Review: OSRAM LIGHTIFY Smart Connected Lighting System

As we rapidly move towards an IP-connected society, with buzzwords such as “internet-of-things” and “sensor networks”, suddenly the concept of home automation has become cool again. No, we’re not talking about the old fashioned expensive, temperamental, slow and clunky power-line based systems like X10 that have existed for the better part of three decades, but instead, we’re looking at a new generation of integrated, wirelessly connected devices and cloud-based solutions.

The OSRAM LIGHTIFY smart connected lighting system is a relatively new entrant to this space, from a pioneering name in lighting, competing with the already established Philips Hue and Belkin Wemo systems amongst others. It tries to distinguish itself by offering new features unavailable on other systems, and a different user experience, as some of the existing solutions have not been well received due to their limitations and unreliability.

While the product has been available overseas, it will be launching in mid-December in Australia. Thanks to the generosity of Voltimum, I was able to receive the “kit” for review under the review challenge terms prior to the launch of the product, and as a result, I’ve been tirelessly working over a few days to fully evaluate the system before it launches.

Update: The OSRAM LIGHTIFY products have been launched, readers can obtain it directly from Voltimum here.

The Products

For this review, a kit of Lightify products was received including one gateway, one CLA60 RGBW LED globe and a CLA60 Tunable White LED Globe.

OSRAM Lightify Gateway AU

20151124-1139-0318The first item is the gateway. This is a plug-in unit which is the controller for all the Lightify lights in the system. The gateway connects to your home network and the internet via Wi-Fi, and communicates using 2.4Ghz Zigbee Lightlink to the globes.

The unit is quite light, and is about the size of a large power brick, and thus will obstruct adjacent power points. This is probably on purpose to ensure the internal antennas have adequate separation from other metallic objects.

20151124-1139-0319 20151124-1139-0320

As this unit is for the Australian market, it has the Australian plug and the appropriate certifications required for operation in Australia (regulatory compliance mark). It is also noted to be a ZigBee certified product for Light Link, indicating that the gateway should be compatible with Zigbee based globes from other vendors. The default credentials for the internal Wi-Fi AP used for intial configuration are printed in human-readable and QR-code form for convenience, and a reset button is also provided. The unit is designed for indoor use only.

20151124-1201-0346 20151124-1217-0349

The unit has an LED on the front to indicate the gateway status. When not set-up or having difficulties connecting to the home Wi-Fi network, it flashes orange. Once it has been properly set-up and connected, this changes to green.

One thing to be aware of is the clutter in the 2.4Ghz band. Operating Zigbee devices around Wi-Fi could potentially result in slight throughput degradation due to transmissions colliding, and reliability of operation can be impacted on by the use of microwave ovens or other high-powered 2.4Ghz transmitters.

OSRAM Lightify CLA60 RGBW and Tunable White LED Globes

20151124-1144-0331 20151124-1140-0322

At the present moment, the Australian page for the Lightify system doesn’t have many globes to choose from, but I suspect that will change in the future as products become approved. The international page shows outdoor and indoor luminaires and switch units also being available overseas. For now, we are left with two main choices for the home user – the RGBW globe which features tunable colour as well as tunable white, and the Tunable White globe that features just tunable white. Both globes claim 810 lumens output with a 60W incandescent equivalency. This is enough for most rooms, although some people would prefer brighter (1000-1200 lumens) for larger rooms. The RGBW globe consumes 10W, whereas the Tunable White is rated at 9.5W for a luminous efficacy rating of 81 lumens/watt and 85 lumens/watt respectively. These are not stellar efficiency figures, but they’re not bad at all especially considering that there is overhead from the control electronics and exceed those of some regular LED globes on the market today.

20151124-1145-0334 20151124-1140-0324 20151124-1144-0332 20151124-1141-0325

The side panels of the box show that both globes achieve an A+ rating on the energy efficiency scale for the EU, with the RGBW globe Made in Italy, and the Tunable White globe Made in China. Both globes claim a warm-up time of 0.35s to 60%, 20,000h operating lifetime, 100,000 switching cycle lifetime, no mercury, operating voltage of 220-240v at 50/60Hz, and operating temperature range of -20 to 40 degrees Celsius. The specifications for the RGBW globe gives the CRI as > 85, which is very good, with the tunable white offering a CRI of 80 which is better than most fluorescent tubes but not commendably so.


