By Matthew C. Wood, P.Eng., Roar Engineering Inc., December, 2018
O Christmas Tree
I cherish my childhood memories of Christmas time and, in particular, the way that the Christmas tree filled our living room with the scent of pine and the glow and twinkle of lights. And so, last winter, it struck a chord with me when I investigated a Christmas tree fire that saw a family displaced, though thankfully unscathed. I resolved then to write this article, a caution for the following season, and an exploration of decorative mini lights and their failure modes.
I have long known that dry Christmas trees are an excellent fuel; many a scouring of the woodpile looking for a hint of red needles to aid in starting a summer bonfire drove that home long before beginning my career as a fire investigator. However, it is apparent to me that many people underestimate or are unaware of the extreme hazard presented by a dry Christmas tree.
When fire investigators and fire scientists want to talk about the “size” of a fire, or its power, we talk about the heat release rate (HRR), which is typically measured in kilowatts (kW). For example, a small wastebasket fire might have a maximum/peak HRR on the order of 50 kW and a metal-framed plastic chair, 270 kW. By comparison, the peak HRR of a Christmas tree fire ranges from 3000 to 5000 kW.
Further, a fire can be characterized by its HRR over time, from ignition, through growth, full development, decay, and finally, extinction. The time from ignition to peak HRR is dramatically brief in Christmas tree fires; in extreme cases, a dry tree may become fully involved in fire in under 10 seconds.
These factors contribute to the rapid growth of a compartment/room fire where a Christmas tree is present and/or the first fuel ignited. For years, when I have explained compartment fire development to people, one of my favourite demonstrative resources has been a video produced by the US National Institute of Standards and Technology (NIST), entitled “Dry Scotch Pine Tree Fire”(1). This video shows a typically furnished living room with a Christmas tree in the corner. The tree is ignited at its base and is fully involved in fire in 9 seconds. The fire quickly spreads to adjacent furniture and the smoke layer rapidly descends, with flashover(2) and full room involvement in around 40 seconds!
NIST continues to produce quality content along these lines: last year they released the powerful video, “Christmas Tree Fire: Watered Tree vs. Dry Tree”(3). Here, two identically furnished setups are shown side by side, with a watered tree on the left and a dry tree on the right and both are ignited. After about a minute of inferno, the readily available fuel of the dry tree has been consumed and adjacent fuel packages (i.e., an upholstered chair and bookshelf) have been ignited, while after 1 minute and 45 seconds (being the end of the video), the watered tree fire appears not to have spread much beyond its point of origin, though notably having produced considerable smoke, as might alert an occupant or activate a smoke alarm and allow for evacuation and timely notification of the fire department. The conclusion to the video is well put, “Don’t get burned this holiday season. Water your tree.”
The difference between the watered tree and dry tree scenarios may very well be survival(4): the US National Fire Protection Association (NFPA) statistics indicate that while Christmas tree fires are relatively uncommon(5), they are particularly deadly. The 20112015 numbers are, on average, 1 death in every 32 fires where the first fuel ignited was a Christmas tree, compared with 1 death in every 143 reported structure fires generally(6).
Aimed at education and prevention, the NFPA has, this year, provided a reproducible “Christmas Tree Safety” tips sheet, available here: https://goo.gl/DbuAHb. I defer to the NFPA and this document concerning prevention and would encourage its distribution.
Let us turn our attention to ignition sources for Christmas tree fires. Here, the NFPA statistics indicate that “electrical distribution or lighting equipment” is involved in the causation of 40% of Christmas tree fires, followed by deliberate acts (24%), heating equipment (15%), candles (8%), and fire play (i.e., children, 6%).
Concerning ourselves solely with the first category, I note that extension cords and lighting sets used for holiday decoration may be more prone to mechanical injury than other temporary household wiring, which could lead to electrical arcing activity and fire. Namely, these products are annually put up and taken down (read untangled and re-tangled, not always with the lightest touch); sometimes routed under furniture or the Christmas tree stand; sometimes improperly secured with staples or other fasteners not intended to protect electrical wiring; sometimes used to hang ornaments from; and, even subject to attack by four-legged family members, among other abuses. But what about failure modes relating to the lights themselves?
A Government of Canada recall search for Christmas/fairy/mini lights returns a slew of recalls(7) for both incandescent and LED products, many coinciding with an apparent crackdown in 2015, but the hazard descriptions don’t provide detailed information, simply indicating some combination of shock, overheating, and fire hazard.
