Surplus electronic parts : https://epartsconnect.com
Stock and Crypto AI Prediction : https://stocksignalslive.com

I was hoping this lamp was going to have similar circuitry to the Dubai lamp, which is not sold in other countries, and while it does have the same concept of under-run LED filaments for long life and higher efficiency, it uses fairly standard circuitry inside.
The one interesting feature of the circuitry, and also the bit that took me longest to reverse engineer because it caught me off-guard, is a novel design twist that reduces the voltage across the chip. The inclusion of transient clipping resistors and filtering should also add to the reliability.
It's been pointed out that the circuit configuration will also increase the efficiency slightly, as the chip current will contribute to the LED current.
I'm surprised that a linear regulator wasn't used as in many other lamps due to the simplicity. It's possible that this lamp has been in the design stages for a while and used buck regulator circuitry as a result of that. The choice of 100V filaments is possibly to cater for 120V and 230V options. Tuning filaments for linear regulators is a bit more complex than the multi LED arrays.
The high cost of this lamp (I paid £12 at John Lewis in the UK) is purely for the life/efficiency features. In the right location it will pay its way by reducing power and maintenance costs. Especially where architects/designers have put light fittings in locations that make servicing them difficult or hazardous.
In the home it could be useful in areas like halls and stairways where the lights are used a lot during the day.
The 3000K white has a slight nudge in the green area of the spectrum, possibly to increase perceived intensity by giving it a peak in the area of human eye sensitivity. I'd guess that aside from the efficiency gain of under running the LEDs, a fair amount of engineering has probably gone into the phosphor choices too.
If you enjoy these videos you can help support the channel with a dollar for coffee, cookies and random gadgets for disassembly at:-
http://www.bigclive.com/coffee.htm
This also keeps the channel independent of YouTube's advertising algorithms allowing it to be a bit more dangerous and naughty.
#ElectronicsCreators

A while ago, i featured the dubai lamp and it caused a lot of controversy because uh this is a lamp that was not available to the rest of us. It was only available to the people of dubai due to a deal between phillips and the ruler of dubai, who said he wanted the most reliable longest lasting, most efficient lamp in the world. They created them custom lamps in one two and three watt ratings. This is a two mark rating and for those of you who said, oh, we have those uh locally at the local supermarket for ages.

This is the one you have. It's got four filaments, it's four watts. This one is actually brighter. It's got 12 filaments.

No, it's got eight filaments and is rated two watts, so basically, four filaments per watt versus one filament per watt and this and it just means that under running the leds and the choice of phosphor makes this much more efficient than this. You have this one. Well, actually, no you don't. You now have this new one by philips, which is for the rest of us.

Let me read: what's in the packaging first 3000k color temperature for this one interesting color i'll, show you it. Uh 4 watts is equivalent to 60 watts. 840 lumen, which makes it slightly more efficient than this one, the original dubai lamp. It has some notable things.

Inside 60, less energy compared with standard phillips led bulb 50 years or 50 000 hours that, if you calculate it out, works to roughly about three hours a day. If you want an actual year, 24 7 life span. It's five point, seven years, fifty thousand hour uh. Fifty thousand switch on off cycles, which uh the circuitry is going to be pretty robust in these i think uh, instant lighting, not dimmable, can be used in enclosed fixtures, not suitable for use in the rain, and then i don't know what the symbol is.

I do not know what this symbol down here is hold on, i'm going to zoom up in it and focus on that, and you can tell me what you think that is, i don't know it looks like maybe an overhead light. Is it just means it's not suitable for illumination of uh walkways or something like that? I'm not really sure it's a new symbol to me. Let me put this back down and zoom out and then focus back on the subject in hand. Right say what let's open it up: let's plug it in so it is quite unusual.

It's got longer filaments than the original dubai lamp. This one is rated. The 4 watts just double checking that yes, and it has eight of these longer filaments, um and uh. Let me plug this in.

Let me show you lit: that's the best bet. Isn't it so i'm going to plug in the hoppy meter here is going to bring it in and show you the power rating and you'll see it's quite an odd shade of white. This is like this is compared to sort of daylight, because that's a typical lighting i use at my bench. The color is a worm white, but it's almost like a shade of cream and i've got a theory about this because they do specifically say eye comfort.

It's optimized for that. I'm just going to give this a swipe and see if it flickers no flicker at all. Let me just uh demonstrate that on screen. Well, actually it does look like it's flickering, but that's probably the rolling shutter effect, uh 4 watts 3.84.

