EEVblog 1453 - Elgato Key Light TEARDOWN

[@mrbmp09]

They added tripod mounts top and bottom so it could be used in Austrailia without all the electrons falling out. I can't nbeleive you missed that one Dave.

[@thextrmntr]

You have 2 of them; Add thermal adhesive to one and compare lifetime

[@jaro6985]

That and thermal cam comparison would be good.

[@radry100]

The margin must be enourmous with such a simple construction for $200...

[@docwhogr]

for $200 you can buy 1 new or 2-3 used 27" monitors and leave them blank to illuminate you. i can't see this shit cost more than $50 to make.

[@v8snail]

It's the high CRI LEDs that cost the $. Cheap LEDs don't show true colours very well.

[@zwerko]

@@v8snail 1M of dual 3000-6500k, CRI≥95 high-quality, high-output strip costs $25 retail, probably half the price if ordered in bulk from the manufacturer. And they are using ~80cm of lower quality one. High CRI does not explain the super-high markup here...

[@RobertSzasz]

@@v8snail Osram doesn't even offer a high CRI product (they top out at 95, the expensive high CRI LEDs right now are around 99)

[@redtails]

@@zwerko now add a PSU that doesn't flicker on camera

[@RickTheGeek]

I’m curious as to what secrets can be revealed by the UART and debug ports. Although I must admit that I’d prefer a couple knobs to set the brightness and temperature as opposed to having to mess with the wifi, especially if you’re going from place to place

[@amahashadow]

UART is connected to the wifi board/ARM, and the specs says it's configurable, so they might have something implemented on it since they bothered adding pad to it, (or futureproffing of the board ?). I guess it could be used for inital programming in factory, but the FW or DEBUG would be better suited for that.

[@stuartcastle2814]

@@amahashadow While I love things being app controlled (particularly home entertainment stuff - no remotes to lose as I always have my phone on me), there are times when I think being app controlled complicates things too much. Small scale studio lighting systems is one. That said, it might be handy if you have to mount the lights in a very difficult to access place, but even then, if you are having to run power to the light in that place, you may be better off running a wired network to it as well.

[@pengiswe]

I'm surprised not to find any inductors, to make a constant current source. For what I can tell, those are just PWM with a dropper resistor, given that circuit board. It would be really interesting to see some oscilloscope traces from the current curve (or voltages, since less intrusive to measure) to the LEDs, both to see how well it regulates brightness, and how much trouble it will give for a camera with rolling shutter. I've looked at that key light myself for quite long, but haven't motivated to get one since I'm not that convinced that it's better than others. Looking for tear downs on others, to see if there are any with proper flicker free current regulation.

[@pengiswe]

Current measurement is quite easy though, since there are series resistors to measure on. But thinking about it, having multiple channels (two per color temperature), could the flickering then be mitigated, but not entirely removed, by driving them out of phase? Would be really interesting to know.

[@adult-songs]

PWM is preferable way to regulate brightness for high color fidelity leds. Other way will result in small color shifts.

[@pengiswe]

@@adult-songs was thinking about that top. But how would that hold up when using high speed shots and rolling shutter? The line speed in 1080p and 60fps is about 64800 lines/second, and that's kind of not pushing anything today? The pwm chip seems to by just above that frequency... I know the idea about rolling shutter is that it has about a frame to collect photons, so it's lot longer time, but we are still working in the same order of magnitude

[@adult-songs]

@@pengiswe You are comparing wrong things.

[@pdrg]

There's a place for hard shadows! Key is usually to throw a shadow for texture, Fill would be less intense to control those shadows, then Back just to help separate you from the rest of the scenery. Anyway Key, Back, Fill is the "three point setup", utterly classic :)

[@EEVblog]

This is KZbin, not Hollywood, so most creators just use the key light in front. Good enough for Australia.

[@irishpotatogamer1850]

@@EEVblog did you see the open source project to get rid off the spam above me here? Linus did a video about it.

[@maximilianmustermann5763]

However the direct LED panels don't throw a hard light, they are hundreds of little hard lights and they produce extremely weird multi-shadows when you use them undiffused. For a good, hard light source you'd need a single-source LED light (Aputure makes those, too) The direct LED approach in LED panels is usually done to give you more output. You can always diffuse them heavily if you need diffused light, but you can take the diffusion of for more power. As LEDs become more powerful, the bigger panels are increasingly all made as side-lit diffused panels because it just looks very nice. The small on-camera light by Aputure (the Amaran F1 he showed at the end) on the other hand is really bright for such a small light - and sometimes that is just very useful in an on-camera light.

[@LordDragox412]

@@irishpotatogamer1850 It's pretty sad that creators need to use things like that because Google, a billion dollar company, is too greedy to deal with thotbots because they see them as profit...

[@tuttocrafting]

1mm thermal pad all around was easy to install and would increase the lifespan of the LEDs.

