Đi đến trình phátĐi đến nội dung chính
  • 6 phút trước
Quá trình chế tạo đèn 1.5kW hoàn toàn thủ công sẽ như thế nào? Mọi người hãy xem và ủng hộ kênh nhé
Phụ đề
00:00This is one of the brightest LEDs in the world,
00:05capable of emitting over 120,000 lumens
00:09and consuming 1.5 kilowatts of power to do so.
00:14Insane.
00:15This is an incredible amount of light,
00:18but 1.5 kilowatts of power generates a ton of heat as well.
00:22And without extreme cooling,
00:24it just will literally burn itself up.
00:27My initial attempt at preventing this from happening
00:29was to use a custom built water cooling loop
00:32to pull heat away from the LED as quickly as possible.
00:35Despite promising results,
00:37it drew so much power that it literally blew up my power supply,
00:41putting a stop to further testing.
00:44It's clearly a specialist component, unique and extremely rare.
00:49So in this video, I want to do it justice
00:51by building a better engineered rig for it
00:54that can not only keep its temperature in check,
00:57but also power it portably
01:00so that we can, for the first time ever,
01:03see what this LED is truly capable of.
01:06We're in for a treat.
01:09So I'm going to need to go to extreme lengths
01:12to stop this LED from burning up.
01:14And by my calculations, I'll need six full copper radiators
01:20to dump its heat in two.
01:22These are really intended for water-cooled PC builds,
01:25but they should provide a huge amount of cooling capacity for the LED.
01:29And as I have six of them, they can make a pretty cool hexagon shape,
01:34which I think will look amazing for the build.
01:36To mount them together like this, I'm making my own custom brackets out of brass,
01:42bending some tabs over to get the required angles.
01:45I've decided to use brass here because I want this build to be aesthetically interesting,
01:49and brass is always a great choice for this.
01:52Now, as each radiator has room for three fans,
01:55I'm going to max things out, giving us 18 total fans,
02:00which is absolutely wild, and we're only just getting started.
02:05At this stage, it is a little bit wobbly, but thanks to geometry,
02:09this is easy to fix with three internal bracing rods,
02:13making it an extremely solid framework to base the rest of the system around.
02:18Now, to begin the water cooling pathway,
02:21I'm just using some rubber tubing to link each of the radiators together
02:24in series, meaning that water will flow out of one and into the next
02:29until it makes its way all the way through them,
02:31with two ports left open as the in and out ports.
02:34Next up for a large water cooling system like this is usually a reservoir,
02:39which stores additional coolant and provides a trap
02:41to capture any remaining bubbles that are left inside the loop.
02:45While there is just enough room in here to squeeze a normal PC-style reservoir,
02:49I want to make my own bespoke reservoir
02:52because I want to have it perform an additional cooling role that a normal reservoir is incapable of.
02:57You see, my LED is split into six individually powered arrays,
03:03each one requiring 250 watts to reach full brightness.
03:07This is a very clever design because it means that the power draw can be spread over six individual
03:12power boards, so they don't have to be particularly expensive.
03:16In fact, mine are simply voltage regulators with current control, and I'll be going into more detail
03:21about them later. But the main thing to note for now is that in powering the LED,
03:26they do heat up quite a bit, so I want to give my reservoir the ability to cool them down.
03:32To make it able to do this, I'm going to construct my reservoir's outer perimeter out of brass box
03:38section with triangular cuts in it, which allows me to bend it into a hexagon, one edge for each power
03:44regulator board. As I want the water to flow through this perimeter, I'm adding a brass barb fitting so
03:51that water can enter it directly. After entering here, it'll be split and flow around the outer perimeter,
03:57and then exit through this top hole and into the central part of my hexagon, which I want to convert
04:03into a water chamber. It will of course need to be watertight for this, so I've cut out a brass
04:07plate
04:08to act as a cap, which too can have its own brass barb fitting where the water will exit.
04:13The beauty of using brass for these parts is that they can be soldered together to make them watertight,
04:19and it's such a cool way of making things like this without complex tools.
04:24Now, I'm definitely not an expert, and I tend to use too much heat, which results in a lot of
04:29charred flux,
04:31but after some cleanup it looks okay and appears to be well sealed.
04:36As for the front, it too needs a cap, but instead of soldering it straight on top and blocking my
04:41view of the inside,
04:43I want to cut a hole in it first to make a window where we can later observe the water
04:47level through a piece of acrylic.
04:49To provide a stronger clamping force for this acrylic,
04:52I'm adding some PCB pillars in each corner for some screws to do this job.
