I just had the capacitor fail on a "quality" Osram lamp after a small fraction of its rated life - interesting aroma - solvent, ozone and charring. Maybe I'll try one of these.
14W and no cooling? It's not going to last long at all. LEDs will fade very quickly. And at barely over 100lm/W they're not event that efficient form the start.
I've been running these Sansi ceramics for about 3 years. I had the design here in my bathroom ceiling cans until they all burnt out the LEDs, and about a dozen of the chunkier 27watt versions with the far more massive conical shaped ceramic heatsink , all installed in track light fixtures. These lamps have such high power density they run too hot. I've written the company criticizing their use of 12 parallel strings in the 27watt version. Once an LED chip burns out , with that little black dot, the remaining strings get more current and soon fail in cascading action. The company was very good and sent me new lamps. I permanently removed the clear plastic covers on the 27watt versions, and on all bulbs I changed the current sense resistors to adjust power to 70% of original. All of mine have a traditional circuit board inside , behind the ceramic block, employing the non-isolated brite power chip + inductor based driver (as opposed to linear regulator). I bought mostly dimmable versions as the power factor is near perfect versus poor ~0.6 measured on the non-dimmables. With mods and open air cooling the bulbs are good and will last. But if you run them stock, they're simply running them too hot and the ceramic heatsink is no match for the dense , searing heat. One LED will burn and then the cascade failure of remaining series strings will completely nuke the whole bulb. Sansi is playing the lumens marketing game and whilst it delivers, 'the bulb that burns twice as bright... '
The ceramic seems like a heat absorber, not a proper heatsink. Designed by default to not run for long and hope it's turned off to cool down. They do look nice to modify though.
@@volvo09 i totally agree with you . ceramic sounds exotic but feels misplaced here. Presumably it must be way cheaper to cast and kiln fire these babies and mount the led chips directly versus using aluminum heat sink plus an aluminum clad circuit board...
Can you imagine having to write to GE or Sylvania (back in the ~good old days) to tell them how to build a lightbulb? The world is just awash with non-repairable, undocumented e-waste built in unionized sweatshops/concentration-camps, with the profits going to morally bankrupt CEOs and pudgy psycho dictators who view democracy as a disease. Enough with garbage LEDs; reverse engineer a time machine so I can get out of here!
Woah, was expecting a ceramic heat lamp not a ceramic LED PCB. Turns out the company that makes those (Shanghai Sansi Technology Co Ltd and Jiashan Sansi Photoelectric Technology Co Ltd) have a number of patents about "ceramic seats" (陶瓷座). It seems the advantage is they can make the PCB, heatsink and frame of a device all one component with ceramics and be non-conductive unlike aluminum/copper PCBs. This bulb is a relatively boring example of this with only a central hole for cooling and an exposed outer ring but the company also makes stuff like street lights with a finned heatsink integrated into the ceramic "seat" as well as a lamp where the entire base is ceramic unlike the plastic of this one. Some of that patents for the actual ceramic material seem to imply it has some additives to increase thermal conductivity but also the far IR emissivity, to much better than bare aluminum. (they're all in Chinese so it's hard to tell exactly)
So Tim. I seem to have trouble locating the information you have presented before the TH-cam masses. I have contacted the companies provided and will be continuing to search for the patent/s you have referred to. For those not aware, in Material Sciences, engineering ceramic is very complex and results in a high performing BUT brittle substrate. . Most ceramic compounds exist between metallic and non- metallic elements. Therefore, interatomic bonds in ceramic are between ionic or predominantly ionic that have some covalent characters. Atomic defects can happen among ceramic compounds, creating both vacancies and interstitials. Therefore, defects for each ion type can occur. There are two types of intrinsic defects that can happen in ionic structures as Frenkel defect and Schottky defect. With this in mind, we can understand why I'm genuinely curious about these patents you refer to.
@Kevin Sellsit perusing some documentation... indicated high emissions via the LED's themselves and through the ceramic substrate with no fins (awaiting response from Sansi). Also, the enclosure prevents debris, airborne dust, and fluids from adding a film to these "breathable LED'S". Therefore, opening and touching the LED'S without gloves or a cleanroom is strongly not recommended.
Very interesting evolution of the lamps, they really are churning out so many different types. This looks like its well made compared to some that we have seen.
Ceramic PCBs have a very long history in the electronics business as the substrate for hybrid circuits. Tracks are often laid down using a "print and fire" technique. An example of an early ceramic hybrid is the SLT circuits used in the IBM System/360 computers of the mid 1960s.
It is my understanding that this process of "firing" is limited with several drawbacks. Consider that the ceramic substrate and the oxide layer between the tracks prevents a positive covalent bond between the surfaces. These aberrations become clear through normal use as the tracks contract and expand with radiation of heat which leads to premature failure or lifted tracks. The truth of the matter is, ceramics are complicated.
@Sundog Interesting, you said that. Indeed, from what I can observe by Clive's video, the manufacturing process seems to have been simplified or reduced to a nominal cost. I recall, a few years ago Taiwan conducted a study using graphene oxide/graphene layered composite that created an atomic layer between the substrate ceramic and the tracks and used an electrochemical sensor/s. This technology may allow for the oxide layer to become sacrificial, layered in a anisotropic chemical eching. According to the the numbers submitted, this improved efficiency: "This highly anisotropic hydrogen etching technology may work as a simple and convenient way to determine graphene crystal orientation and grain size and may enable the etching of graphene into nanoribbons for electronic applications". Preparation, Bandgap Engineering, and Performance Control of Graphene Nanoribbons Hao Luo and Gui Yu Chemistry of Materials 2022 34 (8), 3588-3615 DOI: 10.1021/acs.chemmater.1c04215
I have used Sansi 15 and 40w bulbs for 5 years. they cycle off for only 4 hours every evening. They are ALL still working as they were when they were new. Amazing things.
I've taken apart every LED blub that fails and one of the more common failures in my experience across many brands (including Philips) is the capacitor, most of them are completely fried. Poor ventilation and the typical overdriving of LED's producing too much heat. This one looks great with how much ventilation it has and the distance the main cap has from heat source.
Julian Ilett figured out sticking red electrical tape over his red LED displays makes them easily readable on camera. It would make your light bulb tester a lot easier to read during use. My home test equipment became much more pleasant to use after copying his example. Red tape is much quicker and easier than trying to glue or mount diffusion gel in front as I've done in the past. Another enjoyable teardown! Thanks again Clive.
DiodeGoneWild has used strips cut from black plastic garbage bags for the same purpose. The idea is to reduce reflectivity, so it's more obvious when light is actually coming out of the display as opposed to just bouncing off it.
I've had a set of 16W 2000 lumen versions used in my garage for a little more than 4 years. $10 USD each at the time. Nice and bright, and handle the wide temperature range of the garage in Minnesota (approx -5 to 110 deg F).