[@jackykoning]

They could have done a squeeze system with a single screw and added a small bit if thermal paste or ~0.5mm pad then insert the LEDs and tighten the screw. I actually think they did that on purpose though... the aluminum slightly increases in size as it heats up. So the aluminum on the leds will grow until it touches and then dump the heat until the outer aluminum grows. This will dance for a little while and then settle. LEDs shift color slightly as they heat so its done to make the temperature more stable over time.

[@tuttocrafting]

@@jackykoning Thats why I've said to use a thermal pad and a not rigid solution. Pads are a bit flexible and would handle the expansion quite decently. Probably with a 0.5mm is even better. But with an aluminum pcb the total length would not increase a lot considerign a final temperature of 70°C (starting a 20°C) and a pcb length of 60cm. The final length would be 60.06cm. Not so much to fill those huge gaps.

[@lolilollolilol7773]

It would increase their lifespan by an order of magnitude probably.

[@jtb2586]

1:32 A device that needs a dedicated reset switch, no thank you.

[@MeppyMan]

I’ve not had to use that switch at all since I first set mine up.

[@SianaGearz]

I think it's reassuring to see a reset switch, i actually don't trust some kinds of devices which don't have one, but i'd like it to be recessed and requiring a pen or paperclip or something to push down, indicating clearly that this is not something you need on a day to day basis, but an emergency provision. Because once in 10 years, the device misbehaves, what do you do then? It's not needed in devices which have no configuration storage and where you can claw out the battery or disconnect them from power. In devices which do store data, the reset switch should bring them back into factory configuration.

[@macd7743]

@@MeppyMan And I've used mine about 1000 times. 🤣

[@AnthonyGoodley]

For what Elgato charges for these lights they should be ashamed for the lack of thermal adhesive. Considering how much of a difference keeping LED's cool has on their lifespan this is very disappointing. An obvious upgrade for any tinkers is.... . Well it's obvious so I'm not saying it.

[@iamdarkyoshi]

Am I being too pessimistic? Personally the actual design feels pretty terrible for how much they cost... No constant current switchmode driver to keep the efficiency up, a half assed heatsink, and no manual controls for when the wireless BS becomes obsolete and/or stops working. Two potentiometers would have accomplished the task just fine, then maybe use the wireless to override them if you want application control. I'd be curious to see what temperature the LEDs actually reach at full tilt, and then tested with just a basic glue holding them firmly to the heatsink frame.

[@EEVblog]

The main market for these is desktop application control integration with their popular switch deck thingo.

[@DanielLopez-up6os]

For color Temperature a Potetiontiometer isn't that great when you wan't to match a Bunch of them to the same Color Temperature.

[@ortzinator]

@@EEVblog Elgato is taking a page out of Apple's book and making an ecosystem of products that work well together

[@lolilollolilol7773]

@@ortzinator With the price but not the quality. That's a ripoff. They spared no expense for the outside in order to overharge but the inside is full of petty economies. They didn't even make the software right as the wi-fi control is not properly secured. We must thank @EEVBlog for this teardown and hopefully calling Elgato out, because this product is all smoke and mirrors.

[@zwerko]

​@@DanielLopez-up6os Manual digital setting then. Or a simple sync/link interface (albeit that wold probably end up being proprietary and work only with lights of the same manufacturer). Not saying that software controls are without their merits, but having on-the-light backup for when they are not available would make the lights far more versatile and future proof... Then again, no reason to future-proof something of this build quality-the resistors, caps and LEDs in this setting will not last long anyway.

[@reilleysinventions4155]

Power resistors packed tight are not good for the rated power! Dave, you did not comment on that!

[@_PatrickO]

Airgapping the heatsink has one purpose and one purpose only. Purposely shortening the lifespan so you have to buy another one in 2 years. That air gap is no accident. People who buy these will want to install a thermal pad or adhesive around the edge to fill the gap.

[@drkastenbrot]

Its not as dramatic as you think. That air gap will certainly result in a significantly higher die temperature, but the biggest gradient will be the transition to ambient. Power supplies commonly have a pretty significant air gap between heatsink and plastic case and it performs pretty similar to having them in direct contact. There is just not that much heat here (look at the heatsink area vs air gap area).

[@sleepib]

@@drkastenbrot It's way more dramatic than you think, the thermal conductivity of air is several thousand times lower than aluminum, at least at the case to ambient interface you're getting helped out by convection. There's also a lot more surface area exposed to air outside the case than there is between the aluminum pcb and the case. A led is also a lot more temperature sensitive than say, a power mosfet.

[@lolilollolilol7773]

@@drkastenbrot "There is just not that much heat here (look at the heatsink area vs air gap area)." You are wrong. At full power the LEDs will heat quite a bit, and it can even drift their color temperature. The air gap is an engineering mistake. I don't know if it's on purpose but it's a bad design, that's for sure. It's quite incomprehensible as well as the back is made of die cast aluminum to act as a heatsink, which is way more expensive than plastic. So indeed it could very well be planned obsolescence.