The rear of the box provides some sage advice – namely that the globes cannot be properly used without also purchasing a gateway device and installing the associated app. The package itself has a plastic hanger tab at the top and comes sealed (although mine was opened probably for inspection prior to shipment). A new label can be seen under the barcode showing the Australian regulatory compliance mark, indicating the required testing was completed by the importer. Again, a QR code makes an appearance on the package, allowing buyers to quickly access documentation about the Lightify system.

20151124-1141-0327The box itself hinges open to reveal the globe inside. It really does give a premium feel to the product, which I suppose, is appropriate as it could be quite pricey, although I haven’t been told what the anticipated retail price would be.


Aside from the globe, there is a fold out leaflet with pictorial set-up diagrams to help users get started. Lets take a closer look at the globes themselves.

20151124-1145-0337 20151124-1146-0338 20151124-1142-0328 20151124-1142-0329

Both globes are roughly the size of a regular GLS globe, although slightly longer, and should fit in most luminaires without much difficulty. Aside from the printing, it would not be obvious that the globes are actually any different from a regular LED globe. Regulatory compliance mark stickers have been affixed to the globes themselves.

20151124-1147-0340 20151124-1148-0341

The bases of the two globes do differ, suggesting the products are actually manufactured by different entities. The internal conductor crimp to the shell can be seen in both images, with the RGBW having a small bit of protruding wire, and the Tunable White seeing a “messy” crimp.

20151124-1149-0343 20151124-1149-0344

One give-away that these globes are a little special is that they do weigh quite a bit more than other LED globes I’ve handled. The RGBW weighs in at 143 grams, and the Tunable White at 136 grams.

As these globes contain all the required Zigbee radio and microcontroller logic integrated within them, the whole unit will need to be replaced when it fails as there are no interchangeable or serviceable parts. This can be a problem in the future, should parts for this system become unavailable and the cost of replacing the units could be an issue to be mindful of.

Upon plugging them into power, they turn on with the configured default settings, which allow you to use the globe like a regular globe in an emergency in case you’ve lost all control. This could be annoying though as it means your globes will all be on after a power outage.


The Setup Process

Set-up is as simple as following instructions for the most part. The instruction leaflet has two series of instructions, one labelled A for a new system set-up, and the other labelled B which is for adding a new globe to an existing system. The quick start guide also implies some exciting possibilities behind the Lightify system, which supports a maximum of 50 devices (which is a lot better than the 4-8 “in range devices” commonly seen with Bluetooth based units).


instructions-b instructions-sysdia

It might seem a little cryptic, but A1 is telling you to ensure you have internet access by Wi-Fi and a phone that is connected to the access point. A2 is telling you to install your globe. A3 tells you to scan the QR code, to install the app (alternatively, you can search it inside Google Play / App Store).


The QR code redirects to a webpage, which then redirects you to the appropriate URL to download the app. The app itself does have a few privilege requirements.

lightify-app-market lightify-permissionsreq

lightify-anim-setupFrom there, it’s a case of following the app’s instructions as noted by the pictorial A4. The set-up process has been condensed into a short animated GIF.

During the set-up process, you will need to create a Lightify account with an e-mail address (which will be verified). You will also need to have your Wi-Fi SSID and password so that the app can configure the gateway to connect to your home network. Once configured, the gateway stops broadcasting its own network and connects to your home network, and once your phone re-connects to the home-network and discovers the gateway, it is ready for pairing with the globes.

The pairing process requires you to power cycle the globes to make them “discoverable”. Unfortunately, it seems that this is where the first hurdle was encountered.activation-confirm-email

For some unknown reason, no devices showed up in the searching screen despite having cycled the globes as instructed. Continuing on resulted in a “broken” app which wasn’t responding properly. After force-closing the app, and re-opening, I chose to sign into my newly created account which then resulted in the app working properly, and the globes coming up immediately. As a result, if you encounter some difficulties, do try it again!