For illustration for this article, I opted to explore representative products hands-on, and so broke the bank and headed to the dollar store to pick up two indoor 20-lamp sets, one incandescent (“clear”), and one LED (“pure white”). Both were made in China and imported by the same company. Both had CSA markings on the boxes, instructions, cord tags, and subcomponent parts, and both had 22 American wire gauge stranded copper conductors with the ubiquitous green insulation, rated for a maximum operating temperature of 60°C. Note that these small-gauge conductors can typically carry around 2 amps continuously without exceeding their insulation temperature rating (i.e., this is their ampacity).
Both products had the same style of non-polarized attachment plug, with internal 3 amp fuses on both legs (Figure 1); these fuses appear to be slightly oversized for overload protection but presumably offer adequate short circuit protection, which is not to say that they provide immunity from arcing events. The attachment plugs for both sets included an outlet for straight-line connections (i.e., the provision for connecting multiple sets at the attachment plug end, which does not draw more current through the set wiring).
The LED set terminated in a single outlet, for endtoend connection (which places additional sets electrically in parallel), whereas the incandescent set did not have an end outlet. Where sets do allow for endtoend connection, it is important to limit the number of connected sets to prevent overloading the wiring. Both sets of instructions stipulated that, for endtoend connections, the maximum power was 216 watts (or 1.8 amps at 120 volts). The LED set was rated at 2.16 watts, such that you could merrily connect 100 sets or 2000 lamps together, while the incandescent set was rated at 9.6 watts, allowing for a more modest 22 sets or 440 lamps, had they outlets at their ends. Clearly, the potential for such overloading is more in the province of long outdoor runs or creative misuse.
Being one of the few people ever to have read the instructions for these sets, I can inform you that both clearly contemplated the hazards of mechanical injury, both strain and cutting; thermal damage from the lamps themselves or other heat sources (“Do not let bulbs rest on the supply cord or on any wire.”); unattended use (prescribing unplugging both when leaving the home and when “retiring for the night”); generally, fire, burns, and electric shock; thermal insulation preventing adequate heat dissipation (“Do not cover the product with cloth, paper or any material…”); use in dry trees (“…the tree should be well maintained and fresh. Do not place on live trees in which the needles are brown or break off easily. Keep the tree holder filled with water”); overheating related to failure to replace burned-out lamps promptly (more on this cascade failure later); and, shock or electrocution related to use by standing water.
Diodes are the one-way valves of the electrical world. When a sufficient (forward) voltage is applied across a light emitting diode (LED), it begins conducting and producing light, with the intensity being proportional to the current. However, in excess of this threshold voltage, the current increases exponentially with the applied voltage, and so the current must be limited to prevent self-destruction, typically to a value in the design range of 10 to 20 thousandths of an amp (i.e., 1020 mA). In mini lights, current limiting is achieved with series resistors. This arrangement is a compromise, being at once inexpensive and inefficient (compared with constant current regulation, etc.), with the LEDs running cool while the resistors dissipate excess power as heat.
The LED set that I purchased (Figure 2) is the modular style that uses essentially the same lampholders as traditional incandescent sets. Here, as shown disassembled in Figure 3, the leads of a standard 3 mm LED are inserted into a plastic base then bent up to secure it and a decorative lens is press fit into the base, over the LED. This base is then pushed straight into the lampholder socket so that the LED leads make electrical connection across two flat contacts, one on either side, which are crimped to the ends of the wires that are fitted into notches in the socket. The socket has a keyway to ensure that the LED is installed in the correct orientation, also indicated by the small locking tab. In this setup, standard axial resistors are similarly installed in their own holders and sockets with caps (Figure 4).
The other prevailing style (not pictured) is to have LEDs soldered directly to the connecting wires, separated by small spacers, with current-limiting resistors (where required, and likely distributed throughout the series) soldered directly to one of the LED leads. These assemblies are then either simply wrapped in heat-shrink tubing sleeves or first inserted into a clear holder filled with resin and only then covered in heat-shrink tubing, which is much more robust from a strain relief and water resistance perspective.
There are also various circuit configurations, with the set that I bought being the simplest (and probably most dangerous): a series of LEDs with current-limiting resistors directly across mains power(8). Other sets may be entirely in parallel or include several parallel arrays in series or several series in parallel, etc. Also, the LEDs may be driven by battery packs, mains power (sometimes rectified), or low-voltage power supplies (preferred for reducing shock hazard) and may include controllers with patterning, timing, or even remote-control options.