[@drkastenbrot]

@@lolilollolilol7773 Look, I even gave you a reason why. Look at a cpu heatsink. The cpu to heatsink area is maybe 9-16 cm2, while the actual heatsink has a radiative surface of well over 2000 cm2. This ratio of *125* means that the small heatsink interface needs to be very thermally conductive to permit good heatsink performance. The heatsink shown in the video seems to have a ratio of less than 3. Having an airgap as the thermal interface with that low of a ratio will not ruin the performance of it.

[@drkastenbrot]

@@sleepib I literally do this for a living, and I say its not that big of a deal because it isnt. Go buy one and do a thermal analysis on it. Wouldnt be surprised if that heatsink is more cosmetic than necessary.

[@clemensruis]

It's quite remarkable that they can get 2900k - 7000k out of LEDs with 3000K and 6500K

[@Jonas_Meyer]

If I remember correctly LEDs change there color temperature with PWM so its doable. Also 3000k and 6500k might refer to the typical color temperature at a specific operating point. But I wonder how they get away without some sort of temperature correction or if they just assume heated up LEDs.

[@maximilianmustermann5763]

@@Jonas_Meyer I guess you're overthinking this. The panel just goes from 3K to 6.5K but 2900 to 7000K sounds better. The difference between 2900 and 3000 as well as 6500 and 7000 is really small, too. They just reached out a little to get the 2 and the 7 in these numbers I wouldn't really trust these numbers in any film light under at least 1000 dollars. Maybe if an Arri light for 5000 dollars a piece says "this is 3550K" then you can trust it. Because they're selling it to people who really care about the difference between 3500 and 3550. But that's why it costs a few grand. And this Elgato is not even in the film light category.

[@zwerko]

In theory, the diffuser could alter the color temp beyond that one of the light source (like quantum dots that Samsung uses for their QLEDs) but then their lumen rating would suffer (can't get more energetic light out of thin air). I somehow doubt that's the case here...

[@maximilianmustermann5763]

@@zwerko I don't think the diffuser could both reduce AND increase the color temp at the low and high end at the same time. They're just saying 2.9K and 7K because the difference to 3K and 6.5K is minimal, probably inside the accuracy range of the LEDs, and the numbers sound better.

[@SianaGearz]

Colour temperature depends on actual temperature of the LEDs; but it is pretty difficult to reach above it without subjecting them to the hazard of burning out. Going lower is possible, but there isn't a ton of range, but maybe just enough. I do think it's more likely they decided to just handwave in the extra range, hoping that if someone were to try to nail them to the wall by measurement, they'll just claim that the actual range is within reasonable tolerance (plus or minus a few percent) from the declared one - it is after all a prosumer product, not a precision device.

[@EgonSorensen]

A right to repair would quickly separate the 'sheep from the goats' players in various industries. $200 for *that* - no thanks - and is most likely another reason why some lobbyists are working so hard to ultimately prevent customers from knowing things like this and fixing 'issues'

[@vananything]

Yeah! If only we had some way to see inside this thing…?

[@EgonSorensen]

@@vananything It *was* called a service manual.. - with schematics, parts list and a exploded view of all internal parts. Serious companies would offer these for sale, and could be purchased *before* a sale (to avoid being taken 'hostage')

[@mr_gerber]

I'm not against the right to repair, but for the mass market, this product is probably not worth the time repairing. The gamerkiddies that Dave talks about are most probably not electronics competent enough to do any repairs deeper than module replacement, and a repair shop would utilize their time more cost efficient by discarding. Man hours are more expensive than this doodad. And almost nobody would be ready to pay the price for a proper component level repair. Unless we politically incentivise towards making it cheaper to repair than replace, I struggle to see how the market will regulate itself

[@EgonSorensen]

@@mr_gerber Perhaps 'gamerkiddies' can't fix items on a component level, but why stop them from learning - and hopefully being successive in a repair attempt. Imho once companies learn the PR value in publishing schematics + various documents so their items can be fixed after they go obsolete, the 'kiddie' *might* remember who helped, and which company just were after hard earned money providing a crap product which might keep failing. It's better to have bad/failed support, than none at all. There's a thing as good PR, just as bad PR. It's both PR - but the good pays off in a good way, where as the bad pays off in a bad way - for both company and customer. Try tell that to the PR consultants, bosses and accountants, and those likely having a stroke - might tell you they're into bad PR ;ø)

[@mr_gerber]

@@EgonSorensen I don't think you're right. I don't think the market segment that elgato is aiming for with this product is willing to pay the price for having a service and repair system, because don't be mistaken - it's not for free. Other types of products, sure - I'll agree with you that that balance is different, but a consumer product like this? I don't think so. (And it is a consumer product even though it looks like a "pro"-thing. It's the same with power tools. People buy the cheap crap tools, even though there are pro-tools that are repairable, you can get spare parts and service, and that will last a lifetime)

[@dosgos]

Please upgrade thermal solution in next video.

