I was initially thinking 78M05 (500mA) which wouldn't even need a different casing as those already come in TO220, but these fakes in the video have already faceplanted at just 500mA load
Don't sweat those handful of losers who have nothing better to do than criticize or accuse others of fakery. They are overwhelmingly in the minority, but simply have the biggest mouth. Keep doing what you are doing, because I appreciate and enjoy it!
I think otherwise, I think we should be humble, I think sources deserve to make criticism without accusing those sources of being “losers” A big attribute to failure is not being subject to any opposing views and putting yourself into an echo chamber or egotistical position like you have suggested. This man doesn’t deserve failure, I don’t know him nor have I seen these videos but I think not..
Your setup is indeed good for proving that there are fakes out there. And it is a good thing to bring this issue under the spotlight. For a more precise measurement, you should measure the input voltage at the terminals of the device under test, because there's a voltage drop over the wires from your power supply to the device under test. That's why your first setup was actually better: with the supply at 12V, for sure you won't be below 7.00V at the device under test. With the supply at 7.00V and a voltage drop over your wires, you might end up with for example 6.8V at the device under test, which is out of spec.
I have found exactly the same issues with fake 15-volt type regulators. They kind of work OK up to a couple of 100mA but fall away rapidly above that number. A very long way from 1500mA. I tested mine with 20 volts input. Not an unrealistic input voltage. We always get a couple of people that think you are testing it incorrectly. I don't lose any sleep on such people, nor should you.
I too have experience using these devices, for well over 45yrs, and have many times come across bogus parts, though this tends to be a more recent trend. I categorically debunk the "too high source voltage" or "too hot" claim from the detractors. It takes a hell of a lot to cause the legitimate devices to be off spec, or much less kill them, (having on occasion, in much younger years, had the device too hot to touch) The "operating temperature" is listed well above 100°C(212°F), yet as you clearly shown and stated the DUT was barely warm, and I somehow doubt you would be calmly touching the DUT if it was much hotter than +55°C(131°F), which for a majority people is borderline 'bloody HOT'. Another point is that the 78xx devices are typically rated to Vin 33v-35v for 5v-18v devices(40Vin for 24v devices). You were only feeding with 12v, which puts operating parameters well and truly within spec. Kudos for the basic explanation you provide, I'm sure will help many understand. (Just tell the not so knowledgeable to go learn before splashing bulls..t)
Oh i believe that its definately a bootleg part. I was one of the people that questioned the heatsink mounting..but if the temp wasnt above those numbers..I agree it seems to not have been a contributing factor.. I also know the b+ voltage also was no issue..its clearly just junk part.
Maybe they snuck a die for 78M05 in that package since it current limits at 500 mA or they just had a bad chip copy engineer. I had a bunch of cheap 78L05 that were much worse performance - they actually started to smoke - missing over temp circuit ! They were marked WS78L05 for WingShxx or something LOL
would be interesting to see how the 100mA variant of the 7805 would perform. I think it would be very similar to the fake 1.5A one. Maybe they put a 100mA variant in the case of the 1.5A variant.
you could find different varient of fake 7805 from 20 years ago,some of them only current limits didn't work some thermal limit some need bigger output input cap to be stable and some with very low output current capacity
WOW. I never thought to check if 7800 series regulators were fake. I will now this coming winter. Great winter project. I have tested LM317s to see if they heat regulate. They do. Did not find any fakes thank goodness.
I don't see the usual bypass capacitors on the input and output of the 7805 on your test rig. Try your scope- the 7805 (or whatever it is?) may be oscillating. They lose regulation when oscillating. Thanks!
because of the load through the power lead cable , what's in the input og the 7805?, , you should use little more then 2 volts then the datasheet states , and put a 100 nf cap on the output to precent parasitic oscillation.
An output cap generally isn't necessary for stability but it does improve dynamic performance. An input cap is necessary with long and hence inductive input leads. Two separate test leads are about as bad as it is possible to get. It is much better to use a tightly twisted pair if the leads must be long. Also in a test like this the input voltage should be measured right at the regulator. I'd be looking at everything with an oscilloscope.
I have a bunch of these with a different lot code: MAR 1904, otherwise exactly the same. Mine seem to be 78M05/LM341 devices. They drop out at around 500mA. Was a very annoyed when I found out, but they were _really_ cheap so in retrospect don't mind it very much.
I seem to remember many of those type devices don't act right without some capacitance at the load. Look at the output on a scope for any oscillations. (I'm not saying they aren't cheap knockoffs, or they ever perform to spec!)
Maybe teardown and differences in silicon next? By the looks of it, it does look like it is ams1113'ish LDO just in counterfeit package. Nice vid tho, never trusted those 7805 with those paper thin heatsinks :) .
China source IC's & batteries such as those found on Ebay, Bangood etc are a lucky dip at best going on my own experiences. They work but don't quite perform to the genuine device's expected ability in standard published datasheets. Not all, but some and the only way to sort them out is to test each one individually so a lot of frustration & time can be saved by spending a little extra and ordering genuine devices from local suppliers (in my case it's Farnell, Digi-Key, RS components etc)
I'm actually impressed how many Timmy know-it-alls in the comments don't understand thermal conductivity well enough to prevent them from writing their entire thesis out
I have no doubt that what you have there is a counterfeit (or a small chance of being a batch of factory rejects) 7805. Any decent 7805 will not drop significant voltage before 1A load. Even at 1A, it should not be over 10% of its rating, worst case. The only criticism I have for your previous and this video is not using input and output caps as per datasheet, and not checking for oscillation (or ripple) with a scope. I have an educated suspicion that fake 7805 is oscillating like crazy under load.
I don't understand, this regulator's datasheet says 7 to 35 volts input. 12v seems to be easily on the safe side for a supply! You also kept the temp far below the 125 degree C range that is also allowed. Ignore the naysayers! This fake is obviously not doing the single thing it was supposed to do, regulate! Have a great day.
While a (genuine) 7805 is capable of having much higher input voltages than the measly 7V he's feeding it, you gotta keep in mind that the 78xx are anything but efficient, and the higher the input voltage, the more waste heat these things produce. It's good practice to keep the input voltage as low as you can (with some overhead), otherwise you'll be dealing with absurd amounts of heat to get rid of. These regulators only have two plus points left that justify their existance, one being that they're dummy simple to use (great for people learning electronics), the other is that they don't have much of any noise on the output like a switchmode regulator would (unless you massively overengineer it, which adds cost). This is kind of crucial when you're dealing with audio stuff where you want as little noise as possible that could bleed into the signal path
I would be interested to see these opened up and analyzed for what is in them. I have to dig through my collection of 7805's since I think I may have some of these to test for myself.
Weird, they're the right shape to be ST parts with the indents on the sides, the thickness of the tab could be a clue, recent ST regulators I've bought have been wafer thin, 0.5mm and they were definitely genuine parts. I think it's almost definitely a regulator of some sort but no idea which, the fakers tend to be pretty lazy when it comes to cheap parts like these and it's usually either completely wrong (bipolar transistor sold as a MOSFET for instance) or a re-marked lower spec/different part (I had a bag of parts that were meant to be 78L05 but turned out to be some other 5V regulator with a different pinout) Have you hooked one up to one of the GM328 tester (or derivative) thingies? Fancy knocking the top off one and trying to get a die picture?
Had the same effect with lm317 recently. I can set any voltage but I can not get more than 250 mA on load without a drop down in volts. And here, in Russia, we have our own copy of these stabilisers KP142EH** series. Which also has counterfeit versions. Imported LMs are selling with the same price as locally manufactured. Pure lottery. 😅😊
Have a oscilloscope and check for oscillation of the regulators. These regulators are very sensitive to oscillation. It could be that they oscillate because you did not add 100nF capacitors to input and output. Also have a 10uF tantalium capacitor on the output. Always kKeep the wiring on in and output terminals as short as possible.
