Condensators are used to attenuate the waves, ripples, glitches and other transient perturbations inside a current source line or an information signal. Alas, a part of the high frequency perturbations will not be "swallowed" by one condensator. This is why a full set of big and little condensators, resistors and self coils is often used. This text describes a single condensator structure that will attenuate the perturbations even better than a set of components.

A common big condensator is made of two insulating straps and two conducting straps piled above each other. an electric connection is made to each conducting strap:

The straps are rolled together to become a cylinder. That cylinder will be encapsulated inside a plastic or aluminum case.

The schematic drawing of a condensator for an electronic circuit is the following. The two horizontal fat lines represent the surfaces of the two conducting straps viewed sideways. The two vertical thin lines are the two electric connections.

The reasons why such a condensator does not manage to swallow completely the perturbations are numerous:

• The two electric wires of the condensator act like minute self-inductances. They hamper partially the perturbations to be "swallowed" by the condensator.

• The conducting straps do have a given electric resistance. So the parts of the straps situated far from the connection are partially isolated from the rest by the resistance of the strap.

• Very high frequency perturbations simply fork at the condensator connection. So, roughly put, the condensator cannot absorb more than half of the intensity of the perturbations.

To counter this the solution I thought about is to put four connections on the conducting straps:

The whole can too be rolled to become a cylinder. Yet this time with two pairs of connections coming out of the cylinder:

This could be the schematic drawing for that condensator type:

Using this condensator shape the signal can be forced to go through the entire condensator. The condensator acts like a transmission line, yet wit an opposite purpose: while a transmission line is designed to keep the signal shape unchanged, the four-legged condensator is designed to flatten the signal.

Tests with electrolytic condensators showed the attenuation for high frequency perturbations was virtually perfect. They could not be measured at the output. In one test, a standard two-legged big electrolytic condensator was used to shortcut a TTL signal. Yet the rising and falling edges of the TTL signal that came through still were powerful enough to trigger the TTL input gate at the other side. Then the condensator was opened and two more connections were soldered to the conducting straps, to make a four-legged condensator out of the standard condensator. This time, the TTL input received no more signal.

In some cases one of the two signal wires cannot be split. For example when it is a reference ground wire. In that case there is no choice but shortcut the wires of one of the straps. This reduces the yield of the condensator:

One solution is to put a lot of connections on the "ground" strap. The best way to do this is probably for SMD components: The "ground strap" can be also the case of the condensator, intimately soldered to the ground wire. The signal would go through two little connections entering the condensator, insulated from the case.

Other condensator shapes can benefit too from this idea. For example ceramic condensator made of a pile of a lot of little surfaces all directly latched to one of the two connection wires. Or tantalum sponge condensators. The point is the signal must be forced through the plates or through the sponge.