.

[@FuZZbaLLbee]

Same goes for SD cards. They just change the controller to make ik look larger the is actually is. Of you pass the actual size, the SDcard corrupts

[@AndreasSpiess]

That one should be easy to detect, I think. But you are right. I read about them.

[@JerryEricsson]

@@AndreasSpiess I fell for the low priced cards once, I can't believe that I thought I could get a card that sold locally for 39 bucks for 4 bucks on Ebay. But I did, still have them somewhere around my shop in a junk box I think with a big red X over them to show they are bad. I couldn't even get 4 gig out of the 32 gig cards.

[@uraoshi]

@@JerryEricsson 32 were in reality gigabits. Divided by 8 makes 4GB. The seller used some wordplay. Common.

[@PhG1961]

Really nice video and so was the previous one. I tried to comment on that one too, but for some reason, my comment didn't get through. Also the video (mailbag) with the power supply was very interesting. I've been doubting for a while to buy that same power supply. Not that I really need it (I have more than enough PS's) but it sure would looks nice on my bench. I especially like the way you described (with the schematic) how you build the circuit for testing. A nice fresh-up for things I almost forgot.

[@PhG1961]

Btw, I forgot to mention, I usually buy on BangGood, and unfortunately, I did receive fake parts/equipment that is even dangerous like LiIon batteries, BMS'ses/loadbalancers etc... If something is really to good to be true, it usually isn't !

[@AndreasSpiess]

The new PS is on my bench and I like that I can key-in the desired voltage... I test most of the stuff I get. Especially batteries. And I file a dispute if they cheat.

[@samuellourenco1050]

You hit the nail in the head when you said that the most dangerous condition for a MOSFET is when they are not fully conductive. Actually managed to damage a power MOSFET in that manner, at half of the rated current, because its gate was not being driven sufficiently. It was the consequence of a cascading failure, since the driving MOSFET was also damaged.

[@AndreasSpiess]

Shit happens. Like with my potentiometer. It also dies when the MOSFET died :-(

[@jfkdkslslee]

Which model transistor tester was used? Thank you for another excellent video.

[@AndreasSpiess]

Maybe you watch the video about these transistors? There you find the links.

[@SaltwaterC]

GM328A. I have watched again that bit as have also failed to spot a link in the description.

[@addyvandendoel602]

I also like to know that transistor testers link. May be the link is also not there because of the re-upload?

[@AndreasSpiess]

This is a link to the one I used: s.click.aliexpress.com/e/_srkjb3

[@stevenbliss989]

AO3400 & AO3401 are mostly manufacturer rejects on AliExpress (except for the ultra expensive ones). Many power mosfets are 2 x Rds, but most in specs for Vds. Sellers do this knowing that many mosfets are used < 1/2 full power. But at 1/3 to 1/10 of genuine price, it is often a great buy anyway in spite 2 x Rds.

[@picwiz2]

On the transistors metal case try to put an alchool drop: if the part number disappears instantly, it is a fake! Of course I am writing about metal cases on which the part number is printed.

[@AndreasSpiess]

I ordered a few to check it myself

[@tazman4855]

Aliexpress has strict requirements of their vendors. One complain is a severe black mark on their reputation. My rare complains of undelivered or slight defects, eg scratches caused by delivery are often handled by free reshipment, refund, or partial refund. It'd be very unfitting for a people with long term outlooks to ruin their reputation or that actual business can making inferior products to compete in China unless requested to do so by design/low cost, eg iPhone vs Walmart.

[@AndreasSpiess]

I always use the dispute process. I assume that they treat bad quality differently than lost packages. If I get bad products or damaged products it is important for me to give feedback. Also for the next buyer.

[@kwinzman]

I have ordered a lot of products on Aliexpress and dispute process is far from perfect. For example I had the Aliexpress judge determine that I could only get a refund if I send back the defective item to china. And the postage would have cost more than the item was worth. Also if you want to contact Aliexpress there is no number that you can call. There isn't even an email address on the whole website from Aliexpress. So if you make your dispute too late then there's no way for you to talk to anybody. Moreover lately they changed the system so the order closes automatically after a certain time. Before you could "extend the time for the seller" and give the seller more time to still deliver the package without losing your ability to dispute. All in all I dread the system.