@@rfmerrill Yup, in general an output cap isn't required for stability but does improve dynamic performance. Some of the LDOs are really fussy about capacitance at their output. The cap ESR needs to fall within a fairly narrow range for stability. ESR that is too high or too low leads to problems. This cropped up in some of the sort of "mid generation" LDOs. Most of the newer designs are tolerant of very low ESR so you can use monolithic ceramic output caps (if you avoid those wretched types with horrendously large negative voltage coefficient of capacitance - they'll work but by the time you take the capacitance drop with DC bias into account you might as well have used a good cap in the first place)
@@uploadJ "All characteristics are measured with a 0.33-µF capacitor across the input and a 0.1-µF capacitor across the output." (from the µA7800 series datasheet). Application schematics are inconsistently drawn with the output capacitor.
This is surprising, and news to me. Thank you for your demo. I'm not using these devices on a day to day basis as I once did, but I never implanted them into circuits where I only had 2 volts of headroom over the rated output. I thought 3 was the dead minimum and more = better. Still, I found 78xx devices that could not maintain output even with plenty (5-8) excess volts feeding them, and then they start to get hot. Counterfeit voltage regulators. How long do you suspect this has been going on? Man, I want to go on a rant, is every effing thing potentially a fake?
well it been pretty obvious from the first video too that these specific chips are current limited to 500ma on the output, it is no heat issue either because with the real chips you would burn your finger if they reached thermal throttling. It is a very common practice from china to re-label or re-brand components or even whole products in the "cheap" segment and they can even sell salvaged components that been re-label or marked too. today you can only trust your instruments or quality distributors of these parts. you get what you pay for. a lot you buy even on sites you trust is a cheaper copy or replica of something else often with worse performance. Everything today is about profit
My one criticism is the same for others, I wouldn’t classify this as a “fake”, perhaps “crap quality” or if you’re nice, “out of tolerance from spec”. I’ve had fake chips, they just don’t work as they’re completely not what they’re labeled as. The horrid part is that someone feels the need to make crap versions of something like a 7805. I mean, why.
Sorry I'm late to the party. Don't these regulators require a capacitor on the input? TI seem to recommend 0.22uF if the chip is far from the power supply filter which it is in this setup..
While i wouldn't run 12v into a 5v, 1.5A reg , that's not to say you can't. And in your first video you were able to hold you finger on the reg , which is my "good enough" rule of thumb . So thanks for a second video on the subject , but I wouldn't worry about the nitpickers who studiously ignored the fact that the output had dropped out of spec before thermal derating would have come into play on a genuine 7805. So thanks for the first video and the bonus follow up too.
Fakes are on the up. I bought a brand new GT40WR21 IGBT from eBay which was marked Toshiba but it was dead and the casing looked nothing like a genuine Toshiba part. After showing the seller photos of it next to the genuine one, I got an instant refund.
So you delivered the template for an updated fake part to the seller? :) Once I bought a fake 1,5 V-battery in a backyard shop in China for the camera. The charge was just enough for one flash, and I was flabbergasted how people would fake simple cheap objects like a battery. I should have been alerted by the soft malleable form, but it was dark, I was in a hurry and of course don't speak Chinese...
@@brilanto _"So you delivered the template for an updated fake part to the seller? :)"_ Yes I have bought from the seller many times before and all have been good apart from that. That datasheet clearly shows the dimensions and the physicality of it.Obviously it was the seller who was scammed and was very apologetic. I guess you will be a bit smarter the next time you buy anything .😆
The IC was NOT heat sunk as is required, it takes more than touching the heat sink. How does a genuine IC of that type performed similarly heat sunk? Ron W4BIN
I wonder if they really are fakes or maybe legit parts that didn't meet spec and for whatever reason ended up on the market anyway in cheap bundles and kits.
Farnell data sheet states 10volt min input so it will fall sharply Fixed Output Voltage Nom: 5V Adjustable Output Voltage Min: - Adjustable Output Voltage Max: - Input Voltage Min: 10V Input Voltage Max: 35V Product Range: 7805 Voltage Regulators Output Current: 1.5A
That's why I decided to test them in the first place. I replaced a faulty LM2940 with one of the counterfeits and it didn't even come close to 5 until I put two in parallel.
With the exceedingly poor regulation they would share, but if you have good regulators simply connecting them in parallel works poorly. The regulator with the highest output voltage, even by a few millivolts, will take the majority of the load current. You can use ballasting resistors to improve sharing but that degrades regulation. This is universally true with constant voltage sources. They just don't share well when paralleled because of their fundamental nature.
7V input? That's low for a standard 7805. I'm not too surprised at the results, 78 series regulators where always a bit crap. I'm also not sure your "heatsink" is actually doing anything as its jut the tab that's on the sink. The main body could still be getting a lot hotter and thermally throttling the output. 18:00 - Something like 3W is enough to overheat a 7805 without proper heat sinking.
I thought your last test was just fine!👍 I'm going to say again, these things are total garbage! & again, 4 of these in parallel would not match the performance of a single brand name component!
Obviously fakes. You can increase the output on fixed regulators by adding a Diode in series with the ground, for every diode you add in series, it increases the output by .6v a volt per diode. So with a 12v fixed you can get 13.8 with Three diodes. Ideal for a trickle charger for a Lead Acid Battery. Adding a wafer switch lets you step the volts out.
There are fake NE555 timer chips on the market, so why would anyone think that other common semiconductor parts wouldn’t be fakes if bought on trading sites?
Exactly. china doesn't give a sh*t about what they fake - if it makes them money they will do it. That's the china way, it's all about "how much can I make today?"
You even get fake resistors and wire. For IC's the common thing is to get any thing in the same package, retin the leads, then either sand the top off or blacktop it, and laser etch the fake part number on it.
@@SeanBZAYup. Ordered metal oxide resistors, got basic carbon film made to look like metal oxide ones. And yes, something as simple as wire is now also a problem. The amount of times I've run into CCA/CCS (copper clad aluminum/steel) trash instead of actual copper is insane
@@Knaeckebrotsaege Yes. I use LMR-400 coaxial cable in my work, and a respected wholesaler here is selling "CNT-400". Looks like LMR-400 but the centre conductor is CCA. Made in china of course.
@@SeanBZA There are also fake capacitors being manufactured by china. I despise them and the horses they rode in on for what they are doing to the electronics industry - all so they can make some easy money.
Counterfeit or just crap? I don't know what the logo is, but it didn't look anything like an ST logo. The behavior seems to be consistent with a 500mA current limit in the regulator. I'd be curious if you could figure out what that logo is and whether there is a datasheet from the actual manufacturer.
hi i had some 7805 of ebay i know there fakes out there that don't run to to data paper work i find what the load you put on them i used 5v dc to drive ic logic in tascam 38 i had 40 year old ones go down fake one would blow the fuse if they start to play up i find never had this ever if the 7805 or 7824 is going to do a job were it got to be right go to RS 100% real thing i hope you put a big load on the 7805 the deck would never get this ever i think there's alot of worry about ic's got to be right note the pcb's in decks i repair are coming apart and lifting there copper tracks as they are getting older note tascam 30 line were not made great from the start same as the FOSTEX R8
One small point, your load is not a current draw. The current is what would be drawn if the voltage didn't drop. I'm sure you are aware of that but some viewers might wonder why the current shown on the PSU was less than the load current you implied.
i dont think its fake 7805. its more like unethical selling. the manufactureres of the device separate the defectives from the normal ones. and then decide to sell the defective units as normal devices.
I think some times those are just relabeled cheaper chips. GreatScott have few videos about these fakes from amazon/ebay/aliexpress. Also Julian Ilett had some fakes used in ebay 60V stepdown converter where hi voltage chip was relabeled normal (40V) so it burned where he wanted to use it.