When you mentioned how they might apply the circuit tracks to the ceramic base, I thought back to my time working in an auto-glass factory. We used silk screening to apply the heater lines to back windows. The paint was a ceramic silver mixture that bonded with the glass under heat. The could do something similar with these lights: silk screen the tracks and then use lasers to bond them to the ceramic.
That small addon PCB is so they can solder the wires with a normal iron. You cant solder on ceramic substrates with a normal iron because the ceramic immidiatly conducts the heat away. You would need to heat the whole substate on a hotplate, which is not convenient on the assembly line
So glad your covering these - I'm using the Sansi 36W on a Scotch Bonnet pepper plant and it responds very well - BUT - I've heard some folks have had bulb flicker after 10 months, so disassembled drivers on Big Clive is excellent news❣
Re the thick ceramic - this will give a substrate that is mechanically more robust, but also may increase the yield of good parts during the ceramic manufacturing process. The thinner ceramic parts are the more likely a higher percentage will get broken during drying and firing (and in use, when the finished lamp is being punted around by growers etc.) Looks like a well made unit. Cheers!
Mr. Dayman. It is apparent to me you understand the mechanics of ceramics. Are you aware of any implementation or applications of ceramic fins in sintered substrates? I am currently asking my colleagues who are unaware of these applications in scientific settings as well.
We have a bunch of their lights, the quality is excellent and so is the warranty. I had one of the 40 watt models fail near the end of the warranty period and they replaced it with zero hassle.
Fun fact - if you use a fiber laser to engrave on to a piece of aluminium nitride ceramic, a reaction occurs which produces metallic aluminium and will hence give you conductive tracks ;-)
I worked at a company making ceramic boards for high temperature circuits. The traces are silver powder in a binder put on with a silkscreen and then fired to burn off the binder and fuse the silver. That ceramic is REALLY hard on the machines, constantly grinding down the hardened steel guides.
I have used this type of lamps for my plants and seeds for several years, never had any problems. Also have a customer service team who you can actually use!
I use 2 Sansi C21BB 27watt E27 bulbs for my garage and I lights it up a treat. The ceramic on those is even more beefy but they use a driver on a little board and the centre of the bulb. Was so impressed I bought a Sansi PIR flood light on the driveway. Built to last and not a hint of flicker.
Very informative so much so I have just orderd one of their plant grow bulbs. These look really well made and I will give a little vid how I get on with this lamp and seedraising. Thanks Clive always enjoy your vids and good info.
Ceramic circuit boards were used many years ago by Motorola in cell phones. I worked at a factory in New Mexico, USA, where Motorola had a factory, inspecting them in their green form (not yet fired) for flaws. Of course no solder tracks were on the board yet. These were about 0.16 cm thick.
Just checked at the time when i bought them (hence why i grabbed/bought them) I was getting them in quantities of 6 for £5.83 (for 6) delivered so bought 3 lots this was end summer last yr
A good reason, as you noted, for having included the circuit board is that the ceramic substrate is difficult to soldier to. The tracks, often times a precious metal like silver or platinum, are too soft and ductile to support components other than surface mount components. Additionally, ceramics are brittle, making it difficult to drill any via (blind, buried, through hole, staggered) at risk of breaking or damaging the ceramic. There is a preference for sputtering indium silver alloys with graphene deposits in scientific settings, but equipment and lab conditions make the process far too costly for general manufacturing. In higher applications, Antimony, Tin, Cu and cubic boron nitride are added for vibration resistance and "springy-ness".
If I may also add Clive, the ceramic is a little thick-ums due to the fact that, depending on the cutting technique, the ceramic becomes difficult to cut the thinner it gets. I'm betting this puck was sintered thick for packaging and the occasional bump.
Hey Clive 😊. Wow this is a very interesting design indeed. You are so funny at times that I honestly thought that bulb was going to go pop when you turned it on 😂. Have a blessed day and thanks for sharing.
Considering the thickness and the type of material used, I would suggest that this is a reworked 6” wafer and the tracks are plated on it in a a MEMS foundry.
Sweet. I was looking at this company for a room's lighting. They dont seem that scammy nor that cheap in their build quality. Glad to see this reviewed and opened up. I'm trying to source lights better than the ikea LED bulbs (looking for full spectrum and getting a split between 5600 K and 2700 K)
I just ordered 2 4 packs of the 3K color and 1 4 pack of the 5K, directly from Sansi. They are about 4 USD per bulb. The site claims the tracks are silver. Great find, thanks for a good breakdown too. Cheers!
I love how the circuit photo looks on this one - primary colors and slightly broken symmetry. Not sure if the background being blue was intentional, but it's perfect! Do you archive these photos?
Been a while since I’ve seen ceramic substrate circuits. I remember first encountering them in Sanken audio ICs like the STK433 and RF modules in Philios pagers from the 80s.
Yes yes yes finally! I've been waiting for this! I have the American version of these bulbs at my house and they're awesome but I tried taking one apart and I couldn't get the stupid potting material off . Nor can I find the chip on the internet.
Interesting comment about using a LED chip that was different spec to all the others to balance the voltage rating. Similar to the makeweight chocolate you may get in a tin of Quality Street (600g and still dropping).
Would be interesting to see if tracks get warm, since they are not copper, so probably have a higher resistance. You could also measure resistance on some of those unused tracks.
Current is low milliamps due to the really high voltage of LED strings, and power is current squared times resistance, so the resistance of the tracks doesn't need to be low.
there is a graphic on their official website that calls the tracks 'silver lines' however it's a very simple 'ours vs theirs' comparison type graphic, hard to say how technically revealing they're being
We have PAR30 LED floodlights at work which are 40 Watts / 4,000 Lumens. They have a similar shape to the one Clive is showing, but have a fan inside. They are rated 35,000 Hours lifespan. They have been operating 24/7 for 5 years and not burnt out yet !! Very good stuff.
I may be mistaken, but I think the lifespan rating on most LEDs are telling you how long it will last before it drops below 80% its initial brightness.
I've got something like that in floodlight fixtures around the outside of my house. They're a pain in the butt to service (need a big ladder to reach them over the 2nd floor) so having bulbs that will last is imperative, cost is only a secondary consideration. Currently there are 5 working after 6 years and one that just failed a week ago.
@@MrMartinSchou That is true, we actually took one of the 5 year old floodlights out and changed it with a brand new one, and for sure there was a 20% difference in brightness. But I haven't seen any single burnt out leds yet, so they are wearing out evenly.
The fan is probably the shortest lifespan of the components, if it's a motor-driven one (as opposed to piezo or other "solid state" type). Passive cooling would be better, IMHO
Used to work for a lighting manufacturer over here in the states that specialised in horticultural lighting. We did some research with the Sansi ceramic substrate products. Interesting stuff, but stupidly expensive
Interesting (also some of the comments) as my initial thought was that the ceramic substrate (A) wouldnt be thermally conductive enough and (B) not big enough to dissipate the heat, but I admit I know very little about ceramics. I've noticed generally LED lamps over 10W dont seem to last more than a year or so, which I put down to not being able to dissipate enough heat. I'd be curious to know what the running temperature is compared to more standard lamps.
