LED Flicker 3 - More Electric Boogaloo

[@EEVblog2]

Two things: 1) That input resistor is a fusible resistor. 2) There is a spare footprint for an extra output capacitor to get the ripple even lower.

[@viperwizard491]

LLC resonant half bridge topology. Split transformer bobbin is there only to introduce leak inductance for resonance to happen.

[@felixokeefe]

It seems this kind of LED driver is referred to as "off-line LED driver" in the industry. Quite what off-line means in this case I'm not sure. From line voltage? Many prominent IC makers make dedicated ICs for these LED drivers. Including TI and ST. I have been looking at the datasheet for one from IXYS the "MXHV9910" to be precise.

[@felixokeefe]

Also LT3799 from Linear Technology. TPS92410 from TI.

[@yoksel99]

3) Input filter cap is not directly connected to the output rectifier. There is some circuitry in between, probably some resonant stuff to correct for PF.

[@bigclivedotcom]

Is that a split bobbin transformer? That's so rare in modern power supplies. I wish they use them a lot more for better isolation.

[@EEVblog2]

Does look like it.

[@Audio_Simon]

But why is it the secondary that had a cover?

[@Okurka.]

It's a split bobbin dazzler.

[@SeanBZA]

Probably a combined magnetic part, half is the transformer, and the other half is a secondary side inductor, arranged so that they share part of the ferrite, but at different points in the switching cycle. Explains the 2 diodes on the secondary side.

[@3v1Bunny]

@WannaDIY could have done without the caps..

[@BogdanSerban]

They probably used PFC (that's what the black inductor is for). There's no point in using another driver on the secondary side if the controller has current mode.

[@nrdesign1991]

Right, keeping the current flowing from the inductor! Good catch.

[@markoantesic4362]

Just wanted to comment the same. Probably a boost type PFC. They have (from the looks of it) 2 transistors, one for the boost stage and then the isolation flyback stage. So it would still be a 2 stage solution. If we are right, interesting little 6 pin IC than can do all that.

[@BogdanSerban]

@@markoantesic4362 this or it's a half bridge/forward converter with a gate transformer and the inductor is in series with the power rail to smooth out the current.

[@LaserFur]

looking at the cap trace it goes to the smaller transformer area. so I would concur. A boost converter to charge the cap and then a flyback converter.

[@T3sl4]

FYI, transistors are JSMC 3DD4244DM's -- BJTs, in half bridge, and the small transformer is for base current drive. It's a lot like a classic CFL circuit, with a controller tacked on. Can't figure out where PFC comes from.

[@farktard2740]

That "resistor" in series with the input IS the fuse ;)

[@Rainbow__cookie]

Wow

[@BenjaminEsposti]

They're supposed to be safe, but IMHO, too often the "fusible resistor" is mounted up against a flammable component, meaning when the resistor burns out in the case of a minor to moderate overload, it will melt nearby components and create a carbon pathway. How do I know this? Seen it myself, in those cfl bulbs.

[@AntonBabiy]

I might be wrong but I'm pretty sure that extra choke tranny on the mains side is a dedicated pfc circuit which is why there is 2 switching transistors. If you look at the output of the bridge it doesn't go directly to the big filter cap but through pfc thus high pf

[@piotr86]

Topology is similar to old AT computer power supplies with controller moved to primary side, and there is some kind of power factor correction added. They won't put expensive 450V capacitor if 400V was sufficient.

[@johnfrancisdoe1563]

Piotr Kościński Few brands offer 400V and 450V in the same series and size, because the values are too close. Capacitance is already -20%/+80% tolerance.

[@atmel9077]

Looks like this is NOT a single-stage converter. The 2nd MOSFET and the inductor at the left of the transformer probably are part of the PFC boost converter. So this is basically the opposite of the 48W driver from last video. This one has boost converter PFC circuit followed by an isolated current-regulated flyback. The 48W one from last video had an isolated voltage-regulated flyback followed by a buck current regulator.

[@yoksel99]

Hi, Dave, I think you are wrong, the input cap is not directly connected to the bridge rectifier. Something more complicated is going on. And it is directly related to PF correction.

[@mranderson2809]

Watch out for them failing! I fitted 60 of them in a shop last year, more than half of them have failed since. The replacements are twice the size, same part number. Think they messed up.

[@T3sl4]

Interesting that the panel says 700mA but he measures ~420mA!

[@howarthcd]

I have the 950mA version which looks identical on the underside, apart from the values of the three paralleled resistors (R3, R3A and R3B). The marking on the IC U1 on mine is R6AZ. I can't find any reference to this marking anywhere online though. It would be interesting to find out what this part is.

[@BenjaminEsposti]

I didn't know that CapXon made an "FU" series capacitor, but it seems relevant to their track record! X3

[@1pcfred]

All CapXon caps are of the FU variety!

