Fullscreen HTML5Med Res (??MB)HTML4

The one that got away . . . A RATHER Pointless Project! SILICON CHIP staff design a lot of projects – and 99% of them appear in the magazine. Very occasionally, though, we have one that doesn’t work – or doesn’t work as intended – and in journalistic parlance, it’s “spiked”. Here’s the story of one project that did work perfectly but still didn’t make the grade. Only after it was built and tested did we come to the realisation that there was simply no point! Why then are we publishing it? Because it’s an interesting idea, if nothing else! I So what do you do if they fail? Because they are so cheap, t all sounded really good in theory. Take one of those ubiquitous, cheap 3-cell ultrabright LED torches and most people chuck the old one in the bin and buy another! So the theory behind the project is perfectly sound and add some circuitry that not only increased the battery it may well be that some readers may wish to adapt this life but also the life of the LEDs themselves. And while the project here does just that, it doesn’t make circuit for other uses. Indeed, some may wish to build this any economic sense, given that you can buy these torches for exactly the reason we designed it, just for the sake of for next-to-nothing, either from bargain shops or on line. doing so. It won’t be an expensive project; the parts will be readily Also given also that AAA cells are really cheap, no-one is likely to begrudge replacing them every now and then. available from companies such as element14 (and the PCBs If you’ve used one (or more!) of these torches, you’ve will be stocked by SILICON CHIP PartShop, just in case). One caveat; this was designed specifically to suit those probably found two problems: (a) the batteries don’t last very long and (b) the LEDs tend to fail much quicker than small 3 x AAA torches. It is not suitable for use with higher voltage models (eg, four or six AA, C or D batteries etc). you would expect. Anyway, enough of our sob story. Let’s have a look at the As we’ll explain, both of these problems have the same root cause – to get the very bright light output, the LEDs design – we are publishing both the circuit and compoare driven much harder than they should be, putting them nent overlay just in case you do want to put one of these together. in mortal danger. That extra current has to come from Design by John Clarke Typical torches somewhere, so the batteries Article by Ross Tester and John Clarke Low cost AAA LED torches don’t last long at all. 78  Silicon Chip siliconchip.com.au RESISTOR CELLS A 3 x AAA CELLS 9x WHITE LEDS A  TORCH ON/OFF SWITCH K A  K  A  A K   K K K DRIVER LED ANODES A A A 2x AAA CELLS  A A K  9x WHITE LEDS A  K K  TORCH ON/OFF SWITCH K Fig.1:FIG.1: a typical three LEDTORCH torch CIRCUIT circuit. Some TYPICAL 3 xcell, AAA nine CELL LED torches will have a resistor to limit current but many will have none, with a direct connection between the cells and LEDs. The LEDs are all paralleled. A  A  A K  A K   K A A  K A K  K  K K Fig.2: this howCELL theLED LED driver is used in the LED FIG.2: 2 xisAAA TORCH WITH LED DRIVER torch. It replaces the top (cell 3) AAA cell. The torch then runs from two AAA cells with lower current because the LEDs are not being driven as hard. tend to have a common design. Housed in an aluminium body, the light source usually comprises some nine LEDs with the three AAA cells arranged in a triangle pattern (side by side) within a holder. A switch is located at the bottom end of the torch. The circuit arrangement is shown in Fig.1. The three AAA 1.5V cells are connected in series to obtain a nominal 4.5V supply with fresh cells. The switch connects the negative terminal of the battery to the aluminium case, making connection to the cathodes of the paralleled LEDs. Some torches include a current limiting resistor as shown, connecting between the LED anodes and the battery plus terminal. But in many torches there is no resistor. With the torches we tested, the LEDs were severely overdriven, even including one that has a 2.2Ω limiting resistor instead of a direct connection of the battery to the LEDs. Yes, they’re bright – but they won’t last long. AA cell each would be discharged to a rather flat 0.505V before IC1 stops working. Circuit operation begins with transistor Q1 being switched on via base drive from the Vdrive output. This allows inductor L1 to charge up via the 3V supply, Q1 and resistor R1 . Transistor Q1 has very low saturation voltage (at around 10mV) which minimises power losses. Inductor current is sensed across the 150mΩ resistor. The inductor is charged until voltage at the Isense input (pin 5) reaches 25mV. This is at a current of 166.6mA and the transistor Q1 is then switched off for 2.5µs, allowing