[@kwinzman]

Ebay is 100 times better buyer protection wise and if everything else fails you can still talk to a human in ebay support. Aliexpress is not even close from a buyer safety persepective. I only use Aliexpress because I can't get some parts on Ebay and because of the price.

[@Basement-Science]

@@kwinzman I agree that ebay is much better, the problem is they often dont sell what I need, or are much more expensive. With aliexpress you always have to watch out for expiring buyer protection, check every order etc. I always contact the seller if a package has not arrived yet and tell the seller to extend the purchase protection (they can do that), or if it ends in just a few days, I additionally open a dispute and tell them I will close it if they extend the purchase protection. Aliexpress does have a live chat, but they have not been very helpful for me. So far the Aliexpress judges have been on my side when there was a problem, but yes, the system is far from good.

[@hinzster]

All (yes, ALL) my disputes with AliExpress vendors were about shipping. At some point in 2018 or so one of their usual shipping providers, Yanwen, seems to have gone rogue and just delivered about 80% of the ordered packages. And then there were the clear cheaters that just didn't post anything and gave you a non-working tracking code, or the ones that started a delivery but the shipping company never got the package. All of those were deemed as refund situations by AliExpress, so I can not complain at all.

[@McTroyd]

Thanks for this! I play with enough analog electronics that I think I'll need to build a curve tracer now. Side note: W2AEW is awesome. Have you chatted with him on the air? (I don't know him personally; just curious.)

[@AndreasSpiess]

I am no more active on HF.

[@connecticutaggie]

I am glad to hear we can use eBay to get transistors for maker projects. It is not just the part cost; but, the shipping cost as well. While I can often find free shipping for small quantity orders on eBay, this is usually not the case for electronic distributors.

[@AndreasSpiess]

Minimal order size is sometimes also a problem with distributors

[@stanyoung4974]

Please test Toshiba 2sc5200 transistor from Amazon. People report internal circuit too small to handle rated power. How would you test this device ?

[@AndreasSpiess]

You can test it by applying the needed voltages and measure the currents according the data sheet (as I did). Or, maybe easier, you insert it into the place it belongs in your equipment and check it out.

[@santopino2546]

I read the comments about how Chinese fake parts, I think we should consider that China is a superpower and has a very advance, innovative and dynamic production industry. They are perfectly capable of inventing and producing the most advanced electronic components, but if the market wants a specific component, they will produce that component, why waste resources and intelligence to invent something new if it has no market and creating a market takes further resources and time. Producing new costs must less than rebranding or whatever.

[@uwezimmermann5427]

Look at any video on youtube which shows how the electronics markets in Shenzen look like and you have the answer. Buying old, used equipment for manufacturing transistors or integrated circuits on 4-6 inch silicon wafers is increadibly cheap as the real factories upscale and renew their equipment all the time. Setting up a new factory is thus not difficult. If you then can produce transistors in TO220 packages which almost look genuine but use a much smaller die size, if you skip all steps of quality control and sell everything you produce you might be able to make a living of it - and then you will find it in China. With western cost of labor, premises and quality control this type of business would not be viable.

[@yanfishtwig2356]

Wow Thank you for making this video i have been trying to find an answer to this question at this level for ages !!! keep up the great work

[@AndreasSpiess]

Glad you enjoyed the video!

[@tommihommi1]

I recently purchased a few MRF9060L on AliExpress, and they definitely are genuine. Both frequency response and threshold voltage very close to the datasheet. Guess these more exotic transistors in ceramic packages are too difficult to fake

[@AndreasSpiess]

Thank you for sharing your experience!

[@Basement-Science]

I believe the less common a part is, the more likely it is to be genuine. It's just not worth faking if few people will buy it anyway.

[@KF-bj3ce]

Do not mind your accent, same as mine. Thanks for this explanation. Have been using AliExpress transistors and mosfets for a while and apart of some BD139 / 140 build on soft steel base and a pitiful small chip, all have been usable.

[@AndreasSpiess]

Indeed, standard parts usually are good. There is no reason to fake tehm because they are anyway cheap.

[@dekipet]

I just throw in a basket dozen MAX6675. It's not that i lost some money for them, i will not open dispute, because i don't have time to pursue the seller. It's that i lost time finding out why my board doesn't work as it should. Checking dozen times my circuit, my PCB, my sketch, looking through a magnifier to see did i solder it correctly... all that for the every single chip i bought... And it the end i take one from the MAX6675 module, and it works. It took me days. And that is what costs more then fake chips. Time, frustration... etc.