A Tour of the App

app-iconBy now, you’re probably interested in just what the Lightify system can do. As it turns out, this is a rare instance where a video could really help out, so here’s a brief 6-minute video showing you around the system.

groups-scenesOf course, the video cannot cover everything, so lets take a closer look at the app and its functionality.

The main screen of the app is the Home screen and consists of two main features, namely Groups which allow you to collect either a single, or a group of globes and control them in unison. The right side has a quick toggle which turns the group on or off, whereas pressing on the circle allows you to bring up a detailed controller to change colour (RGBW only), colour temperature and dimming and define scenes (more on this later).

The other feature is the scenes feature which is basically a list of “one-touch” presets for the groups, which allows you to restore the required colour, colour temperature and dimming level to a group at one touch.

At the top center, there is a button which turns everything on or off, and at the left, an indicator is provided to let you know how the app is connecting to your gateway. Three possibilities are possible, namely the following:

connmode-localcloud connmode-local connmode-cloud

In normal operation, it should show local and cloud “both”. When directly connected to the gateway’s Wi-Fi due to failure of the device to connect to your home Wi-Fi, control is possible with “only local connection”. When difficulties arise with communicating with the gateway, or you are not at home on your home network, the app will show “only cloud connection”, which still allows control but with slightly increased latency. The cloud-connected functionality means that you don’t have to be concerned about port-forwarding, protocols, etc in order to get control of your lighting when you’re not on your home network.

devices-menuThe second pane is the devices tab where you can toggle the settings for your devices directly. Pressing on the lamp icon itself brings up the control panel which gives you all the adjustment features.


The adjustments are shown on a planetary style controller dial. Pressing on the center icon turns on or off the globe. The next orbit has the colour temperature control slider, followed by the brightness/dimming slider, and the colour control panel. To adjust these values, you must hold your finger down and slide, making setting precise values quite difficult.

devices-ctchange devices-dimming devices-colour


devices-presets devices-favouritesOf course, the colour control slider is not available on the tunable white globe. Pressing on the heart button allows you to save the current settings as a favourite. The next page has OSRAM pre-configured presets which cannot be changed, with the page after featuring a grid of your saved favourites.

The final sub-page features a camera feature which allows you to take a photo of something, and devices-camerachoose a colour from the photo for the RGBW globe to emulate. It is also possible within the devices menu to configure the default settings for the globe when it is physically powered on.


schedulerThe third pane is the scheduler tab, which allows you to set schedules to apply certain scenes, toggle groups or toggle individual luminares for a given time range. The schedules themselves can be bypassed by toggling them on and off individually.

scheduler-setup scheduler-devices

The settings themselves are fairly simple with a start time, end time and day selection, but not very flexible. For example, there isn’t any ability to ramp the colour temperature over time to simulate a sunset.


Due to the cloud platform in use, creating, disabling and deleting schedules does take a little time as it needs to be synchronized with the server.

settings-paneThe settings tab allows you to find out information about the system and perform some administrative tasks.


For example, account lets you change the password for your account. You can also pair the system with Nest so that it knows whether you’re at home or away and apply scenes.


It also lets you choose time zone and daylight saving settings, although since these are managed by the mobile OS, it results in an incorrect time zone showing up (I am at GMT+11 due to DST, but it should be Sydney, not Vladivostok).

settings-sysupd settings-sysupd-begin settings-sysupd-inprog settings-sysupd-noupd

You can also apply firmware updates in the settings, and interestingly, a firmware upgrade was available for one of the globes. The upgrade is a very lengthy process and took me a total of 9 minutes to complete, but it didn’t cleanly complete. Instead, the bar disappeared but the system was insistent an upgrade was still available. Clicking on Upgrade again resulted in a stalled upgrade. Eventually, I power cycled everything to find that the update did actually take (the globe blinked at the end of the upgrade process to confirm this), but there is a risk of improper upgrade causing permanent damage to the equipment so do take care.

Aside from that, the settings menu allows you to purge all settings and start from scratch.

login-screenOf course, you would probably want to control your lights from more than just one mobile device. Luckily, this is not difficult at all. All you need to do is install the Lightify app on all your other devices and sign them into the same account for all of them to have control over the same globes (up to five in total). As a result, you can have your whole family controlling the house lights from their mobile devices. Generally, the last committed command to the device stands, and updates are periodically pushed to the other devices so they can see the status of the globes (but with a 10-20 second delay).