Perhaps everything would be fine if you only ordered parts from Digikey or Mouser. Using any other 'trading' site is just asking for trouble. If you're not in the USA, then there is a electronics parts vendor in the UK and probably one in Germany. Just saying.
Don't fall for the trolls. You did everything OK in the first video. 12V was a bit on the high side but that is not very uncommon to do that in a real design.
It's beyond me why somebody would want to fake those parts, them being as jellybean as they are. I had to find and watch part 1 of this series to see where you got them from, in this case ebay, which I have seen referred to as "the dumbpster of the 'net". It would seem to me that you bought a batch of rejects there. Perhaps the chip is not properly bonded to the tab? I donno, but I'd not peronaslly be buying those sort of parts from ebay...
Horrible line and load regulation; absolutely a fake ST part. Also looks like the little indent on the case has no raised number, another giveaway. It is curious that there is virtually no quiescent current draw, another giveaway. If I had to guess, there is a MOSFET and a simple bias resistor inside, not connected to ground, designed to give about 5 Volts out at 100 mA with a 10 Volt input. It makes you wonder why anybody would counterfeit a 78L05CV which is a big die and quite simple fab. It couldn’t have saved much if any money to fab a fake part as there is apparently some semiconductor thing in the case. Why not just manufacture a legitimate part and be done with it? 😊
Or they got a box or fifty of reject parts, and simply took them, and marked them. After all, just has to pass a junction test to pass most incoming inspection, and not even that. Probably all the reject parts off a line, that should have been dropped into a grinder and etched out for copper, but instead a few bags walked out the door.
Presumably the guy doing the test knows how his power supply meters perform, but I wouldn't put too much store in apparent zero current with no load. If the meter in the power supply uses a low value shunt for current sensing and the supply is 10 A at full scale, the voltage out of the shunt at a few milliamps will be very low. A shunt delivering 100 mV at 10 A gives 10 µV per milliamp. The quiescent current of an L7805 is about 6 mA max, for 60 µV from the shunt. That is well below typical input offset voltage for a very large percentage of op amps.
@@d614gakadoug9 I agree that it is getting into the weeds for expecting an accurate measurement at 5 mA when full scale is 10 A. He didn’t say if he did more than look at the readout on the power supply for verification. The genuine part showed a small value for quiescent current consumption, versus zero for the fake part. Zero could possibly be a value between zero and slightly less than 5 ma. The best way to phase it is that the fake part apparently has less than the expected quiescent draw, probably less than 5 mA.
@@d614gakadoug9 Yes most PSU meters also are single side opamp based, and thus will read anything below 100mA as zero current anyway, as the opamp will be operating with essentially grounded inputs. Not many will have even a simple 2 diode drop to give a -1V2 rail for the opamp to be able to measure reverse current flow. Even fewer will have a dual supply and range switching, to give both a better resolution at at low currents.
Um. That's not going to work at all well as a heatsink. It needs to be under the package where the chip is, not just under the hole you use to bolt it down. It's almost like you deliberately did the heatsink in a way that would keep your temperature probe cool while still allowing the regulator to overheat.
Since the device was only dissipating about 2W (0.5A x 4V) at most, there's no way its die could reach the thermal limit of 150°C, even if the heatsink wasn't there. The regulator clearly didn't overheat.
@@RexxSchneider You're significantly kinder than I am. Kudos. I had to make sure I wasn't the dummy after I posted my mean replies and tested one of my fake L3808s and though it did better than the 5 and 12 I tested previously the point was a back-to-back comparision of heatsinking methods and there was no significant difference at all. In fact the one held by just the tab read 20mV higher on the highest draw tests. This is a really stupid thing to be a keyboard qualified engineer in the comments over.
A couple comments. First off it looks like a fail on voltage regulation under load. But, are you still getting 7.00 volts at the regulator terminals? How much has been lost in the test leads from the power supply? This is critical as you are right at the 2 volt limit with no load. Second, what is your intent with the heat sink? Because with the thermocouple in between the regulator and heat sink like that you have very little contact area between regulator and heat sink and it will be hard to make this uniform between tests. A fun exercise would be to calculate the loss in the regulator (input power minus output power). Is this the same between a genuine and a counterfeit regulator? It's going to be close! So measure really accurately. If you have a power supply capable of remote sense put the sense leads right at the regulator chip itself then you can be assured of getting exactly 7.00 (or whatever) voltage there.
The test leads from the power supply are pretty good, I believe they are 16 gauge, and they handle the full 10 amps without heating up. I also have some cheap ass leads that are 16 gauge jacket with 24 gauge wire, but that subject is another video all together.
@@DCFluX 16 AWG is a good start but there will still be voltage drops in the leads and connections. Since you already have the DMM there to look at the output of the regulator you might as well move the lead over (common negative) and verify the input is still at 7.0 volts. If necessary you can slightly raise the power supply voltage to get to exactly 7.0 volts at the regulator input terminal. Not only will this give a more accurate analysis of the chip but you can also increase the power supply to see if a little extra input voltage will make it work properly and give you 5.0 volts output. This would be an interesting thing to know, as even if counterfeit, you might find they work but need more than 2 volts at higher currents.
If you have little knowledge of electronics, it is true that some fake regulator transistors are unable to maintain a constant 5v. Most of this type should handle around 1 ampere max. You must also make sure that the power supply is above a certain volt and does not drop below it. Normally you should stay at a minimum of 7v if you are going to have 5v out.
Sorry for my English! I don't see any fakery here. The lm7805CV is not an LDO regulator, it needs 2.0V on average (2.5V maximum) to work, the maximum input voltage is arou 35V, according to the datasheet. It can supply 1.5A maximum. The 7.0V input voltage is a bit on the low side, if I conducting the experiment I would use 7.5V, but again, according to the datasheet, 7.0V has to been enough. The initial 4.8V output voltage is the minimum acceptable voltage, the datasheet says. So, what is your concern?
@@lor0the0fallen0angel This part he has is fake! He tried with a much higher voltage than 7v but still the voltage was far below 5v at 1A draw which is not what it should be. what's the point of a 5v regulator if the voltage is 2-3v ?
I'm not sure they are fake, I use 78 and 79 series regulators since 1970's and I've learned a lot from my experience: NEVER USE ST parts, II do not have anymore available the official response from SGS of those years, stating that te internal protection circuit of these devices is different from other similar (Fairchild, TI, Motorola) and mostly the 79's had an overvoltage protection the did not allowed them to start correctly. My ANSWER WAS : NO MORE SGS or ST regulators I was building audio mixers, and the start -up problems were present also in equipment built by some competitors,: When mains voltage was very high the negative output was ZERO
Not all 78 regs are the same specification - that's why they have a different data sheet - but they generally give the rated voltage over the data sheet specified conditions. This test is not stressing the part (over voltage or over current) so your argument doesn't really hold any water. I've used hundreds of ST regs over the years in digital equipment (input up to about 10V over output voltage) and not had problems in *normal* use. Either these are faulty parts or fake parts ... If you are saying ST never made a working regulator I, and millions of installations, would beg to differ.
And here we have a prime example of a boomer unwilling to admit that he's wrong 🤣 "It *must* be the company that produces nonworking crap for 40+ years, it's definitely not me who's incompetent here!" /s
Not talking about current and ohms law. Talking about the chip output voltage. It should work with 2 volts input margin, If the output voltage drops (lower output = more margin).
I'm not sure he understands how the heat sink is supposed to work. 🤣🤣 Sticking the tip of my pinkie finger into a glass of ice water isn't going to cool my body down on a hot summer day. At a minimum, I will need to either cool down my core, cool all of my extremities, or at least expose a significantly large section of my body to coolness in order to cool down my entire body. It's much the same with this component. Simply touching an extended part of the body (an "extremity", as it were) onto a heat sink isn't enough to efficiently draw that heat away from the body. You're making that heat travel a ways just to reach the heat sink, instead of allowing it to immediately dissipate. Just sayin'....
