Electrolytic caps

How to bring up an old amp without fireworks

Capacitor basics

All capacitors consist of two conducting plates with an insulating dielectric material in-between.

This material stops direct current flowing through the capacitor, and also determines the value of the actual capacity.

Two factors influence the actual capacity.

The first is the dielectric constant of the insulating material used. Some materials enhance or magnify the capacitance more than others and these are said to have a high dielectric constant. With ceramic caps these are called high-K, and while this makes the cap physically compact it also tends to make the capacitance value unstable, thus the values of electrolytic caps are very nominal. Air is taken as the reference.

The second is the thickness of the dielectric material. Capacitance is proportional to the area of the plates, and inversely proportional to the plate spacing - so the thinner the layer the higher the capacitance.

But as the insulator gets thinner so its ability to withstand voltage without breakdown is also reduced. Naturally everyone wants the highest capacity and working voltage in the smallest possible volume.

One way of doing this is to form a dielectric of insulating aluminium oxide on the surface of an aluminium foil electrode using a hydroxide electrolyte. This process is basically electro-chemical and therefore this type of cap is called “electrolytic”.

Forming

During manufacture the creation of this layer is carried out through an electro-chemical process called “forming”.

Initially the capacitor has no insulating layer and so has a low resistance. By passing a limited current through the new cap the insulating layer is grown or “formed”, and the assembly then acts like a capacitor.

In normal operation the operating voltage tends to re-new this insulating layer.

Reforming

In electros that are not regularly subject to a polarizing voltage the insulating layer tends to decay and the capacitor becomes leaky.

This will happen with new caps placed in storage, sometimes sold as NoS, and to caps in an amplifier that has been out of service for an extended period, say a year or more.

When such a cap is re-energised it will typically go one of two ways.

The normal situation is that the loss of dielectric is small, so the resulting leakage current is also small. This leakage current will re-form the layer without excessive internal heating.

What can happen with old caps that have been out of service for some years however, is that the layer has gone and the initial leakage current very high, only limited by the supply.

Excessive current will lead to rapid internal heating and the hydroxide electrolyte boiling. This will cause the cap to explode violently without warning. I've witnessed a few of these, and it comes out of the blue.

Depending on the situation either the can or the innards are shot out. The innards happen to look a bit like duck-down spread all over everything, but it's conductive so it can cause a lot of secondary damage too.

Even a small cap could cause a serious injury, but a large cap exploding can be quite a serious risk. Modern caps often have a frangable vent or special criss-cross burst caps that will split but stay attached, but old caps may burst, fragment, or shoot out a spray of boiling hydroxide.

Limited power

Damage can be avoided and old caps may be salvaged by re-forming. This is much the same process used in manufacture, and applies equally to all electrolytic capacitors, low and high voltage.

There are a couple of ways to apply limited power. If you have a high voltage Megger this will serve to bring the caps up slowly with limited power (but remember a HT line charged by an innocent Megger to 500 volts is still charged to 500 volts and even sick caps can still bite hard).

Otherwise the amp power supply itself can be used. To do this you need to make up a load-limiting lead with a light bulb socket in series with the active.

You will need a selection of different wattage light globes, say 15 watt (or smaller), around 25 watts, and around 40 watts (DO NOT try to use a light dimmer).

Presuming that you have already checked out and made the plug, lead and mains wiring safe, and the power tranny and rectifier are known to be in good order.

Remove all valves to un-load the heater circuit (except a valve rectifier, if used).

Clip a voltmeter across the main filter capacitor (rectifier output) to monitor the HT voltage.

Power-up the amp (power = on, standby = run) with the lowest wattage lamp in your limiting lead.

The limiting lamp should light brightly (and get hot) as the power supply tries to charge up.

If the caps are in good order, or perhaps gone open or low in value, the lamp may dim quickly.

But if the caps are leaky it will remain brightly lit and the HT voltage minimal. Let the amp soak like this while keeping an eye on the HT voltage.

Over a few minutes the HT voltage should start to creep up. This is a sign that the caps are re-forming.

If not, the caps are failing to re-form, and if the HT is still near-zero after about half an hour, then rip them all out and fit new electros throughout. Done. Otherwise...

You can occasionally switch off and watch how the HT voltage decays. Initially it will be fairly rapid, but as the process proceeds the voltage build up will become quicker and the leak-down slower.

Once the HT voltage seems to have stopped rising with a given lamp you can move up to the next wattage and let the electros settle at that level.

A 40 watt lamp should be lighting only dimly and finally giving you a high HT voltage close to the expected value.

---

Low voltage high-value electros such as those from big solid-state amps can be re-formed by connecting to a suitable voltage bench power supply via a 1 meg resistor, and monitoring the cap voltage.

Again the applied voltage is started low and only increased in steps as the voltage on the cap rises to the supply.

It should be noted that in electro-plating an even fine grain result is obtained by using a very low current for a long time. High current produces the thickness in a short time but it tends to be lumpy. This is also true for recharging batteries, and NiCd in particular will have a higher capacity when charged slowly.

See also Elektor May 2006 p72.

<<<OzValveAmps