Capacitor Data

Capacitor Bypassing using multiple capacitors

Paralleling Capacitors

An Integrated Circuit [IC] is normally decoupled using one bypass capacitor from 0.01uf to 0.1uf. However the actual value changes depending on the recommendation from the data sheet. Some data sheets recommend placing two capacitors in parallel, with values that differ by a decade. The example transistor circuit uses both a low frequency and high frequency bypass capacitor. Using the graph below shows the benifit of using multiple capacitors each with a different value. For example; a 0.1uf frequency response would result in the first dip at f1 followed by the dotted single line. Placing a 0.01uf capacitor in parallel with the first cap also provides the second impedance decrease at f2. Both capacitors together, as a pair, tend to increase the overall frequency response. The example transistor circuit uses two capacitors in parallel per supply voltage.

High & Low Frequency Caps

Parallel Capacitor Frequency Response

The graphic above shows the benefit of placing two different value capacitors in parallel.
The impedance is reduced across a band of frequencies, adding the frequency response of the two nulls.
The first null is developed from the larger value capacitor, and the second null from the smaller value capacitor.
Placing a large value Tantalum capacitor next to a smaller value ceramic will cover a wide range of frequencies.
Placing two ceramic capacitors, a decade apart in value, in parallel will have the same effect but over a different frequency range.

Many Application notes require two or more capacitors in parallel each a decade apart in value to produce the desired effect.
App notes for larger chips, FPGA's for example, require compound by-passing.

IC's benefit from using parallel capacitors [compound by-passing] because the smaller capacitor is able to react very quickly to variations in voltage, smaller value capacitors have a lower ESR and lead inductance because the physical size is smaller].
While the larger capacitor is able to hold a larger charge to supply to the IC when required. Of course the two capacitors are also able to operate over a wider frequency range as shown in the graphic above.

Parallel Capacitor Placement Near the IC

Normal interface ICs that drive long traces, heavy loads, or backplanes may also require compound or parallel capacitors.
A bus driver IC would require compound by-passing to handle the two main functions of a by-pass capacitor.
First the smaller ceramic capacitor would suppress fast transients on the Vcc line [Voltage variations].
The ceramic capacitor would also be used to prevent Voltage Droop between nearby IC signaling [ICs on the same board].

While the larger tantalum capacitor would be used to supply larger amounts of current as the IC switched signals over the backplane [or cable].
Because the backplane represents a much heaver load, the IC would tend to switch slower, this would tend to be a benefit because the larger tantalum capacitor would also be reacting slower.
Board-to-Board communications always require a larger capacitor.

Design Hint; Remember to add in the trace and connector capacitance when calculating By-pass capacitors which are driving signals off-board.

So, in this example the Parallel Capacitors solve two design issues.
The capacitor pair both reacts quickly to small transient voltage variations on the voltage plane [by-pass capacitor] and accommodates large current requirements from the Integrated Circuit [IC], [Decoupling Capacitor].

Design Hint; note the placement of the two capacitors to IC 'B' and the location of the via.
Never place a trace via between an IC and its By-pass capacitor.
A via represents a low-pass filter and would inhibit the reaction time of the ceramic capacitor [defeating its purpose].

Back to the main Capacitor Design page, or the Capacitor manufacturer page, or Glossary of Capacitor Terms.

Design Note; the terms By-pass capacitor and Decoupling capacitor are used interchangeably and refer to the same action.
To decouple the IC from the power supply, having the decoupling capacitor providing any required short term current demands, which would otherwise cause a voltage reduction on the Vcc line.
Or to by-pass any voltage variations on the Vcc line away from the IC, so that the line remains at its steady state condition

Design Recommendation; When the design will only result in a deliverable quantity of a few dozen boards, add the larger parallel capacitor.
Adding the tantalum only results in a few extra cents to the design cost.
Better yet, if the tantalum is not required, it does not have to be installed and the ceramic capacitor still ends up next to the IC.
At the same time the pads are still on the board to accept the tantalum in the event it is needed.

However if the deliverable quantity is in the thousands than more analysis should be done to determine if the additional capacitor is required.

The smaller value capacitor should be nearest the IC, followed by the larger capacitor next to it.
A proto-type board would answer any questions regarding the IC's needs for bypassing.

The down-side to adding the second capacitor include the following design issues;
The PWB requires board space for the surface mount pads or drill holes for a through hole component.
An additional line item is required in the parts lists, if the part was not already being used in the circuit.
Another part would be required to be stocked, if the capacitor is new to this or previous designs.
A reduction in the boards MTBF would occur, because of the additional component.
... Assuming the MTBF is calculated by the parts count method.

Home
Distributors	Components	Equipment	Software	Standards	Buses	Design	Reference