My guess as to the “PCR” labeling was “Programmable Current Regulation”, so that would certainly fit. Basically a way to assign a weighted multiplier to the current sense pin.
Since it has a sort of rounded 3D appearance my guess would be screen printing to apply some active alloy on that can directly bond to the ceramic. Alternatively it could be as simple as conductive ink to keep costs down, but not sure how good the thermal dissipation would be. It could be laser sintered as well, but given they are going for bulk production I would guess they are going for something fairly economical. In a way its almost sort of a moot point since they go thru the effort to make it ceramic, but then just plop conventional low temperature components on it rather than the dies themselves, nor bother to use an actual heatsink to take advantage of it. Marketing I guess...
A look into the capacitors rating for temperature and life, aswell as measuring the temperature it is exposed to, would be useful to determine how long it will last. Usually there are several different life ratings for each temperature ratings, also ripple current rating and resistance determine how much heat is dissipated in the cap. All the Sainsburys LEDs I have ever purchased fail at about 1/4 their rated life, I' assuming they are cap failures but I havn't investigated.
That Sansi company makes a ton of different types of LED bulbs and light fixtures. Very nice. I'm getting two new Sansi security porch lights. They can't be any worse than the ones I got that stopped working way too soon.
I use these Sansi as SAD lighting because of their full spectrum output. I have a really massive one in my office ceiling. I really rate them for quality of light. I switched after my previous preferred type started regularly failing by getting the fitting super hot.
I ordered 6 of the smaller 8w lamps, they have no real circuitry on the LED board, just the LEDs. I did manage to get the front off without breaking anything though! Trying to work out if I can lower the power easily...
hello andy. i would be interested to hear what your findings are. as i have ordered 6 of the same today, and would like to do similar things... also in particular if there is any chance of a dc derived low voltage somewhere. (but probably not it seems)
Mine arrived today, and I pulled one apart to test. There are 3 parallel surface mount resistors on the rear of the main/power circuit board. These can be removed to adjust the current. In the default configuration according to my cheap plug in watt meter the lamp pulls 38mA 6.7W. With 1 resistor removed - 27mA 4.7w, and 2 removed 13mA 2.2W. With the board still inside the lamp I think it should just about be possible to clip off the top resistor with a pair of small side cutters without damaging U1 above it.
@@MrKureigu07 just uploaded a video showing the mod, can you check i have cut the correct resistor?! It is still pretty bright, but doesn't get as hot! 👍 (Edit, just measured it by plugging modified lamp into my inverter to see power, it's approx 3w lower than unmodified)
Interesting lamp. The grey color reminded me of the ink used for some of the PCB making machines that use a pen of conductive ink to draw the circuit on to boards. The ink only needs to dry prior to soldering.
The way these were made in the 1960s was to run the substrates thru a printing press, depositing conductive ink. Then you fired the substrate like a coffee cup, which melted all the little silver bits into a continuous trace, and fused them into the top few microns of the ceramic surface. The results were extremely durable.
Typically the ink is composed of actual silver nanoparticles as silver is highly conductive and most of its oxides also remain conductive. Not as conductive as a solid copper trace would be, but still perfectly acceptable if you can accept some extra resistance losses in your traces and for short traces and lower currents/higher voltages those losses are often quite acceptable for many electronics. I have occasionally used such pens to repair or route around damaged traces on PCBs as opposed to routing a bodge wire as the repair just to see how such worked out long term and….20 years later those repairs still work.
@@zilog1 i really cant remember. i found it on the internet a long time ago. liked it. stuck it on my github profile, and only recently added it to my yt profile ;)
It could well be Electron Beam Melting / Sintering for the tracks not laser, I've seen work like this done on Arcam systems. You could load up cassettes / trays of parts and process them very very quickly.
Easier to screen print the tracks with silver loaded ink, and simply fire the ceramic a final time to bond it to the board, as is used on so many ceramic substrates. Low cost, and good adhesion, plus reasonably conductive traces, and as a bonus a second screen after drying of the first of a glaze to cover the tracks aside from the exposed solder pads makes a good solder mask as well.
You won't believe this clive but your fire sprinkler heating video was used in my online classroom module of my firefighter training, from the state fire academy. Specifically the heating bit with the air gun on the bulb for 4-5 seconds. I know it was you because of the phat sooty skidmark on the bench.
it seems like a lot of air gap around the capacitor for airflow... so i wonder how hot it gets? for example an 8w cheap one, if it does not get too hot in there maybe we could... have enough room to put an esp-01 with a relay module inside, and then flash a firmware to control them from home assistant? although i suppose it also would need a stable low voltage dc power for the esp8266 and the relay coil. i am not sure how feasible this is yet
Isn't ceramic a really good insulator (thermally and electrically), and thus probably not great for heat dissipation? Wonder how long it will last before getting the black spot of death.
It would be interesting to see one of those multi white bulbs. Where it changes the white tone every time you power it on and off fast. But when turned off and in normally it keeps the same tone. Normally they are 3 faze, soft/warm/day. It's not just a brightness change like old incandescent 3 phase bulbs.
So where does the 400v come from, with a 240v input? Is that something to do with the rectifier, or..? I may have missed a part of the video, but I don't recall seeing anything about a voltage increase up to the part where you show the schematics and mention 300+v
Capacitors are normally spec'd with a safety factor - quality 12V circuits will typically use a capacitor rated at 35V or higher. With 400V being a standardized component, I expect this is the same.
Your mains supply is measured with RMS (Root Mean Square) to give an average value of the sine wave. The peak voltage is 1.41 times higher. So on a 240V supply the capacitor will charge to about 350V. A 400V capacitor gives a safety margin.
@@bigclivedotcom Ah, perfect answer! Exactly what I wanted to know, as always! Seriously, your ability to explain things in your content is impressive but what is even more impressive is your ability to understand what others are asking and giving them an appropriate answer without going too in depth or providing information which is too broad and shallow. Just signed up for a month of patreon, as thanks. Barely begins to cover the value of information you've given me, but I'm a broke boy so I just gave what I could :p Thanks, yet again, for the help big Clive! Really appreciate it!!
okay my version of this lamp that was bigger and looks too close in design to be a deviation - the glass cover blew up under one year. The resulting "bulb" is a flying insect death magnet. They made his version harder to covert to flying insect death tool but go buy a $9.99 electro death fly swatter and use this LED config to destroy flying insects. Tested in good faith on the shores of Lake Erie for over a decade. I assume this same set-up is still being used as I sold the property before the light source died. I felt bad for the thousands of Crane Flies that didn't adapt. No warmth for dead mosquitoes.
