The LED retrofit market has been growing considerably, both in commercial and residential contexts, as LEDs offer instant-on to full brightness, superior energy efficiency and long lifetimes, reducing labour costs and saving on cooling. Commercial interest in retrofits have been incentivized through schemes such as the Energy Savings Scheme (NSW) and Victorian Energy Efficiency Target (VIC) where business owners can stand to benefit from subsidies and rebates on work done to improve energy efficiency. One key route is the replacement of all fluorescent lighting within a facility with approved LED lighting.
Many commercial facilities currently have suspension ceilings with multi-bay fluorescent tubes which are prime candidates for retrofit. Many of them may have older inductive control-gear with luminaires which are due for replacement due to age, or needing complete re-builds to make them compatible with retrofit LED-tubes. Others may be rather dated 4 x 18w (or 4 x 20w) fluorescent units where retrofit LED-tubes are uncommon.
To address this market, Philips (along with other vendors) have manufactured LED lighting “panels” which are a “drop-in” like-for-like replacement for such fittings. These panels are sized in the standard 600 x 600mm or 1200 x 300mm tile size, and offers better lighting quality with all the benefits of LEDs, plus even illumination and less glare.
In this review, we look at the Philips SmartPanel 2.0, specifically an RC160V LED25S/840 PSU W60L60 (12NC: 911401501121) which is a 32W 600 x 600mm panel with a 4000K neutral-white 2500 lumen output. This review was made possible by Voltimum and Philips Australia which sent a sample for review under the review challenge terms.
The unit itself is packaged in a large flat rectangular box, almost reminiscent of a super-sized pizza box. Despite the box having clear indications not to step-on or apply pressure to the carton, and fragile tape everywhere, the blistering of the cardboard seems to suggest that it has been slightly mistreated in transit and something heavy was placed on top.
The unit is Made in China. There is some confusion as to the product designation, as the Philips 2014/2015 catalogue seems to allude that this code is a SmartPanel LED which implies it is the original version, whereas the same code is referred to by Philips (Peru) as a Smartpanel 2.0. The difference between the two versions include improved luminous efficacy to 80 lumens per watt, and improved reliability from 30,000 hours to 50,000 hours to 70% (L70) at 25 degrees C. A look at the 12 number code (12NC) seems to show that the 12NC of the supplied unit does not match with any of the literature which may indicate that this unit is newer or specific to the Australian market.
At this stage, a look-up on the VEET and ESS databases does not seem to have this product pre-approved for use with the schemes. Those looking to employ this unit in their projects may have to submit an application to get this unit approved, which would be additional work, but as other Philips products have been approved in the past, it seems unlikely to pose significant problems.
The unit is packed inside a protective plastic bag with foam edges for protection. The centre area is left unprotected, and thus, should not have pressure applied. This is not a problem where shipping multiple units of the same sort stacked together, but when shipping one-off units (e.g. for this review), other packages can end up squashing this one. In this case, the “no step” symbol was printed directly above where the current driver is, and the applied pressure resulted in a slight bowing of the base-plate which was correctable. No permanent damage seems to have been sustained.
The first thing that you see is the panel itself. It has a matte white frame edge with a subtle grey Philips logo printed on one side. This is a nice touch, but it also means that installers should take some care to orient the panels so that the logos are not haphazardly pointing everywhere as that would detract from its look. The panel itself is covered with a protective film adhered in the corners with masking tape, and is to remain in place until installed to protect the panel from dirt and damage. The front optics is a matte finished diffusing surface with very little surface reflection.
Also included is a single-sheet installation manual which lists some of the key features of the RC160V unit:
- It is a SmartPanel 2.0, and thus has the higher efficiency and longer lifetime.
- This unit has a fixed current driver (PSU) although units are available with DALI control (PSD) for digital group/individual control and dimming. The acceptable voltage is 220-240v AC at 50/60Hz.
- This unit has a luminous flux of 2500lm, with units of 3400lm also available.
- This unit measures 597 x 597 x 46mm with a weight of 4kg. Units of 1197 x 297 x 46mm are also available to replace 2 x 36w style fluorescent fixtures with a weight of 4.2kg.
- The unit carries the regulatory approval marking, is double-insulated, meets AS60598 and is designed for indoor use only at 25 degrees C ambient, with an IP rating of IP20.