The maximum expected dissipation in the tests is 3 watts. That isn't much. The "tab" on a TO-220 package is normally pure copper, with very high thermal conductivity. The test setup is plenty good enough. Unless the nerves in your fingers are dead if you can hold a finger tip on the plastic part of the package for several seconds the die is operating far below its over-temperature shutdown point.
The datasheet shows a thermal resistance junction-to-case of 4°C/W for the TO-220 package. As the power dissipated in the device never exceeded (0.5A x 4V) = 2W in all of the tests, it's safe to say that the junction temperature would never be more that 8°C above the case temperature. Since a genuine part has thermal shutdown at a junction temperature of 150°C, it should be pretty clear that being able to put your finger on the case without any problems should guarantee that the junction is nowhere near thermal limiting. The heatsink was a bit of an irrelevance in this test.
@@d614gakadoug9 re: "The test setup is plenty good enough. " Tell me again this is good with the thermocouple tucked UNDER the TO-220 tab and next to the heatsink?
@@uploadJ It is plenty good enough for what is expected. If I were doing the experiment I wouldn't even bother with the temperature probe. It _is_ poorly positioned for accurate measurement but it just doesn't matter when a finger can be held on the device. In a "genuine" regulator, overtemperature shut-down is abrupt. It does not cause a gradual drop in output voltage. We saw no such abrupt turn-off. The counterfeit part may behave differently but we have no idea what to expect so we have no idea if temperature is coming into play. Something significantly missing from the experiments in the video. and far more important than a vague idea of device temperature, is measurement of the actual input voltage at the regulator terminals. He also took the "typical" dropout voltage from the datasheet as "the" dropout voltage. You don't get to do that. You must consider worst-case. Against that, a bipolar regulator that uses an emitter follower as the pass device has a negative temperature coefficient of dropout voltage. I'd also be checking with an oscilloscope just to make sure there was no oscillation. back to temperature ... If I wanted to improve the accuracy of temperature measurement in that setup, I'd leave the body of the device hanging over the end of the heatsink and carefully affix the thermocouple to the free part of the "backplate" near the die area (roughly central under the plastic part). Obviously that would limit the allowable dissipation in the regulator, but the maximum expected dissipation based on the design of the experiment is only three watts. Low-cost thermal cameras are a good way to do some of this stuff now. I used to make quite extensive use of temperature "crayons." These are solids with tightly controlled melting point. They only tell you if the temperature is above or below the melting point but that can still be very instructive. I did a lot of mains-powered SMPS design - you don't go sticking you fingers on the power semis. Even thermocouple probes can be a problem because of the noise that abounds in switchers. Sticking with thermocouple probes, I can think of a number of experiments that would require far, far more care and proper design. An example is a comparison of screwing a TO-220 to a heatsink versus using a mounting clip That might even involve careful drilling of probe wells into the plastic. (I'm no fan of the TO-220 because it is hard to mount well; big fan ot TO-247, but it's considerably more expensive).
You 12V critic need to be educated about electronics, ...you did everything right before. NOT THIS TIME, you heat sink is now crap, because the BODY of the reg should be in contact with the heatsink! AS for temp, any "proper" reg in th 78/79 series will operate up to the limit of temp, and if too hit will SHUT DOWN, not voltage droop. Small die pass tranny inside is why the voltage drops at higher currents. ...again, YOU DID EVERYTHING BEFORE!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
The heatsink in this test is plenty good enough. The maximum expected dissipation is only 3 watts. The maximum allowable die temperature for guaranteed operation is 125 °C and the shutdown temperature will be well above that. The thermal resistance from die to ambient only needs to be 40 K/W to keep the die at the allowable max at the full expected power dissipation.
If you have little knowledge of electronics, it is true that some fake regulator transistors are unable to maintain a constant 5v. Most of this type should handle around 1 ampere max. You must also make sure that the power supply is above a certain volt and does not drop below it. Normally you should stay at a minimum of 7v if you are going to have 5v out. This is not a 5v regulator you want in your circuit! Fake !
H̶a̶v̶e̶ ̶y̶o̶u̶ ̶c̶h̶e̶c̶k̶e̶d̶ ̶f̶o̶r̶ ̶o̶s̶c̶i̶l̶l̶a̶t̶i̶o̶n̶s̶?̶ Just watched the first video, Test setup seems stable as the genuine parts worked correctly.
Looking at those results, it would not surprise me if that’s a cheap 78L05 (100mA) copy encapsulated in a TO220 case.
Thinking the same. Enough 5v regulators that have a lower spec. The case is not an indicator of how much power it can handle.
I was initially thinking 78M05 (500mA) which wouldn't even need a different casing as those already come in TO220, but these fakes in the video have already faceplanted at just 500mA load
Try operating without heatsink to check it has been filled with the correct volume of smoke 😊
A legitimate part would simply shut down due to excessive die temperature, probably at around 150 °C
True. I’ve made heat sinks out of 1 inch nominal *copper pipe caps* - had to mill a slot, drill a hole for the bolt to hold the part down, etc.
Always drives me nuts when the datasheets spec smoke in cubic centimeters instead of inches.
@@c1ph3rpunk conversion factor is 2.54^3…
It is annoying, though, just the same.
Don't sweat those handful of losers who have nothing better to do than criticize or accuse others of fakery. They are overwhelmingly in the minority, but simply have the biggest mouth. Keep doing what you are doing, because I appreciate and enjoy it!
Indeed, internet warriors are going to complain anyway. It's a waste of time to hold them in mind for more than half a second.
Absolutely correct.
not to mention these clowns won't bother to watch a 20min proof video either
"Overwhelmingly in the minority" Now that makes sense!
I think otherwise, I think we should be humble, I think sources deserve to make criticism without accusing those sources of being “losers”
A big attribute to failure is not being subject to any opposing views and putting yourself into an echo chamber or egotistical position like you have suggested.
This man doesn’t deserve failure, I don’t know him nor have I seen these videos but I think not..
Your setup is indeed good for proving that there are fakes out there. And it is a good thing to bring this issue under the spotlight.
For a more precise measurement, you should measure the input voltage at the terminals of the device under test, because there's a voltage drop over the wires from your power supply to the device under test. That's why your first setup was actually better: with the supply at 12V, for sure you won't be below 7.00V at the device under test. With the supply at 7.00V and a voltage drop over your wires, you might end up with for example 6.8V at the device under test, which is out of spec.
After watching your first video, tested ones I got from eBay and they act the same as yours. Crap!!! I used two but forgot where I put them in.
I have found exactly the same issues with fake 15-volt type regulators. They kind of work OK up to a couple of 100mA but fall away rapidly above that number. A very long way from 1500mA. I tested mine with 20 volts input. Not an unrealistic input voltage.
We always get a couple of people that think you are testing it incorrectly. I don't lose any sleep on such people, nor should you.
👍🏻🙂
I too have experience using these devices, for well over 45yrs, and have many times come across bogus parts, though this tends to be a more recent trend.
I categorically debunk the "too high source voltage" or "too hot" claim from the detractors.
It takes a hell of a lot to cause the legitimate devices to be off spec, or much less kill them,
(having on occasion, in much younger years, had the device too hot to touch)
The "operating temperature" is listed well above 100°C(212°F), yet as you clearly shown and stated the DUT was barely warm,
and I somehow doubt you would be calmly touching the DUT if it was much hotter than +55°C(131°F), which for a majority people is borderline 'bloody HOT'.
