How hot are LED light bulbs?

A lot of our customers like to touch and handle our bulbs a lot – at times when they otherwise wouldn’t for standard incandescent or CFL bulbs (don’t worry, we do the same too). So one of the popular questions we get is, “wow, your lightbulbs are hot. Is that okay?”

Our VP of Engineering, Marc Alexander, takes a look at this question.

How hot are LED light bulbs?

Hot! Hot to the touch, but not nearly as hot as Incandescent, Halogen and CFL bulbs are.

LED light bulbs are one of the latest and most efficient lighting technologies.

High powered lighting LEDs generate light at a much lower running temperatures than the hot filament used in previous generation bulbs. The hottest outside surface of an LED light bulb is often half the temperature of an equivalent brightness Incandescent or Halogen bulb, and around 20% cooler than CFL bulbs.

Should I touch my LED light bulb when it’s on?

LED light bulbs should be handled by the diffuser – the plastic dome that the light shines out of.

When it’s lit or hot, don’t touch or handle LED light bulbs by the heat sink.

Heat sinks on LED light bulbs are designed to get hot, drawing the heat out of the LEDs and transferring the heat into the air.

It’s the hottest part of the bulb, and for good reason – the heat sink is designed to be the hottest part, while keeping the LED power supply and electronics as cool as possible.

Okay, but how ‘hot’ is hot?

In development and testing, we found that the heatsink of a fully lit LED bulb was around 60°C to 100°C (140°F to 212°F) depending on the make and model of the LED bulb, room temperature and airflow.

Here’s a thermal camera image analysis including some representative top brand-name samples of LED light bulbs – purchased new last week from the supermarket and hardware store. Brighter yellow is a higher temperature.

From left to right:

  1. 9W LED. Compact LED bulb (output: 600 lumens)

  2. 9W LED. Compact LED bulb (output: unspecified, looks 600 lumens)

  3. 13W LED. A19 size bulb (output: 1055 lumens)

  4. LIFX A21. A21 size bulb, full-color range, set to maximum power white (output: > 1000 lumens)

The two 9W compact LED bulbs are the hottest!

Though the compact bulbs are lower power and lower light output, there is very little heat sink area and no airflow management. The power supply electronics are sitting directly inside the sealed 86°C 186°F heat sink.

The 13W LED A19 bulb at third is a little bit cooler. It’s 30% brighter than the two compact bulbs, but it has a larger heat sink area to work with and get the LED heat out into the room air.

The LIFX A21 bulb is the coolest.

It has an airflow management design which uses convection airflow through the body, flowing both inside and outside the heat sink. The power supply and electronics sections are kept as cool as possible in this next generation bulb.

In even hotter room temperatures than the warm 28°C 82°F used in this test, or in semi-sealed enclosures, the LIFX A21 bulb keeps the lowest possible temperatures for the electronics and heat sink system, especially compared to current model LED light bulbs as tested in sockets 1-3.

Why does it feel so hot when I touch the heat sink?

Any object above 50°C 122°F will quickly feel very hot when touching it with your fingers! You will reflexively pull away. It’s a heat sink, so always remember that it is supposed to be hot.

Your fingertips and pain sensors are automatically protecting themselves from the temperature where the proteins in your skin will start cooking. This cooking is also called ‘protein denaturing or unfolding’. This cooking starts from around 57°C 134°F and up.

It’s a reflex response to remove your body parts quickly from anything hotter than 50°C 122°F, and get the message that something is ‘way too hot’.

Could I burn my fingers?

If you don’t reflexively remove your fingers within a few seconds of anything hotter than around 55°C 131°F, yes – you’ll burn your fingers.

Don’t touch the hot parts of any light bulb.

This includes nearly all lighting technologies including the new generation of LED light bulbs with efficient, lower temperature heat sinks.

What temperature is ok for LED light bulb electronics?

We can’t speak for all LED light bulbs, but good quality designs use power supply and driver components rated for 125°C 257°F.

Note that this is the internal electronics temperature, different to the outside heat sink temperature. Done properly, a bulb design can keep the electronics temperature at least 10°C to 30°C lower (18°F to 54°F lower) than the heat sink temperature.

So an LED bulb with a heat sink temperature of even 90°C 194°F could comfortably have an electronics temperature of 60°C to 80°C (140°F to 176°F), both well under the temperature rating for the electronics components.

Does the light bulb position matter?

