EEVblog
Вставка
- Опубліковано 10 жов 2019
- Part 2 of the LED Flicker investigation.
Modding the driver for input and output capacitance and it's effect on power factor correction and thermal performance.
High voltage differential probe: www.eevblog.com/product/hvp70/
Part 1: • EEVblog #1252 - LED Pa...
Power Factor Correction: • EEVblog #273 - Power F...
Forum: www.eevblog.com/forum/blog/ee...
#LED #Flicker
Bitcoin Donations: 38y7DE8HEHNj8fGDtUr4PkCn9nWxiorvvy
Litecoin: ML7oQokTwB38bgzzjLDbRV97HKAHuwRfHA
Ethereum: 0x11AceA38DCA9DbFfB4F35f3F746af65F9dED28ce
EEVblog Main Web Site: www.eevblog.com
The 2nd EEVblog Channel: / eevblog2
Support the EEVblog through Patreon!
/ eevblog
AliExpress Affiliate: s.click.aliexpress.com/e/c2LRpe8g
Buy anything through that link and Dave gets a commission at no cost to you.
Stuff I recommend:
kit.com/EEVblog/
Donate With Bitcoin & Other Crypto Currencies!
www.eevblog.com/crypto-currency/
T-Shirts: teespring.com/stores/eevblog
Likecoin - Coins for Likes: likecoin.pro/@eevblog/dil9/hcq3 - Наука та технологія
pls more power electronics/PSU Related stuff!
At first glance I thought the thumbnail screen said something else.... 🤣
my exact thoughts lmao
Really should change the font here or something!
That Dave is a crazy flicker, isn’t he.
I woke up this morning with bleary eyes and did a comedic double take when I saw it.
We all do, we all do...
Thank you, this gave me a whole new understanding about led driver circuits.
Thanks Dave! Brilliant!!!!
I love your FLICKER series!
i got stuck in building a circuit for a week . finally worked and im so relaxed....
Excellent insight and hands on of LED drivers, and why capacitance is not the answer in efficiency.
Electronics thicko here.
Thanks for sharing and best regards from the.
good one Dave thanks
Thank you @eevblog excellent video. I have some led panels for my office that have the questionable driver, I think it could be a fun project to make my own and pot it
Wow, I'm impressed that your Aussie politicians can even appreciate the subtleties of power factor! It would be a rare day when any US politician could spell his name the same way twice on any given day!
I am also impressed. Here in Germany, the power factor is only prescribed for powerful switching power supplies. As far as I know, the rule only applies to > 150W nominal power.
All mobile phone chargers and similar wall warts have such poor power factors here. Not infrequently, only capacitive droppers are used. I don't need to say anything about their power factor.
Nice DJ!
You should have connected the primary filter cap /behind/ the 2 filter chokes, that would have smoothened out the current peak a good bit, and so improve the PF!
Good to know!
Love the reference in the name. I know what era you grew up in.
Love your vids bro
This one was a super interesting video. I do not understand how you calculate all those results, because I do not know electronics, but the conclusions are very useful. Thanks.
I downloaded that font just because it looks like it says a whole different thing.
Demonetized!
Your main B+ should have went up considerably with the primary filtering you added. That would raise the VCC to the switching IC, you probably need to change those dropping resistors to compensate.
The IC powers itself from the feedback winding, not the 660k resistance.
The problem with big input capacitors is the nasty inrush current seen from the line at power up. That was a nightmare for people that designed electrical protection equipment like ups or automatic voltage regulator based in SCR's or IGBTs, easily a small room with older pc's could produce hundreds of amps.
If you have a small solar panel lying around (e.g. from a calculator or solar led lamp), those are great light sensors for oscilloscopes!
"yeah okay... how do we calculate that?"
Now I have to clean all the beer I just aspirated off my monitor! :) Love ya work mate!
The flicker from that cheapo LED driver is no worse than the good old flourescent strip light that we have been happily using for the last seventy years or so. Admittedly there were always one or two people in the office that complained that the flicker was giving them headaches or other ailments but when the lamp drivers were upgraded to electronic versions switching at 50kHz or so with negligible output ripple, they still complained.
I wonder if there was some design reuse happening here. The primary side filtering is suspiciously similar to what many triac dimmable led drivers use. It seems logical that they might just modify an existing higher end design to just be as cheap as possible. It would simplify BOM inventory and be much quicker to market with half the design already done. Both the board and case being designed for screw terminals yet having wires directly soldered on lend credence to this.