Another point is that the 78xx devices are typically rated to Vin 33v-35v for 5v-18v devices(40Vin for 24v devices).
You were only feeding with 12v, which puts operating parameters well and truly within spec.
Kudos for the basic explanation you provide, I'm sure will help many understand.
(Just tell the not so knowledgeable to go learn before splashing bulls..t)
Oh i believe that its definately a bootleg part. I was one of the people that questioned the heatsink mounting..but if the temp wasnt above those numbers..I agree it seems to not have been a contributing factor.. I also know the b+ voltage also was no issue..its clearly just junk part.
Maybe they snuck a die for 78M05 in that package since it current limits at 500 mA or they just had a bad chip copy engineer. I had a bunch of cheap 78L05 that were much worse performance - they actually started to smoke - missing over temp circuit ! They were marked WS78L05 for WingShxx or something LOL
WS is for WhaleSh1t Semiconductor (Xinjiang province)
78M05 does come in to220 package, i've had them in the past
WS... as in "weak shit" 🤣
would be interesting to see how the 100mA variant of the 7805 would perform. I think it would be very similar to the fake 1.5A one. Maybe they put a 100mA variant in the case of the 1.5A variant.
you could find different varient of fake 7805 from 20 years ago,some of them only current limits didn't work some thermal limit some need bigger output input cap to be stable and some with very low output current capacity
WOW. I never thought to check if 7800 series regulators were fake. I will now this coming winter. Great winter project. I have tested LM317s to see if they heat regulate. They do. Did not find any fakes thank goodness.
I don't see the usual bypass capacitors on the input and output of the 7805 on your test rig. Try your scope- the 7805 (or whatever it is?) may be oscillating. They lose regulation when oscillating. Thanks!
They lose regulation because they are FAKE it's as simple as that.
The yellow cylinders between the resistors on the test jig are 10uF Tantalum capacitors.
because of the load through the power lead cable , what's in the input og the 7805?, , you should use little more then 2 volts then the datasheet states , and put a 100 nf cap on the output to precent parasitic oscillation.
An output cap generally isn't necessary for stability but it does improve dynamic performance. An input cap is necessary with long and hence inductive input leads. Two separate test leads are about as bad as it is possible to get. It is much better to use a tightly twisted pair if the leads must be long.
Also in a test like this the input voltage should be measured right at the regulator.
I'd be looking at everything with an oscilloscope.
I have a bunch of these with a different lot code: MAR 1904, otherwise exactly the same. Mine seem to be 78M05/LM341 devices. They drop out at around 500mA. Was a very annoyed when I found out, but they were _really_ cheap so in retrospect don't mind it very much.
I seem to remember many of those type devices don't act right without some capacitance at the load.
Look at the output on a scope for any oscillations.
(I'm not saying they aren't cheap knockoffs, or they ever perform to spec!)
so... they are actually 5v 200ma regulators that have been labeled as 5v 1.5A.
Yuk.
Maybe teardown and differences in silicon next? By the looks of it, it does look like it is ams1113'ish LDO just in counterfeit package. Nice vid tho, never trusted those 7805 with those paper thin heatsinks :) .
these might be 78L05 (100mA) in disguise. They certainly behave like overloaded 78L05s
The lengths you go to because the internet is populated by know-it-alls. Thanks again, sir!
China source IC's & batteries such as those found on Ebay, Bangood etc are a lucky dip at best going on my own experiences. They work but don't quite perform to the genuine device's expected ability in standard published datasheets. Not all, but some and the only way to sort them out is to test each one individually so a lot of frustration & time can be saved by spending a little extra and ordering genuine devices from local suppliers (in my case it's Farnell, Digi-Key, RS components etc)
CPC is very good and a bit cheaper than Farnell that is the parent company. I have only ever had genuine parts from them.
I'm actually impressed how many Timmy know-it-alls in the comments don't understand thermal conductivity well enough to prevent them from writing their entire thesis out
I have no doubt that what you have there is a counterfeit (or a small chance of being a batch of factory rejects) 7805. Any decent 7805 will not drop significant voltage before 1A load. Even at 1A, it should not be over 10% of its rating, worst case. The only criticism I have for your previous and this video is not using input and output caps as per datasheet, and not checking for oscillation (or ripple) with a scope. I have an educated suspicion that fake 7805 is oscillating like crazy under load.
Don't bother trying to appease the doubters. They'll find creative new ways to doubt you no matter how hard you try.
I don't understand, this regulator's datasheet says 7 to 35 volts input. 12v seems to be easily on the safe side for a supply! You also kept the temp far below the 125 degree C range that is also allowed. Ignore the naysayers! This fake is obviously not doing the single thing it was supposed to do, regulate! Have a great day.
While a (genuine) 7805 is capable of having much higher input voltages than the measly 7V he's feeding it, you gotta keep in mind that the 78xx are anything but efficient, and the higher the input voltage, the more waste heat these things produce. It's good practice to keep the input voltage as low as you can (with some overhead), otherwise you'll be dealing with absurd amounts of heat to get rid of. These regulators only have two plus points left that justify their existance, one being that they're dummy simple to use (great for people learning electronics), the other is that they don't have much of any noise on the output like a switchmode regulator would (unless you massively overengineer it, which adds cost). This is kind of crucial when you're dealing with audio stuff where you want as little noise as possible that could bleed into the signal path
I would be interested to see these opened up and analyzed for what is in them. I have to dig through my collection of 7805's since I think I may have some of these to test for myself.
Where's your input bypass cap? ANd the output cap? What value does the application note recommend?
Weird, they're the right shape to be ST parts with the indents on the sides, the thickness of the tab could be a clue, recent ST regulators I've bought have been wafer thin, 0.5mm and they were definitely genuine parts.
I think it's almost definitely a regulator of some sort but no idea which, the fakers tend to be pretty lazy when it comes to cheap parts like these and it's usually either completely wrong (bipolar transistor sold as a MOSFET for instance) or a re-marked lower spec/different part (I had a bag of parts that were meant to be 78L05 but turned out to be some other 5V regulator with a different pinout)
Have you hooked one up to one of the GM328 tester (or derivative) thingies?
Fancy knocking the top off one and trying to get a die picture?
The thin ST regulators are "single gauge" and the standard tab thickness are "dual gauge" and have a -DG suffix on the part #. They make both.
Had the same effect with lm317 recently. I can set any voltage but I can not get more than 250 mA on load without a drop down in volts. And here, in Russia, we have our own copy of these stabilisers KP142EH** series. Which also has counterfeit versions. Imported LMs are selling with the same price as locally manufactured.
Pure lottery. 😅😊
Have a oscilloscope and check for oscillation of the regulators. These regulators are very sensitive to oscillation. It could be that they oscillate because you did not add 100nF capacitors to input and output. Also have a 10uF tantalium capacitor on the output. Always kKeep the wiring on in and output terminals as short as possible.
a proper 7805 should not require much output capacitance. Output capacitance is typically needed on low-dropout regulators.
@@rfmerrill
Yup, in general an output cap isn't required for stability but does improve dynamic performance.
Some of the LDOs are really fussy about capacitance at their output. The cap ESR needs to fall within a fairly narrow range for stability. ESR that is too high or too low leads to problems. This cropped up in some of the sort of "mid generation" LDOs. Most of the newer designs are tolerant of very low ESR so you can use monolithic ceramic output caps (if you avoid those wretched types with horrendously large negative voltage coefficient of capacitance - they'll work but by the time you take the capacitance drop with DC bias into account you might as well have used a good cap in the first place)
@@rfmerrill re: "a proper 7805 should not require much output capacitance"
What does the National or TI datasheet say?
@@uploadJ "All characteristics are measured with a 0.33-µF capacitor across the input and a 0.1-µF capacitor across the output." (from the µA7800 series datasheet). Application schematics are inconsistently drawn with the output capacitor.