[@BenjaminEsposti]

Just buy a meanwell CC PSU, they use good Japanese caps and are sold at a decent price.

[@timstreeter9795]

I would have thought for better current ripple regulation you would need a higher secondary inductance, which the bigger output transformer would be achieving.

[@watswat123]

Send these to Big Clive, I want to see a full reverse engineering and theory of operation.

[@1pcfred]

If you really want to know how they work send one to diodesgonewild.

[@adriansrealm]

"I don't know what you call them" clips, Dave, we call them clips

[@shickster1]

Latches

[@Graham_Langley]

@@shickster1 Ratchets?

[@shickster1]

Ratchets are serialized latches. This video featured a device with two latches which were independent for retaining a cable. (There were two retention devices in a single housing.)

[@MatthewSuffidy]

What's making it better? A faster or more responsive switching IC? A bigger inductor on the primary side? Larger capacitor caps? Was the ripple caused by the switching frequency? Maybe it is just higher.

[@Jay-fp8iy]

Can't wait for the actual boogaloo. We aren't going down without a fight.

[@Razor2048]

Fuses tend to fail when there is a surge or short. I should make a fuse that can handle a large surge without failing. 😊

[@cmssniper]

Quite funny that the housing says AC-L on the top and AC-N below, but the colors of the cable are connected vice versa. Not that it matters, but I assume that they assembled it wrong in the first place?

[@JoeBee999]

I still doesn't get it how they achive this. Whats the trick to get a better powerfactor?

[@WolfgangMahringer]

There are chips that do power factor correction (google PFC to see how it works). That's why there are 2 power transistors.

[@JoeBee999]

Yea. But I was hoping Dave would make a part 4 out of this. ;)

[@WolfgangMahringer]

@@JoeBee999 Good point, yes Dave, please do a video about active PFCs!

[@EEVblog]

@@WolfgangMahringer Done it: kzbin.info/www/bejne/g3O0n2mupdmibK8

[@ArnaudZANETTI]

Sure looks a lot like an active PFC ; this would explain all the added components: an inductor, a power switch and a pair of low value capacitors (plus a diode hidden somewhere).

[@rocketman221projects]

Now that you posted a video about those drivers, they have gone up to $30-$40 on ebay with shipping.

[@gblargg]

We have suppler issue. You take new Won Hung Lo version instead? Ready to ship.

[@EEVblog]

@@gblargg LOL

[@redtails]

7:23 was waiting for him to say "for all your tranny aficionados"

[@derofromdown-under2832]

Dave, the AC connections are around the wrong way, according to the cover...

[@superdau]

Don't tell me you have never seen a fusible resistor? It's right there at the input.

[@1pcfred]

I've never seen a resistor that wasn't a fuse.

[@tuxontour]

You have to love: 'everything above 0.95 is pretty fine' .... power factor of 0.94

[@Richardincancale]

1:43 it’s a ratchet Dave!

[@schitlipz]

Strain relief?

[@shickster1]

Latch

[@Graham_Langley]

Smaller ones are called mouse droppings.

[@1pcfred]

It's the rack for you!

[@Stefan_Payne]

Hm, could that Lifud be based on an LLC primary topology? That would expalain the two red caps and the additional "Transformer"thingy right next to the big ne.

[@pengiswe]

Did they boost the frequency? Probably flickering at a few kiloherz of something. Which means the filtering will be easier too...

[@eloygc1732]

Hi, maybe this was asked somewhere already... What are those sort of parallel lines on top of the solder mask on the secondary-bottom side? Idk they look like solder... Do they serve a purpose?

[@BineySaurus]

Heat dissipation.

[@craigs5212]

Dave, you should try and reverse engineer this one. There seems to be more to it than meets the eye. Two diodes in in the secondary, are they using interleaved switching phases? Appears to be a second transformer next to the main transformer, possible power factor correction. Don't think its push pull the two FETs look different, would be identical if push pull. Possible resonant topology. The main switching transformer also looks somewhat unique. Possible current feedback winding?

[@konohh]

I go with that. I also think that this PS is worth a closer look. There is... There must be some clever stuff on there to get rid of the ripple and to get the PF correction right.

[@melissaedwards9468]

It would nice to see a KZbin video on component lay out pro's and cons of primary side vs secondary side rectification and filtering. I see plenty of big guns knowledge people here so please some one put up tech clip on the best design power supply.

[@ThePopolou]

How do them square binding post screws work?

[@ebb2421]

blue mov thingy providing some feedback, maybe?

[@Graham_Langley]

No. It's a Class Y capacitor (look it up) that couples the secondary side back to the 'earthy' primary side to stop HF switching noise being radiated from the secondary side wiring. It's on most SMPSUs unless the secondary side is earthed or where the leakage current from the mains would cause a problem, say in medical kit.