[@AndreasSpiess]

This is strange. Usually, chips (exception: Opamps) seem to be ok.

[@pentachronic]

put a series resistor on the wiper of the gate potentiometer and this will prevent you burning out the potentiometer if the transistor is placed incorrectly or there is a short.

[@AndreasSpiess]

You are right!

[@zerog2000]

Curious if you had run calibration procedure on your transistor testers, perhaps in a previous video?

[@AndreasSpiess]

It seems to be simple: Connect all three pins and press start... discovered it later with a destroyed transistor...

[@goamarty]

You can use a series resistor between the pot wiper and the gate. This protects the pot in case of a failed FET.

[@AndreasSpiess]

You are right. I was too lazy...

[@rickhunt3183]

Try some 2n3055's those are very popular..also lots of Toshiba RF transistors are fake and don't have to be tested because they haven't been in production for a long time, but here they are for sale on the internet looking new. When it counts always go with a reliable distributor. Digit-key electronics will sell the real thing if it's in production. If a power transistor sells for 5 dollars or more and a Chinese supplier has them for .75 cents. You probably don't want to waste your time buying them. Always wear your safety glasses when driving parts that could explode. You only have two eyes. I'm sure you know all this quite well. I hope you have a safe and productive day. Oh..and good video..

[@AndreasSpiess]

Looking at the various comments it seems, that rare and expensive parts are far more faked than the cheap and common parts I tested. What I also learned is, that you sometimes have to take your changes if you need an old part and gamble if you get a working one. Fortunately I am not in retro, so I do not have those problems.

[@AntoninKral]

I have received 40pcs of NDP6020P which is P-Channel MOSFET. All IOs are clearly marked as NDP6020P, but they have two different batch codes on them -- 9S12AR (10pcs) and 9J08AR (30pcs). While 9S12AR are behaving as expected (P-channel MOSFET), the 9J08AR don't and they behave like N-channel.

[@AndreasSpiess]

I would record the evidence with a video and file a dispute.

[@AntoninKral]

@@AndreasSpiess absolutely. Already done that couple hours ago so no answer yet. It was just a surprising find to be honest.

[@AndreasSpiess]

Me too. These cheap transistors are usually ok.

[@AntoninKral]

@@AndreasSpiess So I have the dispute resolution already. They have returned 3/4 of the price, but no information about the weird batch.

[@nukiduki2]

@Antonin Kral Here I am, sitting with a batch of 9J08AR P-Channel MOSFETS, and not only do they behave like N-Channel instead, but the Drain-Source connection is always open when reverse biased. In other words, they're faulty, throw them away.

[@junkyard118]

Learned my lesson , NEVER buying semiconductors or other electronic components again on Aliefakepress!

[@AndreasSpiess]

A good decision. I was not so unlucky so far.

[@thiamjoo]

It's too expensive to make fake electronic parts. Most likely low quality rejects which somehow found its way into the market. A friend of mine who works for Siemens Electronics making these type of parts told me that the testing machines grade the parts into more than 10 grades. The testing machine spits into the bins according to the quality. Some fly onto the floor and they are collected and called floor sweeps (another grade?). The highest grades of course are used by big companies or maybe even the military.

[@AndreasSpiess]

You are right. Unfortunately it is not easy to distinguish between them after the sorting (if they are not labelled differently). But fortunately we usually do not use our parts to the limit and often can live with the 9th grade...

[@ABaumstumpf]

The AO3401 very likely IS a fake. Fake by no means implies that the part is not working or something completely different. Most of the fakeing is just relabling like generics as products from a well known company, labeling a 5A part as 10A or selling salvaged stock as new. I have my fair share of fake parts as we have a small local shop that sells refurbished and scrap parts for cheap. For example you can get a bag of roughly 3 000 SMD parts for 3€ - but those are mostly the leftovers from reels, like the last 20 pieces of a tiny transistor with no markings. or some supposed USB 3.0 controllers that micraciously only handle USB 2.0.

[@AndreasSpiess]

If they are labled as "unknown", nobody would probably complain, as in your case with your loca supplier. However, if they are sold as genuine and new and they are not, most of us would complain.

