EEVblog

[@AnotherBrokenToaster]

Anybody else misread the thumbnail? :D

[@WillArtie]

lol! yes - was looking for such a comment. Wouldn't put it past Dave to have deliberately reduced the kerning for such an effect!

[@RubixB0y]

F⌰CKER

[@Schwuuuuup]

OMG, yes... Fonts Matter

[@EEVblog]

Someone on twitter pointed this out, I didn't notice it. I literally googled for "flicker font" and found the "Flickr DEMO" font, it looked good so I used that. So no kerning tweeks. I edit on a bigger screen and juts didn't notice it until it was pointed out. Too funny to change!

[@DrakkarCalethiel]

German speaking folks could read it as Ficker, which also means fucker :D

[@JerryWalker001]

'Flicker Free' from Chinease suppliers just means that they do not charge extra for the flickering.

[@GeneralPose]

Remind me next time i buy lead free solder...

[@HighestRank]

Chineasy does it every time.

[@bostedtap8399]

New South Wales office employees increase in migraine headaches, government cannot see the problem. Thanks for sharing.

[@rocketman221projects]

All of my cheap Chinese LED drivers have a big filter cap on the primary side and they don't have any visible flicker. I'm guessing they left out the filter caps on your driver to increase the power factor.

[@logmeindog]

Probably has a higher switching frequency. IIRC, good ones are in the low megahertz.

[@simontay4851]

Yes, try putting much larger caps on the primary and secondary. 330uF is far too small. should be 1000uF and two of them. Also try removing the blue Y caps.

[@ProdigalPorcupine]

logmeindog - It’s not the switching frequency causing the flicker, it’s the fact these things runs on unsmoothed mains. This causes 100Hz (120Hz in North America etc.) pulsating of the LEDs because the whole circuit rides on the 100Hz waveform. It makes no difference if they use 45KHz, 65KHz or MHz range switching, it’s still modulated with 100Hz.

[@kstxevolution9642]

wouldn't the caps compensate the power factor? or are the lights not considered inductive load for it to happen? or I got it all back to front? legit curious

[@SmithKerona]

@@kstxevolution9642 No the caps wouldn't compensate the power factor because of the extremely distorted current waveform the caps introduce if they are put right after the mains rectifier. It is not your classic power factor issue where the current lags the voltage issue at play here. It is specifically called *distortion power factor* introduced by non-linear loads such as these led drivers.

[@Stefan_Payne]

I'd really like to see an analysis of a decently high end ATX PSU with 80plus Gold certificate and the usual 25mV Ripple or less - and no big coil on the Output ;)

[@AgressiveHouse]

There is a good way to check flickering without any camera because camera would show you only certain frequency range. Even if there is no flickering on camera does not mean there is none. You can connect a speaker to the small solar panel and actually "hear" the flickering up to maximum audible frequency. As it turns out, almost everything flickers (except for the bulbs)

[@C00ltronix]

The PCB material is CEM-1. Inside is paper, but the binder is epoxy, not phenolic, and the outer layer on each side is glassfabric, usually 7628 type. CEM-1 can only be used for PCB without chemical through hole contacts, so it's mainly used for single side PCBs. It's still UL-V0 flame class (hence the red logo, UL-HB material has blue logos), same as FR4. "ZD" is the logo of Jinbao factory from Shandong. China also has a laminate type that uses a CEM-1 layer construction, but with phenolic binder. They call that 22F, it's not an ANSI type and has no UL classification. I guess it's not used in export products. Great video! Thanks!

[@AirCommandRockets]

I remember at our lab at work we spent some time trying to debug an intermittent 38Khz IR sensor problem that would occasionally pick up false triggers. Half a day later we found that it was a bad office flouro light occasionally flickering at exactly the same time as we picked up the false triggers. Replaced the tube and problem went away. I am not sure how that wouldn't have been filtered by the IR sensor, but perhaps as the light flickered it went through a range of frequencies and happen to get past the filter.

[@JWH3]

I'm pretty sure the filters on those things are pretty course.

[@johnfrancisdoe1563]

Air Command Rockets Fluorescent tubes start at a different light frequency, then smear it out to cover the visible spectrum, leaving significant sidebands in the IR and UV ranges.