[@ArixZajicek]

On the secondary side, what's with that trace coming off of the positive of the diodes that wraps back around on itself? Why isn't it just one big block? It looks to just go down to the smoothing capacitors before coming up to the output. Does that really make a difference being laid out like that? Either way the capacitors are going to be the same distance from the output connector.

[@Graham_Langley]

One way of looking at it is it reduces the effective series resistance of the capacitor contributed by the copper track. Another is it forces the ripple current to flow past the capacitor connection instead of taking a short cut. Whichever way you look at it it it's a standard technique on PSUs.

[@ed-jf3xh]

HV NPN with protection diode. Who'd have thought!

[@3v1Bunny]

Does this thing vary the switching frequency depended on the load to get the pf / ripple ratio? Would love to see the switching frequency on 2 different loads.

[@neteagle2k9]

Any news from the custom µsupply LCD ?

[@TheDefpom]

Where’s the “FLICKER 3” thumbnail

[@gblargg]

FLICKER 3: FLICKED

[@EEVblog]

2nd channel zero care factor

[@mmartinm]

FUCKER 3

[@ferrumignis]

'I'VE FIXED THE FLICKER' would be perfect.

[@TheDefpom]

SkyWizardless -LOL excellent

[@thomasboos2470]

Why is there on the secondary side allways only one Diode and never a full bridge rectifier? Doesn't it make sense because of the ratio/waveform of the output from the transformer?

[@LaserFur]

Because it's a flyback design. The output voltage has one direction that is based on the input voltage. the other direction is constant current so it increases to the output voltage.

[@iamjadedhobo]

This one has two diodes and a center tapped transformer. It does the same thing.

[@Ultrazaubererger]

Something doesn't add up here. You show a 700mA driver but your scope says its 400mA? Did you have the wrong range set or am I missing something?

[@timturner7609]

Im guessing you missed 1:10

[@Ultrazaubererger]

@@timturner7609 He mentions a 500mA-Version but still... The scope gives 400mA (4:30).

[@McTroyd]

Perhaps he's running (a lot) less voltage than the max of the constant current supply? That would drop the amperage down too. In fact I'd rather operate that way if it's a light I plan to use a lot.

[@Okurka.]

Change those caps. Green cap with gold lettering is the sign of imminent failure.

[@EEVblog]

I'll just paint them.

[@bazj5392]

If you have issues in the future then I would definitely check those caps

[@BenjaminEsposti]

It's manufactured by "Beryl" which is obviously from the bottom of the barrel - ahem, beryl.

[@johnfrancisdoe1563]

Okurka Nonsense. Some brand may prefer those colors, and some models and years of those may have suffered from that notoriously bad electrolyte formula years ago. The more important thing to check is the specced lifetime of the cap used, adjusted for how much margin there is between the max ratings and actual use.

[@Okurka.]

@@johnfrancisdoe1563 Nonsense. Green/gold combination Chinese caps are notorious for failing fast.

[@JimGriffOne]

Wouldn't an LC filter on the output make ripple a thing of the past? Or is there something bad about doing that?

[@jaro6985]

So for an LC filter at 100Hz, its going to be physically very large, heavy, and the components will be expensive. Its not bad to do it, just not optimal.

[@stclairstclair]

I must be pathetic, IDE watch Dave cook a hotdog part 3, Love from Chicago.

[@aaaaaaaaaassssssssdf]

i-prober sounds like something apple would make when they go off the deep end

[@tamaseduard5145]

👍👍👍

[@maicod]

new component: diodees 😂

[@God-CDXX]

PF .95 wow

[@johnfrancisdoe1563]

GLASS MONSTER What's the efficiency in % output w / input W? What's the 3rd harmonic content of the input current (affects current in Neutral wire)?

[@God-CDXX]

@@johnfrancisdoe1563 watch the first of these videos (( EEVblog #1252 - LED Panel Lighting Flicker Investigated ))

[@Graham_Langley]

No Dave, it's not simplistic. Look it up...

[@funkyironman69]

CapXon :'(

[@Okurka.]

The green/gold one is even worse.

[@funkyironman69]

@@Okurka. What is it? Fuhjyyu?

[@Okurka.]

@@funkyironman69 Wan Hung Lo

[@mforrest85]

I love power supply porn.

[@weerobot]

You get what you pay for so can't complain really??..!!

[@thephantom1492]

What I wonder is how much cheating they used in it. With only 6 pins to drive 2 transistors, get the main power and aux winding, and most likelly a current shunt... this little chip pack alot in a tiny package for sure... What I wonder is if the output current change based on the input voltage, or if it just assume let's say 240V and just go from there... Since it know the output voltage and input current (I assume there is not enought pins to sense the input voltage) it is easy to set the output current... Just a bit of math with the input voltage (assumed to be 240V?), input current, and output voltage, and an efficiency value... ((Input current * 240V) / efficiency) / Vout = current out. Since the efficiency is probably relativelly the same across the output voltage range, it can be asumed to be of a fixed value...