[@jstro-hobbytech]

That little transistor tester is a prime example of shareware lifting everyone up. If we couldn't get that tester from a million places in China and someone patented it, we'd be paying a few hundred for it.

[@AndreasSpiess]

I agree

[@georgegonzalez2442]

I bought some LED light bulbs on Ali Express. Very inexpensive. But when I opened one up, it had a 100 volt capacitor with 160 volts across it, the wiring was loose and could easily short to the aluminum case. Is smoke and/or dying worth saving a dollar?

[@AndreasSpiess]

I do not accept bad quality and open a documented dispute. I always got my money back.

[@polik971]

The Alan channel you provide is really good , since many years. He's amazingly able to teach even hard subjects

[@AndreasSpiess]

Thank you!

[@matheokoning4456]

Is it an idea to test more electronic parts from Aliexpress? I am still fairly new and inexperienced but have now found out that Aliexpress sells more bad / fake parts. For example, I had bought the HT7333 through a video of you and did not understand why the battery was still draining so quickly. Now I have a better multimeter and found out that 9 out of 10 chips used a lot more current. Who knows, we may be able to set up some sort of list of trusted sellers from Aliexpress.

[@AndreasSpiess]

This is quite stragne. Usually the cheap Chinese parts are not faked. If you want to have a reliable Chinese supplier you have to go with LCSC.

[@matheokoning4456]

@@AndreasSpiess Yes that was a bit disappointing. I am still inexperienced and it took me a long time to find out that the part was not right. Ordered some more from different sellers on Aliexpress. am curious if I was just unlucky or if several sellers are not reliable. I had indeed seen LCSC. Would have liked to order there but the shipping costs are very high.

[@marcrives279]

I had problems with supercaps : far off spec internal resistance and self discharge. Unusable. On the other side, even "fake" or "datasheet compatible" IRF3708 are overkill for my use (and 10 time cheaper) ;) However, you may always check prices in well known suppliers first. Sometimes, it's not worth buy in China.

[@AndreasSpiess]

You are right. Here, the problem with distributors are added cost for small orders and shipping cost. Usually prohibitive for makers.

[@wildinthejardin]

Supercaps are sometime like those 8000mAh 18650 😂

[@marcrives279]

@@AndreasSpiess Shipping cost are free when above ~30€, and so much faster! Something to think about. Yet, I'm still mainly buying in China.

[@marcrives279]

@@wildinthejardin I finally bought a nice 2.5F supercap in a well known distributor. Impressive ... and still made in China ;)

[@michim5661]

I had exactly the same problem with supercaps. Perhaps it was the same dealer? Mine were these: www.aliexpress.com/item/32631591419.html?spm=a2g0s.issue_5ptha.0.0.278c4c4d5sRJOY

[@cls9474]

Relabeling can be tested by rubbing the part with a q-tip soaked with alcohol. If color comes off, it most probably has been relabeled.

[@notsonominal]

i find acetone takes off some markings but isopropyl alcohol not so much. but yeah, lots of fake stuff is still fit for purpose for most hobby projects

[@SianaGearz]

They make acetone-resistant fake labels now too. I mean i know for sure that no Yamaha OPL2 chip was made in 2008, but whoever relabelled the chip, apparently didn't.

[@cls9474]

@@SianaGearz seems that we'd have to up the game too. We had a lot of fake chips for repairs of our old products and new devices we simply have in production for years. AES transmitters/receivers and stuff like that - long before AliExpress came along. This business has a long tradition... www.aeri.com/counterfeit-electronic-component-detection/

[@tinkerer67]

I'm curious, would you purchase anything on Aliexpress until this epidemic is over? I heard that corona viruses could survive up to 28 days on hard surfaces at low temperatures. Perhaps leave the packages for an additional few weeks before opening?

[@AndreasSpiess]

I am glad if my packages will arrive before the 28 days. Currently most of China is blocked and I assume it will take quite a sile till everything is working again... So I do not fear anything. Packages at your retailer might be as dangerous because their supply chain is fater than ours.

[@Tocsin-Bang]

Maybe I've just been lucky, but over the last 5 years or so I have bought thousands of semiconductors from Aliexpress, Banggood and eBay, and I've yet to find a fake. Ok some have been close to limits, but not an issue, provided, as you say, you do not run near limits.