Given the construction quality I have to wonder if it'd be worthwhile to replace the bridge rectumfryer with a voltage doubler to allow use here in the USA.
It's interesting you did a teardown of this, since I was /just/ looking at them the other day to replace some outdoor lamps and I thought they looked cool. The advertisement on Home Depot here says that the combination of the ceramic ring plus the hollowed bar-like base is supposed to maximize heat dissipation and I can see that. I can also see, though, that the ceramic would be subject to moisture unless something was added in production to make it less porous. (This is an interesting point considering the bulbs I was looking at are outdoor rated and the grow lamps can be expected to be used in humid environments). I regret not buying them, just because I was in a hurry at the store, though after this I'm quite curious and may get one anyway, haha.
That's incredible! It gets me thinking: could they have a ceramic material with metal pieces (like aluminum or tin) suspended in it, then with a ceramic material that could evaporate or vaporize under a laser which could let the metal particles sinter/fuse together to form traces? That probably isn't how they did it but I'm wondering if it could work, haha.
Ceramic PCBs have been around since the 1960s, and maybe even the 1950s. The standard way to make them was to simply run them thru a printing press that deposited a silver-bearing ink in the surface. Then you fired them in a kiln, just like you would fire a ceramic coffee cup. This burned off the volatiles in the 'ink' and melted all the little bits of silver together so you got a conductive trace. The silver ink was black colored in every instance I've seen, and I noted that these tracks are a grey-black. It is possible they are using aluminum or some other metal than silver, but likely doing it the same way.
Print and fire, not laser sintered; the latter would require long laser times on each bulb and registration of the substrate. The temperature under operation would be interesting. The surface area is small compared to what's achievable with an aluminium heat sink which could fill that empty space in back of the plate.
I found them 'ceramic heat-sinks' to be useless for cooling. the 9W variant still makes a desk lamp to hot to aim at what I'm working on (I measured over 170F in a 70F room after an hour of the bulb being on, and it did throttle back to under 6.5W at that point, and was still getting hotter and thermal throttling more). 4:30 ish, the ones I got, have a sorry attempt of motherboard VRM ornament shaping to them, it's not flat on the back side, just as ineffective. it just bakes away in the bulb. I gave up on the Sansi bulbs after a few hours of looking at them, and went back to fussing with the aluminum backed PCB ones. the sansi bulbs are all gimmicks at best.
A good desk lamp design with LEDs doesn't use retrofit bulbs anyway. You need directional light, but the retrofit bulb's diffuser goes out of its way to send light everywhere, so then you need an inefficient reflector to re-direct the light. Which then cooks the bulb by trapping hot air. Plus so many watts worth of heat are only needed only because of the inefficiencies from the diffuser and reflector. A safe low voltage design with a good exposed heatsink is probably the way to go.
@@rexsceleratorum1632 yeah, I have a few vented back basket-like desk lamps, and I have removed that frosted plastic cover on a aluminum backed PCB bulb, drilled some holes in it's base to let air in behind the PCB, etc. much cooler operating than the Sansi ones even at the same mA/Watts/Power. the Sansi ones just produce way more heat than it should for a 'LED' bulb, and the ceramic heatsink thing is just a fail, lol.
@@Zarcondeegrissom Here in India the mains is a spicy 230V, so I don't want a desk lamp with an exposed mains PCB or even exposed mains-adjacent heatsink. Also any sort of holes are a bad idea in the tropics because it instantly becomes a bug trap.
Dear Big Clive, another random question (i am subscribed to your channel), I am buying a V-tac led spotlight for use over an aquarium. Its 20watts, But these things burn out usually (never had a v-tac one before though), Do you think using a capacitor inline with the live wire (resistor in parallel) would increase longevity of the spotlight ?
Wow, something built to last. Unusual.
Only time will tell. They probably built on purpose at least one weak link into the device so the customer must buy more.
I just had the capacitor fail on a "quality" Osram lamp after a small fraction of its rated life - interesting aroma - solvent, ozone and charring. Maybe I'll try one of these.
And how much does it cost?
14W and no cooling? It's not going to last long at all. LEDs will fade very quickly. And at barely over 100lm/W they're not event that efficient form the start.
Maybe the early ones are built better to build some reputation.... then the cheapening happens
I've been running these Sansi ceramics for about 3 years. I had the design here in my bathroom ceiling cans until they all burnt out the LEDs, and about a dozen of the chunkier 27watt versions with the far more massive conical shaped ceramic heatsink , all installed in track light fixtures.
These lamps have such high power density they run too hot. I've written the company criticizing their use of 12 parallel strings in the 27watt version. Once an LED chip burns out , with that little black dot, the remaining strings get more current and soon fail in cascading action.
The company was very good and sent me new lamps. I permanently removed the clear plastic covers on the 27watt versions, and on all bulbs I changed the current sense resistors to adjust power to 70% of original. All of mine have a traditional circuit board inside , behind the ceramic block, employing the non-isolated brite power chip + inductor based driver (as opposed to linear regulator). I bought mostly dimmable versions as the power factor is near perfect versus poor ~0.6 measured on the non-dimmables.
With mods and open air cooling the bulbs are good and will last. But if you run them stock, they're simply running them too hot and the ceramic heatsink is no match for the dense , searing heat. One LED will burn and then the cascade failure of remaining series strings will completely nuke the whole bulb. Sansi is playing the lumens marketing game and whilst it delivers, 'the bulb that burns twice as bright... '
The ceramic seems like a heat absorber, not a proper heatsink. Designed by default to not run for long and hope it's turned off to cool down.
They do look nice to modify though.
@@volvo09 i totally agree with you . ceramic sounds exotic but feels misplaced here. Presumably it must be way cheaper to cast and kiln fire these babies and mount the led chips directly versus using aluminum heat sink plus an aluminum clad circuit board...
Can you imagine having to write to GE or Sylvania (back in the ~good old days) to tell them how to build a lightbulb? The world is just awash with non-repairable, undocumented e-waste built in unionized sweatshops/concentration-camps, with the profits going to morally bankrupt CEOs and pudgy psycho dictators who view democracy as a disease. Enough with garbage LEDs; reverse engineer a time machine so I can get out of here!
To me ceramic sounds like an insulator material...
Is there likely to be a cost difference from straight aluminum? Might the ceramic cost be more stable and predictable?
Woah, was expecting a ceramic heat lamp not a ceramic LED PCB.
Turns out the company that makes those (Shanghai Sansi Technology Co Ltd and Jiashan Sansi Photoelectric Technology Co Ltd) have a number of patents about "ceramic seats" (陶瓷座). It seems the advantage is they can make the PCB, heatsink and frame of a device all one component with ceramics and be non-conductive unlike aluminum/copper PCBs. This bulb is a relatively boring example of this with only a central hole for cooling and an exposed outer ring but the company also makes stuff like street lights with a finned heatsink integrated into the ceramic "seat" as well as a lamp where the entire base is ceramic unlike the plastic of this one.