[@JWH3]

@@johnfrancisdoe1563 it's not about the spectrum of light it's about it's intensity flickering. IR sensors are actually sensative to a fairly broad range of the light spectrum.

[@Schwuuuuup]

Flicker free LEDs = Afterglow in the phosphors of the white LEDs?

[@whitcwa]

The phosphors in white LEDs work by fluorescence not phosphorescence, so they have only nanoseconds of afterglow.

[@nrdesign1991]

Now that's an idea

[@Schwuuuuup]

@Dave Micolichek no i mean desirable chemical afterglow

[@GadgetAddict]

Last year i.upgraded a bunch of "generic" brand LED bulbs with Philips branded bulbs. I was shocked to see the Philips bulbs caused major flickering on most of my cameras 🤦

[@adrianknott9708]

Add 100uf on the primary side after the bridge rectifier and I'm betting the ripple will go away.

[@AnotherBrokenToaster]

That would definitely solve the problem, 100u at 400v is a pretty bulky cap, not that there isn't space for one :)

[@AntonBabiy]

that'll blow the bridge on start up. then you'd have to add a thermistor. then once the ripple is gone it'll output over the ratred capacity since it's on all the time and over heat itself and or the panel so current limiting will need to be lowered..... I've done these mods myself a few times and it's more involved than it seems. especially if there are (assuming) dozens of drivers.

[@johnfrancisdoe1563]

Anton Babiy Maybe the input filtering coils will lower the inrush at the peaks.

[@jeremylister89]

That will destroy the power factor.

[@markusepple6204]

At 5:37 you see the rectifier "D81". On its output, two solder dots without holes is the PCB follow. Drill them and place a 450 volt capacitor there. Before powering it up, use a resistive load and measure the current it delivers with and without the capacitor. Also check if the drive chip gets unexpected hot. Never touch the operating or connected circuit to avoid electrical hazard.

[@reddcube]

First hand experience, Flickering light can give you headaches. Especially if the duty cycle is only 50%.

[@JWH3]

It's not the duty cycle that's important it's the modulation depth. The crappy one he measured here is 100% modulation.

[@todayonthebench]

Doing constant current on the output without secondary side sensing isn't "too hard", and there are off the shelf PWM based chips for it. The main cheap out of this supply is the fact that it has bugger all in mains side bulk capacitance, two small film caps aren't going to do much. Also, the two capacitors are rated 150 volts and are in parallel, thankfully these types of caps are usually derated by the manufacturer by a lot, but still.... And then they are placed facing each other, making the text harder to read, are they deliberately trying to hide this fact? Strange that they didn't apply glue to it, or would that offset the tiny savings on cheap caps....

[@juststeve5542]

A full bridge rectifier (why does that phrase always echo?) and some beefier caps on the output and things would be much better on that cheap-shit driver.

[@teardowndan5364]

The reason why commercial LED drivers have no meaningful input capacitance is because most commercial and industrial installations pay for apparent power and putting input capacitors would screw up the power factor unless you go through the considerable additional expense of adding PFC circuitry. In some places, there are also surcharges for peak or inrush current and in those cases, putting large capacitors directly after a bridge rectifier is an even worse idea when you have hundreds of them: if each light peaks at 1A and you have 500 of 'em, you end up paying for 500A service (the power grid needs to be sized to deliver those 500 x 1A peaks to your location) just for lighting instead of the ~100Arms you are actually using.

[@johnpossum556]

@@teardowndan5364 I would think you can get around that by using capacitance multiplication schemes. Matter of fact by now I am surprised they are not as common as switch mode power supplies are to outdated wall warts.

[@teardowndan5364]

@@johnpossum556 If what you mean by "capacitor multiplication" is boosting the capacitor's current through a transistor to emulate a bigger one, this does not help here since the output will dip to whatever the input is when the input is less than the output. That circuit is effectively an open-loop voltage regulator with all the inefficiency that comes with one.

[@mustafayasiraydin]

@@teardowndan5364 so actually AC needs to die if possible. We may need to switch DC grid.

[@Golem386]

The second winding on the primary side is not only for feedback in case of primary side feedback controllers but also for supplying low-voltage power to the primary side switching IC.