[@rud]

Yes I bought these as well. My green screen looks much better, I can keep them at lower intensity so they don’t bother me as much. I have started using them other places than my desk because my other lights (viltrox) were unreliable. So I have installed a quick release camera plate on one of them so I can move it around for other projects easier.

[@theoriginalzads]

Great lights except for the glaring security issue. You can use HTTP posts to control these with no authentication. Yeah they're just lights and not gonna destroy your network but it makes you wonder if there's other vulnerabilitiea that can be exploited and used to infiltrate a network.

[@blvckbytes7329]

Imagine building a botnet of lights, lol. But there are probably ways to do so, as this webserver runs on embedded, which I guess is more vulnerable in general.

[@martinmckee5333]

This is a general problem with IoT devices, unfortunately. Security seems to be lax for just the reason you pointed out. They're just small systems and the thought seems to be, "what's the worst that could happen?"

[@imark7777777]

Well of course they're secure just look for the s in the name IOT. Although not having Security on the HTTP request means they will be easily sniffed deciphered and can be controlled by other meme. Which means you're not stuck with it only being controlled with their app.

[@zwerko]

Now you just need to find a streamer who is susceptible to light-induced epilepsy, get on their network and you can DoS their brains via these lights 🙈

[@chrisw1462]

In most cases, the reason for the separate board is RF compliance - FCC in the US. They can change the rest of the product without having to get re-certified, or use the same board in many products with only one certification needed. (And not theirs, since they bought the board from a supplier.)

[@Blitterbug]

Relying on a push-fit and being a bit springy for thermal coupling is horrifying given the price!

[@McTroyd]

@0:30 The compliment to "key light" is "fill light." Fill does what it sounds like -- makes the secondary details visible, but doesn't draw focus from the key. @10:00 (edit) Maybe they're counting on thermal expansion of the LED PCB to fill the difference in the space? Even then, it still seems like they should have used (at least) those flexible thermal pads at US$200. @16:00 (edit 2) That Aperture light might work well for fill, or indirect key lighting. Looks like something that could be bodged together by an enterprising home hobbyist with good LEDs and a 555 timer. 😁

[@maximilianmustermann5763]

In German, the key light is called the "leading light" - it leads the whole lighting setup and is your main light on the subject. If you had to just work with one light, it would be your key light. Then you'll get a film noire look with strong, black shadows.

[@McTroyd]

@@maximilianmustermann5763 That's neat. I love the noire look, but at least here in the 'States, it's more common for a fill light to be low intensity, or dark color, than to omit entirely. Outside of maybe interrogation scenes, there's almost always fill light in use.

[@maximilianmustermann5763]

@@McTroyd Yes, of course. Going completely without a fill is a bit extreme, even for film noir. It's also hard to do in real life because your key light is going to reflect somewhere - and there is your fill!

[@McTroyd]

@@maximilianmustermann5763 Excellent point. I stand corrected.

[@electrosoundaust]

I have 4 Nanlite Compac 688 lights with manual controls for brightness and temperature. They run quite hot at full power, (especially when it gets over 30 C) but turning them down to around 80-90% they run substantially cooler. I've used them most days over the past 6 months without problems. No problems with flicker @ 24, 25, or 29.95 fps. I'm certain they are not upto Arri standards, but then they cost less than $300 each. Things are expensive here in Australia because of low volumes, and "supply chain difficulties". I think I'll avoid these, but with some minor mods they could be quite useful in the studio.

[@M0XFXUK]

It looks like they function pretty well but I'm not so sure on the build quality. Apart from the poor heat dissipation capability of the LED's I'm wondering how warm those 5W resistors get with prolonged operation and also how the "questionable quality" electrolytic caps hold up in the long term with that potential heat from the resistors within such a small enclosure. As you say they may have done there calcs but I guess only time will tell. Great vid as always.

[@Muny]

I'm still a bit confused about why they have the 5W resistors there. Couldn't you drive the LEDs with a constant current source, etc? I'm pretty sure I've seen chips that will do that with dimming capability.

[@casemodder89]

they've done their calculations to let it die after 2 years. nothing else was intended.

[@M0XFXUK]

@@casemodder89 I think you have hit the nail right on the head. There is however one other thing to consider. What is the life span of the LED's? I'd say if they are as they claim OSRAM then I would guess that the LEDs would outlast the the rest of the unit.

[@M0XFXUK]

@@Muny Me too I am also confused, surely there must be a better and more efficient way to regulate the energy to the LED's presuming that they are being controlled by PWM. Could be something to do with flicker reduction maybe? I really don't know.

[@jaro6985]

@@Muny a constant current source means either FET dissipation or a complex IC, probably cheaper for them to use the resistors.

[@ryoandr]

A few wedges at key points could dramatically increase surface in contact with heatsink...

[@jimellis5604]

You could put a white opaque plastic diffuser sheet from a broken LCD monitor or TV over your old light to make it not so harsh. I did that to my shop light over my bench and it worked great.