Online sources indicate that a potential fault relating to LED lighting sets is the use of current-limiting resistors with an insufficient power rating, with one source reporting thermal imaging temperature measurements as high as 100°C above ambient where tiny 1/8th watt resistors were run at around half a watt. The products in question were purchased from China from an ecommerce site, thus bypassing import safety standards; however, it is not unfeasible that such products could find their way into the marketplace as substandard counterfeits. Despite that the typical resistor failure mode is open, I hypothesize that overheating resistors in the soldered/sleeved light sets could result in shorting across the connecting wires, effectively removing the resistor and LED from circuit and increasing power across the remaining resistors and precipitating further failures and, if unchecked, overload.
For its part, my LED set spread out its power dissipation across three 1800 ohm (1.8 kΩ) half-watt resistors, which were not being overpowered in this application. After several minutes of operation on my bench (in free air, without any thermal insulation), temperature testing with a Type-K thermocouple probe applied to the surface of the resistor pot showed a modest temperature rise of around 5°C above ambient.
And, finally, to the good old tungsten mini lights of my youth, which remain in wide circulation and continue to be sold, although one would assume that their retail days are numbered. A trip to the local thrift store will reveal that the second-hand market is currently rife with tungsten decorative light strands as people upgrading to LEDs donate their used (and possibly abused) sets.
The operating principle of incandescent lamps is that current passed through a tungsten filament heats it to around 2400°C, causing it to emit visible light—almost as an afterthought, considering that greater than 95% of the power consumed goes straight into heat energy. The electrical resistance of the tungsten filament has a positive temperature coefficient (i.e., as temperature goes up, resistance goes up), with the cold resistance being about 1/15th of the operating temperature resistance. Accordingly, when they are first turned on, for a tenth of a second or so, incandescent lamps experience a high inrush current of around 15 times their operating current; this is the reason that, at end of life, lamps tend to burn out when first turned on.
Miniature decorative light sets came about in the 1970’s, quickly outstripping the existing and more power-hungry mains-voltage holiday lighting: whereas their 5-10 watt predecessors were wired in parallel, with 120 volts across each lamp, low-voltage miniature lamps typically dissipate about half a watt and are wired in series, such that the 120 volts applied across the set is divided amongst the lamps (e.g., 120 v supply/2.5 v rated lamps = 48, with a typical set actually being 50 lamps(9) for a more marketable number; and, for the 20 lamp set that I purchased, each is rated for 6 v).
However, unlike parallel lamps, a problem with series connected lamps is that if one burns out, the circuit is opened and all of the lamps go dark, making the task of finding the failed lamp a potentially frustrating and time-consuming one. Luckily, this was by no means a new problem(10) and solutions were ready to hand: for example, a 1938 patent by Davis discusses what is referred to as a shunt, which is simply a means to allow for the continued flow of current through the circuit once the lamp filament is ruptured (i.e., bypassing a failed lamp). The described invention is a fine copper or tungsten wire within the bulb that is wrapped around the lead-in wires, which support/supply the filament, after they have first been oxidized or lightly varnished to insulate them. In operation, current follows the path of least resistance, which is through the filament; however, upon rupturing of the filament the circuit opens and the shunt wire is exposed to the full supply voltage, which is much greater than the lamp operating voltage and so causes the (dielectric) breakdown of the insulation or oxide coating, such that the shunt becomes conductive and bypasses the lamp(11).
Despite several patents concerning novel shunt designs, every mini lamp that I’ve inspected has a fine wire shunt as described above, with the only difference appearing to be that aluminum is the shunt wire material of choice, with an aluminum oxide coating acting as an insulator. Figure 6 shows the filament and shunt wire between the two lead-in wires in one of my set’s 6 v lamps, with the supporting glass bead on right.
Although a brilliant solution (pun intended), the shunt mechanism does allow for a particular type of cascade failure, a fact not lost on Davis:
“In all globes [lamps/bulbs] of this nature, hooked up in a Series, as a globe burns out the voltage is increased proportionally to the remaining lights in line. This Overloading will cause them to [burn brighter and hotter and] deteriorate more rapidly and the failure of each in turn will so increase the Voltage that the last two or three globes to remain in operation will flare up and burn out simultaneously, and in so doing, set up a complete circuit through the shunt systems. The result in this case is either a ruptured house plug-fuse or a fire at the globe.”
Davis’ solution was simply that his shunt wire acted as a fuse and would rupture with an abnormal increase in current and again open the circuit. My testing also found the shunts acting as fuses (see below).