the inductor current to flow into the paralleled LEDs. For the nine LED torch that current is 18.5mA per LED. After the discharge into the LEDs, transistor Q1 is again switched on and L1 recharged. Note that the LEDs are pulsed rather than continuously lit. Inductor (L1) has a lower inductance and a higher DC resistance than is optimal for minimal power loss but was selected so that it would fit in the AAA cell space. A higher value inductor or one with a low DC impedance would be physically larger. Power from the AAA cells is bypassed with a 1µF capacitor. A Schottky diode connected across the supply is there to protect the circuit should the cells be inserted into the cell holder with reverse polarity. The diode shorts the battery voltage, restricting reverse voltage across IC1. Our driver The LED driver replaces one of the AAA cells. The circuit arrangement is shown in Fig.2. The design is such that the bottom of the LED driver PCB has contact with the plus side of cell 2. The LED driver negative connection is made with a length of wire to the torch case via the LED cathodes of the torch. The top end of the PCB is the anode output for the LEDs. Circuitry for the LED driver is shown in Fig.3. This is based around a single cell DC-DC converter, IC1, a low saturation voltage transistor (Q1) and inductor, L1. Supply for IC1 is directly from the two series connected AAA cells. The IC can operate down to 1.1V and this means that each Construction We’re not going to go into a lot of detail on construction because we don’t think many will be built. L1 47H 2x AAA CELLS 4 3 K D1 SM5822B S1 A (TORCH ON/OFF SWITCH) SC 2013 1F 1 Vdrive Vcc RE EM IC1 ZXSC100 GND BAS Isense FB 7 AAA CELL LED TORCH DRIVER 8 2 B WHITE LEDS CONNECTED IN PARALLEL C Q1 FMMT617 E A A  5 K A  K  K K 6 0.15 (150m) FMMT617 ZXSC100 8 Fig.3: based on a single cell DC-DC converter, inductor L1, is first charged 1 with Q1 conducting and when current reaches 166.6mA (25mV across the 150mΩ resistor), the transistor switches off and the inductor current flows through the LEDs. siliconchip.com.au A  D1 K C 4 B E A LEDS K A March 2013  79 16102131 Q1 IC1 C 2013 8mm OD FLAT WASHER R1 – AAA 13TORCH 120161 L1 LED DRIVER PCB TERMINAL PIN D1 TORCH CASE M3 x 10mm SCREW 470 UNDERSIDE These two oversize photos show the same view as the diagrams above; ie, of the top and bottom sides of the PCB respectively. They clearly show the way we mounted the washer and screw which form the connections to the torch. LED Current (mA) LED Current (mA) TOP OF PCB SILICON CHIP 1F Fig.4: component overlay diagrams showing both sides of the double-sided PCB. Q1 and the 150mΩ resistor are surface-mount components soldered to the top side of the PCB. The SMD diode (D1) mounts on the underside of the PCB. + Typically, low-cost AAA LED torch manufacturers do not power the LEDs correctly, applying excessive current with fresh cells. They apply this excessive current either via a direct connection of the LEDs to the AAA cells or via a low value resistor. The reason for over-driving the LEDs is probably so that the torch appears to be very bright. But this brightness is at the expense of the LEDs. So what are the consequences for the LEDs? The graph at LED forward voltage against Current right shows the 120 typical forward 100 voltage of the LED with current. For a 80 direct connection 60 of a 4.5V battery to the LEDs we 40 can expect some 20 120mA through each LED. In prac0 2 2.5 3 3.5 4 4.5 5 tice the current Forward Voltage (V) does not quite reach this extreme due to the internal resistance of the battery. With nine LEDs, the battery cannot deliver 120mA to each of the nine LEDs, just over 1A total. LED current is therefore not quite so severe. Actual LED current will depend on the cells, whether alkaline or zinc-carbon, and the cell voltage. Another complication with paralleled LEDs is that they do not share the current equally. The differences between each LED’s forward voltage with current will mean that some LEDs will draw more current than others. That imbalance is made worse as the higher current drawing LEDs increase in temperature and draw even more of their share of the current. For equal current sharing, the LEDs should be connected in series and driven from a higher voltage current limited driver. We use the words severe and extreme when mentioning the LED current because 5mm LEDs are just not rated for the current they are subjected to. Absolute maximum for any 5mm white LED that we can find among 10 well-known LED manufacturers is 30mA. And that maximum current is at 25°C ambient temperature. But at a room temperature of 25°C, we measure the LED housing temperature at some 36°C when each is driven at 20mA. This temperature rises to as high as 53°C at 100mA per