Some of that patents for the actual ceramic material seem to imply it has some additives to increase thermal conductivity but also the far IR emissivity, to much better than bare aluminum. (they're all in Chinese so it's hard to tell exactly)
That makes a lot more sense than in this specific product.
Wow they actually invented something
Wonders whst the symble for emissivity looks like.
€*///
So Tim. I seem to have trouble locating the information you have presented before the TH-cam masses. I have contacted the companies provided and will be continuing to search for the patent/s you have referred to. For those not aware, in Material Sciences, engineering ceramic is very complex and results in a high performing BUT brittle substrate. . Most ceramic compounds exist between metallic and non- metallic elements. Therefore, interatomic bonds in ceramic are between ionic or predominantly ionic that have some covalent characters. Atomic defects can happen among ceramic compounds, creating both vacancies and interstitials. Therefore, defects for each ion type can occur. There are two types of intrinsic defects that can happen in ionic structures as Frenkel defect and Schottky defect. With this in mind, we can understand why I'm genuinely curious about these patents you refer to.
@Kevin Sellsit perusing some documentation... indicated high emissions via the LED's themselves and through the ceramic substrate with no fins (awaiting response from Sansi). Also, the enclosure prevents debris, airborne dust, and fluids from adding a film to these "breathable LED'S". Therefore, opening and touching the LED'S without gloves or a cleanroom is strongly not recommended.
Very interesting evolution of the lamps, they really are churning out so many different types. This looks like its well made compared to some that we have seen.
Interesting! Thanks Clive 👍 🇬🇧
Ceramic PCBs have a very long history in the electronics business as the substrate for hybrid circuits. Tracks are often laid down using a "print and fire" technique. An example of an early ceramic hybrid is the SLT circuits used in the IBM System/360 computers of the mid 1960s.
It is my understanding that this process of "firing" is limited with several drawbacks. Consider that the ceramic substrate and the oxide layer between the tracks prevents a positive covalent bond between the surfaces. These aberrations become clear through normal use as the tracks contract and expand with radiation of heat which leads to premature failure or lifted tracks. The truth of the matter is, ceramics are complicated.
@@LeviGoldwing Indeed, I had to deal with a lot of hybrid failures in the past. Hopefully this new technology will be better.
@Sundog Interesting, you said that. Indeed, from what I can observe by Clive's video, the manufacturing process seems to have been simplified or reduced to a nominal cost. I recall, a few years ago Taiwan conducted a study using graphene oxide/graphene layered composite that created an atomic layer between the substrate ceramic and the tracks and used an electrochemical sensor/s. This technology may allow for the oxide layer to become sacrificial, layered in a anisotropic chemical eching. According to the the numbers submitted, this improved efficiency:
"This highly anisotropic hydrogen etching technology may work as a simple and convenient way to determine graphene crystal orientation and grain size and may enable the etching of graphene into nanoribbons for electronic applications".
Preparation, Bandgap Engineering, and Performance Control of Graphene Nanoribbons
Hao Luo and Gui Yu
Chemistry of Materials 2022 34 (8), 3588-3615
DOI: 10.1021/acs.chemmater.1c04215
I have used Sansi 15 and 40w bulbs for 5 years. they cycle off for only 4 hours every evening. They are ALL still working as they were when they were new. Amazing things.
I've taken apart every LED blub that fails and one of the more common failures in my experience across many brands (including Philips) is the capacitor, most of them are completely fried. Poor ventilation and the typical overdriving of LED's producing too much heat. This one looks great with how much ventilation it has and the distance the main cap has from heat source.
I have been using Sansi leds for about 6 years. They are super bulbs. The larger 200-250 watt equivalents are very bright with little heat.
Julian Ilett figured out sticking red electrical tape over his red LED displays makes them easily readable on camera. It would make your light bulb tester a lot easier to read during use.
My home test equipment became much more pleasant to use after copying his example. Red tape is much quicker and easier than trying to glue or mount diffusion gel in front as I've done in the past.
Another enjoyable teardown! Thanks again Clive.
DiodeGoneWild has used strips cut from black plastic garbage bags for the same purpose. The idea is to reduce reflectivity, so it's more obvious when light is actually coming out of the display as opposed to just bouncing off it.
I've had a set of 16W 2000 lumen versions used in my garage for a little more than 4 years. $10 USD each at the time. Nice and bright, and handle the wide temperature range of the garage in Minnesota (approx -5 to 110 deg F).
When you mentioned how they might apply the circuit tracks to the ceramic base, I thought back to my time working in an auto-glass factory. We used silk screening to apply the heater lines to back windows. The paint was a ceramic silver mixture that bonded with the glass under heat. The could do something similar with these lights: silk screen the tracks and then use lasers to bond them to the ceramic.
Sansi makes some pretty good lights. Very popular choice for keeping plants indoors.
That small addon PCB is so they can solder the wires with a normal iron. You cant solder on ceramic substrates with a normal iron because the ceramic immidiatly conducts the heat away. You would need to heat the whole substate on a hotplate, which is not convenient on the assembly line
So glad your covering these - I'm using the Sansi 36W on a Scotch Bonnet pepper plant and it responds very well - BUT - I've heard some folks have had bulb flicker after 10 months, so disassembled drivers on Big Clive is excellent news❣
Wow! I've never seen tracks and components mounted on ceramic before!!!!
Re the thick ceramic - this will give a substrate that is mechanically more robust, but also may increase the yield of good parts during the ceramic manufacturing process. The thinner ceramic parts are the more likely a higher percentage will get broken during drying and firing (and in use, when the finished lamp is being punted around by growers etc.) Looks like a well made unit. Cheers!
Mr. Dayman. It is apparent to me you understand the mechanics of ceramics. Are you aware of any implementation or applications of ceramic fins in sintered substrates? I am currently asking my colleagues who are unaware of these applications in scientific settings as well.
@@LeviGoldwing Short answer - no , not aware, sorry.
What a well designed and manufactured product - makes a nice change!
We have a bunch of their lights, the quality is excellent and so is the warranty. I had one of the 40 watt models fail near the end of the warranty period and they replaced it with zero hassle.
Fun fact - if you use a fiber laser to engrave on to a piece of aluminium nitride ceramic, a reaction occurs which produces metallic aluminium and will hence give you conductive tracks ;-)
I worked at a company making ceramic boards for high temperature circuits. The traces are silver powder in a binder put on with a silkscreen and then fired to burn off the binder and fuse the silver. That ceramic is REALLY hard on the machines, constantly grinding down the hardened steel guides.
I have used this type of lamps for my plants and seeds for several years, never had any problems. Also have a customer service team who you can actually use!
After you offered this review on your Patreon, I ordered a couple of these in the smart version. Looks great!
I use 2 Sansi C21BB 27watt E27 bulbs for my garage and I lights it up a treat. The ceramic on those is even more beefy but they use a driver on a little board and the centre of the bulb.