Yes. Light bulbs positioned straight up or straight down will generally run cooler than sideways. The hot convection air flow flows past more of the bulb length, so it cools a bit more effectively. Sideways is still acceptable though and tested to be within normal operating temperature ranges.

In testing, the heat sink on a sideways mounted LIFX A21 bulb in a ceiling mounted semi-enclosed fitting was around 85°C 185°F, but more importantly, the power supply electronics temperature measured 75°C 167°F, and the driver control electronics was kept at 53°C 127°F, a good advantage.

Room temperature was 28°C 82°F for this and the CFL heat comparison.

For comparison, a similar lumen output CFL lamp in the same test was running a glass temperature of 120°C 248°F and electronics temperature of 85°C 185°F.

32°C higher than the LED based system.

What safety testing is done?

Light bulbs for the USA and all customers worldwide have been through extensive independent laboratory testing and certification processes.

This includes UL safety testing for lighting products, which includes checking the high temperature materials and components used in the light bulb’s design and production, and a sealed ‘heat box’ test.

We’re serious about our technology, so the LIFX A21 bulb has an additional safety feature beyond what’s required to pass UL and other certification and safety tests: 

The power supply is self protected and will turn itself off at an internal temperature of 105 to 115°C. It’s not possible to trip this protection in normal operation across the extended temperature range, but we put this protection in the system as an extra feature.

How hot are other light bulbs that aren’t using LEDs?

Smoking hot! Don’t even consider touching other kinds of light bulbs.

Incandescent and Halogen bulbs shown in this test were as hot as 181°C 357°F, and sections of the glass on a CFL bulb were as hot as 131°C 267°F.

Though you can safely handle most LED light bulbs from the plastic diffuser without getting burnt, do not under any circumstances touch an incandescent or halogen bulb.

And as detailed here, although the LIFX A21 bulb design runs cooler and safer to touch than this representative sample of current model LED light bulbs, any heat sink on an LED light bulb shouldn’t be touched. Even if the bulb looks good in its fitting and it’s tempting to touch and handle it!

Questions, comments and feedback? Drop us a line at support@lifx.co

Software lights up the world

The LIFX bulb has been a significant effort of hardware engineering but it’s also a significant feat of software engineering. As with many “Internet of Things” devices, the LIFX bulb is largely driven by software: the software that controls the bulb (e.g. smartphone app) but also the software inside the bulb that makes it come alive.

We asked our CTO, Daniel May, to talk tech about the software under the hood of the LIFX bulb, as well as covering some of the big questions that you’ve been asking.

So … what’s the deal with all this software?

Your LIFX bulb has much more in common with your Bluetooth speaker than your regular analog lightbulb. It’s an electronics device with chips, memory, radio and networking components. It’s pretty much a small networked computer. And like all computers, it runs software that makes it do the things it does.

The software inside the bulb – called firmware – does a lot of things: it controls the LEDs, processes all the messages from apps, it remembers what colour the bulb is, connects to Wi-Fi networks and much more. We don’t have a whole lot of space on the chips in the bulb, so there’s a lot of work to make sure it does the job as efficiently as possible, works properly and can be updated in the future.

Then we’ve also got the smartphone apps. First, you need to spend a lot of time designing the app. What does a smart bulb app do? How does it look? How does it behave? There’s no precedent for this: it’s a new category of product. And then you have to implement the app. This is complicated because the app has to talk to the lights – so both firmware and app teams have to constantly work together and be in sync. (And it’s much trickier than a standard smartphone app that just needs to talk to a web backend.)

On top of that, we’ve also been working on LIFX Cloud, our web service on the Internet where you can store your bulb details, schedule events and remotely access your bulbs. It’s a web site but it has to work very fast and scale to handle hundreds of thousands of simultaneous users and bulbs.

Because our product is integrated, firmware + app software + cloud software all need to work together to deliver a satisfying experience for our customers.

There’s a lot of moving pieces.

Yup 🙂 That’s one of our biggest challenges.

Anyone who has worked on software of any real complexity will tell you that it’s hard.

Also many of us will have experienced that growing a team can be challenging. It’s no different for us. We’ve scaled up a world-class team of engineers to tackle our workload but can’t scale infinitely, instantly. So we are constantly prioritising what we need to deliver to our customers: sometimes, making difficult decisions in the short-term to give the best outcomes in the medium and long-term.