Please make a video on Power Factor Correction in general.
hopefully we go for dc mains soon :P
Adoro os seus vídeos ,não perco um ,e uma pena não ter legendas pra o português Brasil ! ;(
9:15 - are you forgetting someone did? That shed full of salvaged parts!?
I would of loved this when I was in college for education. Could assign groups to experiment and see who could come up with the best PF and lowest ripple.
@@Okurka. lol I took Engineering classes and I'm far more comfortable with numbers than grammar =P
"would have". It's due to the contraction where it is written or spoken as would've, so people think it's 'would of' but it's not.
@@sw6188 Thanks =D
@@sw6188 I would have course like to concur.
missing link to part one in video description? I mean, it's not hard to find :) but to have to do it manually?
There is an un-soldered leg on the input side of that six pin chip.
Love these videos. I wonder why there aren't (easily found) room-level solutions. i.e. if you have multiple LED panels in a room, why not have one main AC/DC converter with good PF, efficiency, and EMI, rather than 10x mediocre AC/DC supplies. Then each panel can have its own DC CC driver.
It's actually quite common, called 12V LED strips driven from a 12V PSU.
@@NiHaoMike64 They aren't really good solutions for large scale lighting though. Too much power is wasted in the current limiting resistor in each section of a 12V LED strip. For example, one that I have lying around has a 150R resistor per 3 LEDs. At 12V, 25% of the power is wasted in the resistor.
@@JohnBurgessMusic Same with your case of have dc-dc per panel and single ac-dc. you would either waste power in wires between them or waste money on thickass cables. Anyway it is not very efficient and a lot less redundant. I imagine that one faulty dc-dc could at least trigger protections (potentially cause damage) on parent ac-dc and render all the panels useless until the fault is isolated from the rest of the setup.
@@Mr.Leeroy it's more about simplifying the install. You cauld have a 240V current source and connect as many panels in series as required up to a limit of 240V but then you have to run all panels at once. The converter per panel is a compremise on functionality and ease of install for electricians that are not neccessarily that educated in electronics and can be quite thick like the lovely guy i use. I had to help him instll my solar batteries as it was his first time with that kit.
I watch these videos yet I'm able to follow very little of the technical details
17:18 this IS the place to go into details.
What's the RFI look like from one of those cheapies? We have a problem in the states where cheapo LED lighting is so noisy, that it blanks out radio reception.
Does the leading power factor of a capacitor help to balance out the more prevalent lagging effect of all the inductive loads on the network (including the network cables themselves)?
The lousy power supply has a massive ripple, so generating an output that is solid with the same voltage range would require more current and an up voltage inverter? Which would overheat the components of the supply? Does the supply have like panel number limits?
Nice Video, I don't know if you noticed it already but jlcpcb now offers an assembly service for a very very low price. Can you make a video about that and what you think about the quality.
Maybe put a filter cap after the input chokes? And lower the capacitance? It's overkill 47uf...
It's all very nice to have flicker-free light, but there is one small point... The old standard inductive-ballast troughers (and for that matter any AC-powered gas-discharge lamp) will all have 100Hz flicker, so the green-subsidy-gobbling suppliers have effectively replaced like-for-like (electronic ballast upgrades notwithstanding). The move to smart meters will result in billing based on apparent power, so it's in the customer's best interests to maximise the PF.
As someone who is easily bothered by flickering I never found fluorescent lights mounted in pairs (out of phase) to be particularly problematic... guess they have a good amount of decay time, similar to CCFLs. That's why PWM backlight dimming only became an issue with the transition to LEDs, which often seem to respond much quicker.
I thought that board looked a bit toasted around that diode...
On the cheap circuit, simply add three diodes to the output to make a full-wave bridge rectifier.. It'll double the output current capability and improve the ripple :)
No. This is a Flyback topology, it relies on having one (exactly 1) output diode (per output rail). Ripple is there from having essentially no input filtering
I am still trying to figure out why any excess capacitance is a problem, just about every home and industry is inductive. In Florida the power companies have capacitor banks (at the 13.5 kV level that they can switch in and out via 900 MHz links, all over the place.
Actually Dave the new smart metres can measure power factor, and according to bigClive, power companies can now charge you for that bad power factor you return to the grid!
Hey Dave, what's the best way to learn DSP as a newbie?
Any news from the custom usupply LCD ?
considering the power draw is pretty much constant
wouldn't a *linear* power supply make sense?
next Dave project...an adjustable voltage, ripple-free LED power supply pcb project.