Further concerning protection against cascade failure, my understanding is that UK sets contain a “fuse lamp”, which does not have a shunt and is rated at a slightly lower voltage than the remainder of the lamps in the set. Canada, however, does not appear to have any such requirement and no set that I have inspected includes such lamps, which, in any event, could be easily bypassed by the end user by removal and installation of a standard replacement lamp.
This cascade failure mode, which is why manufacturers prescribe prompt replacement of burned-out lamps, is of particular relevance to the fire investigator in considering whether mini lights could be a competent ignition source in and of themselves.
Very informal testing with my 20 lamp set spread out on the bench (i.e., basically in free air with no impediment to heat dissipation) found that it normally operated at 78 mA, or approximately a ½ watt per bulb, and bulb surface temperatures of around 35°C, measured with a Type-K thermocouple probe resting on the side of the horizontally-oriented lamp at the approximate location of the filament and with an ambient temperature of 24°C. To simulate multiple lamp failures, I simply physically removed five lamps and holders from the circuit, giving a 15 lamp set; this drew 90 mA, or approximately ¾ watts per bulb, and yielded bulb surface temperatures of around 50°C. I also failed the removed lamps by subjecting them to overvoltage and used one of these to successfully observe shunt operation. Finally, I removed five more lamps, giving a 10 lamp strand (i.e., 50% failed with double the rated voltage per lamp), resulting in a current draw of 117 mA, or around 1½ watts per bulb, with bulb surface temperatures around 80°C, which is notably in excess of the wiring insulation rating, again, without any thermal insulation over the lamp. After operating the 10 lamp strand and recording measurements, I disconnected power to make changes to my metering and, when relit, the set promptly failed as multiple lamps burnt out; further inspection revealed that some shunt wires also ruptured while the 3 amp fuses remained intact. And so, my testing was abruptly concluded.
Luckily for all of us, more rigorous and scientific studies have been conducted on this matter: Possible Fire and Electric Shock Hazards from Hot Lamps in Miniature Christmas Tree Light Strings and Decorations, by Fulcomer, is a 1979 report commissioned by the Consumer Product Safety Commission and is an excellent read on this subject and details rigorous testing and findings. Here is the abstract:
Failure to replace burned out lamps in series-constructed miniature Christmas light strings, or replacement with lamps of incorrect voltage rating, can lead to very high power dissipation by some or all of the lamps in the series string. Hot spot surface temperatures as high as 470°C were measured for lamps subjected to simulation of the above conditions. Additional testing showed that contact with surface temperatures above 390°C can cause glowing ignition, within two minutes, of cellulose material (e.g., tissue paper, decorative cotton) often found in the vicinity of Christmas light strings, and can cause flaming ignition in some samples of absorbent, untreated cotton. Furthermore, the hot spot surface temperature of normally operating lamps (no excessive power dissipation) can, particularly if the filament is off center, be higher than the melting temperature of some commonly used plastic insulating materials such as polyethylene, Contact between a hot lamp and insulating material can thus cause a shock hazard due to exposure of current carrying parts.
Concerning the melting of lampholders, such as by lamps contacting other lampholders or by the lamps being encompassed in decorative materials, the study was solely concerned with the shock hazard presented by exposing live parts; however, I put that such damages are a concern for the initiation of arcing and also present a fire hazard.
The study concluded with two recommendations. First, to reduce the potential for ignition, the study recommended a maximum wattage/time specification. Namely, that the lamp filament should be so designed as to burn out in a brief period of time (2 minutes) if operated at a dangerous wattage (determined by study to be 55.5 w), thus limiting the ability of the lamp to achieve dangerously high surface temperatures and transfer heat to adjacent combustibles.
Second, it recommended higher melting temperature insulating materials be required for lampholders.
This report informed the development of safety standards, such as UL 588 Standard for Safety for Seasonal and Holiday Decorative Products and CSA C22.2 No. 37 Decorative Lighting Products.
Unfortunately, I have been unable to source any studies on lamps manufactured to modern and improved standards and so I cannot conclude that direct ignition remains possible with modern incandescent sets, unless they are substandard. I aim to conduct further research along these lines and perhaps next Christmas will yield another article about same.
In the meantime, I hope that this article has been illuminating and I wish everyone a happy and safe Holiday season.