What about all those features in the Kickstarter video? I want my features!

I really feel your pain. They’re coming!

Our goal has always been to deliver on the Kickstarter features at minimum, no questions. Our focus on this first quarter of 2014 is to keep delivering these features. All of our customers are important to us, but our Kickstarter customers are extra-special to us: they were the ones who believed in us first and it’s our highest priority to honor the commitment we made to them.

Creating something new usually means unexpected challenges and delays. We also made some big decisions along the way like creating a bulb that was much more powerful and premium than we had originally planned: so effectively, Kickstarter customers would be getting a better product for their money. We made the call to release a product that first delivered core features to our customers: so that our loyal supporters could get something in their hands now, while we progressively shipped updates post-launch.

We realised that we could use the app update features of smartphones today, as well as the capability for our bulbs to be updated. These will be standard features of Internet of Things devices, so why not use them?

Now the Android app. It’s pretty painful to use …

Yup, this is one of those unexpected challenges and delays that I mentioned. It’s hard to get the right people who can do this properly.

We had issues that we couldn’t resolve with an external developer who put together the current Android app and we parted ways just before launching the product. The Android app feels like it lacks features because, to be honest, we couldn’t complete those features at launch.

You’re dealing with a smartphone app that is not your standard app that has a simple user interface and queries a web back-end. There’s a lot of technical challenges and subtlety that’s needed to communicate effectively with the bulbs. Plus connecting this to a user interface that also has to behave in a sophisticated way.

We had a lot of internal debate about whether to delay Android app launch til we found someone who could revise the code. Again we made the call to release so customers would have something in their hands now, while we worked on fixing the situation.

It’s not an easy situation: we feel our customers’ pain – and we’re busy fixing the problem to take away that pain – but the fix doesn’t happen immediately and also feeling your pain all the way too. We’re using that to keep ourselves focused and to solve the problems.

The big win out of all of this is through all of this, we found our senior Android engineer, Jarrod Boyes, who is well into building a robust Android app. Previously, Jarrod wrote the top-reviewed Android version of the Footy Live app, also chosen by Google as one of the best Australian Play Store apps for 2013. Releasing a proper Android app is one of our top priorities and this will happen in the first quarter.

This video features our upcoming Android app and the new firmware in the bulbs, with improved messaging rate and reliability. You’ll be able to update existing bulbs over-the-air through our smartphone apps, once the new firmware is released.

Why don’t you release an API? Don’t you want people to write apps for LIFX?

Most of the LIFX team have a maker/hacker background. We love APIs. So the LIFX API has always been – and still is – a critical part of our product. We absolutely want people to write apps for LIFX bulbs.

It comes down to priorities. We have a long list of things to get done – even though we’d like to do everything straight away 🙂

Developing an API actually takes a lot of time and effort. We started working with customers and partners on preview versions of our API last year. Every time the product changes, you have to update the API and documentation – e.g. our protocol is in its third significant revision now. And there’s a lot of design decisions to be made around an API: how do you want to expose this functionality? would a developer find this or that approach helpful?

We realised that the more time we were spending on the API, the less time we could spend on the product itself 🙂 So again, we decided to deliver a product that all our customers could experience, and then start focusing on supporting our developer customers so they can create even better experiences for everyone.

People often ask how I feel about all the LIFX customers that are reverse-engineering our underlying protocol and create all kinds of apps and libraries. We love that spirit of hacking! There’s something special about someone building and extending something new on the foundation you’ve created. It gets us excited because this is the start of our community of developers that we want and need to support, as we keep growing this year.

In the first quarter of 2014, we’ll start to release API language bindings that will wrap around the protocol, simplifying a lot of complexity and bringing LIFX to a much wider developer audience. This approach will also insulate developers from future protocol changes. Our app libraries will also be open source and royalty-free. Our goal is to make LIFX available and accessible to as many customers and developers as possible.

So what’s next?

I wish I could say 😉 We have a lot of things in the works.

The best thing is to stay tuned to our blog. I know a lot of our customers already are – so thank you for your truly inspiring support. Now we have shipped, we have turned a corner and will be doing better at sharing our journey and keeping you updated.

Our main focus right now is to keep delivering those features we promised and live up to the trust that our customers have invested in us. We feel privileged to reinvent the lightbulb and create experiences that will be an intimate part of everyone’s everyday life.

Questions, comments and feedback? Drop us a line at support@lifx.co

Manufacturing begins!