[@Stoneman06660]

Helpful info, Dave. Looking at replacing the single globe in my garage come lab with some LED lighting (of some sort) and stuff like this means I'm not going to make as many rookie mistakes during acquisition of parts. Building an LED driver looks like a reasonably approachable project, however. Not sure I'm keen to play with mains, but yeah, interesting stuff.

[@samchan5251]

Why can't you increase the output capacitance to reduce the ripple? Because larger cap tends to have hight serial inductance and resistance? or just too expensive.

[@adamkekow6558]

Hi! I wonder, why this video is in 60fps. I know that eariel videos are in 50fps for a reason. It elminates 100hz flicker.

[@onradioactivewaves]

14:00 how is it AC coupled, do you have additional circuitry on the probe? The scope says ch1 is DC coupled.

[@Tupeutla]

37-44V with a 50V capacitor ? wow ! waranty is 3 months ?

[@stephenbell9257]

A fair amount of that 100Hz ripple current is going to go through the output capacitor as well, further reducing the capacitor life

[@superdau]

Even worse, it looks like the primary side one is a 50V type as well.

[@EEVblog]

@@superdau The primary side electro is not filtering the 100Hz output, it's the power supply for the chip.

[@heinrichhein2605]

the problem is the primary capacitor, and it is a Flyback that is why you only have one diode. And from the PCB the diode seams to get quite hot

[@dhpbear2]

Dave, you mention the first circuit generating "100% Ripple", Would that be a peak-to-peak value of ~38 volts (?) (17:36)

[@michaelpiotrowicz6100]

Some information on the practicalities of identifying and swapping out drivers in another video would be awesome

[@gandalf87264]

Is that a Negative Feedback transformer? Suggestion: Why don't you build your own LED light driver and show us how you would do it? That would be a treat for me.

[@ricardoelectronicsrepair]

transparent case for driver box would be nice, can see if circuits quality on site before buy😀

[@ArlenMoulton2]

I have a RECOM RACT25-500 driver, 25w (50v@500mA) which would do the same job as this, I was VERY impressed with the build quality of it!

[@samuelfielder]

Sorry, I don't understand Dave's account at all. Any ripple on the secondary side of the transformer would be at the switching frequency, not at 100 Hz. A fatter cap on the secondary side won't fix that 100 Hz ripple. It must be caused by lack of smoothing on the primary side, surely?

[@threeMetreJim]

Looks like a very small smoothing capacitor on the input. Do they do that to improve the power factor?

[@vant4888]

Also it (flicker) increases perceived brightness because peaks of that ripple are very bright which in turn allows to decrease PSU power and "do it for cheap" even more.

[@Graham_Langley]

Yes.

[@steverobbins4872]

You would need a two-stage converter to maintain low output ripple current and high input power factor. If you only have one stage then it's a trade-off. For example, the cheap one is single-stage. For high power factor they must use a relatively low input filter cap, and the supply that runs the chip drops to zero every line cycle. You could eliminate the output ripple by increasing the input cap, so the converter runs continuously, but then you would have really low power factor. Increasing the output cap won't do any good because it is filtering the much higher switching freq, not the line freq.

[@steverobbins4872]

Actually, now that I take a closer look at the schematic, they do have an aux supply that keeps the chip powered continuously, but I think the output voltage from the bridge rectifier still drops to zero twice per line cycle, and of course the output conversion must stop when the voltage gets too low.

[@EEVblog]

2nd videos coming. You can improve the output by increasing the output cap, but not by a huge amount. Input cap does the trick of course, but PF goes to crap as expected. You can't squeeze blood out of the wrong topology.

[@johnfrancisdoe1563]

EEVblog Would having multiple primary side taps with active FET rectification allow lower loss operation with 10ms of energy stored in a primary side 30V cap charged and discharged from one of the lower primary taps? Would require a more complex switching chip to shape the input load near a 100% PF while keeping the output current near DC. Ideally the voltage drop in the active rectification switcher should be less than the usual 1.4V plus.

[@volodumurkalunyak4651]

@@johnfrancisdoe1563 there is a simpler solution to output ripple: lineal (switching also works) post-regulator on the output.

[@skulisveinsson]

The driver case look the same as Sonoff only longer.