04:57Now, again, it is pretty messy, but after some cleanup it looks so cool and is a reminder that
05:04you don't need expensive tools to make something interesting.
05:07Anyway, to add the window and finally make it watertight, I've got a little gasket ring to go
05:13around the opening, on top of which my acrylic glass can sit. It is quite a thick piece to make
05:19sure that it's nice and strong, and when it's clamped down onto the gasket it makes a really tight seal,
05:25which I confirmed with a pressure gauge.
05:28As I mentioned earlier, the intention is to have this reservoir cool down my six voltage regulator
05:33boards, and as they each have a metal bottom plate into which they dump all of their generator's heat,
05:39just having each board touch the outer perimeter rim through some thermal pads
05:42is all that it takes to wick away enough thermal energy from them to keep them properly cooled,
05:48and it's a good compact way of getting both a reservoir and power system combo.
05:53As for mounting it inside the outer shell, well, I don't have any proper metal bending tools,
05:59but using a vise and a hammer I've been able to wrangle a sheet of brass into a hexagon shape
06:04to act
06:05as a cover, with some brass PCB pillars soldered to the front to act as mounting points.
06:12This basically forms a front facade for my reservoir, allowing it to be lowered into the
06:16main shell and mounted securely, and it looks absolutely insane. This is going to have a wild
06:23aesthetic when it's finished. So with the core of the water cooling system now in place,
06:29it's time to mount the LED. One of the problems though with water cooling this particular LED is that
06:35it is just so physically large. I can't find any water block that's big enough to cover
06:41its entire backside and keep it cool. Thankfully though, last time I tested this thing out,
06:47I actually built my very own large format water block to do this very task. I made it by getting
06:55a solid copper heat sink and using brass to make a watertight chamber around it. This forces the water
07:01to flow through the heat sink's fins and is incredibly effective at pulling vast amounts of
07:06heat away from the LED. Getting this ready, the LED itself needs to be thermally coupled to the block
07:12using some thermal compound and clamped down with some pretty thick copper brackets. If you've been
07:18wondering how I've been making all of these individual pieces by the way, my new method is to design them
07:24in
07:24some CAD software and then print them as a thin dimensionally accurate 2D layer on my 3D printer.
07:31This gives me a template that can be used to precisely mark out the sheet material including
07:36any hole positions, allowing it to be simply cut out with a jigsaw. I love this method because it's a
07:42rapid way of making these simple 2D shapes and if care is taken it's possible to get them more than
07:48precise enough for projects like this and it's a fun, fast and cheap alternative to using a CNC,
07:54in true DIY perks fashion. Anyway, to hold it in front of the cooling unit, my plan is to use
08:02some
08:02copper pipes to make a supporting framework for it. By screwing these pipes to the radiators it allows me
08:09to use 45 degree elbow joints to pull the pipe ends together into a point, which then in turn go
08:15into a
08:16little custom bracket made from end caps. The beauty of using copper pipes like this is that all of the
08:22joints can be soldered together just like the reservoir, making for a very strong supportive
08:27structure, one that's plenty strong enough to hold the LED and its water block. Oh yeah, this is looking
08:35really cool. So with the LED in place at last, it's time to hook it up to the water loop
08:41and get it filled
08:41with coolant. To ensure the longevity of the loop, I'm using a pre-made coolant mix that includes
08:48corrosion inhibitors and a biocide, so it should last for years before needing to be replaced.
08:54It's worth noting too that I've closed the loop at the back with a water pump that pulls the water
08:59directly out of the central hole in the reservoir, and as you can see, the area above this hole is
09:04used
09:05to capture any bubbles and keep the system operating efficiently. If this area gets too full of air,
09:11which it did several times during the filling process, I have included a little bleed valve
09:16to get rid of it. This is simply a threaded hole in the back of the reservoir with a flat
09:21head screw
09:22and mini rubber gasket to keep it sealed. It's rudimentary, but it's surprisingly effective.