Wouldn't a full wave rectified secondary on the turd help with ripple ?
Ive resubscribed after a 3ish year break. I dont know if its because youre rambling less or i understand more. maybe both? Keep it up
Could you run the whole string of lights on DC ? Would the electrical code care ?
Гений
ну так
Yes, but is it full wave rectified?
50 to 60 hz ac input, rectified, no filter capacitor, will be reflected to the output.
I liked this turd polishing video. More turd polishing please! :)
Without heat sink OR air vents on the secondary side, I give it 6 months before it goes POOF!
👍
I remember when I made my bare minimum shitty led driver from an old TV power supply and a 150watt 230/24v/8a transformer. The ripple was 200%. Straight from the 230 mains haha. It "worked" but it was awfully regarding flickering lol (no sh*t!) It was just for testing no longer than a few minutes. It where like 200 leds drawing like almost 2 amps from the mains
Glancing at the thumbnail I was sure it said 'fucker 2' 😂
polished turds sell very well...may start a crowd funding project for a turd polishing device?
Would it be worth getting a single large properly smoothed DC supply and running all the LED lights from that instead of having a separate convertor for each light?
that would require an install by install design as no one can predict how many lights each location will have. Eletricians can be surprisingly thick and are just builders that have read the electrical instalation manual as in how to physically install stuff. Few understand the underlying principles so any install have to be simple enough.
No It would not. As it would be a low voltage installation and so covered by the LVD the installation could be carried out by the user. Anyway that was not really my point. My point was from a technical (Not Legal) perspective, would it be better and more cost effective to have the lights installed using a single supply.A suitably rated converter needs to be selected anyway so it would be easy to have supplies rated for power, which they already are. Eg a 500W supply could be used to feed 10x 50W lights. The number of lights for a particular installation is not an issue as a suitably rated supply could be selected. I am not suggesting that a bespoke supply be designed for each installation and that is certainly not required. There are literally thousands of existing supplies which could be used. Understanding of the underlying principles is not needed.
@@JerryWalker001 And i answered the tecnical aspect not the legal one. You would have to put them in series. This means that you need a driver for "x" amount of panels. You could in theory have a 240V current driver that would allow up to 240V worth of panels in series. You can't put the panels in parallel because they are LED's. it also means you have to turn the whole lot on or off and there can be an energy savings case for being able to independantly switch the panels on and not the lot. The simplest trade off is one driver per panel although one driver of better quality per 2 panels would not go amiss.
Hmm, I am not sure if you are intentionally ignoring my point or you just don't understand it. Dave's video was about the inner workings of LED drivers and my comment was SPECIFICALLY about the technical advantage (or otherwise) of driving them all from a common source. Powering multiple LED's from a single supply is not difficult and makes no difference to my point. Also are you saying that in office light installations there is a switch for each individual light? That is NOT how lights are normally installed, Users would not want 20 switches for the lights although there is no reason this could not be done for a single source driver. It would be identical to any other approach so I have no idea why you think they would all NEED to be on a single switch. Normally they are installed in 'banks'. They would NOT need to be installed in series, there are many ways to create very simple single source multiple drive supplies. Technically it is NOT difficult. However my point was NOT about that, it was about the technical advantage of such approach. Finally, YOU are the one who mentioned the need for particular understanding from installers and this is not the case any more than any other system or other LED driver approaches. My company installs such centralised systems and they work out far more cost effective and reliable than distributed modules. A similar example would be the advantage (or otherwise) of single (or multiple) invertors for solar panels compared to micro invertors fitted to each panel.
@@JerryWalker001 Solar systems are large and expensive (£11'000 for mine) and designed to a degree for the building. i went through much negotiating with my installers as the inverter had to be matched to the panels and with 15 panels in series I had nearly 500V which the 4KW inverter I already had could not take even though it was 4KW of panels so i upgraded to an 8KW inverter just to get increased voltage input, it required some consideration. I am not ignoring your point you are missing mine. The office i work in has 3 banks of lights for 10 people. Yes we could have 3 series strings of lights. You cannot connect the LED panels in parallel, they have to be connected in series. I don't know if it is practical to design a supply that can cope with multiple voltage ranges. It also means that you need several power supply modules available to the installers as they choose the one required for the number of panels. These supplies were clearly designed down to a price so a single unit capable of such a range of output might not be practical in the cost constraints. Remember that efficiency and power factor must both be high. Having a bigger transformer than required will not be good financially as it's the most expensive part. i don't know what the install rules are. put it this way, i am sure they considered all angles during design given how costed down the install is. Yes it would make sense to at least do panels in pairs or say 3 or 4 but it may mean the installs have to be done in blocks of however many panels the supplies work with.
why use a dedicated constant current led driver specialty chip if basically a lowpass filter a say 78xx series regulator and a couple of pass transistors should do just fine too i guess?