Just wanted to let you know the good news that the certification wrestle is over!

All tests are now passed (pending paperwork) and we’ve started the manufacturing in earnest! Thousands of heat sinks have been forged this week and Marc is headed to China Thursday to oversee the set-up of the assembly line and installation of the end of line testing gear (more pictures to follow).

This means that as long as the paperwork all comes in as we expect and Marc’s happy with the end of line testing – the line begins, along with the rolling ship date. I’ll be joining Marc in China, so will come back with more photo’s of how it’s progressing, the exact ex-factory dates; and your tracking numbers!

In the meantime the apps are complete, and this week has mainly been about bug testing and finalising, and we’re submitting the iOS app on Thursday, with the Android app to follow very soon after. The firmware guys are even starting to feel the weight lifted too, with our soft launch / test day coming up early next week and most aspects looking under control – it’s a nice position to be in.


Whilst everyone else on the team has been working hard this week I’ve been spending a bit of time with a KBS, who came out to film a story about LIFX for the Korean version of 60 minutes. Strange, but fun, and lot of the questions gave me pause for thought on what an incredible year it’s been. We’re not done until we get these to you but we’re damn close now, and really grateful to you folks for being a part of it.


Ticking the boxes.

Just wanted to let you know that things are travelling smoothly on the certification and manufacturing front. We’ve since passed the factory’s in-house UL testing which has also been lodged for official lab verification.

The timeline is now mostly dictated by a third-party certification laboratory: we’re pushing hard that we’ll have the official nod of approval in the coming weeks. This could also be sooner, so we’ll share this news immediately as it comes to hand. Being cautious (as I’ve learnt to be through this process) I’d say 4 weeks to get all documentation. In the meantime, as we pass their sub-tests, the factory is readying the relevant, individual sub-assemblies for large-scale production e.g. heat sinks, plastic molding, LED heads.

Once we hit production, our output ramps up to 7,000 bulbs per week over the first four weeks- allowing us to get everyone’s (long awaited) order out quick smart.

On the software front, we’ve been working on functionality to allow direct control of a bulb via it’s own, inbuilt WIFI access point. Previously you’d connect directly to the bulb purely for the setup procedure of getting it to join your home/work WIFI network. This new feature will be handy if you don’t have a WIFI network (for whatever reason), and also gives us the peace of mind of having the ability to issue firmware updates directly from app to bulb.

Our other task has been ensuring app compatibility with the upcoming iOS 7 release from Apple and getting the apps into the app stores in the coming weeks. We’ve also started testing the API in a closed beta and will release more widely asap. On the business side, we’re about to start the retail discussions, as a lot of forward planning goes into this. Of course as backers you’ll get the bulbs first, and when the shipping has taken place and feedback has been received we’ll go deeper into the retail discussions. For now, it’s just making people aware we’re close, certification is nearing completion and manufacturing is about to start.

WIFI and mesh radio test setup in the certification lab’s anechoic chamber

Huzzah! EMC and Radio tests all in acceptable range.

The 200 control boards needed for the certification and sample test process.

Marc getting his hands dirty with the certification batch. 

GU10 prototype ready for certification post A21 certification. 

Now we’re talking!

We’re nearing the final bend now, and it’s hugely exciting to be at this point! Yesterday the production line fired up and produced the first full batch of C-Samples and here they come!


thumbnail-10large-8 large-9large-7  thumbnail-9   large-2



To get to this point, is incredibly exciting and I’m hugely grateful to everyone in our team who’ve been here with me in China, and the team at home for continuing to refine the hardware and the app at home.

Marc in particular has had a huge month spending all of it between Shanghai (over seeing manufacture of the control boards) and here in Shenzhen doing anything and everything required to support the production team at the factory and get the bulbs working perfectly.

And boy do they work well! This for me is the most fun part, taking them back and testing them in the hotel and having all the parts working seamlessly together. This is the part I look forward to sharing most, so stay tuned on this one!

The next big step for us is the certification process, CE, UL and we sent the control board away for FCC (wireless) certification today! In total this process will take (3-6 weeks) then the moment they’re approved we hit go on the production run and then off they go to you.

One thing we’ve done that will expedite this too is pre-purchased all the necessary components so that as soon as certification is passed we can hit go, as opposed to then ordering them all after certification. So, for us right now, amongst the excitement and jubilation at having a small batch in our hands it’s still in fingers crossed and over to the certification bodies to do their thing!