To be precise, um power factor of 0.5, which is fairly typical for led lamps. It's not so critical because they're, not a super duper high load. So the color - i would say this is - is a soft white biasing very slightly towards the green area of the spectrum, which makes sense, because the human eye has a peak sensitivity of around about 555 nanometers, which is a sort of um, aptly green color. And they might be aiming for that uh just to improve the efficiency, but also to make it softer in the eye, because that's where a lot of the visual data in the eye will be resolved right, tell you what uh we're here to take this apart.

We do that a lot in this channel stuff gets taken apart, so uh, i'm gon na try popping the little stud at the bottom. Here i shall zoom in a bit for this, so we can watch the blood gushing from my hands. When i have a terrible incident, it can happen. Blood has been shed in the channel, it occasionally happens, it's called workshops.

So that's a little stud out. The end am i going to be able to nibble into this. This is where people also suggest. I use a dremel.

The dremel does work, but it makes a bit of a mess. I'm just going to nibble it away like this and see where we get. It might be a disaster i did say i'd zoom in i did not zoom in there we go. I have fixed that now for those of you who also brought up the subject of the phoebus cartel, how there's a lamp in america in a fire station that has burned for ages, and they could have made lamps that last forever.

In reality, the 1 000 hour thing for lamp life, the original tungsten lamp. So this has got a plastic shell in here. That is making this quite hard to nibble into, but the phoebus cartel the 1 000 hour. Life is a compromise.

You can make a tungsten lamp that uh lasts much longer the building site, the rough service lamps last a lot longer by under running the filament. So, instead of being a nice cold, crisp white light, it's a deep golden light, but it puts out a lot of energy in infrared area of the spectrum and not the visible layer of the spectrum. So you end up with a lamp that lasts for ages, but uh exponentially longer, but the efficiency tails off dramatically. So if you consider the uh 100 watt lamp of the original tungsten ones, if you ran that 24 7 um and you only needed to say, for instance, you only need about half that amount of light, but you use the the longer lasting, lower running lamp.

That would cost you about 100 euros, 100 pounds a year in electricity cost. This is this has got a little shell in it. That is different to the last one, or i should mention this is made in china. Where else you would expect to make all stuff made in china for everybody who criticizes china uh your your computer's made in china most likely everything is made in china yeah, it's not an ideal situation, but uh, but it is what it is yeah.

This is a quite uh tricky i'm going to persist, though, i'm going to keep peeling away at this. Oh now we're getting into the uh the glue at the base of the lamp, the bit that holds into the lamp cap. This is where it bursts forcibly. In my hands so yeah, the phoebus cartel was a standardization apparently between the manufacturers.

I don't know, did it really exist but um it was. It wasn't just to fix the life of lamps and make profit. It was based on efficiency of the lamps that compromise between the light output versus um the fish. This is more suction than i was expecting.

This looks like i can see an inductor in here. Is this a switching power supply in this one? That's a change. A little buck regulator, perhaps let us peel the last bit off and slip this off. Let's give this a squeeze and a big pair of pliers, scrunchy scrunchy squeeze the plastic, lift it off.

We have what looks like a little buck regulator. Right say what it's time for some reverse engineering: a little bit filtering circuitry as well uh. One moment please it has been reverse engineered. But before i show you the schematic here is a bizarre coincidence.

I decided to power up the lamp just to measure the voltage across the filaments, because i needed that for a specific reason in the schematic that took the longest time to reverse engineer just for this strange off data sheet oddity. But i connected the lamp to this. Capacitive dropper, which is based on a 680 nano farad capacitor, watch the power rating and the voltage across leds, but the power factor and the current and everything. Let me just plug this in so the lamp lights up.

This is it with a capacitive dropper, um 3.86 watts, which is the other one was 3.84 29 milliamps milliamps, the other one was 29 milliamps and a power factor of 0.53, which is very similar. This thing has gone over range, hello, let's uh select or i shall reset it and uh select the voltage and it's showing 200 volts. So each of these filaments there's clusters of them in parallel and then in series. So each of these filaments is dropping 100 volts, but there's two sets of three as far as i can see in series, and that means that that it's 200 volts across them, okay time for the drawings and continue so on one side of the circuit board.

We've got an 18 millihenry inductor, which is used to actually limit the current through the leds. We've got. The incoming supply comes via 2.2 ohm resistor as a fusible resistor, there's a 6 to 8 nano farad, 275 volt ac capacitor across the incoming supply and a metal oxide rest on the other side. Then it goes through the bridge rectifier and it charges this capacitor.