_Of course_ you could use something LM317-ish as a constant current source on the secondary side... but efficiency would be pretty shit once there is a lot of voltage to get rid of.
LED lighting is such a big market that rolling custom chips is easily worth it if it promises lowering BOM cost, complexity and development effort.
why not have a full wave or bridge rectifier on the output side? wouldn't that make it easier to filter?
This will not work with the Flyback topology. It's all about how the energy being transferred from the primary to the secondary side. Wikipedia has a pretty good explanation of Flyback converter operation. en.wikipedia.org/wiki/Flyback_converter
Are you just joking with the "diodee" or is it the same as "cache"/"cayche"?
The video thumbnail title looks like a cuss word
1:25 nope i would recommend taking it to bits and then toss the mortal remains of it in the bin... and keep the good parts of them (y-class caps and others)
Why does adding input capacitance effect the power factor? What IS power factor anyway?
Oh, there's a name I haven't seen in a while. Anyway, instead of reinventing the wheel I'll just recommend starting with the Wikipedia article:
en.wikipedia.org/wiki/Power_factor
Adding input capacitance (or even input inductance) effects the phase angle between the voltage and the current. If you have a pure resistive input impedance then the phase between the voltage and the current is zero and the ratio of real power to apparent power where apparent power is measured voltage times measured current (VoltAmpere)and real power is measured voltage times current in phase with the voltage (Watt)is 1, i.e. the power factor is 1. As the voltage vector is common to both calculations it essentially cancels out and the calculation can be expressed as current vector (I) times the cosine of the phase angle (Resistive current) divided by the current vector (Apparent current). Again the current vector is common so power factor is cosine of the current phase angle.
If the input impedance is such that the phase angle is 30 degrees then power factor (PF) is cosine (30 degrees) = 0.866.
So why does this matter? Firstly your electricity meter measures the current vector so you have a higher current demand then if you had a PF of 1 i.e. you are paying more then you need to. Secondary reason to have a good PF is that your wiring and switch gear does not need rating for the higher current that would be present with low PF.
Also with non-linear circuits ( in this case the diode bridge) you get harmonics from the current waveform distortion, particularly in high power (kVA circuits) this could lead to overheating in the supply wiring ( look up skin effect and litz)
Which is the model of the termal camera?
Flir E8
@@EEVblog thanks, always usefull, great video
@DiodeGoneWild dodgy...
Its worth noting that the phrase: "POWER FACTOR" can be rearranged to spell: "WET PROOF ARC". Now I'm not much on conspiracy theories, but shouldn't this be looked into? 🤗
4:41 - I have this little ticklish in my mind about the power factor discussion here. PF (as of my understanding of Wikipedia definition) is a ratio of apparent average power in .vs. real power consumed. How can this device return some power to the main if its got a bridge rectifier at the front end? No power can go back to the main if diodes are feeding current in only one direction. Could it have to do with the presence of C1 and L3, at least partially? By the way those two parts are absent from the schematic. So can anyone explain to me why the power factor is not plainly 1 ?
A low power factor can not only be caused by a phase shifted current through inductors/capacitors. It can also be caused by having a non-sinusoidal current, which any bridge rectifier with sufficiently large filter cap will produce. In this case, power factor can be low, even though there is no power fed back into the mains.
The wiki definition of power factor is only valid when dealing with perfect sine waves
@EEVblog dont you have enough a power factor to change this government contract? 🤭
Is there actually some benefit to having a full bridge on the secondary side or doesn't that really matter if you don't generate a proper full wave input to the coil?
You don't use full-bridge rectifier on secondary of flyback converter. Having a full bridge there will cause a) high inrush currents when first switching on, can blow up the primary side circuitry; b) higher loss in operation due to two diode voltage drop instead of one.
The operation of flyback is highly asymmetrical. During one polarity, the transformer builds up energy in its magnetic loop. During the other, the energy is "discharged" into the secondary side. The actual voltage values in these phases may be somewhat different.