09:29So at this stage, I could technically power the LED on and test it out. However, that would be a
09:36little bit risky. You see, if the pump fails or the water stops flowing for whatever reason,
09:41the LED would very quickly overheat and potentially die. And as it's a $1,300 part, that would be a
09:49very
09:49expensive mistake. So what I need to do is build in some sort of protection methods that will cut the
09:55power to the LED if it detects anything that would harm it. And to do that, I think that I
10:01can use
10:02a little programming board. So in an effort to protect the LED from burning up if the cooling
10:08system were to ever fail, I'm going to use an Arduino programming board to monitor three different
10:13sensors. A water flow sensor, a thermal sensor, and the pump's rotation sensor. If all three of these
10:22sensors report data back that falls within my set parameters, my Arduino board will trigger a bank
10:27of relays that allow my voltage regulator boards, through a sensor line, to each start outputting
10:32voltage to the LED. In any other scenario, like the temperature of the coolant being too high or
10:38the water not flowing quickly enough, the Arduino will deactivate the relays and the LED will remain
10:44off. And it fits all inside the unit on a simple 3D printed part. So things are looking really good
10:52now, but how on earth am I going to power this thing on? Well, a massive LiPo battery should do
10:58the
10:58trick. It's intended for high power RC models and is capable of delivering approximately 5kW of power,
11:06which is plenty for my 1.5kW LED. As this is basically a pack of raw battery cells though,
11:12it's necessary for me to complement it with a battery management board, which will protect the
11:17cells from any dangerous scenarios like short circuits or over discharge and excessive current
11:22straw etc. As the cells only have a thin layer of shrink wrap around them though, physical protection
11:28is required as well. My first attempt at this was to make a shell out of aluminium. While it protects
11:35the battery nicely, it does look horrendously out of place and completely destroys the build's vibe,
11:42like seriously, what on earth was I thinking? Anyway, scrapping that, I'm switching back to brass,
11:49a simple folded perimeter, soldered straight onto a flat base. Sanded and polished, this looks really
11:56quite smart, with the base seam not even visible. This shell should protect the battery cells nicely,
12:03and by joining my copper pipes together on the back to provide a mount for it, it ends up looking
12:08so much better than my previous attempt. Now a cool extra that came with my battery management board
12:15is this data screen that can display various battery parameters alongside a power button. Pressing this
12:21power button activates the battery, and we can see its charge state on the screen as well as the active
12:27current draw, which is presently 1 amp. This current draw is mostly the fans and water pump, with a
12:34little for the Arduino. Now thankfully, all of the sensors that the Arduino is monitoring appear to
12:40be reporting values within my set parameters, as the Arduino audibly flips the relays into their on
12:46state with a satisfying click. But the LED remains unlit. This is because the voltage regulators,
12:56which provide power to the LED, depend on adjustable trimmer potentiometers to set both the output voltage
13:02and output current. And as I don't have access to these potentiometers, what with the boards being
13:08buried inside the shell, I earlier decided to take them off and extend them with wires to make them
13:14adjustable remotely. However, presently these wires that come off the voltage regulators aren't
13:19connected to anything, they just form an open circuit, which means that each voltage regulator
13:24defaults to its lowest possible voltage, hence why the LED doesn't even illuminate. So the plan is to hook
13:31each one of these extension wires up to its own new external potentiometer. Each one of these will
13:37control the output voltage of the board it's connected to, allowing them to control the LED's brightness.
13:44To keep things neat, I've mounted them inside a little 3D printed shell with a brass cap to match
13:50the aesthetic of the rest of the build. And this is where I thought I'd have a little bit of
13:55fun.
13:56You see, while it is necessary for me to have six individual potentiometers, one for each powerboard,
14:01it isn't necessary for me to have six individual knobs for them, because I can simply link them
14:07together mechanically using some brass gears. These make the potentiometers rotate together,
14:13and although the rotation direction is inverted on half of them, this doesn't matter at all because
14:19potentiometers provide two points of reference for resistance, making their direction of rotation
14:24irrelevant so long as they're wired up correctly. To add an overall control knob for it, I'm including
14:30a couple more cogs. One smaller one to gear the thing down, and a much larger one for the knob
14:35itself,
14:36purely for aesthetics and feel. This looks so cool, and while it is perhaps a little bit eccentric,
14:44you've got to have a bit of fun with the design sometimes, and this certainly did it for me. So,
14:49after screwing it tightly in position and hooking it up to my loose wires, it really finishes off the
14:55build nicely, and provides a sort of weighty and significant feel for controlling brightness,
15:00which is just as well considering how bright this thing will be getting.
15:04Speaking of weighty, the whole thing weighs about 25 kilograms, and I'm not going to lie,
15:11this is pretty heavy for me, so to make it a little bit easier to handle, I'm adding some leather
15:16straps that go to a copper pipe grip, which does help quite a bit.