[@maximilianmustermann5763]

You can also get different types of diffusion gels from Lee or Rosco. They make them in steps from 1/8th to full depending on how much diffusion you need. If you find a store that sells them in smaller pieces (instead of the 7 meter rolls), they're quite cheap, too.

[@vgamesx1]

The craziest thing about the power supply for this thing is that, I bet those beefy 5W power resistors actually cost more than some cheap off the shelf $2 constant current thing.

[@ZeroInDaHouse]

They should have made the housing split in two pieces and mated them with screws, would have made adding thermal adhesive much easier too.

[@NoLandMandi]

they could use the non-adhesive thermal pad. if it was possible to turn this on without the back aluminum panel, you might be able to see the heat dissipation with a thermal imaging camera?

[@zaphodelektra960]

Nice one Dave! I also like it that you used a Kiwi term "Yeah Nah". Kia Kaha from Auckland!

[@VikasVJois]

The double-sided thermal tape (commonly used for Raspberry Pi heatsinking) might be used between the LEDs and outer heatsink, since the thermal paste can get into the optical area

[@lolilollolilol7773]

Did you even watch the video ? No you didn't. There is no double-sided thermal tape, he said it twice. There is simply no heat path between the LEDs and the heatsink. This is an incredibly faulty design.

[@RedwoodRhiadra]

@@lolilollolilol7773 I think Vikas is suggesting to add some tape as a relatively simple fix.

[@VikasVJois]

@@lolilollolilol7773 my proposal is to add thermal tape to fix the issue that Dave found. He also mentioned that the reason thermal paste might not be used might be because it can be messy and squeeze onto the front of the LEDs. So the thermal tape is a good fix I think

[@VikasVJois]

@@RedwoodRhiadra yes exactly

[@lolilollolilol7773]

@@VikasVJois I don't think it's possible because the leds are fixed on a susbtrate that looks very insulative.

[@alanmon2690]

Dave, love your blogs, they've reawakened my long dormant interest in electronics. Back in the day I worked on old computers and I mean old, with ferrite cores memory, paper tape input/output, the whole shebang. Even built a couple from components, as in transistors, diodes, capacitors, resistors, inductors and some ICs. I became a wiz at using oscilloscopes (storage if possible) to track down mysterious spikes. Anyway, I think you were wondering why they used a separate pbc with Wifi; Isn't this because equipment with built-in wifi needs to go through detailed approval testing, whereas using a previously approved module is cheaper, much cheaper (in line with the overall design of the light?) ?

[@FireStriker_]

Interesting seeing inside this, it’s a real shame that there isn’t any thermal interface material between the strip of leds and the heat sync frame.

[@DedmenMiller]

$200? Looks more like 80 or 120

[@N0gtail]

PWM MOSFETs and a bunch of power resistors in something this price? I expected some sort of constant current setup, especially for video. I've designed a couple of LED strip drivers for use with video and even in the 10's of kHz you get banding and flicker for shorter exposure times.

[@WizardTim]

Wow, that looks like complete garbage for US$200, I was expecting two proper DC-DC converters for the two colour temperatures not a pile of power resistors and PWM MOSFETS! I guess Elgato intends them to be solely used with 30 fps webcams otherwise I'm sure there's flicker. The lack of thermal interface material on those 45 W LED strips surely has to be a manufacturing defect, no way you'd design all that and charge US$200 only to skim on a few cents of thermal tape and Chinese labor. Speaking of which where's the required country of origin marking "Made in China"? (correction: it says made it Taiwan, no clue how I missed that) And what happened to the spec claiming 2900 - 7000 K when the LED strip is marked 3000 - 6500 K? Maybe they're relying on the dies and phosphor to overheat and drift the colour temperature into spec? 🤣 Might want to confirm the colour temperature of them before you go manually locking your colour temperature to what it claims. Also why is that reset switch so user accessible? It's almost like they expect it to see a lot of use... I also question how they bend that aluminum PCB given the quality of the rest of it, I was expecting a flex PCB, if they didn't calculate their bend radius right it might lead to some problems in manufacturing and life expectancy. Thanks for posting this, I was looking at buying some other stuff from Elgato but had much higher expectations of good design for the price, at least from the outside they look like US$200 lights.

[@adult-songs]

Your comment makes no sense, as lot of pro lights use 15-45Khz PWM. Even cheap Chinese panels that are much better than this branded thing use such frequencies. Lack of thermal paste is common and lot of brands do not care. Elgato certainly has no idea how to make lights, they just asked manufacturer to make them one.

[@WizardTim]

@@adult-songs The lights I have use constant current DC-DC converters (along with the ones I designed myself). I also have some 1 kHz room lights that cause no end of problems, at 15 - 45 KHz you'll still get problems with certain high shutter speed photography and high frame rate videography scenes but for webcam stuff I can see why they don't care.

[@adult-songs]

@@WizardTim How about you show all this issues for real? You'll get 40-50khz light and show me issues for video at 480fps and below.