[@AndreasSpiess]

So we are 2 ;-)

[@santorcuato]

Andreas Spiess 3, I’m on board.

[@jaynottelling5892]

I'll 4th that, 7yrs of buying on alibaba/AliExpress, ebay bangood etc. worse I have had (eBay) was trying to get BME280 temp/press/humidity sensors, 4 different sellers all sent me BMP280 temp/pressure, all also gave me full refunds when I complained. O and often cheap batteries with 50% of stated capacity.

[@Tocsin-Bang]

@Repairman diymore Alice 1101983 ic chip Store Super Mall of Electronic Components: are three recent ones.

[@MACYNET323]

@@Tocsin-Bang So these are reccomended sellers or not?

[@gkdresden]

I think, for the standard transistors the is no need to sell fakes as the wafers of their dies are produced in large numbers. So they cost nearly nothing. The fakers only need to buy the wafers, devide them into dies, do the bonding and encapsulation at low cost as well by second hand mashinery. You can find real fakes mostly in the case of more complex devices. I know the cases, where TDA2030 were sold as TDA2050. In most cases it is not critical, but if you go to the operating voltage and current limits, these devices failed.

[@AndreasSpiess]

I agree with your finding. The more expensive or rare, the bigger the chance for a fake.

[@FantaBH]

Well I indeed was always lazzy to test it so detailed as you did, so for me always was enough what good tester says. Neither one component ever failed in my devices , which comes from ebay or aliexpress . But your test is now final word, that they do not sell fakes to us , regarding electronic components.

[@AndreasSpiess]

If you read the other comments you can get fakes if you buy special or more expensive transistors

[@alejandrotrujillo5679]

As always very useful videos!!! Love your work!!

[@AndreasSpiess]

Thank you!

[@MalagasOnFire]

Very usefull information and setup to test before hand the reliability of transistors -> FET. The best chance here is to use 12V 1A due to limit of external power supply for DPS5005 , until get some 48V PSU .

[@AndreasSpiess]

As I mentioned: Your current power supply is good for testing.

[@MalagasOnFire]

@@AndreasSpiess Thank you. First batch will be International Rectifier IRF540N , purchased on local store.... I'll update the Drive folder as i go: drive.google.com/drive/folders/1H5iCqIbPty6IsIaEM8H9dOra2z6zHRn_?usp=sharing For now Transistor testers had almost same results for two mosfets of the same batch / purchase. For the current setup the RDS values will be writen in a spreadsheet

[@MalagasOnFire]

The first one gave around 0.036 Ohm, which is not far from the datasheet ( 0,044 with Vds=100V). By the way the HEXFET did not get hot at all and the turn on effect as mentioned in the video did occur at 4V. Fits for purpose and complies with test setup

[@Atelierul29]

The difference is in the life span,not the specifications!

[@BruceNitroxpro]

Atellerul29 , Please explain how you would monitor this.

[@ABaumstumpf]

The difference can be everything.

[@Atelierul29]

@@BruceNitroxpro You can't really monitor it but you have two options: 1.build a test bench that simulates the wear at normal working conditions over a certain period of time or just compare it when let's say..you replaced a transistor in an amplifier which was never broken since 1985 and the replacement part only lasted 2 years...this is the long way but an obvious one so...it is not impossible to monitor it

[@BruceNitroxpro]

@@Atelierul29 , Thank you. Excellent example! LOL

[@aytacdede81]

I bought some mosfets from aliexpress to change them into an inverter. I'll tell you that the circuit originally uses CEP76139 (30v 75A N Channel) I wanted to put Irf1404 and they didn't work, then I bought CEP83A3 and it didn't work either and then I bought CEP93A3 and it didn't work. Finally I bought IRFP7540 from another platform and it worked. I don't know if it's because of an imitation mosfets or for another reason. These mosfets don't work. Do you have any experience? Please leave your comment. Thank you very much.

[@AndreasSpiess]

The CEP76139 have a very low VGS voltage of around 2.5V. I only checked the IRF1404, and it has one of 5V. So it is well possible that this is the reason. I did not find the IRFP7540 datasheet

[@aytacdede81]

@@AndreasSpiess So do you think this could be the reason why irf1404 doesn't work instead of cep76139? Please comment why irfp7540 worked instead of cep76139!I told him that during this time I have tried other MOSFETs as well instead of cep76139They are CEP83A3, CEP93A3, P80Nf12, and none of these worked, I was thinking of buying HY1906 to test as well.On the gate pin of each mosfet there is a 10Ω resistor. I changed the value for testing from 2Ω to 30Ω but the result was the same.I made a circuit to test the status of the IR2153 integrated circuit which works as a pwm oscillator for the mosfet! The integrated circuit is operational! I thank you in advance for your time and support. Blessings!