Was so impressed I bought a Sansi PIR flood light on the driveway. Built to last and not a hint of flicker.
Very informative so much so I have just orderd one of their plant grow bulbs. These look really well made and I will give a little vid how I get on with this lamp and seedraising. Thanks Clive always enjoy your vids and good info.
Ceramic circuit boards were used many years ago by Motorola in cell phones. I worked at a factory in New Mexico, USA, where Motorola had a factory, inspecting them in their green form (not yet fired) for flaws. Of course no solder tracks were on the board yet. These were about 0.16 cm thick.
I have about 20 of them in a garden lighting string up in the loft they are excellent!
Just checked at the time when i bought them (hence why i grabbed/bought them) I was getting them in quantities of 6 for £5.83 (for 6) delivered so bought 3 lots this was end summer last yr
A good reason, as you noted, for having included the circuit board is that the ceramic substrate is difficult to soldier to. The tracks, often times a precious metal like silver or platinum, are too soft and ductile to support components other than surface mount components. Additionally, ceramics are brittle, making it difficult to drill any via (blind, buried, through hole, staggered) at risk of breaking or damaging the ceramic. There is a preference for sputtering indium silver alloys with graphene deposits in scientific settings, but equipment and lab conditions make the process far too costly for general manufacturing. In higher applications, Antimony, Tin, Cu and cubic boron nitride are added for vibration resistance and "springy-ness".
If I may also add Clive, the ceramic is a little thick-ums due to the fact that, depending on the cutting technique, the ceramic becomes difficult to cut the thinner it gets. I'm betting this puck was sintered thick for packaging and the occasional bump.
Extremely well designed & produced lamp!
Clive is always infectiously enthusiastic for modern light bulb replacements. :)
Definitely interesting and very nicely designed. A shining example indeed.
it really looks like it is meant to last for once. just the amount of LEDs gives a clue about it. looks like you might have found a gem there Clive :)
1. The PCR is for Constant Power Control. It is also usable for dimming.
2. The electrolytic cap probably makes the power factor somewhat low.
Hey Clive 😊. Wow this is a very interesting design indeed. You are so funny at times that I honestly thought that bulb was going to go pop when you turned it on 😂. Have a blessed day and thanks for sharing.
Quite apart from the ceramic pcb, it's just nice to see some proper cooling slots for a mains led lamp for a change :o)
For most consumers, 'it not coming apart easily' is usually a selling point... not so much for Clive 😄
PCR is for dimming.
Ranges from 100% brightness @0.8V to 0% @1.6V
These look well made. It's a pity they are not easy to disassemble.
No! No Disassemble!
Shame you tore it apart, looks very well made, of course now we now how well it's actually made.
@@albear972 Newton Crosby !
Considering the thickness and the type of material used, I would suggest that this is a reworked 6” wafer and the tracks are plated on it in a a MEMS foundry.
Sweet. I was looking at this company for a room's lighting. They dont seem that scammy nor that cheap in their build quality. Glad to see this reviewed and opened up.
I'm trying to source lights better than the ikea LED bulbs (looking for full spectrum and getting a split between 5600 K and 2700 K)
I just ordered 2 4 packs of the 3K color and 1 4 pack of the 5K, directly from Sansi. They are about 4 USD per bulb. The site claims the tracks are silver. Great find, thanks for a good breakdown too. Cheers!
Thank you for this teardown. I almost bought one. I like the way they think.
Have you considered playing with those extra features on the power chips?
I love how the circuit photo looks on this one - primary colors and slightly broken symmetry. Not sure if the background being blue was intentional, but it's perfect! Do you archive these photos?
I do store all the images.
Been a while since I’ve seen ceramic substrate circuits. I remember first encountering them in Sanken audio ICs like the STK433 and RF modules in Philios pagers from the 80s.
Yes yes yes finally! I've been waiting for this! I have the American version of these bulbs at my house and they're awesome but I tried taking one apart and I couldn't get the stupid potting material off . Nor can I find the chip on the internet.
Impressive! Those ceramic PCB's aren't cheap or easy to make.
They must have for sure found a way to manufacture them quite cheaply per unit, no doubt at massive scale and high initial cost.
Interesting comment about using a LED chip that was different spec to all the others to balance the voltage rating. Similar to the makeweight chocolate you may get in a tin of Quality Street (600g and still dropping).
VERY FINE PIECE OF TECH ! Nicely done ! I'll give it an 9+ !
Why the through-hole resistor?
With the rest being so polished, I wonder if there is an actual reason. Heat dissipation, maybe?
That's not through-hole, that's MELF.
Would be interesting to see if tracks get warm, since they are not copper, so probably have a higher resistance. You could also measure resistance on some of those unused tracks.
Current is low milliamps due to the really high voltage of LED strings, and power is current squared times resistance, so the resistance of the tracks doesn't need to be low.
there is a graphic on their official website that calls the tracks 'silver lines' however it's a very simple 'ours vs theirs' comparison type graphic, hard to say how technically revealing they're being
We have PAR30 LED floodlights at work which are 40 Watts / 4,000 Lumens. They have a similar shape to the one Clive is showing, but have a fan inside. They are rated 35,000 Hours lifespan. They have been operating 24/7 for 5 years and not burnt out yet !! Very good stuff.
I may be mistaken, but I think the lifespan rating on most LEDs are telling you how long it will last before it drops below 80% its initial brightness.
I've got something like that in floodlight fixtures around the outside of my house. They're a pain in the butt to service (need a big ladder to reach them over the 2nd floor) so having bulbs that will last is imperative, cost is only a secondary consideration. Currently there are 5 working after 6 years and one that just failed a week ago.
@@MrMartinSchou That is true, we actually took one of the 5 year old floodlights out and changed it with a brand new one, and for sure there was a 20% difference in brightness. But I haven't seen any single burnt out leds yet, so they are wearing out evenly.
@@tncorgi92 These do last a long time, and at 4,000 lumens they are extremely bright, but not the cheapest at 28$ /ea or 150$ for a box of six.
The fan is probably the shortest lifespan of the components, if it's a motor-driven one (as opposed to piezo or other "solid state" type). Passive cooling would be better, IMHO
Used to work for a lighting manufacturer over here in the states that specialised in horticultural lighting. We did some research with the Sansi ceramic substrate products. Interesting stuff, but stupidly expensive
Nice one Clive. That's a step forward in LED Lamp design. I hope they're available for the home soon. Cheers Clive!
They are on ebay for general use.
Interesting (also some of the comments) as my initial thought was that the ceramic substrate (A) wouldnt be thermally conductive enough and (B) not big enough to dissipate the heat, but I admit I know very little about ceramics. I've noticed generally LED lamps over 10W dont seem to last more than a year or so, which I put down to not being able to dissipate enough heat. I'd be curious to know what the running temperature is compared to more standard lamps.