- It is a non-insulation contact (Non IC) luminaire and clearance must be maintained from insulation and building structures (SCI = 50mm, MIC = 150mm, SCB = 250mm, HCB = 250mm).
- The unit is capable of suspension mounting from four individual mounting points. A suspension kit must be ordered separately.
- An in-rush current of 6A peak for 55us is claimed, allowing for 40 units to be installed on a 16A circuit protected by an MCB.
- No pressure should be applied to the rear or the driver of the unit.
- Two eye-let hooks are supplied for mounting to a safety chain, but the chain itself is not included.
The rear of the panel is mostly sheet metal with a current driver mounted in the centre. Four sockets are provided for mounting to the suspension kit or eyelet hooks for the safety chain. The unit is supplied pre-wired with a (roughly) 1.5m flexible cord.
The plug and cable assembly is from Kenic, with a two-pin plug. Despite the metal used in the assembly, the unit is not earthed as the current driver is double insulated and its output is separated extra-low voltage (SELV). The cable is a 2 x 0.75mm^2 sheathed round cable.
A closer look at the rear shows the specifications of the unit, which claims to be 32W, made of two strings of 40 x 0.3W LED chips. It claims a power factor of 0.9 and a current consumption of 0.17A. The stock is dated 25th May 2015. The current driver itself, however, appears to be capable of driving 36W of output, and is specificed for an ambient of up to 50 degrees C and case temperature of 75 degrees C.
Despite being pre-constructed for drop-in installation, there was one peculiarity where a cover screw was not fully tightened and its head was partially stripped. This might be a one-off quality issue, but is unlikely to result in significant safety issues as the output is SELV.
As promised, you also get two eyelet hooks for passing through a safety chain.
Readers may be interested to know just how one of these lighting panels is constructed, so here I go taking things apart.
The panel itself can be easily accessed by removing all the outer screws, so that the clip-together plastic frame can be removed. Underneath, you will find the panel itself.
The panel is basically made very similarly to an LCD TV (or monitor) minus the LCD itself. It basically consists of a PET white reflector in the rear, a piece of PMMA clear perspex-like clear middle and a PS translucent diffuser sheet on the front. The light is channeled into the PMMA at two ends.
The strip of LEDs appears to have a thermal connection to the sheet metal casing, and is positioned to channel its output into the light guide. A black tape covering is used to force stray light to be absorbed so as to avoid bright glare when viewed at an angle (similar to backlight bleeding issues with LCD screens).
The two LED strips seem likely to be connected in parallel, which is acceptable so long as both are fairly well matched for current consumption and are operating under similar conditions. Such a configuration puts the current driver away from the LEDs which reduces heat build-up – a key factor in improving the lifetime of LED products.
I did not take the panel apart any further, as doing so may permanently compromise thermal and optical performance.
The current driver itself is in a plastic case, and the first job was to open it up to view the connections. Push-fit terminals are used for quick connections, and an earth terminal is provided on the driver but left unused. The whole driver can be removed from the baseplate.
The rear of the driver has an authenticity label which would not otherwise be seen. This can be validated by visiting this site. Authenticity checking is done by a simple two-step process.
First, you enter all 16 digits in four groups of four. Then you click on the submit button and await the results.
As the product was supplied by Philips themselves, there was no doubt to the authenticity, but it is nice to see the result confirmed.
The driver itself has its output current adjustable by a setting resistor which is mounted in two terminals of the push-fit connector. Changing its value will likely change the output current and should not be done as it could compromise lifetime or output. However, it should be noted that my unit came pre-fitted with a 2.7kohm resistor between Rset and SGND.
Testing and Subjective Opinions
In this segment, I will report on the results of various tests undertaken on the sample unit, and offer some commentary on the results and of my experiences while operating and researching the unit. As I am not an official test lab and I do not have the equipment to perform all tests to industry standard, please consider the results obtained with this in mind.
Power Consumption and Power Factor vs Voltage
This was a check to confirm the power consumption of the unit, and its power factor. This was measured over a range of voltages to ensure the luminaire performs correctly over the stated range and to determine its operational limits. This was measured with a variac and my Tektronix PA1000 Power Analyzer.