[@maximilianmustermann5763]

The difference between 2900K and 3000K as well as the difference between 6500K and 7000K is really small and is probably within the normal variation of these LEDs. They just picked the lower and higher number for advertising. I wouldn't trust any film light under at least 1000 dollars to be that exact with color temperatures. I work with dedo lights and SWIT and Aputure light panels and professional cameras - and in the end none of them match exactly with their numbers. Even putting another lens on the same camera can shift the measured white balance in the camera by 100-200K. If you really need to match the numbers that exactly, you'll have to buy Arri lights for a couple thousand bucks. Professional gaffers still only use the displayed numbers as a reference and use their eyes and maybe a colorimeter as their final point of judgement.

[@lolilollolilol7773]

It's written "Made in Taiwan", Dave even mentions it.

[@SianaGearz]

Elgato is actually headquartered in Munich, Germany. Originally right above a really fancy Italian ice cream café, but they moved into a nearby skyscraper eventually. Elgato is an anagram of gelato, which is Italian for ice cream. They are a long established company which was making USB TV tuner cards (cable & terrestrial) at first, and was bought by Corsair just a couple years ago. But i suppose not all of the product range marketed by Corsair under the Elgato brand is managed by the Munich team now.

[@RC-1290]

I bounce one of these of the wall and ceiling (which are white) for an even more diffuse effect. Also means I don't have to squint. I use them as my general lighting, switching between day and night mode with the app. And then when I have a meeting, I set them to a specific color temperature that works well with my camera settings.

[@DumahBrazorf]

So basically you risk to reset the bloody thing every time you want to powering it up? Genius!

[@higihups]

To reset them, you have to hold the reset position for 10 seconds and it springs back (as seen in the video). You will notice that. Also most people will just leave them plugged in and switch them on via WiFi, as the software is your only way to change color or brightness.

[@barrieshepherd7694]

I suspect that, given the settings are vis WiFi (or BT?) I bet the unit resets at power up anyway. Maybe Dave can comment.

[@higihups]

@@barrieshepherd7694 I have two of those here. You connect them to WiFi once and they keep the last settings you chose, even when power is lost. The only time you would really depend on the software, is when you want to change settings.

[@keithking1985]

you could put on a defused layer from an old flat screen TV for the harsher LED one you had there Dave..

[@Arek_R.]

I'm confused, they get extra odd 13V PSU then have to use those massive dropper resistors, why? Get 12V or whatever is close to Vf of the lights and have only those small mosfets. Or am I missing something?

[@Arek_R.]

Regarding the strips not contacting the enclosure, I'm guessing they designed it such that tension of the aluminium strip keeps it right against the housing all around but once it went into mass production they didn't get the tolerances quite right and it doesn't work as intented anymore, but re-tested it and it worked fine without excessive overheating and just left it. You probably could fix it by wedging something where the strip splits so it springs out on the longer endges where the gaps are and closes them.

[@DLSDKING]

13Vin - Thats an odd voltage rating i have ever seen.

[@lolilollolilol7773]

$200 for that faulty construction is a bit of a ripoff. Never gonna buy that product that will fail in a couple of years. This thing shouldn't sell for more than $100 and I am being generous.

[@MeppyMan]

The Elgato Stream Deck is really good too. I don’t stream much, but it’s super handy for productivity and task switching.

[@Jonas_Meyer]

Maybe the strip is loose because of thermal expansion? If the aluminium PCB is heating up is its maybe pressing against the case/heatsink. Would be interesting if you get a thermal probe on these and actually see how hot they get.

[@clashwithkeen]

You might want to try Rosco or Lee gels for a diffusion on that light you say is too harsh. We use them in live entertainment lighting. But harsh light has its place and purpose as well.

[@Psychlist1972]

Security: A casino was once broken into through the fish tank internet-connected thermometer. So super important, even if you're not worried about people flashing lights during your stream.

[@twoina]

How do they mix the 7000K with 3000K and 6500K LEDs? Is is possible to drive LEDs with a negative voltage to suck some light out of the equation? ;-)

[@casemodder89]

you simply overdrive them by 50% additional current and they'll colourshift to blue ! 🤘😅

[@zwerko]

@@casemodder89 ... and if you keep at it the lights gain a whole new function of a smoke machine.

[@twoina]

Don't let the magic smoke escape the LEDs!

[@shuypnini9715]

love the ray reference at the end

[@andarthome142]

It's from the nederlands my home, my self i have the ledgo lg-e268c (2x) and thy are also a good product.👍👍😎

[@rmfeld]

Because I prefer to have a physical switches and dials for my Keylights in some situations, I am currently building a "remote control unit" using an ESP32-S2 that has rotary encoders attached and communicates with the Key Lights' web service. (If you search the internet you will find some code examples to control it.) Apart from some bugs and missing features (I want to have the color temperature follow the color of the sunlight outside for the color to match the natural light in my room) it is a working prototype.

[@evanstedman7405]

"... none of that aluminum rubbish..." That got me 🤣🤣🤣🤣 5:52 could be a very dangerous thing to do, especially if they are a photosensitive epileptic and the flashing hits the frequency that triggers that...