Because of the cloud-platform in use, all of the defined groups, scenes and schedules are kept in sync across your devices, so there is no need to redefine them.

I have also installed and tested the iOS version of the app, and it looks and functions almost completely identically to the Android version of the app, so it was not really worth covering again. The iOS app is designed for phone use and does not render as nicely on a tablet, and it features a slide-out panel to switch between the control screens which can be a little touchy to aim for, but other than that, OSRAM have done a good job in making sure both major platforms are catered for in a homogeneous way.

Quirks, Oddities and Concerns

From my experience with the app, it seems the features provided are still somewhat limited. For example, there is no easy way to ramp the colour temperature over time to simulate a sunset, and with the products on offer, there isn’t yet a way to achieve feedback-based control (e.g. dimming based on the ambient light level in a room). Control outside of the app is also not available, say for a desktop PC as the API has not been released, although some reverse engineered libraries have been made. Schedules are a little limiting as well, with an enforced “start” and “end” time and day of the week, making once-off “turn on x group on dd/mm/yyyy and leave it on until someone manually turns it off” difficult to schedule.

Using the system really requires a change of habit as well, in the sense that especially where you haven’t migrated a significant number of globes to a smart-connected system, you will still instinctively find yourself reaching for the light switch which only serves to bring the bulb up on its set default settings (and potentially confuse the gateway). Of course, if they had a physical Zigbee light switch, that could be installed and allow both smart and “less smart” control. This can be particularly annoying when you’ve decided to walk out of a room without your mobile device, or if you wish to quickly get the light on and off for a quick check of something, as the loading of the app can sometimes take a few moments to synchronize the groups, scenes and schedules to the cloud.

The app itself has misbehaved more than just on a few occasions. In fact, while filming the video above, several malfunctions had occurred which caused the need for a re-take. Noted issues and malfunctions included:

  • Loss of colour temperature control of one of the two globes during a scene restore.
  • RGBW globe flashing on a colour temperature command, changing to an incorrect temperature and then refusing further colour temperature adjustments.
  • RGBW globe getting stuck after colour adjustment and not restoring colour temperature from a scene.
  • Loss of power on/off control of a single globe, claiming it is disabled, where it still responds to a global on/off command.
  • Delayed command response or out of sequence responses resulting in two globes showing different states when commanded to a single scene preset.
  • Problems with schedule deletion resulting in application crashes, schedule disappearing from the list on a long press or account being logged out.
  • Scenes descriptions are limited in length, with no way to easily group or re-order them.
  • After deletion of all scenes and groups, recreation of another group can result in old scenes randomly appearing and crashes on reconfiguring.
  • Saved device presets for RGBW globe does not save colour temperature based preset properly, only saving the colour preset, resulting in restoring a colour rather than a colour temperature.
  • Deleted schedules that are not visible still continue to run without being able to be removed.

quirk-lockeddialsSome issues were so perplexing that they deserved a screenshot, for example, the RGBW globe losing both colour temperature and colour controls, only featuring dimming and on/off control. quirk-duplicatereq

And of course, getting dumped out during deleting a schedule as the sole user of a given account – the message doesn’t make much sense.

Some issues were correctable by closing and restarting the app and were generally “non-fatal”, others required power cycling of the globe, and the most insidious ones required resetting the gateway too. Other issues were correctable by manually going to the devices screen, changing say from colour to colour temperature mode, toggling the light on and off, then re-loading the scene.

I suspect that you will run into such issues more frequently during the initial “configuration” phase of the system, where the novelty will result in constant playing around with adjustments and flipping the lights on and off. Once you have settled down with a set of groups, scenes and devices and you only perform infrequent manual adjustments, the system generally behaves. However, if you plan on demonstrating the system to wow the audiences (or let kids play ruthlessly with the system), do prepare for a hiccup here or there.

From my user experience, I would also have to rate the app down on the lack of intuitive operation. For example, pressing the + at the home screen allows defining a new group, but it is unclear how one defines a scene from there. You can’t do it by pressing the + key, and the icons and conventions are sometimes not immediately obvious. For example, you might be fine with controlling your globes via the devices page, but to create a scene, they must be a member of a group first.