On the whole, it seems very stringent active regulation kept the power consumption fairly constant between 30 to 31 watts throughput 130v through to 272v. Between 40v and 130v, the power consumption falls almost linearly. As a result, it seems that the unit is relatively immune to power quality problems with respect to voltage. As for power factor, the unit remained above 0.9 for all voltages below 268v, meeting and exceeding specifications. Between 230v and 240v, the power factor was 0.91 to 0.93 which is very close to unity.
The power consumption evolution during the first 15 minutes from cold was also measured and showed a characteristic decrease in power consumption (common with LED products as they warm up) but of a very small magnitude. Initial power was at 30.525W and stabilized at 30.112W after warming up. This was less than the claimed 32W nameplate rating which may represent mis-setting of the current driver, a more efficient LED/current driver than expected, or reduced output below specifications. What is actually the case can only be verified with measuring the actual radiometric output (which I am not equipped to do).
The claimed in-rush current was 6A for a duration of 55us. An attempt was made to verify this with the PA1000 in inrush current mode, by powering the panel on and off ten times in succession.
The maximum measured in-rush was 48.81A which was significantly higher. This may be due to the way the PA1000 measures in-rush, but also could be due to the fact that repeated power cycling may have rendered any in-rush current limiting ineffective (e.g. thermistor based).
Output Waveform and Efficiency
Unfortunately, as I do not have two PA1000 units, measuring input and output power simultaneously to derive efficiency was not possible. Instead, two sequential measurements were taken – one of input power, and one of ballast output through physically intercepting the output wires.
Using this method, it was determined that the efficiency is roughly 86.8%, which is fairly good for a current driver unit.
It also allowed for measurement of the current output parameters, which were:
- Vcf = 1.022
- Acf = 1.264
- Vrms = 29.88v
- Arms = 0.895A
- Apk+ = 1.131A
- Apk- = 0.6145A
As a result, it can be concluded that the output is not entirely ripple free, with the current swinging between 615mA and 1131mA but averaging close to 900mA. However, this is still pretty much DC at 29.88v on the output landing safely in ELV territory, and is expected not to produce any perceptible flicker.
A slightly problematic Ocean Optics USB2000 spectrometer was used to measure the spectral output. This unit is known to have response issues in the blue wavelengths, and a correction curve based on the sun was developed but not without flaws. The resulting figures should be taken with a grain of salt and should only be compared with other spectral results on this site only.
It seems that the LEDs have a very wide band emission in the green/orange/red region which either indicates a mix of LEDs of different colour temperatures (e.g. 2700-3000k + 6500k) or specially produced phosphor mix to achieve the neutral white 4000k colour temperature. Once corrected, the colour temperature computed was 4831K (it is known to over-report), with a CRI of 88.8 (potentially over-reported), whereas uncorrected, it was computed to have CCT of 4058K with a CRI of 78.6 (probably under reported due to loss of blue sensitivity).
To get a second opinion, I used my Nikon D3300 camera and took a photo of the light source in RAW. I examined the file using Adobe Lightroom and used the white-balance tool to adjust the white balance on the light itself. It came back with a figure of 3900K.
All things considered, it’s very likely that this panel does have a colour temperature of 4000K and a CRI of 80 or slightly above.
The uniformity of the panel was not something I was perfectly equipped to measure, but subjectively speaking, it appeared remarkably uniform to the eye and was reminiscent of the light-boxes used to view X-ray films. Good uniformity reduces glare and makes the panels more pleasing to look at by reducing hot-spots.
I did take a photo of the panel itself using a Nikon D3300 + Tamron 17-50mm f/2.8 lens although some of the non-uniformity in the image could be the result of lens vignetting. However, the image does show very even results, and it’s not obvious that the panel is only edge-lit from two sides. Having a construction that is similar to LCD screens, where uniformity is a key parameter, really helps.
Even at oblique angles, the LEDs were never visible, and their arrangement was completely unseen. This is due to the panel over-reaching into the frame a significant distance, and the use of black taping over the edge to prevent direct viewing of the off-angle LED output. A small fraction of light does leak through the seams of the frame.
In fact, the uniformity was so good that I took a photo of a conventional light-globe sitting on the surface of the unit and it looks quite similar to a light-box image. I would have to say that this is exemplary performance, and no doubt, aided by the matte front diffuser.
