Saturday, 15 February 2014

Power consumption - A issue to deal with in hand held units

Compared to embedded designs few years back, the stress on power savings has been more these days. May it be the increase in hand held products or whatever your circuit consuming lowest power is a very important feature. And this power saving is very critical in battery driven applications. As we know, most of these hand held devices are driven by batteries. With the Li-ion battery becoming prominent because of it's advantages over Ni-Cd battery it is being used in the hand held units. So, for any designer of a handheld unit first he has to justify his battery usage. More the power consumption lesser the battery life and also it is the other way. 

To a no-voice in embedded designs, the power consumption may be confusing. Consider a battery which has a rating of 1000mAh, this rating means that  a battery can supply an amount of 1A continuously over an hour. So, if your circuit draws for example 1A, then your battery lasts for only 1 hour. In designs, let us assume if you can vary your circuit current consumption, as per the need of the hour then there is always a chance that you give your battery more life before next recharge.

So, stressing the battery as per the need always helps battery to last long. To understand more, let us take the example of your mobile, you have a setting in your mobile to power off the back light after some specified amount of idle time. The lesser you make the idle time before the back light goes off the more it is helpful to increase battery backup. This is very high level from a user point of few. But from the designer end, he has to take to care of many things in his design before an end user can start getting the benefits like the ones we mentioned above.

Let us take a design into consideration, where we have a processor and it's associate circuitry on the board which are powered by battery. The below list indicates the opportunities a designer has to reduce power consumption.
  • Use different states of processor efficiently: Sleep, idle, standby, normal, run
    • The more you keep the processor in idle state the more you save power. You can observe some of your gadgets sudden having a back light enabled when you press a key or because of some other action. What happens in this case is that your processor will be in idle state with only some peripherals working. So, when you press a key, the processor gets an interrupt and comes to normal mode.
  • Try to optimize system clocks and processor clocks. Clocks are one of the major contributors in the circuit in terms of power consumption and using them efficiently can help you in reducing power consumption.
  • We see these days inter ic communications being operated at a low signalling levels which helps reduce power (Ex: LVDS)
  • Reduce the capacitance on your board as much as possible
  • Take care that your circuit doesn't draw excessive power during start ups.
  • Isolate the logic from processor and use other circuitry which consume low power (Like implementing digital logic in low power device like CPLD)
  • Wherever possible implement your logic in interrupt mode rather than polling.
  • Understand the glue logic requirement on your board and implement it in a efficient way.

Sunday, 9 February 2014

Braking resistors

Most of us must be familiar with the normal resistors like the Chip type, MFR type, etc. There are some of these applications where we have to handle high power. In such cases you use power resistors. There are many applications in which we use a molded power resistors in high power applications. We have to remember that most of these resistors are metal clad and mainly aluminium housing. Aluminium is most preferred as it is economical and also environmental friendly. Some companies even make custom resistors, shape and value desired as per the end user application. 

There are some applications where the braking of the device generates electrical energy which can be absorbed at that instant or it flows back into the supply system. We call it regenerative braking or rheostat braking. Braking resistors are one kind which find application in motors. the main purpose of the braking resistors is to slow down the motor and drain out excess voltage. This helps the motor to be within safe tolerances. Under normal operating conditions of motor, it converts electrical to mechanical energy, but when motor is powered off, the magnetic field generates back emf which must be controlled for smooth power off of motor and keep it within safe tolerances. The braking resistor is used for the purpose. As we know, DC motors have permanent magnets which creates this electrical energy. In a AC motor, the generated magnetic field creates this effect.



Selection of braking resistor is a design challenge. The heat generated by the resistor on power dissipation is high and requires heat sink arrangement. this adds up to the cost. If the requirement is to stop the motor quickly, the resistor value must be less and dissipates more heat as it allows large current to flow through.

General ratings:
  • Resistance (in ohms)
  • Power dissipation (kW)
  • Wire wound resistors

Saturday, 1 February 2014

Understanding Ethernet terminology - Part 1

Most of us must be familiar with terminologies like switchers, routers, LAN, WAN, VAN, etc., and most of us might be in this domain trying to understand the network topology, switch designs, router usage, IP addresses, MAC addresses and others. Let us try to understand the basic terminologies used in Ethernet in this series of articles:

Before going to other terminologies, we have to  understand the OSI layer stack. OSI is nothing but open system interconnect defines the system architecture. The 7 layers are Application, Presentation, Transport, Session, network, Data link and physical layer. If you are hardware engineer you must be familiar with physical layer which defines the electrical characteristics. Firmware engineers will be more familiar with the Data Link, network and other higher layers. The transmission happens in packets and protocol is defined by TCP/IP. The layer naming convention is as follows:

Layer 1: Physical Layer
Layer 2: Data Link
Layer 3: Network
Layer 4: Transport
Layer 5: Session
Layer 6: Presentation
Layer 7: Application

The layer 1 receives packet from the medium and hands it over to layer 2. In other words Layer N+1 requests data from Layer N and data flow happens in that sequence.

For example, Switch is basically a Layer 2 device and Router is basically a Layer 3 device.