There is also the question of what happens when there is no internet connection. While the app (where signed in already) can operate on local control only, and the desperate can just flick lightswitches and use the globes like traditional lights, the internet connection is definitely necessary for setting up, defining/removing groups, scenes, schedules, and signing into accounts. Of course, should the online platform be withdrawn, you could be left with a hub that is no longer functional, although the Zigbee LightLink system will likely allow them to operate with other hubs (e.g. Belkin WeMo, albeit without tunable colour/colour temperature for now).

Compared to other platforms, the number of devices currently available in the Australian market is relatively few and using products from other vendors can lead to uncertain compatibility. That being said, it is early in the product lifecycle and it is expected that improvements to all points will be had in the future as overseas products gain the necessary regulatory compliance for Australia.

Electrical Performance

The electrical performance of the globes was tested with a Tektronix PA1000 Power Analyzer using a pure-sine wave inverter + variac synthetic power source for stability reasons. In many test cases, a choice of colour temperature was made to try and maximise the power consumption, but the accuracy of this setting cannot be clearly confirmed.

Power Consumption vs Voltage


The RGBW globe operated correctly over the 220-240v rated range with a power consumption of about 10.6W and a power factor of 0.59. The power consumption is marginally higher than expected, and while the low power factor is generally inconsequential in residential applications, it would be nice to see higher power factor globes become the norm to reduce any transmission losses. The light output of the globe remained stable down to about 130Vrms, although the operation of the remote control feature wasn’t checked. At 110Vrms, the globe flickered and changed to yellow, and between 80-100Vrms, it flashed yellow. As a result, it should handle practically all voltage variations on 230V networks with no visible consequences.


The Tunable White globe puts out an even more commendable performance, with a power consumption just under 9W, and a much better power factor of 0.82 in the rated voltage range. The operational voltage range is very wide with stable light output throughout the range down to about 50Vrms, and linearly reducing output down to 8Vrms. That is rather amazing.

Power Consumption vs Colour Temperature and Dimming


A common question is “Do you save power by dimming your lights?”, and the answer is pretty clearly “Yes.” For the RGBW globe, dimming the globe to 1% results in about 1.2W power consumption compared to 9W at full power operation. Changing the colour temperature also has a bearing on power consumption, and likely also has an impact on the perceived brightness, as the 810lm rating is the maximum, and it’s likely that other combination of operating settings will not achieve this.


A similar result was found with the Tunable White globe, although the variation due to colour temperature is much less.

IEC 62301 Ed. 2 Standby Power Consumption

The IEC places relatively stringent rules on how standby power consumption tests are performed, including uncertainty limits, a 15-minute minimum test window, stability requirements on measured power readings, and specific needs for the quality of input power. These tests aren’t made by using your variety store power meter and are highly accurate.

A summary of the results is given in here:

lightify-stdby-power-tableAt a rate of $0.25/kWh, the gateway costs you $2.68 a year to run, and each globe costs about $1-2 a year to run in standby. This cost can add up somewhat, especially if you have a large house with 20 globes or so, but this is the price of convenience. It is somewhat disappointing that in an era where mobile phone chargers are often <0.1W for standby energy, and BLE devices can run off coin cells for years, that the standby power of the globes is not closer to 0.1W. The added power consumption does cause the body of the globe to be noticeably warm, even when not in use, which can shorten the lifetime of the electronics and add to cooling bills.

The individual test reports are provided, in case you’re interested, in PDF format (Gateway, Tunable White Globe, RGBW Globe).

Other Comments

A test of the gateway range was also performed, by plugging in the gateway in the opposite corner at a lower floor of the house, approximately 10m away from the nearest Wi-Fi AP, and 20m away from any of the globes through a floor, several walls and a metal luminaire. No major problems with command reliability were seen, and it performed just as well as having the gateway inside the same room. Due to the lower bitrate of Zigbee and the mesh networking ability, it doesn’t appear that the system will have problems even with moderately large houses.