[@BikerWildRat]

How can they be 2900-7000 K with only 3000 K and 6500K diodes inside?

[@maximilianmustermann5763]

Magic and marketing! 😂

[@jaideep1337]

Looks like junk construction to me. Also the fact that those wires are directly in contact with the power resistors, insulation won't last long.

[@pnjunction5689]

Interesting design decisions, for sure 🙂I'd love to see some actual temperature measurements on the PCB and the case.

[@GreeceUranusPutin]

I like how they've got those big 5W resistors close enough to keep each other warm. 12:37 it's right there in the documentation Al-U-Min-Um, they got it right.

[@jimellis5604]

Maybe they made the LED strip just the right size so that when it heats up it expands to press against the heatsink. Turn it on and let it heat up and find out? I bet it becomes very tight.

[@qubitser]

My thoughts exactly

[@gregfeneis609]

The al-u-minieum frame with the fins is just aesthetics to lend the product an appearance of being heavy duty and rugged. The thing of it is, this Elgato is 85% of a modern LCD display. I wonder if a better alternative is to buy an LCD display and rig up an embedded system to make it display a certain color solid screen and then set the brightness to max? BTW, the manufacturer is out of the SF Bay area, so the name Elgato (El Gato) makes some sense.

[@rudge3speed]

Some day light manufacturers will include a power supply that doesn't dangle off the ground when set up on a stand. Making that cord longer should be possible, but the technology hasn't matured yet. The locking connector helps, but a longer cable has to be possible.

[@maximilianmustermann5763]

Professional video lights usually have a way to fasten the power supply to the light and then you can just run a long mains extension cable up there. You can't really extend a low voltage cable that much (like if you need it to be 2 meters or three meters long)

[@zwerko]

They could've invested a few pennies to add a thermal tape at the back of the LED strip for an effective heat dissipation to increase the expected operational life a few times over. Not sure how much they are taxing the LEDs to begin with to discern if heat is a real issue, but seems to me that a $200 panel light should have that covered just in case.

[@IanScottJohnston]

Threads on all side........very nice. On my light I had to cobble together a wierd bracket arrangement.

[@flomojo2u]

Really disappointed you didn't try running them for a while with a temperature probe on the LED strip.

[@Ni5ei]

No app = No play I'd never buy anything that doesn't offer phone-free control.

[@Ni5ei]

@George Mathieson Doesn't really matter. You still need an external device to control basic functions that could easily have been integrated by just 2 potentiometers.

[@orionsmith7668]

Temp feedback on that setup seems futile. Where are you going to sense it and is that the worst case spot in all units? Maybe some spacers to keep the leds against the outer shell??

[@victortitov1740]

app-only control? pwm instead of proper smooth current? bad led cooling? what an utter piece of garbage! What's the frequency on that pwm, btw?

[@gravlygravy]

I just looked up one of their light strips and it had pwm frequency of "1.5 mhz variable". I guess as you lower the brightness you could get flicker at the lowest range but probably they don't go low enough to interact with ordinary frame rates.

[@victortitov1740]

@@gravlygravy well, that's nice and high, i'm retracting the pwm frown then.

[@leonkernan]

Love the groovy etching.

[@zacharytaylor8523]

Seem like typical gamer gear Made to a decent standard everywhere but where it counts.

[@freespam9236]

I personally have light problem with elgato hardware as i have experienced some troublesome cost cutting from past. Know 1 case where out of the box capture card cable was not up for the task, causing random problems that got solved by replacing with non manufacturer provided cable. Tho their keypads are overprices - the price you pay is for their software in the end And as we see here... you are locked into their ecosystem - buy lights, download our software to control them

[@Huvada]

It needs an app? I prefer a different description “piece of software that will make you dependent on the company to keep supporting the product”

[@uwezimmermann5427]

Does one really need a core M3 to control a LED ligth, if the LEDs themselves are already controlled by an integrated LED driver chip?

[@jaro6985]

Its for WIFI, you won't get much less powerful IC than that running wifi.

[@uwezimmermann5427]

@@jaro6985 EPS8266 comes to mind... If now Wifi is really necessary...

[@lambda7652]

ooof resistor "current source" in a 200€ lamp.

[@msylvain59]

For the price, they also included a fingerprints magnet back side :-D

[@gabest4]

Shame about the lack of thermal coupling, but you can always buy a replacement led strip for very cheap.

[@lolilollolilol7773]

You are not supposed to do that. It's a huge design mistake at best, a downright scam at worst.

[@Xuzyy]

nice lights, just find the price a bit too high for the product

[@fishyerik]

So, essentially the equivalent of a simple background light for a LED back-lit monitor, for the price of a monitor with better background lighting?

[@jimellis5604]

Would it be to get rid of the wi-fi and make it controllable with just a brightness knob and color temp knob? Those drivers are tiny. It's pretty crazy how much power PWM driven LEDs can save.

