Choosing Proper Ferrite beads

Choosing Ferrite Bead

Points to consider while choosing Ferrite beads for EMI suppression

• Ferrite bead is a frequency dependent resistor and the resistance of ferrite bead varies according to the frequency of signal passing through the ferrite bead
• Often Ferrite beads are the only solution for EMI problems
• Ferrite Beads with Low Q are absorptive beads - very lossy and hence absorbs the high frequency current noise and dissipates as heat
• Things to consider for choosing ferrite bead
• Unwanted frequency range of the signal flowing through the ferrite bead
• EMI source
• Expected amount of attenuation by the ferrite bead
• Electrical condition for the Design - DC Voltage, DC Bias currents, Maximum operating current, filed strengths etc
• Available board space to place the Ferrite bead
• Beads chosen without proper planning may become a source of EMI problems
• Often the impedance of ferrite bead drops to 20% of stated impedance when the current flow through the ferrite is high. The current saturates the ferrite and makes it to lose its inductance property
• AC resistance (Z) must be greater than the inductive resistance (X_L) of the ferrite bead. If this is not considered while choosing the ferrite beads, following problem in the figure is very common to occur:

Figure: Illustrating the negative effects of choosing ferrite bead without proper input

•  Refer to impedance chart of the ferrite bead. Normally, in the quick specification only at a spot frequency, the impedance value will be mentioned which is quite not a useful information. Referring to figure below, variation of impedance is so high though all spot impedance are the same (120 Ohms) for the 5 different ferrite beads.

Figure: 5 Ferrite beads with different impedance curves sharing common spot impedance value

CISPR 22 Limits for EMI/EMC test reference 

Table: CISPR limits for reference

References:

http://www.altera.com/literature/an/AN583.pdf

http://www.vishay.com/docs/ilb_ilbb_enote.pdf

http://www.digikey.com/Web%20Export/Supplier%20Content/TDK_445/PDF/TDK_InCompliance_Aug2010.pdf?redirected=1

Switch debouncing circuit for MCU Reset Pin

Check whether Reset Pin has internal Pull-up.
Always prefer to Pull it up using exernal resistors

Place a nominal value of 1 to 10 uF capacitor between ground and the Reset Pin. a 20 pF capacitor for ESD protection. Here it is assumed that Reset Pin of MCU is active low. It means, to reset the MCU, user has to give a low pulse (0 V) on Pin RESET for a specified duration. THe reset duration of MCU will be specified in the data sheet (Minimum). Always providing a reset until all power supplies have reached 90% of their nominal value is a good practice.

A resistor between power supply and RESET pin of MCU is required to delay in the reset pulse for the MCU further and also provides pull-up. The delay will be 4 times RC. A schotty diode with its positive terminal to Reset Pin and negative terminal to the Power supply will be helpful to kill out the positive going spikes which may occur during switch operation.

If space and price options are flexible, there are dedicated IC from Maxim, TI and other vendors which serves both voltage good detection and reset generation along with switch debouncing circuit, all in one IC.

Pictures to be uploaded.
waveform to be uploaded

Calculating LED series resistor Value

How to calculate value of series Resistor value?

Following are the the assumptions made in this post for example sake:

• Supply voltage - 12 V
• LED minimum forward voltage - 2.3 V ( From LED datasheet )
• LED maximum forward voltage - 3.1 V ( From LED datasheet)
• Maximum current allowed to flow in the LED - 20 mA ( Care has to be taken here by considering the maximum operating temperature, if relevant, as normally there will be a derating in the amount of current allowed to flow through the LED at higher temperature)
• Size of LED - Not applicable

With all the above assumptions, the resistor value can be calculated as follows:

R min = (Supply max - Forward voltage min)/ Maximum Allowed LED current;

Rmax = (Supply min - Forward voltage max)/10% of Maximum LED

 

RC delay in PCBs

RC delays are used to introduce a lag in the behavior of high speed signals such as SPI clock, Ethernet lines emmc lines etc. Some times RC are put in the circuit to have control over reflections and also to have minimum over shoot and undershoot.

To calculate delay inserted in the digital line,say SPI clock , for example, if a series resistance have been used of value 33 ohm and a capacitor near the clock input pin of slave SPI ( who receives the clock) of value 47 pf which are normal to see in many boards.. Delay is

5 * RC,
= 5 * 33 * 47
=~7.5 ns
it should be read as: SPI clock now takes an extra of 7.5 ns to reach low to high compared to previous case.
it follows from the same way that while reaching low the signal will take 7.5 ns more.. Some times, this will have impact on maximum operating frequency.. Too much RC is also not good as it increased the signal's include state for more duration causing more power consumption. There should be nice balance between minimum rc for signal integrity and maximum required operating frequency.
 

Series termination for SPI lines

Assuming both Master and slave SPI are CMOS devices. Few things to know about are the trace length of the SPI lines between two devices, Whether the master and slave are on two different boards, and maximum operating frequency, and rise time and fall time of the driver Pins (MOSI, CLK and CSEL on the Master and MISO on the slave). Series termination has worked quite well for devices on same board with the trace length of around 6-8 inches. Assuming driver internal resistance of 15-20 Ohms (if not available in datasheet), placing a 33 ohm standard resistor immediately next to the driver pins to match with the PCB impedance has been shown very much improvement in controlling reflections. Here a PCB impedance of 50 ohms was assumed. PCB impedance varies from board to board and a fine tuning can be done by changing the value of series resistor. Care should also be taken while measuring the overshoot, undershoot and reflections as some times the ground clip of the measurement probe if lengthier, the overshoot and undershoot on the SPI lines will be magnified a lot with ringing effects too. Hence, for measurement uses a ground clip very short (say 1-2 cm) and have ground reference very close to the measuring point. Assuming 20 MHz maximum operating frequency, the high and low level duration will be 25 ns. Assuming 4 time constants, the allowed time constant is 6 ns. Hence the capacitor value can be at max of 6 ns / 33.2 ohms =  37 pF. Assuming hidden 8-10 pF of capacitor, the capacitor which is being placed cannot be more than 20 pF. Capacitor value can be tweaked for best results for the typical application.