On the whole, the system was commendable in the fact that it didn’t put out any detectable level of radio frequency interference to my shortwave/HF gear. Unfortunately, the globes aren’t entirely without their annoyances

The RGBW globe seems to have a current driver that likes to make audible noise which is noticeable when the room is very quiet, or if it is used in fixtures with proximity to people (e.g. desk lamps) and can become quite irritating over time. The audio sample was recorded in my room with three desktop computers and an air conditioner running. The Tunable White globe does not have this issue, however.

It was also used on my regular line voltage where a 17Vrms ripple control signal at 1050Hz is known to cause certain globes to have flicker issues, but it seems that these globes are somewhat immune to this.

Photometric Performance

As a hobbyist devoting his spare time, unfortunately, I am not as well equipped as others when it comes to measuring photometric performance. I managed to borrow a spectrometer which is very finnicky to use and demands Windows 98 to run, and has seen about 12 years of abuse before it landed in my hands. While I was able to verify its wavelength calibration (i.e. the grating was still in its proper alignment), and I was able to obtain the manufacturer’s grating performance curves to compensate for grating efficiency and a script to calculate CCT and CRI from the spectral data, it still seems that the unit is severely lacking in sensitivity in the blue region of wavelengths resulting in wacky CRI figures and underestimation of colour temperature figures.

As a result, these results are to be taken as preliminary, and the *absolute* figures are likely to be horribly wrong, but it was still worth a try for relative comparisons between the two globes. (Update: Slightly more accurate spectral results can be found here.)

Spectral Output


The RGBW lamp achieves a high CRI mainly due to the presence of discrete LEDs for red, green and blue ensuring that our eyes are able to see the full spectrum. Depending on the colour temperature selected, the “mix” of drive currents to each colour of LED is adjusted to tune the colour temperature.


By using the colour feature, and tuning to just one of the LED colours, it is visible that the wavelengths appear to be 465nm Blue, 545nm Green and 623nm Red (standard LED wavelengths). The Green LED seems to have an especially broad spread of wavelengths.


The tunable white globe achieves its tunability in a different way, and appears to have a mix of both 2700K “warm white” LEDs and 6500K “daylight white” LEDs, and varies the effective temperature by changing the drive current between these two groups of LEDs. As white LEDs derive from a blue LED (~455nm) coated with a phosphor that emits the yellow-reds, the CRI from such a solution is typically less.


Please remember that due to issues with the blue response on the spectrometer, it’s likely that these values are not correct on an absolute basis, but it does provide us some ability to compare. For example, the warm colour temperatures seem relatively good, and the co-ordination between the RGBW and Tunable White globes seems pretty good in the 2700-3200k region. Even at 6500k, it seems both have a very similar colour temperature, but the intermediate settings seem to see some disagreement. The CRI values do not change as expected (they should get better at higher colour temperatures), but they do show that the RGBW globe is generally better (which is expected).


Another feature worth assessing is the dimming feature, for two reasons. Changing the drive current to LEDs can result in spectral “shifts” in chromacity, resulting in a perceivable change in colour temperature as the light is dimmed. Another issue is that of dimming range – just how far can the globe be dimmed.


On the whole, dimming the RGBW did not reveal any significant shifts in the peak wavelengths of the constituent LEDs.


The same can be said of the Tunable White globe, and from my subjective experience, which seems to indicate the globes do hold their colour temperature well as they are dimmed.

To assess the dimming level, the peak intensity at each dimming level was used, with the results normalized to the peak 100% value. This may not be quite photometrically accurate, but should be pretty close.

lightify-dimming-linearityThe results show that the globes generally have a range of about 8% upwards with better accuracy at the higher percentage dimming levels. It might not be quite as low as other products however.


In order to try and quantify the issue with flicker, a red LED was repurposed as a photodiode to measure the light output with a Picoscope 2205A USB Oscilloscope.

lightify-rgbw-1pc lightify-rgbw-25pc

From the scope traces, it was determined that the RGBW globe utilizes pulse width modulation dimming with a pulse repetition frequency of 500Hz. This is beyond perception, however, the standard caveat behind PWM dimming is that the stroboscopic effects can be visible when dealing with rotating devices (e.g. power tools) and can be a safety issue. It does explain the strobing you can see in the videos.

lightify-tw-1pc lightify-tw-20pc

Similarly, it was determined that the Tunable White globe also utilizes PWM dimming with a higher frequency of about 1Khz (~983Hz) making it even less likely to be perceivable.

