Fullscreen HTML5Med Res (??MB)HTML4

Can your USB port take the heat? DEAD SIMPLE USB BREAKOUT “BOX” The USB port has made PC expansion so delightfully simple it’s a wonder no-one thought of it before . . . but it has its limitations. Most PC users are completely unaware of this and wonder why the computer starts giving error messages or the USB devices themselves either stop working or misbehave. Here’s a really simple way to find out what those devices are doing. . . C omputer interfacing via Universal Serial Bus (USB) ports, in either 12Mbps (Version 1.1) or more recent 480Mbps (Version 2.0) offerings, has deservedly become such an indispensable connection method that it’s hard to realise USB has only been in use for just five years. Apart from seamless “smart” data connections for digital cameras, flash RAM dongles, WiFi adapters, modems and mice, etc, the availability of a regulated 5V DC supply at relatively generous currents has also lead to such diverse “dumb” devices as mobile phone chargers, coffee cup warmers, cooling fans and inspection lights. Dumb, of course, refers to the hi22  Silicon Chip tech USB data lines (middle pins 2 and 3) being ignored and just the low voltage DC being exploited at (outer) pins 1 and 4. The PC’s USB port can supply up to 500mA <at>5V; however, downstream ports on USB devices are generally limited to 100mA maximum. Although it may be considered frivolous to use a $1000 PC to just power a light or charge batteries, the computer may be on anyway and equipment powered by the 5V USB supply is often conveniently associated with one’s needs at the time. by Stan Swan* Cold coffee, when trying to tame a late night spreadsheet macro, can easily ruin one’s concentration! The USB specification allows up to 5m of connecting cable, since signal timing issues may give data corruption with longer lengths. However, basic DC electrical issues also arise, with heavier currents giving unacceptable supply line voltage drops that infringe typical 5V ±0.25V load electronics needs. As an example, Ohm’s law tells us if 250mA is flowing through a wire of resistance 0.5W, then a tolerable drop of I x R = 0.25 x 0.5 = 0.125 Volts will occur. This wire with the 500mA maxisiliconchip.com.au mum load, however, would drop 0.25V and deliver only a borderline 4.75V to the load, which may therefore work unreliably. An additional issue relates to the power needs of the numerous items now in use. Even with short cable lengths, multiple USB devices (and up to 127 are possible!) can eventually demand more current than is permitted, with resulting port shutdown. Unpowered hubs are particularly prone to this, which explains the need for powered USB hubs that will cater for multiple energy-hungry add-ons. Alternatively, a simple USB power injector could be used with an unpowered hub to achieve the same result. SILICON CHIP published such a project in the October 2004 issue, capable of supplying 5V DC <at>1A to a USB device. Need more power than this for other devices? Add more power injectors! Even though itself now threatened by other advances in technology, Bluetooth is finally making a solid showing. Particular interest in verifying the output power of USB Bluetooth adaptors has therefore arisen. The three Bluetooth classes have decreasing wireless power and range, and innocent purchase of a low power Class 2 adaptor may frustrate when Pin 1: +5V Pin 3: Data 2 Pin 2: Data 1 Pin 4: 0V The connections to a USB type-A plug (the one that goes into your PC). range has to be maximised. The most powerful Class 1 is typically good to 100m, with Class 2 some 10m and Class 3 just a metre “across a motherboard”. Although not specified in classes, Here’s Stan’s prototype USB breakout “box”, potted in some hot-melt glue. The idea is to poke your multimeter probes into the terminal block (avoiding shorts) for voltage measurements; for current measurements, you remove the header pin shorting block and connect your multimeter in series. (instead usually having power quoted in dBm [15dBi = 101.5 ~ 40mW]), increasingly popular USB WiFi adaptors also differ significantly in their transmitter output power. This is especially important when some distance from an Access Point since you may be able to “hear” its strong signals – but it may not sense your weak out-going ones. Given these issues, it’s surprising USB supply breakout adaptors haven’t become readily available, since simple current and voltage measurement of assorted loads can be extremely revealing, especially if device specs are being stretched. Make your own! With the cheapness of short USB M-F cables and DMMs, a simple supply breakout adaptor can quickly be organised by just cutting the cable. Position of the cut is unimportant – BRAID, GREEN & WHITE CABLES REJOINED (SOLDERED) AND HEATSHRINK INSULATED CABLE CLAMP USB SOCKET (M) USB CABLE TWO-WAY TERMINAL BLOCK siliconchip.com.au CABLE CLAMP RED & BLACK