15:24So, with the brightness control in place, the LED will actually illuminate once it's gone through
15:29its initial start-up checks, as we can see. However, before testing this out at full brightness,
15:35one problem that I'm foreseeing is that this is completely unfocused light, which means that
15:41it'll illuminate close subjects very brightly, but anything far away will appear much dimmer,
15:47so I need to pull those light rays and throw them forwards using some optics, which will help it to
15:54illuminate distant subjects much more brightly. As the surface area of the LED is just so large,
16:00I'm using a deep half-sphere lens for this. It's not perfect, but the light from the LED is
16:06significantly more focused now, but before I add another lens to focus it yet more, I want to
16:13have a little experiment. So, I've got some matches here. Will it light it? So, I've got a pen,
16:21just in case I need to make it black, but let's try it without the pen.
16:26Even with sunglasses. Oh my goodness! Oh my goodness, it's smoking. Will it light? Spontane?
16:34Oh, no way! I have just lit a match using light. That is crazy. Wow. What have I built?
16:56What's so remarkable about this is that we aren't observing the effects of infrared heat energy here,
17:02because the LED barely emits any infrared radiation at all. No, what we are seeing here
17:08is visible light photons at such a high intensity that they themselves cause the heating. Wow.
17:15Now, the second and final lens that I want to add is a glass Fresnel lens. This has a much
17:21greater
17:21diameter than the first lens and helps to pull the light rays together even further.
17:27So, with this lens system in place, it should illuminate distant objects much more brightly.
17:32How brightly? Let's go and find out. Oh, it's heavy.
17:39So, I've come out here to the forest, and just for comparison, I've got a parabolic flashlight here.
17:46Generally, it's pretty bright, but as you can see, it's light, but it's not crazy. Let's get crazy.
17:57Are you guys ready?
18:01Let's light this up. Oh, my goodness. Oh, my goodness me. Look at that. Oh, my goodness. Wow.
18:13I've never seen a light like this before. That is wild.
18:17Let's do this.
18:47It looks like this before…
19:07the amount of light bursting forth from the led is truly spectacular and just to give you some way
19:14of contextualizing it here is a shot of my car's headlights on high beam these are those stupidly
19:21bright led headlights that really have no right to be on the roads but they look like cheap dollar
19:27store flashlights in comparison to my led supernova it's just in a completely different league
19:33absolutely wild so my dream of utilizing this one-of-a-kind led is complete
19:43so those are absolutely mind-blowing results i cannot believe how bright this led really is i've
19:52never seen a light quite like it absolutely incredible however how well is the cooling
19:58system actually performing at keeping the led cool and how long does it last for on the battery
20:05well both of these questions can only be answered with a stress test and i'll be honest with you
20:11this is going to be a nerve-wracking experience because having the system powered on at full
20:16brightness for as long as the battery will last will push everything to its limit if the cooling
20:22system can't keep up the temperature will gradually increase until the led explodes which would be a
20:28very sad end to this project however i don't think that that's going to happen so let's give it a
20:34go
20:35the first five minutes are the scariest because up until this point i've had the led on for only
20:41perhaps two minutes at a time in extremely cold weather so anything above that is uncharted territory
20:47particularly as we're in a warm internal room my first observation is that after this five minutes
20:53the system becomes almost fully heat soaked with the coolant temperature at 32 degrees and the fans
21:00ramped to 74 percent i can feel the temperature of the room getting hotter and hotter as i do this
21:06test it's uh genuinely slightly uncomfortable and a little bit nerve-wracking as i said um this is
21:13scary this fear though is thankfully unwarranted for the remainder of the test the coolant temperature
21:18only creeps up a single digit and there's plenty more cooling capacity available as the fans are far from
21:24their maximum speed the battery however is showing a consistent and unchecked increase in its
21:30temperature and at the tail end of the test it's likely at least 50 celsius internally which is
21:35approaching danger territory for a lithium battery it's going to manage it but barely
21:42look how long we've been going for it's almost a quarter of an hour that's actually wild wow
21:51after a total of 15 minutes the battery is depleted and the bms clicks off not a bad runtime considering
21:58the consistent 1500 watt draw it's off it's off it's reached the low point let's check the thermals
22:06amazingly the maximum temperature of the led's thermal substrate was 65 degrees celsius 65 celsius
22:14this is well below the maximum allowable declared on its datasheet and it's really very impressive given
22:20the amount of power flowing through it so i'm very happy to report that the led is now being
22:26successfully cooled and is ready to go so i hope you've enjoyed seeing this led out in the wild it's
22:35been truly mind-blowing to see just what it is capable of and it's been hugely rewarding to finally see
22:41it
22:41after all of these years now i actually don't believe that this is categorically the world's
22:48brightest flashlight believe it or not however i do think that there is potential to take it there
22:56so make sure you're subscribed because things are going to get wild but other than that i'm matt
23:04you've been watching diy perks and i hope i see you next time goodbye for now
Bình luận

Được khuyến cáo