Gateway Connectivity

I didn’t have that much time or impetus to explore the gateway connectivity extensively, however, security of IoT devices is a key issue, and a bit of snooping around is warranted. After the gateway connects to your network, it gets an address with DHCP and responds to pings. It appears to also issue MDNS broadcasts to announce its presence on the .local network.


An NMAP scan results in just one open port, TCP 4000 which is used by the app for local control. By using a special secondary AP and sniffing the bridged link with Wireshark, it seems that no real “encryption” is used, and a lightweight protocol (whose meaning I have not attempted to decypher) consisting of sequential packets of raw values is used.

lightify protocol

No big security issues are seen, as the device does not insert any uPnP entries into the router, and is generally resistant to “probing” on the port from local network devices with only the one port open.

The cloud services are enabled via the Arrayent IoT platform with the device requesting a DNS resolve for, which resolves to a CNAME for, resolving to After a UDP connection is established to Port 80 on the’s servers, replies are also heard from (an adjacent IP). This is what enables the “cloud” connectivity for controlling away from the home network and is a platform used by other corporate IoT solutions (e.g. garage door openers). Only small packets are exchanged and relatively infrequently.

Unfortunately, OSRAM have not made the API or documentation available yet, however, some people have already reverse engineered the protocol and produced their own implementations and libraries.


It wouldn’t be a Gough-style review if I didn’t try to break something. Lets see what the system is made out of. It is not recommended that end users take apart equipment as it will void warranties, can cause damage to the product and can expose them to danger from lethal mains voltages. I will not be held responsible for any damages.

The first victim is the gateway unit, which comes apart if you strategically pry at the right spot and release the glossy front cover.


Immediately, the front PCB is visible, with two soldered down shielding cans for the two radios. If I had to guess, the left one is Zigbee, and the right one is Wi-Fi. Three antennas are populated, two folded metal units and one PCB trace antenna. The PCB is connected by standard 2.54mm header pins, so the PCB can be pried out to reveal the other side.


The other side is home to a Winbond serial flash chip which likely houses the firmware. The PCB design date is 4th August 2014, with a production date of Week 46, 2014.


The rear housing features the power supply, which curiously is model KSP20A0330100 WEMO Bridge. This suggests this product is a cousin of the Belkin WeMo. The power supply features a few CapXon capacitors, which are not particularly known for their quality.

Opening the Tunable White globe was fraught with difficulty due to how well the dome had been sealed on. Ultimately, I had to nibble the dome away using some side-cutters resulting in permanent damage, as no amount of breaking the plastic with a blade or screwdriver seemed to help.


As I had expected, the LED configuration consists of two colour temperature LED strings. There is a total of 24 LED chips, 12 daylight white (bright yellow) and 12 warm white (orange). Each of the 12 LEDs are arranged in a 3-parallel x 4-series configuration. Multiple parallel LEDs are generally not a preferred solution as there is no guarantee that the LED will share current evenly with its neighbours, so accelerated LED failure could be the result, but is a common “feature” amongst many LED products today.

Construction quality wise, there is something to be desired. While it’s commendable that they used proper thermal paste, the application wasn’t entirely adequate resulting in excessive paste residue being spread over the MCPCB and over the front surface of the LED chips resulting in diminished light intensity. There is also one solder ball, which could detach over time and cause problems if it were to migrate into the current driver electronics behind.

The product code L-A806ST tells me that this is a Leedarson product, likely a customized version of the A806S. Some other OSRAM consumer products are OEM’ed by Leedarson, so this is not unexpected.

Further deconstruction was hampered by the need to desolder the wires from the MCPCB – due to the thermally conductive nature of the board, de/re-soldering the wires can be quite difficult, however a quick unscrew and peek suggests there is evidence of potting of the driver circuitry behind.

Taking apart the RGBW globe was not as difficult, as after peeling some of the side plastic apart, the top cover easily popped off.


Internally, we find another internal plastic shroud used to diffuse and better “mix” the LEDs to produce a homogeneous colour. This too, pops off easily.