[@jeremylister89]

The second driver will be less efficient. The ic and transistor on the secondary is just an active linear filter which means power loss. True cure is a two stage switching driver such as PFC flyback to CV output with CC buck tagged on the output.

[@Direkin]

That flickering would drive me nuts. Some people wouldn't see it in person, but I do.

[@JohnBurgessMusic]

I've been dealing with this exact issue. I bought some lights with inbuilt drivers. The output had an average DC of 780mA, but 350mA RMS of ripple on top. It had the exact same 100 Hz quasi sinusoidal wave on the output. The primary side only has 330nF of input capacitance. I believe it is to increase the power factor. This is a 50W driver too btw. The efficiency was ~97% and the power PF was 0.98 at full load, so it does work. However the lights were way too bright for me and the flicker was noticeable to my eyes. Luckily the driver chip has an analogue dimming pin that was unused by the manufacturer, so I just modded them. There's still plenty of light running them around 15W, and the ripple on the output is now around 35-50mA RMS. No more flicker, but the efficiency and PF goes down. The weird thing is that these look like quality drivers. Lots of attention to filtering EMI, good quality PCB and soldering. I guess the banner specs that sell are high power, high efficiency and high power factor. The output ripple is considered "fine print" territory.

[@donreid358]

If your lights flicker at 100 Hz then the camera anti-flicker will try to set the exposure to a multiple of 1/100sec. That can only work when the exposure is longer than 1/100 sec. If the light level is too high or the aperture is too big then it cannot avoid the flicker.

[@uwepolifka4583]

I don´t understand where the flicker comes from. The chip uses a high switching frequency so where comes the 50/100 Hz flicker at the secondary from. Only the primary voltage could modulate the chip somehow.

[@johnfrancisdoe1563]

Uwe Polifka Bingo!

[@wolkman2157]

Uwe Polifka To achieve a high power factor there is little to no buffering on the primary side and in this case also insufficient buffering on the output. The out put consist of both 100Hz and to a lesser degree the switching frequency (as Dave shows when he zooms into the current measurement). A more expensive dual stage driver will have practically no 100Hz ripple at the output.

[@clazy8]

Dave, did you have a choice of lighting supplier, or did the NSW government select Sunled? I can't help but wonder if the company exists purely to take advantage of the ESS scheme. Sunled's website suggests as much, and the bargain-basement electronics would make sense for a company that is less in the lighting business than in the making-money-off-the-government business.

[@TheVirindi]

That's what I was thinking. There was probably some rule about buying from a local company, designed so they could skirt it by slapping their name on someone else's product and taking a cut for the privilege. But that way, others could not compete.

[@EEVblog]

Sunled literally came knocking on my door, demo panel in hand. There are many other suppliers.

[@timg375]

Could you put together some cheap current regulation and splice it in after the main driver box?

[@NICK-uy3nl]

Realistically, what did you expect from FREE government giveaways ?

[@johncoops6897]

It is NOT paid for by the Government. The supplier "trades" deemed energy savings (over a ten year period) for carbon offset "certificates". One certificate per tonne of CO2 abated. The certificates are sold on the open market to energy retailers.

[@reburne2012]

Primary side Cap is responsible for the 100 Hz ripple, ..Fix that and secondary side will only have switching frequency to filter....

[@volodumurkalunyak4651]

Fix can't be done without a active PFC with sophisticated inrush current limit. Thank the government for power factor regulation and power company for charging customer money based on peak power. What can be done and can be done easily: use 3 phaze power and 3 of this drivers (on different phazes) with outputs parallelled (connected obviously to 3 panels)

[@derstreber2]

@EEVblog 13:06 I know"concentration only for you" is a bad translation of something, but did this mean that the schematics are only meant for research purposes but somebody decided to copy it and put it on a board anyway?

[@beachsidetech]

And Dave they are a common source of RFI, affecting phones, radios and TV's due to poor filtering.

[@Very_Dark_Engineer]

Yes, at nowadays that's a big problem for all kind of radios.

[@superdau]

The primary side electrolytic is a joke, right? Is that 4.7µF and 50V?! Its capacity is almost pointless for 24W. And at that voltage rating will make a nice pop at some point.