Seriously, that title is psychological clickbait. You know what it reads on the thumbnail, right? xD
What happened to David?
yeah.......no
I feel better now knowing I am not the only one who read "f#cker"
This is Shenyang Guangcheng Technology Co., Ltd. specializing in the production of industrial Internet gateway, PLC
them samwa caps, REMOVE NOW
A single stage may work with some real current measurement feedback instead of a pseudo assumed feedback. Would almost need to be on a switching frequency cycle by cycle basis.
That L and I look like one letter in the thumbnail.
.. period ton..
woot? :-D
I was looking at the closed caption because you kept talking about the BOM costs (yes, Bill Of Materials), and yes, youtube things your talking about a b o m b. No strikes yet?
The UA-cam algorithm is probably intelligent enough to take context into account, along with the nature of the channel.
shoehorn it!
shoehorn it!
just do it! :-D
Could think of a bigger and proper led driver at 500 watt and hookup 10-11 led panels on that and your chicken diner is ready.
These are designed to be retrofitted into existing lighting setups. IE: replacing florescent tub lighting with LED panels. 240v is already present at each fixture, making it easier to replace individual panels. Also makes it easier to change a panel or two if different k color is wanted.
Why do the power companies scoff at a low PF due to capacitance? I thought the majority of PF losses were due to inductive components like motors. Wouldn't these offset those losses? Or are they worried the PF losses would shift in the opposite direction? (From mostly inductive to capacitive)
CTB: The power factor of other devices is irrelevant. Every device should have reasonably high efficiency.
There is a difference betwen phase-shift from big inductors and caps directly across the mains, and current distortion from big caps behind a bridge rectifier. But both just increase the losses for every kwh of real powwer deliverd
Capacitors shift the power factor in one direction, inductors shift it the other way, which is why large motors with high inductance are usually pared with a beefy capacitor in industrial applications to keep the power factor fairly close to 1, as industry are billed on apparent power, rather than real power
Power companies don't hate big capacitors in appliances. Here what makes them unhappy: full bridge rectifier, followed by big capacitor. Reactive power that this setup generates is particularly hard to filter
🚀dec SpaceX launch date looking good the starship 👍hi dave
if i where to ever order from that company i would order without a psu i will tell them i already know of the poor quality of the psu
If the government is going to use your taxes to replace lamps.....they could atleast make part of the lights locally.
goddamn race to the bottom
Aren't we doing kerning anymore?
It's highly overrated
Somebody isn't saying FULL BRIDGE RECTIFIER properly!
Hint: not me.
In the good old days all you needed was a filament. No PF problem, inherently low flicker (that worked), good red and infrared, no blue peak, output for good health and they warm the house in the winter saving on the central heating bills. Low embodied energy in manufacture and disposal compared to solid state and cheap as chips! LEds, who needs 'em? Oh, the corporations so they can sell us more stuff....
That's quite enough thinking please go and purchase some medication to control your issues.
Just do Muntzing and cut capacitors until it dies 😂😂😂
It was done by manufacturer jet :)
@@volodymyrzakolodyazhny7740 hhhhhhh :D
8 Australian likes..... Hmmmm.
Please, you are an educational institution, academic exercise is wholly appropriate here.
Besides the output capacitor being pathetic, so is the inductor sitting there, it wouldn't be much of a stretch that they together form an output filter and you could get a lot of leeway out of a beefier capacitor by replacing that inductor with something much more inductive.
I'm all for polishing the turd in this case. You can still chuck it into the bin after making the videos, i sure don't wish you to put it under your roof.
Around 3:12 you made an error. Capacitors in parallel follow the same scheme as resistors where the value is divided by the number of components in parallel. Two capacitors at 150 nF in parallel would result in 75 nF, not 300 nF. Two capacitors in series would result in 300 nF.
I have a bit of experience with this because I've had to fudge capacitor values by using two higher value or different value capacitors to get a non-standard specific capacitance, or when I need a non polarized capacitor by placing the anodes/cathodes together. NP electrolytic capacitors can be expensive or not available, using two normal polarized caps is a good alternative, especially when space is at a premium.
No, two capasitors at 150nF in series would result in 75nF not 300nF
@@barnenlenovo681 My capacitor tester must be lying to me then. It's one of those ebay specials so no telling who actually made it.
Admit it you all read “Fucker” first didn’t you
What is fucker?
"FUCKER 2" (small thumbnail)
Fucker 2?
Good sciencewise but in practice nonsense