Pin 3 is "Constant power adjust" or "Light adjust" according to data sheet
My guess as to the “PCR” labeling was “Programmable Current Regulation”, so that would certainly fit. Basically a way to assign a weighted multiplier to the current sense pin.
Since it has a sort of rounded 3D appearance my guess would be screen printing to apply some active alloy on that can directly bond to the ceramic. Alternatively it could be as simple as conductive ink to keep costs down, but not sure how good the thermal dissipation would be. It could be laser sintered as well, but given they are going for bulk production I would guess they are going for something fairly economical. In a way its almost sort of a moot point since they go thru the effort to make it ceramic, but then just plop conventional low temperature components on it rather than the dies themselves, nor bother to use an actual heatsink to take advantage of it. Marketing I guess...
Just ordered some 9w and 13w US versions of these (120v) and will have to see how they last!
How? Are they shipping to US? I tried ordering from the Netherlands but it's a no-go.
The manufacturer (Sansi) sells them directly both on eBay and Amazon
I like how after the dremel you didn't even bother to brush the plastic dust off your hand.
That's the production quality I come for haha.
A look into the capacitors rating for temperature and life, aswell as measuring the temperature it is exposed to, would be useful to determine how long it will last. Usually there are several different life ratings for each temperature ratings, also ripple current rating and resistance determine how much heat is dissipated in the cap.
All the Sainsburys LEDs I have ever purchased fail at about 1/4 their rated life, I' assuming they are cap failures but I havn't investigated.
I've featured the hacks that can be done on the cheaper lamps to multiply their life expectancy greatly.
1:25... Your 7 segment display needs a red filter to see the digits more clearly Clive.
I wonder if these are leftovers from a Gov contract? Unusually high build quality.
That Sansi company makes a ton of different types of LED bulbs and light fixtures. Very nice. I'm getting two new Sansi security porch lights. They can't be any worse than the ones I got that stopped working way too soon.
1:20 It's natural white/daylight. Cold white starts around 6000k.
I use these Sansi as SAD lighting because of their full spectrum output. I have a really massive one in my office ceiling. I really rate them for quality of light. I switched after my previous preferred type started regularly failing by getting the fitting super hot.
I ordered 6 of the smaller 8w lamps, they have no real circuitry on the LED board, just the LEDs. I did manage to get the front off without breaking anything though! Trying to work out if I can lower the power easily...
hello andy. i would be interested to hear what your findings are. as i have ordered 6 of the same today, and would like to do similar things... also in particular if there is any chance of a dc derived low voltage somewhere. (but probably not it seems)
Mine arrived today, and I pulled one apart to test. There are 3 parallel surface mount resistors on the rear of the main/power circuit board. These can be removed to adjust the current. In the default configuration according to my cheap plug in watt meter the lamp pulls 38mA 6.7W. With 1 resistor removed - 27mA 4.7w, and 2 removed 13mA 2.2W. With the board still inside the lamp I think it should just about be possible to clip off the top resistor with a pair of small side cutters without damaging U1 above it.
@@dreamcat4 just uploaded a video to my channel showing how to open the smaller lamp and modify it easily... Using the info Craig C wrote. 👍
@@MrKureigu07 just uploaded a video showing the mod, can you check i have cut the correct resistor?! It is still pretty bright, but doesn't get as hot! 👍 (Edit, just measured it by plugging modified lamp into my inverter to see power, it's approx 3w lower than unmodified)
@@andyreact thanks so much man! appreciate it. and yeah they do seem to have quite a few variations of design. for different type ones
Interesting lamp. The grey color reminded me of the ink used for some of the PCB making machines that use a pen of conductive ink to draw the circuit on to boards. The ink only needs to dry prior to soldering.
The way these were made in the 1960s was to run the substrates thru a printing press, depositing conductive ink. Then you fired the substrate like a coffee cup, which melted all the little silver bits into a continuous trace, and fused them into the top few microns of the ceramic surface. The results were extremely durable.
Typically the ink is composed of actual silver nanoparticles as silver is highly conductive and most of its oxides also remain conductive. Not as conductive as a solid copper trace would be, but still perfectly acceptable if you can accept some extra resistance losses in your traces and for short traces and lower currents/higher voltages those losses are often quite acceptable for many electronics. I have occasionally used such pens to repair or route around damaged traces on PCBs as opposed to routing a bodge wire as the repair just to see how such worked out long term and….20 years later those repairs still work.
That's a very nicely designed LED lamp. Wow. Huge unit, but it will probably last forever, even without hacks!
what is your profile pic from? ive seen it before but i dont know where.
@@zilog1 i really cant remember. i found it on the internet a long time ago. liked it. stuck it on my github profile, and only recently added it to my yt profile ;)
Going to pick up some, 4pack 150w EQ for 12$ USD which seems like a great deal honestly. Ill try and remember to update this when I get them.
It would be interesting to solder a potentiometer across those programming pads and check what effect it has on the output power.
It could well be Electron Beam Melting / Sintering for the tracks not laser, I've seen work like this done on Arcam systems. You could load up cassettes / trays of parts and process them very very quickly.
Easier to screen print the tracks with silver loaded ink, and simply fire the ceramic a final time to bond it to the board, as is used on so many ceramic substrates. Low cost, and good adhesion, plus reasonably conductive traces, and as a bonus a second screen after drying of the first of a glaze to cover the tracks aside from the exposed solder pads makes a good solder mask as well.
You won't believe this clive but your fire sprinkler heating video was used in my online classroom module of my firefighter training, from the state fire academy.
Specifically the heating bit with the air gun on the bulb for 4-5 seconds. I know it was you because of the phat sooty skidmark on the bench.
I'm fine with my videos being used in educational environments.
@@bigclivedotcom Fair enough. It was just pretty surprising seeing a few seconds from a video of yours in my online module.
Excellent teardown and explanation as always. Just curious if the plastic light dome will yellow over time as so,e plastics tend to do.
It may do.
it seems like a lot of air gap around the capacitor for airflow... so i wonder how hot it gets? for example an 8w cheap one, if it does not get too hot in there maybe we could... have enough room to put an esp-01 with a relay module inside, and then flash a firmware to control them from home assistant? although i suppose it also would need a stable low voltage dc power for the esp8266 and the relay coil. i am not sure how feasible this is yet
Is that PCR pin Power factor CoRrection ? .. modulate the current with the supply voltage (slightly)
I'd love to see you do a review on fancier bulbs like Philips hue to see how they compare to cheaper options.
How good they are for the plants growth? I heard the plants enjoy specific wave length, so this cold light will be waste of energy?
That's a very controversial area.
Isn't ceramic a really good insulator (thermally and electrically), and thus probably not great for heat dissipation? Wonder how long it will last before getting the black spot of death.
This ceramic is optimised for thermal conductivity.