[@EEVblog]

Yeah, pretty piss weak. The other one only has the 4.7uF input cap, but of course they do the output right.

[@pichacker]

@@EEVblog So an approx 44V output with 51V unloaded and a 50V cap. Talk about pushing to the limits.....

[@Blitterbug]

Brilliant. Thoroughly enjoyable, Dave.

[@Graham_Langley]

If a driver or lamp says it's mains dimmable it will flicker at the double the mains frequency, but it of course it may still flicker if it's not stated as being non-dimmable. However these drivers don't appear to say either way.

[@DrRChandra]

I don't get it. Why is it on PSU inputs there is usually a full-wave bridge, but on output sides, there is usually only a single diode? Wouldn't the ripple be a lot better (less) if it were a full wave on the output too?

[@home-space]

You should contact the government organisation who is supplying them perhaps with a link to this video, so they can understand the problem.

[@EEVblog]

Oh, you think they'll care, how cute :->

[@IlBiggo]

@@EEVblog Since when is this a reason for NOT ranting about something? SOMEBODY IS WRONG! I HAVE TO TELL THEM! .-D

[@freda5344]

@@IlBiggo you got the url of this video, so get on your own horse

[@IlBiggo]

@@freda5344 I tried, but photos.app.goo.gl/EVy4gXQMWSCRm5st7

[@edcooper2396]

I guess to meet the requirements for government funding they had to provide a good power factor. This is the cheap way of getting an acceptable power factor. In Europe you can’t beat eldoled drivers.

[@wolkman2157]

That may be true for multiple channel dmx drivers intended for stage and entertainment lighting, but for general lighting Philips by far beats Eldoled.

[@lambdaslab6485]

it appears the problem is missing primary side capacitance as indicated bey the 100ħz riple and nor the missing secondary side capacitance? also can you explain how the secondary side current regulation with the extra transistor works?

[@God-CDXX]

he did

[@vihai]

Yeah! I expected a much higher frequency on the secondary. Seeing a 100 Hz ripple was unexpected!

[@EEVblog]

it's not an extra transistor on the secondary, it's specific LED driver constant current driver chip.

[@vihai]

The main transformer should IMHO however run at a higher frequency than mains, it seems too small to run 24 W at 50/100 hz

[@God-CDXX]

@@vihai it is running at 50 KHZ the 100 HZ ripple is from the fact it doesn't have a primary side bulk filter cap

[@stusue9733]

How many houses will they burn down this time?

[@seansvid]

THANKS for this great video. This could explain why so many LED fixtures trigger my migraines.

[@nope1906]

My dollar store GE brand 15W AC (75 tungsten equivalent.) LED has like triple the components. Wow.

[@billflake8015]

i think if you added a cap to the bridge ps it would reduce the ripple in the secondary...

[@stanislavhajek1085]

I don't know, but wouldn't be better to increase the input filtration capacity, there isn't almost any, and when you'll not have 50Hz ripple on the primary/input side, where from would come the ripple to the secondary/output side.

[@GfastGao]

Hi Davi, Your early subscriber here. It's all ready lighting better then before. But I still feeling a video about 20min long is still a little bit "too heavey" for a hobby level guy like me. I saw many compter programming tutorials under 15min. It's always my better choice when your vids and them sitting in my playing list together. You have to admit, most of the time, people I planed to learn from you tube vids, but they actually do not have much time for it. I shorter vid will help people like me a alot.

[@jskratnyarlathotep8411]

12:17 how it is bridge rectifier, if there are only three diodes? what kind of bridge rectifier it is?

[@jskratnyarlathotep8411]

it looks like it is just 3 parallel diodes. Is it legal to call that scheme "bridge" rectifier?

[@rfmonkey4942]

Hi there Dave, it the business with the spectrum analyser that is of most concern with these products ! and you did not check

[@Petertronic]

I hope the government department responsible sees this video!

[@simontay4851]

Dave should email them and link to this video.

[@whuzzzup]

He should inform a newspaper.

[@donnierobertson3088]

Great job again like always

[@redsquirrelftw]

What causes the current to ripple way more than the voltage? Even though it's constant current I would not expect the current to drop so much for each cycle if the voltage is staying relatively high.