3.3. Microfarad 400 volts up to about 330 volts this inductor and a resistor one millihenry three ohms is actually in series on the dc side. With that, that's the component side of the circuit board. Well, that's the through-hole component side, here's the other side! So here's the incoming supply live and neutral.

There's the capacitor pins, there's a metal oxide varistor appears to be, and then it goes to the direct fire comes out. There is a track on the other side uh going over to here, where we've got the inductor and series of the positive going to the capacitor and then a metal oxide varistor across that i shall just add one small minor, pointless technicality to this. There. That's a minor, pointless technicality has been added the circuitry, that's switching, it is a uh pt4554 uh buck regulator with a couple of our parallel sent resistors and it's got the classic capacitor across the leds, plus the diode, and then a couple of discharge.

Very high volume discharge, resistors and then they're just basically going across the leds, which are here over to here and that's about all there is to see it should be a simple circuit. It wasn't a simple circuit to reverse engineer, just one little thing that went horribly wrong in the reverse engineering of this i'll. Give you a look at the manufacturer's data sheet: powtech uh. Pt4554C.

I don't know what the last letter was because trying to scratch the lamp cement off. I rubbed across that it looks like an oh but uh. It's that's not a valid component. It's either c or d.

I think it's a c. Here's, the circuitry very simple bridge. Rectifier smoothing capacitor the usual arrangement, but notice the hv line, which is powers this chip going straight up to the full supply rail. That's the bit that screwed me up.

Here's the actual circuit diagram. The incoming supply goes via that 2.2 ohm resistor, there's a metal oxide varistor across it to clamp any spikes and then there's a six to eight nano farad 275 volt ac capacitor. It goes through the direct fire comes out the other side. The positive gets filtered through a 1 millihenry choke with a 5.1 k resistor across it, and it charges up this 3.3 k at 3.3.

Four invoke capacitor, which also has a metal oxide wrist across it just to protect against spikes and glitches. So this will in circuitry terms, be round about the zero volt reference, and this will be about uh. Well, what would that be? Let's go. Let's go for a euro voltage 230 volt 230 volt times 1.41 will give us uh that's going to charge up to say 325 volts.

This is important to know, because there's a weird deviation from the normal circuitry here is the little chip. That's doing all the work and it basically turns on uh in the sense that it pulls this pin d for drain down to negative current flows through the leds from the positive through the inductor, and because it's a got up. It's not got a magnetic field at that point in time. It pushes back the back emf and it limits the current flowing once it reaches a certain threshold sensed by these two parallel resistors 39 ohm and 12 ohm, which gives about 9 ohm equivalent.

It turns off, but because this was positive - and this was negative - then it turns this becomes negative. This comes positive and actually as the magnetic field in that collapses. It goes through this freewheel diode so that both putting the magnetic field into it and when it collapses and it's turned off - are both used to power. The leds.

There is a 220 nano farad capacitor across the leds, that's about 200 volts across here, and also these very high value 2.2 mega ohm resistors to them in series uh to make sure those leds go out quickly and don't fade away slowly. I presume, but here's the oddity, the data sheet shows the hv connection going up to the 325 volt rail and i honestly i i was like probing about - and i was saying oh that'll, be to the positive and i was probing to the positive, not getting anything And it just nothing was making sense. I ended up removing the component uh, which was quite tricky because it was well glued to the circuit board and discovered that it was actually connected to here instead, so that isn't used that line there isn't used it's actually connecting to the bottom of this circuit, and I thought can they do that and, i suppose, ultimately, 325 volts minus the 200 volts across leds means instead of this chip seeing 325 volts, it's now only seeing approximately 125 volts across it. They may have done that as a clever trick just to take strain off this chip.

That's quite neat, if that's what they intended to do, i don't think it's a mistake. I think they actually meant to do that, but it's not showing the data sheet. I wonder i wonder how they made that decision to use in a non-standard way like that, a bit worrying to actually use a component in an off-label way, but i guess they probably just decided they probably sculpted everything and just decided. Well, that's going to take a lot of strain off this chip, but there we go.

That is it! That is the uh dubai lamp for the rest of us. It's the philips! Well, what do they call it? Uh phillips white, our ultra efficient light. Oh, this was expensive by the way this was about 10 pounds plus v80, so it came up to 12 pounds this lamp, which is a lot more than a typical standard lamp, but i suppose ultimately the place that this is going to win the application. It's going to win is, if you use it either as a if it's you find a nice color for reading, but also locations where are is quite hard to access lights or you want a decent amount of light, and it runs a lot of times say you Had a corridor in the house that you had a couple of lights in it that might make sense to put these in i'm a bit surprised, i'm slightly disappointed at the circuitry.