This is surprising, and news to me. Thank you for your demo. I'm not using these devices on a day to day basis as I once did, but I never implanted them into circuits where I only had 2 volts of headroom over the rated output. I thought 3 was the dead minimum and more = better. Still, I found 78xx devices that could not maintain output even with plenty (5-8) excess volts feeding them, and then they start to get hot. Counterfeit voltage regulators. How long do you suspect this has been going on? Man, I want to go on a rant, is every effing thing potentially a fake?
Hate to do this to you, but what's the voltage at the clips/input? It's more a curiosity thing than a doubt thing.
well it been pretty obvious from the first video too that these specific chips are current limited to 500ma on the output, it is no heat issue either because with the real chips you would burn your finger if they reached thermal throttling. It is a very common practice from china to re-label or re-brand components or even whole products in the "cheap" segment and they can even sell salvaged components that been re-label or marked too. today you can only trust your instruments or quality distributors of these parts. you get what you pay for. a lot you buy even on sites you trust is a cheaper copy or replica of something else often with worse performance. Everything today is about profit
My one criticism is the same for others, I wouldn’t classify this as a “fake”, perhaps “crap quality” or if you’re nice, “out of tolerance from spec”. I’ve had fake chips, they just don’t work as they’re completely not what they’re labeled as.
The horrid part is that someone feels the need to make crap versions of something like a 7805. I mean, why.
are you checking regulators or probing screws?
some one else pointed out these can unstable with long leads and no bypass caps 0.1 ~bypass ect a scope to check wold be good
I thought your first video was quite conclusive, especially since you duplicated the test with a known legit 7805 to show the difference.
Sorry I'm late to the party. Don't these regulators require a capacitor on the input? TI seem to recommend 0.22uF if the chip is far from the power supply filter which it is in this setup..
You should be using thermal grease! all the metal of the package should be in contact with the metal heatsink!
Please see the follow up video:
ruclips.net/video/a0ffW9QgVDQ/видео.html
100mA 78L05 in a TO220 package? Yes?
While i wouldn't run 12v into a 5v, 1.5A reg , that's not to say you can't. And in your first video you were able to hold you finger on the reg , which is my "good enough" rule of thumb .
So thanks for a second video on the subject , but I wouldn't worry about the nitpickers who studiously ignored the fact that the output had dropped out of spec before thermal derating would have come into play on a genuine 7805.
So thanks for the first video and the bonus follow up too.
Fakes are on the up. I bought a brand new GT40WR21 IGBT from eBay which was marked Toshiba but it was dead and the casing looked nothing like a genuine Toshiba part.
After showing the seller photos of it next to the genuine one, I got an instant refund.
So you delivered the template for an updated fake part to the seller? :)
Once I bought a fake 1,5 V-battery in a backyard shop in China for the camera. The charge was just enough for one flash, and I was flabbergasted how people would fake simple cheap objects like a battery. I should have been alerted by the soft malleable form, but it was dark, I was in a hurry and of course don't speak Chinese...
@@brilanto _"So you delivered the template for an updated fake part to the seller? :)"_ Yes I have bought from the seller many times before and all have been good apart from that. That datasheet clearly shows the dimensions and the physicality of it.Obviously it was the seller who was scammed and was very apologetic.
I guess you will be a bit smarter the next time you buy anything .😆
I was just about to order some regulators and now see that eBay is awash with these fakes.
The IC was NOT heat sunk as is required, it takes more than touching the heat sink.
How does a genuine IC of that type performed similarly heat sunk? Ron W4BIN
Hi should there be a 10mfd tant on the output for stability?
It IS there! The axial orange caps.
Why did you mount just the tab? Instead of using an insulator and thermal grease and mount the body on the heatsink also.
I wonder if they really are fakes or maybe legit parts that didn't meet spec and for whatever reason ended up on the market anyway in cheap bundles and kits.
Farnell data sheet states 10volt min input so it will fall sharply
Fixed Output Voltage Nom:
5V
Adjustable Output Voltage Min:
-
Adjustable Output Voltage Max:
-
Input Voltage Min:
10V
Input Voltage Max:
35V
Product Range:
7805 Voltage Regulators
Output Current:
1.5A
I need to convert 50v to 5v DC what would you suggest, needs to be able to deliver 1a
Try hooking 2 or 3 up in parallel just for the fun of. lol.
That's why I decided to test them in the first place. I replaced a faulty LM2940 with one of the counterfeits and it didn't even come close to 5 until I put two in parallel.
With the exceedingly poor regulation they would share, but if you have good regulators simply connecting them in parallel works poorly. The regulator with the highest output voltage, even by a few millivolts, will take the majority of the load current. You can use ballasting resistors to improve sharing but that degrades regulation. This is universally true with constant voltage sources. They just don't share well when paralleled because of their fundamental nature.
7V input? That's low for a standard 7805. I'm not too surprised at the results, 78 series regulators where always a bit crap. I'm also not sure your "heatsink" is actually doing anything as its jut the tab that's on the sink. The main body could still be getting a lot hotter and thermally throttling the output. 18:00 - Something like 3W is enough to overheat a 7805 without proper heat sinking.
I thought your last test was just fine!👍
I'm going to say again, these things are total garbage!
& again, 4 of these in parallel would not match the performance of a single brand name component!
Obviously fakes.
You can increase the output on fixed regulators by adding a Diode in series with the ground, for every diode you add in series, it increases the output by .6v a volt per diode.
So with a 12v fixed you can get 13.8 with Three diodes. Ideal for a trickle charger for a Lead Acid Battery. Adding a wafer switch lets you step the volts out.
There are fake NE555 timer chips on the market, so why would anyone think that other common semiconductor parts wouldn’t be fakes if bought on trading sites?
Exactly. china doesn't give a sh*t about what they fake - if it makes them money they will do it. That's the china way, it's all about "how much can I make today?"
You even get fake resistors and wire. For IC's the common thing is to get any thing in the same package, retin the leads, then either sand the top off or blacktop it, and laser etch the fake part number on it.
@@SeanBZAYup. Ordered metal oxide resistors, got basic carbon film made to look like metal oxide ones. And yes, something as simple as wire is now also a problem. The amount of times I've run into CCA/CCS (copper clad aluminum/steel) trash instead of actual copper is insane
@@Knaeckebrotsaege Yes. I use LMR-400 coaxial cable in my work, and a respected wholesaler here is selling "CNT-400". Looks like LMR-400 but the centre conductor is CCA. Made in china of course.
@@SeanBZA There are also fake capacitors being manufactured by china. I despise them and the horses they rode in on for what they are doing to the electronics industry - all so they can make some easy money.
Counterfeit or just crap? I don't know what the logo is, but it didn't look anything like an ST logo.
The behavior seems to be consistent with a 500mA current limit in the regulator. I'd be curious if you could figure out what that logo is and whether there is a datasheet from the actual manufacturer.
Doesn't the data sheet specify decoupling capacitors on the input and output of these regulators?
Those caps are just to ensure low noise and no transients on the in and out....
The yellow thingies on the board between the resistors are tantalum capacitors, as he mentioned elsewhere in the replies
For such poor thing worse than fake, a TIP341+ 5V6 Zener + 100R Vcb resistor would make a 5V capable of 1A and short resistant to 4A.
Why would anyone accuse you of cheating on your video? It’s not like you’re claiming to have invented desktop fusion or a perpetual motion machine.
Under those videos are opposite comments (conspiracy of governments, Big Energy etc.)
hi i had some 7805 of ebay i know there fakes out there that don't run to to data paper work i find what the load you put on them
i used 5v dc to drive ic logic in tascam 38 i had 40 year old ones go down fake one would blow the fuse if they start to play up
i find never had this ever
if the 7805 or 7824 is going to do a job were it got to be right go to RS 100% real thing i hope
you put a big load on the 7805 the deck would never get this ever i think there's alot of worry about ic's got to be right
note the pcb's in decks i repair are coming apart and lifting there copper tracks as they are getting older
note tascam 30 line were not made great from the start same as the FOSTEX R8
I have a Tascam 34 and Otari MX-50 at the radio station I engineer for.