1 An official version appears to be unavailable from NIST, but the video can be viewed here: https://youtu.be/IwBiZtfjioU
2 Namely, transitioning from a fire in a room, to a room on fire.
3 Available at https://youtu.be/26A-49Wb2F4
4 Then again, in practice, the difference may well be whether or not ignition occurs at all: the authors of various studies referenced herein indicate that it is particularly difficult even to ignite a watered tree, with self-extinguishment being common.
5 With an estimated annual average of 200 home structure fires compared with about 30 million trees sold.
6 While stark, these numbers are an improvement on their 2002-2005 counterparts of 1/9 and 1/75 respectively.
7 Do check to see if you have any recalled products at http://www.healthycanadians.gc.ca/recall-alert-rappel-avis/index-eng.php
8 Which, it may be interesting to note, is only lit for part of the positive half cycle of the AC sine wave.
9 Longer strands, such as 100 and 200, will usually be composed of series of 50 lamps grouped in parallel.
10 Street lighting engineers had employed other solutions to this problem prior to developing film cutouts—which are functionally equivalent to the above-described shunts found in today’s mini lights—sometime around 1886(!).
11 Several online sources notably misreport that it is the heat of lamp burn-out/filament rupture that activates the shunt in Christmas lights, missing that it is in fact the elevated voltage that develops across the shunt.
Ahrens, M. (2017). Home structure fires involving Christmas trees. Quincy, MA: National Fire Protection Association. Fire Analysis and Research Division.
Ahrens, M. (2017). Home structure fires that began with decorations. Quincy, MA: National Fire Protection Association. Research, Data and Analytics Division.
Bigclivedotcom. (2014, December 28). Everything you didn’t need to know about tungsten Christmas lights. Retrieved from https://www.youtube.com/watch?v=Vt9wG8um7Rs
Brain, M. (2000, April 01). How Christmas Lights Work. Retrieved from https://people.howstuffworks.com/culture-traditions/holidays-christmas/christmas-lights.htm
Davis, G. B. (1938). U.S. Patent No. 2,211,884. Washington, DC: U.S. Patent and Trademark Office. Lamp embodying a shunt mechanism.
Engineering ToolBox, (2003). AWG Wire Gauges Current Ratings. [online] Available at: https://www.engineeringtoolbox.com/wire-gauges-d_419.html [Accessed 2018-12-05].
Fulcomer, P. M. (1979). Possible Fire and Electric Shock Hazards from Hot Lamps in Miniature Christmas Tree Light Strings and Decorations (Rep. No. NBSIR 79-1716). Bethesda, MD: Center for Consumer Product Technology; Center for Consumer Product Technology.
Lamm, W. (2014). Street Light Information Sheet – Understanding Shunts. [online] Kbrhorse.net. Available at: http://www.kbrhorse.net/streetlights/understanding_shunts02.html [Accessed 12 Dec. 2018].
Light Emitting Diode or the LED Tutorial. (2018, February 24). Retrieved from https://www.electronics-tutorials.ws/diode/diode_8.html
National Fire Protection Association (2015, December 09). Christmas tree fires can turn devastating and deadly within seconds. Retrieved from https://www.youtube.com/watch?v=xr6b9b8FYKk
National Fire Protection Association: Christmas tree and decoration fires. (2018). Retrieved November 28, 2018, from https://www.nfpa.org/Public-Education/By-topic/Seasonal-fires/Winter-holiday-safety/Christmas-tree-fires
National Institute of Standards and Technology. (2017, November 21). Christmas Tree Fire: Watered Tree vs. Dry Tree. Retrieved from https://www.youtube.com/watch?v=26A-49Wb2F4
NFPA 921: Guide for fire and explosion investigations. (2017). Quincy, MA: National Fire Protection Association.
Stroup, D., Lee, J., DeLauter, L., & Roadarmel, G. (1999). Scotch Pine Christmas Tree Fire Tests. Gaithersburg, MD: National Institute of Standards and Technology.
TheDanster999. (2008, December 04). EXCLUSIVE Christmas Tree Safety Tips from NIST (PSA). Retrieved from https://www.youtube.com/watch?v=dxaUj_rfkN4
Thompson, K. D. (2017, October 11). Video: Fire Safety for the Holidays. Retrieved from https://www.nist.gov/el/fire-research-division-73300/firegov-fire-service/video-fire-safety-holidays
Wikipedia contributors. (2018, December 11). Incandescent light bulb. In Wikipedia, The Free Encyclopedia. Retrieved 23:45, December 11, 2018, from https://en.wikipedia.org/wiki/Incandescent_light_bulb