I must admit i was hoping it was going to be the dubai circuitry. I thought. Maybe even in this day and age, they might possibly have gone for the linear regulator. Approach like many of the others do, but this one.

It just strikes me as it's not get that extra double layer of redundancy the other one has. But having said that, hopefully it will be staying fairly cool um, but it is ultimately just a standard buck. Regulator like you might find possibly not this lamp, but it's quite interesting, strange creamy color, but uh there we have it uh. The 4 watts equals 61 840 lumen, oh color, rendering index 80, so not great, but that's common with the led lamps.

The color rendering index is how well it renders colors across the full spectrum with uh the led lamps. You've got blue leds and stimulating phosphorous. To add the red and the green components, it's not a perfectly linear spectrum, so you don't get absolutely amazing, color rendition, it's just like the early fluorescent tubes as well, but that is just one of these things. It's the compromise you make for the efficiency, but a very interesting lamp.

It was certainly well worth taking apart and exploring and uh. The construction is very different and uh that little twist in the circuitry that took me by surprise was quite intriguing and took a bit of working out, but there we go, that is the philips ultra efficient lamp.

15 thoughts on “Inside a new philips ultra-efficient led lamp with schematic”
  1. Avataaar/Circle Created with python_avatars Mikael Edlund says:

    I just bought this Philips lamp in Sweden. Eight filaments, the same box with the same bar code. The only difference is that mine has an EU flag on the backside where yours had an UK flag. I paid 19 sek, around £1,6

  2. Avataaar/Circle Created with python_avatars jh5kl says:

    the symbol is clear, you don t walk out of a movie just because you don t like the first half, you commit, common sense says you had the initial 9 minutes, after that you watch everything, some exceptions may apply if a detestable actor gets introduced unexpectedly at some point, if the monster is stupidly made and you can see wires moving it, otherwise you see EVERYTHING, you dont give a bite on a sandwich and leave it on the fridge, you don t walk out in the middle of a movie, the symbol is clear

  3. Avataaar/Circle Created with python_avatars Hola! Chris Wilson says:

    As I recently commented on your "Dubai Lamp" video, I want to get some of these but they aren't available in a bayonet fitting. I emailed Philips as well as asking on their YouTube channel if they have any plans to do a bayonet version but no one could be arsed to reply. I know you can get adapters but they are no good in certain fittings where there isn't enough clearance.

  4. Avataaar/Circle Created with python_avatars Mark Moore says:

    This schematic is a start. It needs an ARM processor running Linux and a Bluetooth 5 chip to connect to your smartphone to monitor bulb performance and draw graphs. The deluxe version would have a cell phone modem to send you an alert wherever you are in the world to let you know if the bulb has been turned on or off. An environmental sensor would alert you if the bulb got wet.

  5. Avataaar/Circle Created with python_avatars carlu bambi says:

    the symbol means not for motion detectors ! .There is a huge scam from all lamp manufacturers .Color of lamp is so that they can claim higher lumens per watt.eyes are sensitive to specific areas of light ..Nothing beats a low pressure sodium light but the color is brutal !Good for growing plants ,not for working or socializing around !They make everything in China because they have no patent laws ,no human rights and slave labour and no environmental laws !

  6. Avataaar/Circle Created with python_avatars WillTell says:

    oh nice, it's got an A on the new eu energy scale. Most common LEDs get E or F rating (since as you've explained before, they're not actually made to be efficient), and I've seen one get a D.

  7. Avataaar/Circle Created with python_avatars Jan Kindl says:

    3000K is white light? in what universe? some companies claim 6000K as a White. Both are wrong. 6000K is cool white, 3000K is Warm white.
    White light (for me at least) is 4000K… not warm, not cold.

  8. Avataaar/Circle Created with python_avatars brostenen says:

    I might get a good number of these in the future. But I still miss lightbulps that have the exact same intensity and spectrum as a glow light. Sorry. i dont know what the old ones that we used to have is called in English. But I miss that exact same quality of light for my workspace so I can solder and build vintage computers at night. My eyes are not getting younger you know. 😔

  9. Avataaar/Circle Created with python_avatars attanamir says:

    I would guess that that symbol means to not install the bulb in fixtures that have the bulb "upside down". Some LED bulbs don't handle the heat well and overheat the circuitry that's usually in the socket area, drastically reducing its lifetime.