That being said, I’m absolutely chuffed to be at this point and incredibly grateful to all you folks for bringing this to life, being part of it all, and encouraging us along the way!

I could go on for hours about these past two weeks and all the steps involved to get to this point! We’ve been busy redesigning power supplies countless times; working on our GU10 downlight version and tooling; staying up till 4am each morning talking to the app guys at home as they push through one last iteration (I gave them a reprieve from my June 30 deadline by a few weeks while we run through certification); doing end of line testing and making sure each and every unit meets our quality control spec but I think the end-result is what we’re all looking forward to, and that end result is going to rock! So much so, I’m wearing this hat, and I’m still smiling.


Good to be back.

We touched down Saturday in Hong Kong, made our way back to Shenzhen and have been diving right back into production since Monday. It’s fun to be back, and you almost don’t realise how close you’ve become with a team until you see them again after being away.

Taxi's were hard to find in Hong Kong but somehow we managed.

Taxi’s were hard to find in Hong Kong but somehow we seemed to manage alright.

After a few tea’s with John (the factory owner) we started to discuss the roadmap for the next few weeks and locking in some choices like bulb colors, types of finishing and most importantly the GU10/Downlight.

The trickiest aspect with the downlight is because you have a smaller space to work with, naturally the heat sink is smaller, and as you’ll know from following along heat is our big enemy. The other tricky part is we have to fit all the LIFX technology including control board into a smaller space! The good news however is that the world’s most popular downlight is a 50w (~600 lumen) rather than our 900+ lumen standard bulb.

The reason for this is that down lights are often installed in greater numbers than standard lights. If you’ve got down lights at home, say in the Kitchen, you’ll know exactly what we mean. So with this in mind, we realised the heat sink won’t have to be as big, and that the GU10, doesn’t have to be as bright as the A21 bulb you’ve all seen. We don’t have exact lumens yet, our goal is to 50w which I’m sure we’ll be able to exceed!

This also means when you receive the backer survey you’ll be able to select not only the color of your LIFX bulb, but the type you want. The A21 or the GU10 downlight.

Working with the team on the GU10 downlight design and CAD, before prototyping.

Working with the team on the GU10 downlight design and CAD, before prototyping.

While we’re on the topic of colors, after the survey it’s been decided that the choices will be Pearl White & Gun Metal Grey, which you’ve seen renders but we’ll show you the finished versions of soon. Guy and I even headed to a paint factory nearby today to look all the finishes available and the exact colors we’ll use to create the finish.

In the meantime too, we’ve managed to navigate the tricky process of getting a local sim cards as roaming was becoming a pain, and even managed to get a glimpse of the Children’s umbrella manufacturing business they had going on as a sideline project.

The challenging process of buying a local sim card.

The Sim Card shop also makes and sells umbrellas for Children.

So, all in all, it’s early days into our visit here and I’m keen to get my hands on the solid works prototype of the A21 design this week, and start thermal testing on this GU10 design, but happy with the progress so far.

On the ground this time are myself, Guy, Jake, Dan and Marc and I’m starting to feel like the tour guide as we navigate our way around. It’s great to have a big part of our team here which enables us to get more done, and take ridiculous photo’s of me drinking what I thought was Strawberry Milk, but found out it was Coogee.

Not strawberry milk, but not bad!

Cheers LIFX’ers!

Watching the line

[Guest post by Jake]

As all the manufacturing staff were packing up ready to head home for the evening, 5 of us stood huddled around a monitor watching a tiny red line creep up the screen.  The moment we saw the line level out, smiles broke out across our faces. The line represented the temperature level for the LIFX bulb under simulation testing and we had come in under the cap.

The day started with breakfast in the hotel dining room as the 3 of us ran over what we needed to achieve for the day.  Then it was off to the factory to look at the simulation tests from the heat sinks.  Under simulated conditions the bulb was still running a little hot and required some adjustments to get it down to the required level.  Whilst Marc met with the power supply team, Phil and I ran through all the different possibilities for bringing the temperature down.

thumbnail-1  thumbnail

Chen (the CAD designer) dragged him computer into the meeting room for the afternoon so we could tune the design as we went along. The solution? We ended up slightly modifying the design to increase the size of the heat sink, relative to the enclosure.

We ended the day on a bit of a high and looking forward to more of the same tomorrow.