[@AzaB2C]

7:24, why use massive heat/energy wasting resistors instead of high efficiency synchronous buck converter?

[@silverywingsagain]

I don't understand what part of this costs $200? Am I tripping in thinking i could build this with a raspberry pi w, led strip, a piece of sheet metal, and some frosted glass? What would that cost? $30 maximum?

[@macd7743]

I would love to know more about what it would take to replace that 2.4ghz (Realtek rtl8711am?) chip with a 5ghz alternative. A lot of people (myself included) have serious problems with pairing the 2.4ghz Elgato key lights. It would be amazing if one could solder in a 5ghz chip and connect reliability. Honestly, as it stands, I wish it instead connected via a USB cable. Unfortunately, for me it's a $120 paperweight.

[@rickjljr11]

doesn't it look like the aluminum pcb is clamped between the aluminum back panel and the aluminum frame with the 6 screws. those edges might conduct a lot of heat

[@Psychlist1972]

So Dave, would you say this looks like a good, solid bit of kit? ;)

[@kbfreedom]

Is it possible that led strip expands under the heat of running causing a closer fit between them and the heatsink? It is aluminum i think.

[@hermannschaefer4777]

Well, I'd call Stone caps prob. second tier ones. Not really bad, not really good. Like Teapo or Taicon.

[@Chriss120]

well to be honest, without the thermal adhesive they are at least easy to take apart and if needed easy to tepair.

[@casemodder89]

and WITH thermal adhesive you dont even need to take it apart !

[@Chriss120]

@@casemodder89 electronics can and will break eventually. by your logic no device should be servicable.

[@casemodder89]

@@Chriss120 by my logic the leds would last their usual 20-30K hours (IF properly cooled). and the electronics will not. so thermal compound/glue would be the answer for a long lasting device. designing something crappy and make it in theory serviceable is bad when the design forces you to service it regulary but that is hard to do if nobody supplies these alu led stripes. so you end up with a servicable device that died prematurely and no parts available. so my approach is way better eh ? EH ?

[@Correalian]

Could cut some thin acrylic to make the bound with the back of the led strips to the actual unit itself a lot better. Seems like a easy solution they could have done when making them.

[@franktkalcevic5342]

Wouldn't the heating of the LED strip cause it to expand and contact the heatsink?

[@1tzelG_]

i'm curious how ring light is inside and it's designed. just to replicate him.

[@MattyEngland]

I prefer a 600w HPS myself... dank and sticky 👍

[@WacKEDmaN]

Bright light. Bright light! - Gizmo interesting power input specs...guess it had a chunky brick?

[@rebsdioramas]

So they spend the money on the aluminium PCB then don't even bond it... Surprised the didn't just go with the bog standard flex PCB and adhesive. Could've saved manufacturing costs while still charging the same, least then the aluminium frame would be of some use. The wireless crap is completely unnecessary seeing most people will set a colour and that's that. Glad I slapped together some lights for me and friends instead of forking out for any of the Elgato crap.

[@nameredacted1242]

Um Dave, reducing the life of the light way below bare LED reliability WAS a design goal...

[@MeppyMan]

Sweet. I just bought one of these…

[@TwithGazz]

Only dogs can hear Dave when he gets upset.

[@nigozeroichi2501]

I don't like not having a manual control on the unit

[@viperwizard491]

expect 1000-2000h . same for a bit cheaper LED TV they stick led strips with basic double sided tape and they don't last

[@MatthewSuffidy]

PWM may not be great for film work because it may introduce strobe artifacts.

[@BobKnarly]

What size is the bit on the screws for the elgato key light

[@hempbear]

I'm surprised this landfill light has physical power switch and not "cloud" version

[@petersage5157]

7:31 It warms my heart to see this kind of power trace routing! My only criticism is the rotation of a couple of the connectors; I'm assuming this is for ease of assembly? Maybe I'm just too "old school," but I'm hoping someone could explain to me how it is objectively better to PWM control a couple strings of LEDs with a micro and software rather than just a 556 and a couple of pots?

[@AJ-wf1vh]

I guess it hooks into the computer so the streamer can sync the lights with different stream scenes and have them adjust automatically, in sync with each other and without reaching behind them. Other than that, much lower component count for much better accuracy, lower failure percentage. The whole WIFI pcb is probably a jellybean part that can be ordered en masse.

[@martinmckee5333]

@@AJ-wf1vh As I see it, that's the benefit - multiple lights can be synchronized to change together.

[@Mobin92]

200$ for that seems like a scam tbh... It's basically a 10$ LED strip with some diffusor and heatsink.

[@docwhogr]

hey dave you can use your dumpster dive monitors as key lights, and if you remove the lcd you can get twice the lumens. also can anyone explain me how with 3000K and 6500K LEDs you can produce 2900K to 7000K? do they think we are idiots?

[@Etna.]

Doesn't look any different than those daylight therapy lights you can get for around 30...40 bucks. These are meant to shine your face during winter months so you won't get winter depression. They work great for filming imho. Regards, Etna.