  10. Avataaar/Circle Created with python_avatars Steven Clark says:

    With that kind of drop why bother with a driver circuit? Couldn't you just use the strands as their own rectifier and maybe have some high-watt resistors or resetting fuses if absolutely necessary to handle spikes?

  11. Avataaar/Circle Created with python_avatars Kromaatikse says:

    The use of a switch-mode regulator is a clear efficiency win versus a linear regulator. A linear regulator is, from a power efficiency standpoint, equivalent to dropping the voltage through a resistor, just one that self-adjusts its value to fit the control setpoint. A switch-mode regulator is not 100% efficient, but it generally dissipates significantly less power through its operation than a linear. The reduced heat output associated with higher efficiency will in itself be good for longevity.

    I don't read Chinese, but I did pull up the PT4774D datasheet. It's a current regulator, with an internal current-sense element which is compared to the setting expressed by the external low-value resistors; it doesn't care about the voltage for control purposes. The HV pin is used only to derive a relatively low-voltage internal VDD supply for the chip's control circuits. Taking it from a lower voltage would reduce the power dissipated across that internal regulator (which will be a linear). It's a matter of a couple of dozen milliwatts, I think, based on the quoted typical current consumption from the internal VDD rail. So this deviation from the example by Philips is reasonably smart.

    In this context, the high-value resistors bypassing the LED stack perform two functions: firstly, they ensure that the internal capacitor is discharged when the bulb is removed from the power. This is not actually about ensuring the light goes out promptly – which would only be an aesthetic choice – but to reduce the risk of high voltage being present at the bulb terminals upon exposure to human fingers. The second function is to ensure that the HV pin of the regulator chip is pulled high enough to reliably start up. This is likely in any case, but it "makes sure". A downside of the presence of this bypass path is that it reduces power efficiency, as a parasitic load on the regulated side of the circuit. This suggests to me that the safety aspect is more important to Philips, as they would rather not have this parasitic load dragging down their efficiency rating.

  12. Avataaar/Circle Created with python_avatars Tim Nilsson says:

    The lifespan hours I simply don't believe in one bit.
    I saw one study with x amounts of lamps from different brands and 50 % of them shit the bed before the 10k hour mark.
    Some of them were quite expensive Philips models.
    As already mentioned in the comments the LEDs themselves might last 50k hours but the electronics won't.

  13. Avataaar/Circle Created with python_avatars Hola! Greg Benwell says:

    Phillips LED light bulbs are total JUNK!!! Two years ago I threw down $60 to buy a bunch of these new Phillips LED bulbs and replaced ALL the light bulbs in my house with them!! Within a week one by one they started to burn out!! Now I am running on an old 50 amp electric service that was installed in my home back in the early 1970s!! So these lights SHOULD have lasted a very long time!! But instead one by one over the next six months every bulb I replaced, even the replacement of the LED bulbs with MORE LED bulbs burned out!! And in short I wasted $60 on a few packs of light bulbs that didn't even last two years out of the 6 packages of LED light bulbs I bought!!!

    AND BEFORE YOU THINK I AM BEING HARSH CONSIDER THIS!!! Around the same time I bought those piece of crap Phillips light bulbs, I bought cheap $14 a roll LED light strips for my shed (this was 6 years ago)!! Today my house is lit by 60 watt incandescent light bulbs and my shed is STILL lit by those cheap LED light strips and an ATX computer 100 watt power supply!! So the $30 a package I wasted on the crap Phillips light bulbs was money down the drain, and from now on cheap LED light strips are the way to go!!

  14. Avataaar/Circle Created with python_avatars Tisimo says:

    CRI 80 color rendering, 3000K and biased to green are all tricks to get a more efficient LED. Unfortunately it also means the lamp will be a different color and render human skin as if one is ill. Lamps achieving CRI 95+ and 2700K without a green bias will be far superior to this lamp with regards to natural rendering of colors and people looking healthy. Unfortunately those will use about twice the energy as well..

  15. Avataaar/Circle Created with python_avatars Steve Jones says:

    By my calculations, one of those bulbs would save about £1 per year per daily hour compared to the more usual ones at the same output at current UK electricity rates. So if a bulb was used on average 4 hours per day, that's about £4 a year and any extra cost would pay back in a year or two. However, if it's only turned on occasionally, then it's going to take decades. The lesson here is to put the most efficient bulbs into the fittings which are used most if you want to optimise your saving and expenditure.

    Note, I've not considered any savings from the bulbs lasting longer, but that might not be so important for those which aren't used a lot anyway.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.