Oh I have a ceramic lamp, and it was not cheap. Quite powerful. I’ll have to take a look at it.
i use some ceramic heatsinks (on G9 lamps and e27 ones) as HV insulators, or as ozone generator insulator
Nice design and execution. If they had low-power switch!
Would love to know how warm the ceramic base gets after running for a while.
It would be interesting to see one of those multi white bulbs. Where it changes the white tone every time you power it on and off fast. But when turned off and in normally it keeps the same tone. Normally they are 3 faze, soft/warm/day. It's not just a brightness change like old incandescent 3 phase bulbs.
Yeah I wlike like to see That too
Should last along time looks great thanks Clive
So where does the 400v come from, with a 240v input? Is that something to do with the rectifier, or..? I may have missed a part of the video, but I don't recall seeing anything about a voltage increase up to the part where you show the schematics and mention 300+v
Capacitors are normally spec'd with a safety factor - quality 12V circuits will typically use a capacitor rated at 35V or higher. With 400V being a standardized component, I expect this is the same.
Your mains supply is measured with RMS (Root Mean Square) to give an average value of the sine wave. The peak voltage is 1.41 times higher. So on a 240V supply the capacitor will charge to about 350V. A 400V capacitor gives a safety margin.
@@bigclivedotcom Ah, perfect answer! Exactly what I wanted to know, as always!
Seriously, your ability to explain things in your content is impressive but what is even more impressive is your ability to understand what others are asking and giving them an appropriate answer without going too in depth or providing information which is too broad and shallow.
Just signed up for a month of patreon, as thanks. Barely begins to cover the value of information you've given me, but I'm a broke boy so I just gave what I could :p
Thanks, yet again, for the help big Clive! Really appreciate it!!
For a change I'm impressed!
okay my version of this lamp that was bigger and looks too close in design to be a deviation - the glass cover blew up under one year. The resulting "bulb" is a flying insect death magnet. They made his version harder to covert to flying insect death tool but go buy a $9.99 electro death fly swatter and use this LED config to destroy flying insects. Tested in good faith on the shores of Lake Erie for over a decade. I assume this same set-up is still being used as I sold the property before the light source died. I felt bad for the thousands of Crane Flies that didn't adapt. No warmth for dead mosquitoes.
Wouldn't mind having a few of those to try out.
Nice little lamp, may have to have a look for listings
That housing has a well known Chinese ACL system aka "Anti Clive Lock" system
I have since worked out how to open them cleanly for hacking.
@@bigclivedotcom Don't tell the Chinese !!!
Given the construction quality I have to wonder if it'd be worthwhile to replace the bridge rectumfryer with a voltage doubler to allow use here in the USA.
The doubler could possibly be housed down in the base to provide DC directly to the ceramic panel.
Very interesting (as usual) ... and useful I've ordered some of their lamps as I'm looking for better illumination for some high speed photography.
How easy is it to solder on those ceramic PCBs? Looks very repairable either way.
Not sure how easy it would be to solder on them. They'd probably need preheated.
It's interesting you did a teardown of this, since I was /just/ looking at them the other day to replace some outdoor lamps and I thought they looked cool. The advertisement on Home Depot here says that the combination of the ceramic ring plus the hollowed bar-like base is supposed to maximize heat dissipation and I can see that. I can also see, though, that the ceramic would be subject to moisture unless something was added in production to make it less porous. (This is an interesting point considering the bulbs I was looking at are outdoor rated and the grow lamps can be expected to be used in humid environments). I regret not buying them, just because I was in a hurry at the store, though after this I'm quite curious and may get one anyway, haha.
That's incredible! It gets me thinking: could they have a ceramic material with metal pieces (like aluminum or tin) suspended in it, then with a ceramic material that could evaporate or vaporize under a laser which could let the metal particles sinter/fuse together to form traces? That probably isn't how they did it but I'm wondering if it could work, haha.
Ceramic PCBs have been around since the 1960s, and maybe even the 1950s. The standard way to make them was to simply run them thru a printing press that deposited a silver-bearing ink in the surface. Then you fired them in a kiln, just like you would fire a ceramic coffee cup. This burned off the volatiles in the 'ink' and melted all the little bits of silver together so you got a conductive trace. The silver ink was black colored in every instance I've seen, and I noted that these tracks are a grey-black. It is possible they are using aluminum or some other metal than silver, but likely doing it the same way.
Print and fire, not laser sintered; the latter would require long laser times on each bulb and registration of the substrate. The temperature under operation would be interesting. The surface area is small compared to what's achievable with an aluminium heat sink which could fill that empty space in back of the plate.
I found them 'ceramic heat-sinks' to be useless for cooling. the 9W variant still makes a desk lamp to hot to aim at what I'm working on (I measured over 170F in a 70F room after an hour of the bulb being on, and it did throttle back to under 6.5W at that point, and was still getting hotter and thermal throttling more).
4:30 ish, the ones I got, have a sorry attempt of motherboard VRM ornament shaping to them, it's not flat on the back side, just as ineffective. it just bakes away in the bulb.
I gave up on the Sansi bulbs after a few hours of looking at them, and went back to fussing with the aluminum backed PCB ones. the sansi bulbs are all gimmicks at best.
A good desk lamp design with LEDs doesn't use retrofit bulbs anyway. You need directional light, but the retrofit bulb's diffuser goes out of its way to send light everywhere, so then you need an inefficient reflector to re-direct the light. Which then cooks the bulb by trapping hot air. Plus so many watts worth of heat are only needed only because of the inefficiencies from the diffuser and reflector.
A safe low voltage design with a good exposed heatsink is probably the way to go.
@@rexsceleratorum1632 yeah, I have a few vented back basket-like desk lamps, and I have removed that frosted plastic cover on a aluminum backed PCB bulb, drilled some holes in it's base to let air in behind the PCB, etc. much cooler operating than the Sansi ones even at the same mA/Watts/Power. the Sansi ones just produce way more heat than it should for a 'LED' bulb, and the ceramic heatsink thing is just a fail, lol.
@@Zarcondeegrissom Here in India the mains is a spicy 230V, so I don't want a desk lamp with an exposed mains PCB or even exposed mains-adjacent heatsink.
Also any sort of holes are a bad idea in the tropics because it instantly becomes a bug trap.
Dear Big Clive, another random question (i am subscribed to your channel), I am buying a V-tac led spotlight for use over an aquarium. Its 20watts, But these things burn out usually (never had a v-tac one before though), Do you think using a capacitor inline with the live wire (resistor in parallel) would increase longevity of the spotlight ?
It depends on the light. I use V-tac "grill lights" in my live stream area, but they are all hacked to lower their power.
Finally! An LED lamp that’s built to last
So is this a direction other led makers should follow or will it increase the cost to much to sell well?
The cheaper disposable lamps will always exist.
Nice one Clive. Thanks. Looks 3d printed to me.
The Spec Sheet shows a Cap' across the load/leds. Inrush !?
;)