August Issue of The Beacon is now available

The August Issue of The Beacon has been uploaded and now available for your information and entertainment.

Comments

Charger Article Errata

Some comments regarding the charger article:
This was a good learning exercise but not a design that others should copy, yet. But it is good to see someone building something and making the effort to lay out a PCB.

  • Beware of counterfeit LM2576 Simple Switcher regulators, and boards using counterfeit regulators, on ebay and elsewhere. These parts are lower efficiency and overheat at a lower current than the real parts. If you see a part for less than $2 , chances are it is a counterfeit. Qty 1 prices for real parts tend to be in the $2-$7 range.
  • Most commercial chargers use the much cheaper MC34063 switching regulator "jellybean" (cheap/ubiquitous/multisouce) IC instead of the high priced TI part. This is a $0.47 (mouser) or $0.56 (digikey) qty 1 part. Or $36/100 (digikey). It also has a current rating that matches standard USB connectors. Unless you use one of the fixed voltage versins of the LM2576 or go over 1.5A (which requires an external transistor), the LM2576 doesn't have a lower parts count than the MC34063 so "simple switcher" is mostly marketing gimmick to get you to use a $3 part instead of a $0.50 one. Though it does switch at a higher frequency allowing use of smaller inductors (though that doesn't seem to have been taken advantage of here).
  • You can buy a 12V/13.8V car cigarette lighter to 5V USB converter for $1 at the dollar store. The PCB may easily be removed from this device and it can be outfitted with powerpoles. See note below about filtering, though
  • The charger board described in the article does not comply with USB standards but instead with the apple bastardized charging scheme. Instead of the resistors to +V and GND from D+ and D-, you should short D+ and D- together to charge standard devices. At the very least, the board should have had provisions for a zero ohm resistor or jumper between d+ and d- in addition to the sites for the non-standard apple resistors. Cell phones, unless made by apple, do not generarlly use the apple method. Many devices ignore charger signalling entirely and just draw what they want. Some comply with the USB standard. And some use a chip that tries to detect both USB standard charger signalling and non-standard apple signalling. Between those that try to detect the charger type and those that completely ignore charger detection, you can easily be fooled into thinking you have a working charger when you don't. Most hobby circuits completely ignore charger detection.
    USB Battery Charger Spec 1.2: http://www.usb.org/developers/devclass_docs/BCv1.2_070312.zip
    USB Battery Charger Spec 1.1: http://www.usb.org/developers/devclass_docs/batt_charging_1_1.zip
    Most USB microcontrollers do not have charger detection built in and many hobby circuits don't have a charger detect circuit so they draw power without asking permission.
  • like the $1 charger, the K4WLE constructed charger has no significant EMI filtering (just the input/output caps). No inductors, no ferrites, etc. So it will splatter all over the HF spectrum. Recommend at least one filter inductor and an extra cap on the other side of the filter inductor on both the input and output.
  • This K4WLE switcher signals to the downstream devices how much current to draw, using a non-standard method no-less, and then depends on them actually obeying. And even gives one port permission to draw more current than the connectors can safely handle. It is better to have a separate converter for each USB output. This way a greedy device doesn't drag down the voltage for the other device. USB charger spec relies on current foldback by the regulator when current has been exceeded. Also, the circuit shown delivers 2A into one connector, which exceeeds the standard connector rating of 1.5A and can deliver 3A into either connector into a short or a device that ignores non-standard apple signalling. Current should be limited to 1.5A on each connector.
  • No PCB pads for PC mount powerpole connectors.
  • PCB layout was not shown for switching version. Poor thermal design on one shown. Trace widths and heatsinking not adequate on linear PCB. Breadboard did have a heatsink but it looks inadequate for the 13.2W dissipated at 13.8V and 1.5A. Even with the heatsink pictured, roughly equivalent to a thermalloy 5903, which has a rating of around 10degC/W in still air, would result in a temperature increase of 130degC.
    Linear PCB had no apparent provisions for heat sink mounting. For the switching supply, a well designed PCB with lots of copper area would be much or all of the heat sink. However, being inside a constricted plastic case doesn't help matters. And it reportedly overheated in 60 seconds at less than full load.
  • This could have been used as an easy intro to surface mount construction but through hole components were used instead.
  • Charger detection ICs for USB devices include: FAN3989, USB375x, MAX14578, TPS65921, FSA831, STCC5011. Some of these will detect non-standard apple chargers. Also when designing usb devices note that there is another situation for a downstream device to look for: charging downstream port possible on hubs and computers which permits devices to draw charging power when not enumerated by the computer. Also an ACA dock where your cell phone or tablet is the master device but power comes in on the USB port.
  • There is at least one IC that can be used in a charger to try to simulate either a USB standard charger or an apple non-standard charger depending on what device is connected. However, it may misbehave when given a device that draws very little current as it switches between modes until the devices draws substantial current and this could result in anomoulous behaviour such as a device notifying the user: charger has been connected/charger has been disconnected, ad infinitum or not letting the device draw current when it needed because it was in the wrong mode. The TI TPS2511 is one such device:
    http://www.ti.com/product/tps2511
    Here is an application note that uses it: http://www.ti.com/lit/an/slva464d/slva464d.pdf
  • If you are building your own charger, consider putting a DPDT switch on the D+ and D- lines between a short circuit (standard compliant) and resistor dividers (non-standard apple devices). Use a 3PDT on-off-on switch and you can also turn off using the same switch: ON-OFF-APPLE switch.
  • User's need to be aware of apple's bastardized chargers as chargers you see in stores may be standards compliant or apple bastardized and they rarely tell you which and this can lead to some unpleasant surprises. You can tell which by cutting a USB cable so you can probe the 4 USB leads with a multimeter. A short between D+ and D- means you have a real charger. Resistor dividers between V+ and GND on D+ and D- mean you have a non-standard charger. And neither is another form of non-standard charger with no signalling. If you have an apple charger which can be opened and resealed (or which you plan to use the bare board), you can fix it by removing the offending voltage divider resistors and shorting D+ and D- with a solder blob or jumper.

    EEVblog Episode 110 focused on the MC34063.