[@rklauco]

I have very positive experience with Aliexpress and transistors. However, very negative one considering 1-wire devices and i2c devices. Especially DS1820 - none of them supports correct precision switching, parasite power does not work on 50-60% of them, etc. Have multiple batches. Same goes for i2c sensors - even had a BMP280 that reported constant values no matter what. Multiple other BMP/BME280 reported totally different values - 50% of the batch went straight to bin in the end. Still cheaper than any source in Switzerland, though :)

[@AndreasSpiess]

Cheaper would only be relevant if they work ;-) Another viewer also wrote about bad DS1820s

[@rklauco]

@@AndreasSpiess In fact, the remaining ones worked, but not in parasite power. In the end not a problem - just one more wire to the mix, but it took me days to pinpoint the issue :( Those days (and angry nights) cannot be returned. But at the same time - I've learned I should verify at least sample of my purchased parts - no matter what's the source.

[@birnodin]

Aliexpress logistics AI: if (username = "Andreas Spiess") or (username = "David Jones") { send_genuine_transistors} else {send_fakes}

[@donpalmera]

In most languages your code, if it was valid, would always evaluate the first condition as true because its testing assignment and not equality.

[@Moonbase59]

😄

[@hopje01]

= = ==

[@zerog2000]

donpalmera pseudocode?

[@idiotwithasolderingiron]

The only problem I have had with "Fake" transistors is when i try to find RF transistors for the final output stage of a transceiver. Most of the ones from China are fake. When they shoud output say 5 watts you may only get half a watt.

[@AndreasSpiess]

Comments suggest that the more special and more expensive the transistors, the higher the chance to get a fake.

[@janisvaskevics93]

Thank you for confirming and nice explanation! Super useful.

[@AndreasSpiess]

Thank you!

[@hopje01]

The problem is when fakes get into critical equipment like medical. Saw a documentary of a crashed airplane, a metal part was broken in two. It was made of inferior metal, supplied via authorized company, including the fake test report. Supplier was not aware of this.

[@AndreasSpiess]

You are right. But this is another story, I think. Much more professional than I am...

[@smusmusmusmu]

I bought some J201 JFET through hole transistors on Ebay, the surface where J201 is printed has obviously been sanded. For my application, they turned out to work okay. They are different from genuine J201, I suspect relabeled J113 or similar.

[@AndreasSpiess]

Some other commenters also mentioned JFETS as being faked. It seems we have to pay attention to this category of parts

[@felenov]

I have a scanning electron microscope. I decap almost every I order. From tiny transitors to specialised ASICs. I am just very curious

[@AndreasSpiess]

That seems to be a perfect application of such a wonderful device!

[@simonbaxter8001]

The Peak Altas DCA75 Pro (with the USB output and PC application) is great at characterising your devices .. and before you ask, No, I don't have any affiliation to them other than as a very happy and long term customer!

[@AndreasSpiess]

A little pricey for a maker, maybe...

[@juststeve5542]

Nice video Andreas, hope you didn't burn your fingers on any MOSFET tabs, that's my bad habit (damn those things can heat up quick!).

[@AndreasSpiess]

You are right. And some even died because of that because I did not pay attention...

[@pd8559]

In the industry companies order parts from electronics manufacturers specifying the acceptable failure rate and other deviation factors from the product they are ordering for the BOM. The tighter the contractual purchase obligations are written the more expensive each individual part ordered becomes as the price scales with the testing and defect rejection work done at the manufacturer side. A lot of Chinese companies order parts without the same stringent contractual stipulations for a much lower cost per part to them because it is expected that the buying company will do the testing and binning work themselves. This is how come Chinese phones ended up being rediculously cheap for the features they had. They ordered without testing and used the parts without testing. 9 times out of 10 the parts perform identically to the more expensive part but every once in a while you will get a dud part or a deviations in the part and in the case of the Chinese phones they shifted the burden onto the end customer. If the end customer had a problem with their phone the Chinese company simply replaced it with another. This to their calculations cost them less money than buying all their BOM at higher prices for already tested and binned parts.