This unit is clearly different to the unit above and utilizes a board connector for connecting to the driver unit behind. As it is thermally adhered to the heatsink, it was not possible to remove it without compromising the thermals of the globe. Despite this, it seems this globe is more cleanly constructed, and has each set of LEDs in a series string with a common pole. The PCB itself is a Rayben PCB (Hong Kong) product, but the globe itself is of uncertain origin. The PCB code ZNN3028255-00 doesn’t lead me anywhere too specific. The PCB date is Week 35 of 2014.


In this timely (but thoroughly) prepared review, we have taken an in-depth look at OSRAM’s LIGHTIFY offering. In a market where Internet-of-Things and home automation are the boom areas, the Lightify system tries to offer something different by providing adjustable colour and colour temperature globes with their own user interface design but also remaining true to industry-standard Zigbee Lightlink protocols.

On the whole, aside from the buzzing of the RGBW globe and the limited lumen output overall, both globes performed rather well from a photometric point of view. Pulse-width modulated dimming at high frequencies ensures flicker is kept to a minimum, and interference with RF was not experienced at all. RF range of the Zigbee system proved to be ample even across the house through walls and floors. Electrically, the RGBW globe could have done with a higher power factor, and the standby power consumption could have been reduced, but these are relatively minor points.

As the products were reviewed prior to their official release, it seems that the system is still in its early days and could use some refinement to iron out sporadic issues of delayed command response, incorrect operation and application crashes. Furthermore, it could use some more intuitive design, more accessories (e.g. sensors, hardware buttons) and features in the software to better utilize all the available capabilities in a more convenient way and better integrate with other systems. As a whole, with the limited amount of products from OSRAM on offer for the system at this time in the Australian market, it is a little bit basic, and without knowing the final cost of the units, it is hard to judge whether this will be a hit with customers especially as they will have to replace the whole globe with another Lightify or Zigbee LightLink profile product if they wish to maintain smart-connected features.

That being said, it does work most of the time, and having the ability to control and schedule the lighting around the house from multiple devices (up to five) on the home network and on the road could be quite appealing. Having on-board dimming is also a big plus as it can avoid the tricky situation of trying to match LED globes to physical dimmer switches to find one that actually performs to your liking. If it’s any consolation, many of the competing systems have similar criticisms surrounding their reliability and robustness, and future software updates are likely to improve the system, so keep that in mind.

More importantly, to avoid disappointment, I encourage the reader to think about their needs, and how such a system would benefit them and integrate into their lifestyle before deciding on purchasing one regardless of the brand.

Update: Interested readers can obtain OSRAM LIGHTIFY products directly from Voltimum

Thanks to Voltimum for supplying the items for review. Voltimum is a distributor of various lighting and test equipment products, so be sure to check out Voltideals for your needs.

About lui_gough

I'm a bit of a nut for electronics, computing, photography, radio, satellite and other technical hobbies. Click for more about me!
This entry was posted in Electronics, Lighting, Tablet and tagged , , , , , , , , , , , , . Bookmark the permalink.

3 Responses to Review: OSRAM LIGHTIFY Smart Connected Lighting System

  1. rasz_pl says:

    wow, that was a comprehensive writeup

    ANT3 appears to be a fake antenna? 😮 there are no feed lines leading to it, it connects to ground
    a counterweight for ant2??

    • lui_gough says:

      Yeah you’re right – well spotted.

      I wasn’t looking carefully, but it looks like it may have been a printed trace antenna in the design initially, but they probably just deleted the necessary band-pass filters/connections/etc and opted for folded metal antennas instead in this model and modified the PCB without changing that detail. The reason I say that is that the ground plane is clear on the top area on the other side of the PCB even away from the folded metal antennas, so it’s likely that was intended to be a printed trace antenna initially.

      After all, the throughput requirement is low, so no-diversity or MIMO is no big deal.

      – Gough

  2. Roger Zelinsky says:

    The review was truly thoughtful and well researched. Unfortunately, the bad things have not been addressed and resolved. I spent most of a week (after a dozen attempts to get the gateway and bulb to talk to one another) trying to get a schedule to work. I can control the light with my phone but cannot get a schedule to work consistently. I will investigate other alternatives.

Error: Comment is Missing!