[@xmcx7021]

The proper way to fix these drivers is to hook them up backwards. Once the Magic Smoke is out replace them with higher-quality drivers.

[@phinok.m.628]

But if the flickering is mainly 100 Hz, wouldn't a nice and big capacitor after the rectifying diodes fix that? I mean, sure you have a little ripple from the high frequency switching, but to me that doesn't seem to be the problem. I figure if you had clean DC mains on the input the flicker would be pretty much gone, so where's the problem?

[@wolkman2157]

Phino K.M. That will work, but the power factor will drop to 0.5 or even lower and you might need some inrush limiting. For most consumers the lower power factor will actually not really be a problem, but manufacturers are required to have a power factor above 0.7/0.9 depending on the power rating and the region the product is intended for.

[@phinok.m.628]

@@wolkman2157 That's true, though I've seen many switch mode power supplies without PFC which do more or less the same. I guess it'll only become a problem once power companies start charging apparent power to normal consumers. But if you want a better power factor you can always add an inductor in series with the input or whatever. I was just a little confused as to why Dave was constantly talking about the secondary side and acting as if it weren't worth it to fix. I mean it's not like a simple electrolytic capacitor for the input would be so terribly expensive.

[@ibycus314]

Ive often wondered with these types of things if you’d be better off with on big beefy supply for an entire run, or the individual drivers on each panel...

[@Rickbearcat]

Can you provide feedback to the company of manufacture to help improve the design? Or would they just reject your offering as out of hand for the cost restraints that they are under?

[@JWH3]

It is beyond clear from the design they have no interest in selling a quality product, just the cheapest possible one that functions.

[@EEVblog]

Feedback and link has been given.

[@matarra-fy7ix]

you can see the flickering using the high speed mode or slomo of you cellphone.

[@DjResR]

MeanWell makes some LED drivers where you can trim output voltage and current. Fixed industrial floodlight with one gutted from another fixture with faulty LED COB. _

[@LEO-xo9cz]

I would love to see the internals of those light panels. Are they edge lit?

[@BimmerFordDude]

My Costco flush mount LED panels don’t even have a LED driver! They flicker like crazy!!! Any ideas how to fix?

[@nicolaedumitrache9246]

Increase the primary side filter cap (not the secondary one).

[@n3r0z3r0]

Maybe add some big cap after main ac rectifier ?

[@capnthepeafarmer]

Do you have a good recommendation for a flicker free driver or schematic to build your own? If you made one i would buy a few because my LED lights wreak havoc with my high speed cameras.

[@pchris6662]

Thing is, a lot of people are actually sensitive to that flicker and don't realize it. It can cause headaches and even migraines. Just because you can set your cameras to ignore the flicker doesn't mean it's a good thing to live with.

[@EEVblog]

Indeed.

[@reddragon27284]

Put a big filter cap on the primary. I think that would help.

[@EEVblog]

See part 2.

[@tsusec]

Can someone explain this to me... Isnt secondary side still AC referenced to mains earth via those capacitors between primary and secondary side of transformer? and is this/that small current harmfull to the scope?

[@mrjohhhnnnyyy5797]

Those caps are such value that the most you can get is a tingle, under "ideal" circumstances. The scope is fine too.

[@PetreRodan]

For real now, the video looks infinitely better without those lights.

[@saarike]

Nice investigation!

[@BogdanSerban]

I never knew that the output current waveform is 100Hz sinusoidal in switchmode power supplies

[@wolkman2157]

That is not always the case, it is due to it it being a low cost single stage high power factor driver. A more expensive dual stage driver will have little to no 100Hz ripple.

[@Brian_Of_Melbourne]

Dave, I don't think you said in the video, how old are thes PSUs? Have they flickered from day zero or have they gone downhill over time? Do you have any 'warranty' on their suitability (fitness for purpose)?

[@lauril1377]

Flickering on camera depends on the shutter speed. If you lower the shutter speed to 1/25s you won't see flicker as it averages out. But increasing shutter speed to 1/100s or faster "draws the flickering out" better. My trick is to use my phone's camera in "PRO" mode where I can manually change the shutter speed, increase it and see if lights flicker or not. It has nothing to do with quality of camera or sensor - you just have to have fast shutter. With automatic shutter it happens automatically if you shine the light straight to camera, it starts to increase the shutter speed to lower the effective brightness. 50Hz and 60Hz modes are a bit different story. They are there to avoid a bit different type of flicker which occures because 60 doesn't devide by 50 and vice versa and this gives temporal flickering. Also this effect is reduced when shutter speed is very low (low light!) and increased when shutter speed is fast (bright).