It did appear a little strange to me that the panel itself has a metal back and no earth connection, however, the reason for this is because the unit itself claims to be double-insulated. The current driver itself is housed within a plastic enclosure (which is insulation) and the secondary is separated (another layer of insulation). Even if a fault should occur in the LEDs and the output was shorted to the case, workers would be exposed to SELV at the most, rendering the unit safe.
But just to be sure, I tested the insulation resistance using my Keysight Technologies U1461A Insulation Resistance Tester. Test parameters included a 10 minute test at 500V DC. The full report is here, although a summary of the results shows that all readings exceeded the necessary 1 mega-ohm by several orders of magnitude:
- Active to Case – 49.41 gigaohms
- Neutral to Case – 35.51 gigaohms
- Active to +ve output – 15.95 gigaohms
- Active to -ve output – 16.13 gigaohms
- Neutral to +ve output – 16.30 gigaohms
- Neutral to -ve output – 16.23 gigaohms
As a result, we can conclude that the unit is safe and exhibits proper insulation between primary and secondary, and the metal casing.
While in use, the SmartPanel 2.0 was not audible at all, making it a good choice for office environments. The light did have a slight delay from power on to illumination, however, this delay was sub-one-second and generally not a problem.
Aesthetically, the panel has a clean look, with subtle branding, and its interlocked frame seams are very regular making sure it does not stand out. The diffused front also produces very smooth light, with arguably less glare than many fluorescent fittings (UGR < 19). The thin pressed metal rear does have some play in it which is noticeable during installation as it “breathes” slightly when the eyelets are installed, but does not affect performance. No flicker was evident while in use.
While the SmartPanel 2.0 seems to be well made, and its performance is well specified, its specifications can still be beaten by other products on the market (but not necessarily in Australia). Consulting other major lighting manufacturers, it seems that some of their panels are available with SDCM of 4 or less (compared with 5 for the SmartPanel 2.0) and lifetimes of 50,000 hours to 85% (as opposed to 70% with this unit). Another manufacturer claims a slightly better CRI of 82 and above. Rather disheartening is that there are many unscrupulous dealers which are selling similar products without proper specifications (e.g. lifetime of 50,000 hours but the end-point is not specified, or LED lifetime rather than luminaire lifetime, and missing CRI/SDCM/UGR specifications). As a result, it probably pays to check the specifications against your requirements and pricing before you commit to a purchase.
Another concern seems to be with the non IC rating clearances which were more generous than some other manufacturers. While such LED panels generally boast slimmer profiles, this is mostly moot due to the requirements to keep generous safety clearances. That being said, when retrofitting existing bulky multi-bay fluorescent fittings, it’s unlikely to be a major issue as their bulky fittings would have likely ensured sufficient clearance. On the whole, the panel edges where the LEDs were mounted were warm to the touch, and the current driver only slightly warm. It seems the clearances are more to err on the side of caution, but must be followed to ensure safety.
During testing, no RF interference issues were noted with shortwave/HF reception, nor regular broadcast band reception.
The Philips SmartPanel 2.0, like other Philips Professional LED products, shows clear LED-first design with excellent performance. The output light uniformity was excellent, with no visible LEDs even at oblique angles, and exhibited no flicker. The current driver did not exhibit any noticeable noise, and had reasonably good efficiency and excellent power factor. The thermal design appears to place the LEDs away from the current driver ensuring maximum heat dissipation by preventing the accumulation of heat. Aesthetically, its simple, consistent design and understated branding was hard to fault. From a safety point of view, it claims to meet Australian Standards and has the Regulatory Compliance Mark.
Specification-wise, the product is rather similar to many of the LED panels on the market, with some others claiming slightly better lifetime, SDCM and CRI figures. However, the specifications issued by Philips could be conservative, and are very solid figures nonetheless, which should satisfy most business and office needs.
One unfortunate thing is that this product is not currently on the VEET and EES databases as approved products, so those looking to use these in their retrofit projects will need to submit for approval. This is unlikely to be a major issue, as other Philips products have previously received approval, but does add a further complication which would not exist for other products already approved.
The actual light output could not be verified, but the power consumption figures were below the claimed 32W level, and the current supplied to the LEDs was only sufficient to provide about 26.2W which could imply that there may be power consumption variations from unit to unit and/or that with improved quality LEDs, that lower-than-label-rating power consumption with equivalent performance is a possibility.
Teardown, PCB images removed as per Philips request.