Interesting domestic LED design circuitry

Jan 13, 2021 43 Replies

e:

runs cooler and lasts longer. This was at the urging of the government of t hat country and now these light bulbs are mandated but likely are two or th ree times more expensive than the units we get that have to compete on pric e resulting in less efficient lamps that burn out quicker.

f
.

e
t

Sounds like you would have been better off not asking then. Why do you car e? Why are the zeners any issue? The point is the gadget gives a signific ant efficiency boost. It has much more redundancy than you would expect in a consumer product, but that seems to be what was required to get a better product to market.

In competitive profit driven markets this device would have never seen the light of day. By the same token, CFLs and LEDs have benefited by many gov ernment promotions and regulations because when you could buy an incandesce nt bulb for $0.25, no one would want to pay $5 for a more efficient bulb. Now it's hard to find an incandescent bulb because the public is educated a nd understand the cost effectiveness of a $5 CFL or $8 LED bulb. But they don't understand the $12 LED bulb with two or three times the number of LED s is more efficient and longer lived. I just wish I didn't have to change out all my ceiling fixtures to use them. Funny how the emphasis used to be on keeping cool the structure supporting the hot light fixtures and now it is keeping the bulb cool and not burning itself up.

Rick C. ++ Get 1,000 miles of free Supercharging ++ Tesla referral code - https://ts.la/richard11209

I like the idea, the e-caps will probably be limiting.

If they filled the bulbs with something other than air they could get a bit more heat conduction but it looks like the obvious choice (H- cheap and 7x better than air) is not acceptable for other reasons.

If they drilled some holes to allow convection...

Cheers Clive

It is well known that running LEDs close to maximum current (such as "1 W" LED with Imax=350mA) is harmful. First of all, the efficiency in lm/W drops at least 20 % from optimal values. For white LEDs, the phosphor degrades very quickly in a few thousand hours. Also thermal design gets complicated and the ambient temperature nay have to be derated.

However running a LED at Imax/2 or Imax/3 allows optimal efficiency. Even after 50000 hours, the phosphor works nearly as well as in the beginning,

Of course reducing the current will also reduce the light output, so

2x or 3x times more LEDs are needed to restore the total light output, which is more expensive. Since the efficiency increases by about 20 % some more light will be available, so the total power input can be slightly reduce, but not by very much.

Running three LEDs at Imax/3 each will have a lower junction temperature when mounting on a common heatsink, which has the same size as a single LED at Imax.

Thus running multiple LEds at reduced current is nothing new, so claims of extraordinary performance is not believable, However the greatly increased economical lifetime (phosphor light output) may justify the higher initial cost because of the longer time between replacements.

I don't care.

It could also be nobody wants to pay an extra $5 outside of desert kingdom of nothing for chinese bulb that saves 1 watt. Talk about wasted effort to shave 30% off a thing that uses a few watts, for fractions of a day anyways. The slaves in the UAE don't care how many extra bulbs they have to change. The entire thing is stupid.

That was fun, thank you. I wish big Clive would look at the signals with a 'scope. (He measures ~dc voltage across the current reg. and I'm sitting here wondering if the zeners are turning on and clipping the tops of the voltage every cycle. ) I also wanted him to change the temperature... I guess I'll have to do my own exps. :^)

So I've got some dimmable LED filaments. Is it just a bridge rectifier and series current limiting cap inside?

George H.

10 uF and ~30 (or so) ohms. RC time of ~300 us. that's not going to do much filtering of the AC. I find that AC led's have a lot of flicker. (at 120 Hz here in the US.)

George H.

Manufactures make bulbs (leds) run hot so they burn out sooner. This also results in more waste heat (per watt of light) GH

Do the bulbs have air inside? They could fill 'em with helium if they wanted to keep them cooler.

George H.

You are welcome.

I wish big Clive would look at the signals with a 'scope.

I guess so. I can't remember if he's previously looked at this sort of thing with a scope. Might be some info on his YouTube channel.

It would probably depend on the bulb. If you google "dimmable" +"led"

+"teardown" you'll get several hits, such as , which, if you don't like too many videos, links to .
Jeff

The two 12V zeners don't make sense to me.

And I hope that's a protected-gate fet.

Ouch. that has way too many components inside! I say a tear down on bigclive. looked very similar. series C, rectifier, current reg with ~0.6V drop.

I'll dremel my GE 'relax' led. 120 V, 60 Hz, 45 mA

Hmm well (I took some pics but can't seem to find the button to share the files on dropbox. Grumble, why do they change things :^) There is a 33 ohm resistor (TH) on the input that goes to the rectifier and

10 uF 160 V electro on the dc side of the bridge. There is an 8 pin soic numberless. and ~10 other bits and pieces. No other big caps or such. I didn't try and trace or measure things... too much yellow crud on the pcb.

George H.

I thought they might turn on to limit power dissipation in the fet.

12V @ 20mA.

Hmm well doesn't the npn protect the gate. I've used a very similar circuit... well not at 100's of volts.

George H.

They also keep the fet from being a current limiter.

There is probably a regime where there's not enough current to turn on the npn, but enough voltage to blow a fet gate.

The circuit doesn't make sense to me. I've seen LED current limited by ac-side series caps, and by dc-side transistors or ic's, but not both.

At 20mA regulated, the ripple on the filter caps is ~10Vppk, ripple current in the lamps

It seems LED traffic lights are made rather cheaply. I can't tell the number of times I've seen a display with anywhere between 1/4 and 3/4 of the LEDs were burned out/nonfunctioning. Maybe the existence of about 4 parallel strings is considered to be enough redundancy.

(aside: The burned out portions are usually irregularly shaped. When they are made, do people stuff the LEDs on strings into holes on the signal face by hand so that they are different every time?)

(aside2: There's one LED traffic light which may have been a 'field test' unit, it was the first one I ever saw, and the only one I knew of for about 5 years. I never saw it partially burned out like that. Also it had all 3 colors LEDs. For quite some time they only replaced the red displays with LEDs, yellow and green left incandescent. Now nearly all traffic signals are LEDs. So for maybe 10 years it was the only one I knew of which had yellow and green LEDs)

that wouldn't last 5 years, helium leaks too much. one of the heavier noble gasses would probably be better.

Jasen.

One makes sense on the gate, not on the drain.

But the whole circuit is strange. Maybe the Dubai royal family gets a cut.

It's not an unusual circuit, save for the combination of capacitor power limiting and HV reservoir - usually the power limiter is used to generate much lower voltages - one of it's loss-free characteristics.

The limited input range, constant-load application, make it practical.

RL

yeah why are there two different values of capacitor (total of 4 caps) in series with the input power?

Assuming Dubai is 240 volts, is a 330 volt rating high enough? (don't forget AC sine wave peak voltage is quite a bit higher than RMS voltage)

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required