LED. The lower graph shows that maximum LED current at 36°C is about 25mA and Maximum LED Current Derating below 20mA at with Temperature 53°C. This graph 50 is typical of most 45 5mm white LEDs 40 and shows that 35 30 the LEDs when 25 directly driven 20 from a 4.5V bat15 tery are severely 10 over driven when 5 compared to the 0 0 10 20 30 40 50 60 70 80 90 100 recommended current of 20mA. Temperature °C However, we have shown the component overlays for both sides of the PCB (Fig.4) just in case. . . One point to note is that inductor L1 is mounted unconventionally – it fits within a rectangular cutout in the PCB. And if you’re trying to shoe-horn the PCB into a torch housing, you’ll almost certainly need to lay over the 1µF capacitor to give clearance. 16102131 Driving LEDs in low cost torches 80  Silicon Chip And here’s how the PCB fits inside the 3xAAA battery holder, replacing the top cell. The green wire emerging from the PCB connects to the torch case forming the negative connection. The most convenient connection point is actually the copper track for the LED cathode connection points on the PCB shown above right. siliconchip.com.au Fig.6: this tiny PCB is found in most mini torches and is the way the LEDs are mounted, to connect to the battery pack. Not all torches have the series LED CENTRE ANODES resistor – but even in those CONTACT SPRING that do, it doesn’t achieve a great deal! FIG.5: REAR OF TYPICAL LED ARRAY PCB RESISTOR (IF INCLUDED) LED CATHODES Parts List – LED Torch Driver 1 PCB coded 16102131, 42 x 10mm 1 9-white LED 3-AAA cell torch 1 47µH 1.1A 230mΩ inductor 6x6mm SMD (L1) (Murata LQH6PPN470M43L [Available from Element14 Cat. 178-2814]) 1 M3 x 10mm pan head screw (head diameter 5mm, 2mm thick) 1 flat washer 8mm OD 1 PC stake 1 25mm length of 0.7mm tinned copper wire 1 50mm length of medium duty hookup wire Semiconductors 1 ZXSC100N8TA single cell DC-DC Converter (IC1) [Available from Element14 Cat. 113-2759] 1 FMMT617 NPN switching transistor (Q1) [Available from Element14 Cat. 952-6420] 1 B320A 20V 3A or SM5822B 40V 3A Schottky diode (D1) [Available from Element14 Cat. 185-8605 or Jaycar ZR1025] Capacitor 1 1µF monolithic multilayer ceramic Finally, the way the whole thing is assembled (whether it has the standard three cells or two cells and our driver circuit). And while our driver works fine, at the price these mini torches sell for it’s hardly worth the effort! Resistor 1 150mΩ 250mW 1206 SMD (Yageo RL1206FR-7W0R15L [Available from Element14 Cat. 806-7597] (for a 6-LED torch use 220mΩ 250mW (Yageo RL1206FR7W0R22L). [Available from Element14 Cat. 8067600RL] Tests Rela�ve Light output (%) Current (mA) We ran some discharge tests using some commonly available Light output was measured by shining the LEDs onto a 1.5V torches including two torches that had a direct connection between solar cell panel using a jig that held the light beam in a consistent the battery and LEDs and another (Xtreem brand) that included position and that prevented ambient light entering the panel. The a 2.2Ω limiting resistor. The discharging was continuous, mean- output from the solar panel was measured by placing an 18Ω load ing that the LEDs were driven until the cells became flat. You can across the terminals and measuring the voltage across this resistor. expect more life from the batteries in normal use when the torch The measurement effectively is short circuit current flow. Output is only run for short periods. voltage from the panel at the 100% level was 153mV. To enable direct comparisons we set the 100% reference light The graphed LED current is that calculated from the total torch output level as the initial value for the Xtreem torch using Extra current for an individual LED. Heavy Duty (Zinc-carbon) cells. That’s also the light output from So for a nine LED torch, the total torch current was divided by the same torch after 10 minutes using Alkaline cells. We arbitrarily 9. To calculate the current drawn from the AAA cells multiply the deem the batteries flat when the light output reaches 50% of the individual LED current by nine. 100% level. LED current was calculated by measured the voltage The discharge curve shows the point where each AAA cell Unbranded drop across a 0.1Ω resistor in series with the LEDs. 9-LED Torchreaches 1V. At that voltage the cell can be considered flat. SC 155 150 145 140 135 130 125 120 115 110 105 100 95 90 85 80 75 70 65 60 55 50 45 40 35 30 25 20 15 10 5 0 using three AlkalineAAA AAAcells cells Unbranded 9-LED Torch using 3x Alkaline (0 series resistor) (0Ω series resistor) Deemed Flat 230 minutes Light Output Single LED Current Maximum Allowed LED Current Recommended LED Current 0.1 1 10 Time (minutes) siliconchip.com.au 100 1000 Cell Voltage=1V March 2013  81