One small point, your load is not a current draw. The current is what would be drawn if the voltage didn't drop. I'm sure you are aware of that but some viewers might wonder why the current shown on the PSU was less than the load current you implied.
what is a technician worth who thinks that oh's and zero's are the same ?
It's a sign of sheople.
Got the same power supply some noise there so 8 volt cover it for me , Think this test is fine by me. Not worth have IMHO .
You need a screwdriver tork wrench
i dont think its fake 7805. its more like unethical selling. the manufactureres of the device separate the defectives from the normal ones. and then decide to sell the defective units as normal devices.
They wouldn't use a different laser-engraver and logo just for the defective ones. They're fake.
I think some times those are just relabeled cheaper chips. GreatScott have few videos about these fakes from amazon/ebay/aliexpress. Also Julian Ilett had some fakes used in ebay 60V stepdown converter where hi voltage chip was relabeled normal (40V) so it burned where he wanted to use it.
Perhaps everything would be fine if you only ordered parts from Digikey or Mouser. Using any other 'trading' site is just asking for trouble. If you're not in the USA, then there is a electronics parts vendor in the UK and probably one in Germany. Just saying.
I thought it was 3 volts differential , not 2 volts.
Don't fall for the trolls. You did everything OK in the first video. 12V was a bit on the high side but that is not very uncommon to do that in a real design.
So how did the naysayers explain how the non counterfeit regulator WORKED under the same conditions??? Yeah figured they didn't...
Anyone who says that your test is invalid is probably selling these things
Why are you calling them fake? They aren’t fake they are just really bad quality. They lied yes but they are still 7805. Not fake just shitty.
It's beyond me why somebody would want to fake those parts, them being as jellybean as they are. I had to find and watch part 1 of this series to see where you got them from, in this case ebay, which I have seen referred to as "the dumbpster of the 'net". It would seem to me that you bought a batch of rejects there. Perhaps the chip is not properly bonded to the tab? I donno, but I'd not peronaslly be buying those sort of parts from ebay...
Instead of L7805 it behaves more like 78L05
Horrible line and load regulation; absolutely a fake ST part. Also looks like the little indent on the case has no raised number, another giveaway. It is curious that there is virtually no quiescent current draw, another giveaway. If I had to guess, there is a MOSFET and a simple bias resistor inside, not connected to ground, designed to give about 5 Volts out at 100 mA with a 10 Volt input. It makes you wonder why anybody would counterfeit a 78L05CV which is a big die and quite simple fab. It couldn’t have saved much if any money to fab a fake part as there is apparently some semiconductor thing in the case. Why not just manufacture a legitimate part and be done with it? 😊
Or they got a box or fifty of reject parts, and simply took them, and marked them. After all, just has to pass a junction test to pass most incoming inspection, and not even that. Probably all the reject parts off a line, that should have been dropped into a grinder and etched out for copper, but instead a few bags walked out the door.
@@SeanBZA Agreed. The odd thing is that the admittedly small sample shown here had the same peculiar failure characteristics.
Presumably the guy doing the test knows how his power supply meters perform, but I wouldn't put too much store in apparent zero current with no load. If the meter in the power supply uses a low value shunt for current sensing and the supply is 10 A at full scale, the voltage out of the shunt at a few milliamps will be very low. A shunt delivering 100 mV at 10 A gives 10 µV per milliamp. The quiescent current of an L7805 is about 6 mA max, for 60 µV from the shunt. That is well below typical input offset voltage for a very large percentage of op amps.
@@d614gakadoug9 I agree that it is getting into the weeds for expecting an accurate measurement at 5 mA when full scale is 10 A. He didn’t say if he did more than look at the readout on the power supply for verification. The genuine part showed a small value for quiescent current consumption, versus zero for the fake part. Zero could possibly be a value between zero and slightly less than 5 ma. The best way to phase it is that the fake part apparently has less than the expected quiescent draw, probably less than 5 mA.
@@d614gakadoug9 Yes most PSU meters also are single side opamp based, and thus will read anything below 100mA as zero current anyway, as the opamp will be operating with essentially grounded inputs. Not many will have even a simple 2 diode drop to give a -1V2 rail for the opamp to be able to measure reverse current flow. Even fewer will have a dual supply and range switching, to give both a better resolution at at low currents.
Should send it. The full 35V 😁
Um. That's not going to work at all well as a heatsink. It needs to be under the package where the chip is, not just under the hole you use to bolt it down. It's almost like you deliberately did the heatsink in a way that would keep your temperature probe cool while still allowing the regulator to overheat.
much genius. wow
Exactly
Since the device was only dissipating about 2W (0.5A x 4V) at most, there's no way its die could reach the thermal limit of 150°C, even if the heatsink wasn't there. The regulator clearly didn't overheat.
@@RexxSchneider You're significantly kinder than I am. Kudos.
I had to make sure I wasn't the dummy after I posted my mean replies and tested one of my fake L3808s and though it did better than the 5 and 12 I tested previously the point was a back-to-back comparision of heatsinking methods and there was no significant difference at all. In fact the one held by just the tab read 20mV higher on the highest draw tests.
This is a really stupid thing to be a keyboard qualified engineer in the comments over.
@@RexxSchneider Well, your first video showed beyond reasonable doubt that they are fake, just because they were so bad compared to genuine part.
A couple comments. First off it looks like a fail on voltage regulation under load. But, are you still getting 7.00 volts at the regulator terminals? How much has been lost in the test leads from the power supply? This is critical as you are right at the 2 volt limit with no load. Second, what is your intent with the heat sink? Because with the thermocouple in between the regulator and heat sink like that you have very little contact area between regulator and heat sink and it will be hard to make this uniform between tests. A fun exercise would be to calculate the loss in the regulator (input power minus output power). Is this the same between a genuine and a counterfeit regulator? It's going to be close! So measure really accurately.
If you have a power supply capable of remote sense put the sense leads right at the regulator chip itself then you can be assured of getting exactly 7.00 (or whatever) voltage there.
The test leads from the power supply are pretty good, I believe they are 16 gauge, and they handle the full 10 amps without heating up. I also have some cheap ass leads that are 16 gauge jacket with 24 gauge wire, but that subject is another video all together.
@@DCFluX 16 AWG is a good start but there will still be voltage drops in the leads and connections. Since you already have the DMM there to look at the output of the regulator you might as well move the lead over (common negative) and verify the input is still at 7.0 volts. If necessary you can slightly raise the power supply voltage to get to exactly 7.0 volts at the regulator input terminal. Not only will this give a more accurate analysis of the chip but you can also increase the power supply to see if a little extra input voltage will make it work properly and give you 5.0 volts output. This would be an interesting thing to know, as even if counterfeit, you might find they work but need more than 2 volts at higher currents.
@@flyer617
Agreed!
The very first rule of voltage measurement is to measure it where it counts, which in this case is right at the regulator input.
You have been cheated, sir.
Bullshit oder you.
If you have little knowledge of electronics, it is true that some fake regulator transistors are unable to maintain a constant 5v. Most of this type should handle around 1 ampere max.
You must also make sure that the power supply is above a certain volt and does not drop below it.
Normally you should stay at a minimum of 7v if you are going to have 5v out.