[@electronicsNmore]

I had the same thing happen in the past. Annoying but correctable. :-)

[@pahom2]

You can measure flickery of a lamp if you connect a capacitor and a small solar panel element in sequence to headphones or a mic input of a phone. The noise is very distinct in those lamps that flicker. See the video 9h4_W25qP64

[@drruncmd]

I have seen some interesting text on PCB's. Mainly power boards. One I remember recently had on the underside of the board, labelled "HOT" for primary side and "GOLD" for the secondary side. Hilarious!! May try and dig it out and do a video on it. Was very funny indeed. Great video Dave, I'm a sucker for power supply circuit boards and their theory of operation!!!!!

[@RaithUK]

Lol amusing font indeed. Great video Dave, shame you didn't hack it.. always fun when ya do that. Keep it up.

[@lincolnmicrophonellc]

Art of Electronics spotted. Do you also have the workbook... I need to poke at it someday :D

[@EEVblog]

No I don't. Kinda been a long time since I was a student...

[@lincolnmicrophonellc]

@@EEVblog never stop learning

[@juweinert]

That's a real shame! Even if you don't film and not really see the flicker, it's not really good. As you said, it's bad for the eyes and more so for your head. Observe if you get headaches more often now...

[@samuelschwager]

The contract went to the lowest bidder with the best government connections.

[@devttyUSB0]

Could calculate a rough estimate for that tube light replacement plan in NSW. Your video overlay stating you're actually paying through taxes triggered this idea. Say they find a cheap seller of those panels, there's an average per company, removing and installing is a quick job, shouldn't take hours and hours of work? I don't think it should be that expensive per company, and it would save tons of electricity switching to LED. Perhaps more of an Engineering Explained thing, this sort of calculations, but you also have a whiteboard!! ;)

[@amazac2078]

hayia mate! i think you miss the point why its flickering! a new study showed that steady LED lights are not good for eyes and i suppose the supplier is aware of that fact so they aiming to leave it to fliker for eyes health. nothing to do with saving money nor lower quality issues.

[@tommyboy9998]

NVM the flicker - how's the RFi????

[@hallcrash]

You could add a smoothing board between the driver and the LED.

[@simsys_outdoor]

Also this "kind-o-driver" makes a lot of noisy and some ugly sound frequency. Even is the panel is off (on standby). I do not mention about EMC.

[@butchvandyk1051]

Just a question what happens when the led gives you a "zebra stripes" on white surfaces or any tables ext.... hope you can help

[@starpuss]

Less then $2 Each they could have made them better... But at the same time... IF they have 100,000 on an order... that ad's up!

[@simontay4851]

Well just charge $2 extra to the customer, they're not going to notice.

[@dhpbear2]

13:09 - "All your LED panel drivers are belong to us." :)

[@cody5495]

Hey dave, have you seen the new CREE PoE, RJ45 Light fixtures?

[@Tim3ru]

Could you convert to a half-wave rectifier so that it would flicker at 50[Hz]? :))

[@neteagle2k9]

Any news from the custom µSupply LCD ?

[@EEVblog]

Listen to this week's AmpHour that will be released in a next few days.

[@twocvbloke]

"Eco-friendly Upgrades", as dictated by the government accountants of course, with their "That'll do" attitudes... :P

[@MollyWi]

"Concentration only your you!"

[@redsquirrelftw]

So random lol. Actually I get spam on my forums like this. It's just these completely random weird sentences.

[@GeeWillikersMan]

It's from the manufacture's about page: 'With "Concentration only for you" as the purpose of enterprise,use professional technology,focused attitude and dedicated spirit, we provide world class LED green products as our own highest pursue. '

[@MollyWi]

"All your base are belong to us"

[@davidmcclare443]

hi would a LC filter remove the ripple?

[@TorgeirFredriksen]

Great video! How about making a video about how to design a proper LED driver? I