[@AndreasSpiess]

I agree. There is also a positive example: In his biography, Elon Musk "commanded" to use commercial grade chips instead of military grade chips for his rockets. Obviously it worked and he got it a lot cheaper. I think, the purpose dictates the selection of parts. For my prototypes I can live with dead on arrival parts. For an iPhone it would be ca catastrophy...

[@BerndFelsche]

So there is a form of "triage" to perform when receiving goods. Quality controls in formal manufacturing works in much the same way. For "making" and prototyping, you want to do at least the simple compnent test. I've had resistors with incorrect markings (colour band) ... it took hours to debug one circuit with a bad multiplier band colour.

[@AndreasSpiess]

I even cannot read the colors anymore. The colors are often very bad to distinguish (for me). So I always measure the resistors before I use them.

[@elioalcala4371]

I WANT TO buy a mosfet tester likes that/any recomendations?

[@AndreasSpiess]

There is a link in the description.

[@springwoodcottage4248]

I have often noticed that a number of people always have to pay a high price before they are happy, the classic: "You get what you pay for!" My experience has been different & that price is no indication of quality. In 2013 I bought a used c-class Mercedes. It has been the worst car I have ever worked on, nothing like the quality of the previous Volvo I have owned & the Volvo that replaced it & extremely expensive parts. Most of the Chinese stuff I have bought has been low price, but most of it has worked well for me. Having watched many of the video of bigclivedotcom where he takes apart Chinese stuff there are certainly some products which are fakes &/or dangerous, but generally a lot of the Chinese stuff that he reviews is good enough for what it is intended for. Almost all my test equipment, power supplies, soldering tools etc is now Chinese & I have been very happy with most it & most of the Chinese components I have bought. The great blessing has been the low prices. Thanks for sharing your test results.

[@AndreasSpiess]

I share your view about Volvos (we own and like our 16 years old one)

[@5Komma5]

NEVER accept that the seller will send you the good parts if you can prove that the part is faulty! They will not send them and you will be past the warranty period of the platform. Insist on a refund and say that you will purchase the same part again as soon as you get your money. Ask them to please ensure to send out good parts this time. Use the platform escalation service if they refuse a refund and you will get your money back from ebay/aliexpress in a short time. When they piss me of I will purchase the same part multiple times and claim the warranty each time if they still send out fakes 😁😁😁. Some actually learn and start sending out good parts after a while. Bunch of crocks! I have a small box full of fake Mini360 switching regulators. The ones without the "200" marking on the coil draw WAY too much current when idle. They are all bad and most where ordered from "the good guys" on ebay. Yeah LOL

[@AndreasSpiess]

So far I have very good experience with AliExpress. No problem with refunds if clearly documented.

[@AS-oq7io]

Another Mini-360 victim. Welcome to the club ;-) As you pointed out, there are so many Mini-360 buck converters out there which are using the wrong inductor. In the meantime I avoid them, as you don't know what you get. MP2315 based buck converts are almost as small as Mini-360 and I had no issue with them so far. Unfortunately it's hard to find them when searching for MP2315 as they usually don't mention it. But they are offered as "Adjustable 1.8V 2.5V 3.3V 5V 9V 12V" and there is a soldering jumper for each of this fixed voltages and a potentiometer for adjustable voltage. The fixed voltage options are really cool as you can't change them by mistake.

[@tubewatcher68]

Thank you for the video, though I have not had as much good luck. I have bought the P-channel NDP6020P from two different suppliers on AliExpress, and both lots were fake. This caused me much time and frustration trying to get these to work in a circuit, I could not work out what I was doing wrong. I then bought a transistor tester and it turns out that both these lots were in fact N-channel MOSFETS, so no wonder... I ended buying some FQP27P06 which worked fine. I have ordered some NDP6020P from your link Mr Spiess, I hope this time they are good !

[@AndreasSpiess]

The 6020 is very easily mixed with the 6020p. Happened to me, too

[@tubewatcher68]

ahh ok, I did not realise there were N channel versions of this. However, looking back at my orders and on the the actual MOSFETS, they are clearly marked having a P. Could have been a some dud batches with incorrect labeling.