PAINT (TEXTA) HEADER PIN BASE USB SOCKET (F) just ensure the data wires (usually green and white) and the braid are neatly resoldered and heatshrink covered, with no mischievous whiskers from the braided shield! A small terminal block makes for convenient voltage test points across the red and black supply wires – even though a DMM won’t be worried by reverse polarity, we painted our terminal block red and black with a Texta pen to identify “+” and “–”. For current measurement (which of course must be in series), we included a 2-pin header pin base in the +ve (red) wire. To measure current, the shorting block is removed revealing a handy pair of terminals for our multimeter clip leads. Overleaf are some sample measured currents, using a short 600mm cable to typical loads, supplied from a mainspowered Toshiba laptop: This diagram shows how it can be done – slightly more permanently than the glue version above. We used a small block of timber to mount it on – but just about any non-conductive base would be fine. The header pin base was glued to the timber using hot melt; small wood-screws hold both cable clamps and the 2-way terminal block in place. October 2005  23 Device Measured load current (may vary with load demands) Pocket 40GB USB Hard Disk 300mA Atmel “b” WLAN adaptor DSE XH7947 (2002) 270mA DSE “b” WLAN adaptor XH6822 (2004) 90mA Genius “g” WLAN adaptor GW-7200U 63mA Logitech Quick Cam(2000) 60mA “My Flash” 256MB Flash RAM DSE Class 1 BlueTooth dongle (2003) XH4104 50mA 48.5mA “Itsy-Bitsy” USB LED lamp (ref SILICON CHIP, March 2002) 30mA Genius Mini Traveller USB mouse 10mA Prolific USB-serial D9 adaptor (Picaxe use) 8mA Olympus C-1 digital camera < 1mA (64MB Smart Media). Perhaps camera batteries supplying too? This laptop PC’s unloaded USB supply was measured at 5.04V, which dropped when loaded by the powerhungry Atmel WLAN adapter (drawing 270mA) to 4.88V with a 600mm cable and just 4.75V when at the end of a 5m USB 2.0 extender. This 5m cable was measured as having 0.5W resistance in the supply wire (therefore 1W, considering return too), which accounts closely with the example above. Adding another 5m extender dropped the load voltage to almost 4.5V, with the PC then reporting “ USB device not recognised”, presumably because of timing and low supply voltage issues. Given the lowered supSILICON CHIP’s ply voltage at the end of USB Power Injector these lengthy USB cables, from the October 2004 it’s feasible that heavier issue. It doubles the amount of power (paralleled?) DC supply available from a USB port. cables could be used if longer runs are needed and data propagation timing is not gised via cheap CAT-5 cable. an issue – perhaps to serve a rooftop Aside from the tedious energy isUSB webcam or “sweetspot” WiFi sues, ultimate cable lengths really are adaptor. limited by the data speeds. Although Some of the now-common external such signals travel near the speed USB hard disks and CD/DVD writers of light in cables, even a few extra actually have two USB connectors metres of conductor can delay things at the PC end, wired in parallel to unacceptably. achieve the currents required. They The recently-announced Wireless rely on the fact that (hopefully!) most USB still only offers modest ranges PCs these days have several USB port (3-10m) but given the lack of wires(!), sockets close together. there certainly won’t be any voltage Remote DC supplies via completely drops – or convenient 5V supplies. separate lines and a 7805 voltage Active Extender and doubled “Bus” regulator, perhaps with higher voltage power cables (Jaycar XC4839 and (9V?) initially, may serve to stretch WC7750 respectively) may of course runs as well. better suit demanding USB setups. This is precisely the scheme inBut if only modest extensions are volved in the SILICON CHIP USB Power needed, cheap passive extending Injector mentioned above; however it techniques are well worth consideralso included line sensing to turn the ing – if only to keep your coffee warm power on and off in sympathy with the while lounging some distance away power at the USB port itself. from the PC! Perhaps even a small photovoltaic With the breakout adaptor shown solar panel could be used to charge a above and a DMM, you can at least rooftop battery, providing a regulated be easily informed of your loads’ DC 5V supply. It’s rather akin to “corpodemands, something that currently is SC rate” power over ethernet (POE) aplargely unknown. proaches – well known for rooftop/ masthead WiFi Access Points ener* s.t.swan<at>massey.ac.nz This active USB 5m extension cable uses some of the power available from the USB port to amplify the data signal. Up to 5 can be connected in series. It’s from Jaycar (Cat XC4839). 24  Silicon Chip Where you have power-hungry USB devices, a USB Bus Power Cable can give them a boost, connecting to two USB ports to double power. Also from Jaycar (Cat WC7750). siliconchip.com.au