Sorry for my English! I don't see any fakery here. The lm7805CV is not an LDO regulator, it needs 2.0V on average (2.5V maximum) to work, the maximum input voltage is arou 35V, according to the datasheet. It can supply 1.5A maximum. The 7.0V input voltage is a bit on the low side, if I conducting the experiment I would use 7.5V, but again, according to the datasheet, 7.0V has to been enough. The initial 4.8V output voltage is the minimum acceptable voltage, the datasheet says. So, what is your concern?
@@lor0the0fallen0angel This part he has is fake! He tried with a much higher voltage than 7v but still the voltage was far below 5v at 1A draw which is not what it should be. what's the point of a 5v regulator if the voltage is 2-3v ?
@ford1548 you're absolutely right. The question is not for you. It's for the person who shouted cheating.
To shut down the ignorant simply put in a known genuine regulator and repeat
let em scream..
"but we are getting heat!" - maybe it is not a regulator, maybe it is a cheap Chinese heater :P
I'm not sure they are fake, I use 78 and 79 series regulators since 1970's and I've learned a lot from my experience: NEVER USE ST parts,
II do not have anymore available the official response from SGS of those years, stating that te internal protection circuit of these devices is different from other similar (Fairchild, TI, Motorola) and mostly the 79's had an overvoltage protection the did not allowed them to start correctly. My ANSWER WAS : NO MORE SGS or ST regulators
I was building audio mixers, and the start -up problems were present also in equipment built by some competitors,:
When mains voltage was very high the negative output was ZERO
Not all 78 regs are the same specification - that's why they have a different data sheet - but they generally give the rated voltage over the data sheet specified conditions. This test is not stressing the part (over voltage or over current) so your argument doesn't really hold any water. I've used hundreds of ST regs over the years in digital equipment (input up to about 10V over output voltage) and not had problems in *normal* use. Either these are faulty parts or fake parts ... If you are saying ST never made a working regulator I, and millions of installations, would beg to differ.
And here we have a prime example of a boomer unwilling to admit that he's wrong 🤣
"It *must* be the company that produces nonworking crap for 40+ years, it's definitely not me who's incompetent here!" /s
Thanks for testing some useless piece of crap :)
Please take the time to completely clean up your bench to make videos. Ron W4BIN
Of course the current will not go up, with only 2V across a FIXED resistor load! ....DAAAAAH!!!!! :) ....OHMs LAW!!!!!!!! :)
Not talking about current and ohms law. Talking about the chip output voltage. It should work with 2 volts input margin, If the output voltage drops (lower output = more margin).
I am referring to your LOAD, with the output voltage going down, the current through the LOAD resistors will also go down.
I'm not sure he understands how the heat sink is supposed to work. 🤣🤣
Sticking the tip of my pinkie finger into a glass of ice water isn't going to cool my body down on a hot summer day. At a minimum, I will need to either cool down my core, cool all of my extremities, or at least expose a significantly large section of my body to coolness in order to cool down my entire body.
It's much the same with this component. Simply touching an extended part of the body (an "extremity", as it were) onto a heat sink isn't enough to efficiently draw that heat away from the body.
You're making that heat travel a ways just to reach the heat sink, instead of allowing it to immediately dissipate.
Just sayin'....
That's a long way of telling us you're retarded but good for you, bud. Keep it up 👍
The maximum expected dissipation in the tests is 3 watts. That isn't much. The "tab" on a TO-220 package is normally pure copper, with very high thermal conductivity. The test setup is plenty good enough. Unless the nerves in your fingers are dead if you can hold a finger tip on the plastic part of the package for several seconds the die is operating far below its over-temperature shutdown point.
The datasheet shows a thermal resistance junction-to-case of 4°C/W for the TO-220 package. As the power dissipated in the device never exceeded (0.5A x 4V) = 2W in all of the tests, it's safe to say that the junction temperature would never be more that 8°C above the case temperature. Since a genuine part has thermal shutdown at a junction temperature of 150°C, it should be pretty clear that being able to put your finger on the case without any problems should guarantee that the junction is nowhere near thermal limiting. The heatsink was a bit of an irrelevance in this test.
@@d614gakadoug9 re: "The test setup is plenty good enough. "
Tell me again this is good with the thermocouple tucked UNDER the TO-220 tab and next to the heatsink?
@@uploadJ
It is plenty good enough for what is expected. If I were doing the experiment I wouldn't even bother with the temperature probe. It _is_ poorly positioned for accurate measurement but it just doesn't matter when a finger can be held on the device. In a "genuine" regulator, overtemperature shut-down is abrupt. It does not cause a gradual drop in output voltage. We saw no such abrupt turn-off. The counterfeit part may behave differently but we have no idea what to expect so we have no idea if temperature is coming into play.
Something significantly missing from the experiments in the video. and far more important than a vague idea of device temperature, is measurement of the actual input voltage at the regulator terminals. He also took the "typical" dropout voltage from the datasheet as "the" dropout voltage. You don't get to do that. You must consider worst-case. Against that, a bipolar regulator that uses an emitter follower as the pass device has a negative temperature coefficient of dropout voltage.
I'd also be checking with an oscilloscope just to make sure there was no oscillation.
back to temperature ...
If I wanted to improve the accuracy of temperature measurement in that setup, I'd leave the body of the device hanging over the end of the heatsink and carefully affix the thermocouple to the free part of the "backplate" near the die area (roughly central under the plastic part). Obviously that would limit the allowable dissipation in the regulator, but the maximum expected dissipation based on the design of the experiment is only three watts.
Low-cost thermal cameras are a good way to do some of this stuff now. I used to make quite extensive use of temperature "crayons." These are solids with tightly controlled melting point. They only tell you if the temperature is above or below the melting point but that can still be very instructive. I did a lot of mains-powered SMPS design - you don't go sticking you fingers on the power semis. Even thermocouple probes can be a problem because of the noise that abounds in switchers.
Sticking with thermocouple probes, I can think of a number of experiments that would require far, far more care and proper design. An example is a comparison of screwing a TO-220 to a heatsink versus using a mounting clip That might even involve careful drilling of probe wells into the plastic. (I'm no fan of the TO-220 because it is hard to mount well; big fan ot TO-247, but it's considerably more expensive).
Counterfeit or Qc rejected?
never buy the lowest price, there is so many dishonest chinese sellers.
You can tell its fake just by the looks of it
The multimeters are also fake meters .. looks like fluke..... Try another one
You 12V critic need to be educated about electronics, ...you did everything right before. NOT THIS TIME, you heat sink is now crap, because the BODY of the reg should be in contact with the heatsink! AS for temp, any "proper" reg in th 78/79 series will operate up to the limit of temp, and if too hit will SHUT DOWN, not voltage droop. Small die pass tranny inside is why the voltage drops at higher currents. ...again, YOU DID EVERYTHING BEFORE!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
The heatsink in this test is plenty good enough. The maximum expected dissipation is only 3 watts. The maximum allowable die temperature for guaranteed operation is 125 °C and the shutdown temperature will be well above that. The thermal resistance from die to ambient only needs to be 40 K/W to keep the die at the allowable max at the full expected power dissipation.
If you have little knowledge of electronics, it is true that some fake regulator transistors are unable to maintain a constant 5v. Most of this type should handle around 1 ampere max.
You must also make sure that the power supply is above a certain volt and does not drop below it.
Normally you should stay at a minimum of 7v if you are going to have 5v out.
This is not a 5v regulator you want in your circuit! Fake !
grow up man, the world uses Kelvins or Centigrades....
-10% line regulation at only 1/3 (500mA) of rated current. They're super fake.
H̶a̶v̶e̶ ̶y̶o̶u̶ ̶c̶h̶e̶c̶k̶e̶d̶ ̶f̶o̶r̶ ̶o̶s̶c̶i̶l̶l̶a̶t̶i̶o̶n̶s̶?̶
Just watched the first video, Test setup seems stable as the genuine parts worked correctly.
I can hang a scope on there but I'm working on something else at the moment.