Since the subject of filter caps came up, I thought I'd share an article I wrote for the MetroAmp Wiki page so here it is. Filter Caps Explained By Jon Wilder of Wilder Amplification Filter capacitors (aka filter caps) are to electrons like buckets are to water. Just as buckets store water, filter caps store electrons. The larger the bucket - * The greater the volume of water it takes to fill it. * The longer it takes for a pump of a given flow rate to drain it. The larger a filter cap's value - * The more stored electrons it takes to fully charge the cap. * The longer it takes for a circuit of a given amount of resistance to discharge it. Like all other capacitors, filter cap values are measured in microfarads (1uf, or 1 microfarad is equal to 0.000001 Farad) and unlike other caps, they are only measured in microfarads. This is because any value in the nano or picofarad range is simply too small a value to store enough electrons for the cap to perform efficiently as a filter. The larger the value of the cap - * The more electrons it can store * The more electrons it will take to charge it up to the supply voltage. * The longer it will take to discharge * The more current it can supply over a given time constant before fully discharging Notes on filter capacitor configuration Capacitors work backwards from resistors in that - * Cap values are additive when hooked up in parallel (C1 + C2 + C3...etc etc = Total Capacitance) * Cap values are halved when two equal values are placed in series (2 100uF caps in series gives a total 50uF capacitance) * "Dual" filter caps of equal value placed in parallel with themselves, then in series with each other, gives the value of one of the dual caps. Example, connecting a 50uF + 50uF cap in parallel with itself, then placing it in series with another 50uF + 50uF cap connected in the same fashion (parallel to itself) will give 50uF total capacitance * 2 or more filter caps in parallel will have the same voltage rating as the cap with the lowest voltage rating * Voltage handling is additive with filter caps in series (i.e. 2 filter caps in series that have a 500WVDC rating will increase voltage handling to 1000WVDC) * "Bleeder resistors" are typically two equal value resistors arranged in a "voltage divider" configuration (see Series Circuits) across each cap for voltage balancing (i.e. to make sure that each cap sees the same voltage and one is not going over voltage). They also serve as a quick means to "bleed" the charge from the filter caps upon power down, hence the name "bleeder" resistors. What Is Happening In a filter circuit, filter caps act as "temporary batteries". When the DC pulse is rising in voltage, the cap is charging. The cap reaches full charge when the pulse hits its peak. Imagine your house is hooked up to the power company with a backup battery. As long as the power from the power company remains on, it powers the house while also keeping the backup batteries charged. Once you experience a power outage, the household electrical now runs on the battery backup. While the house is running on backup battery power, the batteries are discharging at a rate dependent upon how much electrons they have stored versus the amount of electron current being drawn by electrical devices in the house (i.e. supply & demand). As long as the household current draw does not exceed the amount of electrons the batteries have stored, the batteries will continue to supply power to the house until the power from the power company is restored. Once power company power is restored, the house will now run on power company power again and the power from the power company will recharge the backup batteries until you experience another power outage. This is exactly what is happening in the filter circuit. The only difference is that due to the pulsed DC from the rectifier, on a tube amp that runs on 60Hz line frequency, a full wave rectified power supply experiences about 120 power outages per second (120Hz = ripple frequency of a full wave rectifier). The illustration above shows the pulsed DC output of the rectifier circuit. You can see that there is a gap between the pulse peaks. These gaps represent the "power outages" that frequently happen so many times a second in a tube amplfiier. When the pulsed DC output of the rectifier circuit is at its peak, it charges the filter caps up. Once the pulse starts to fall from its peak and start its way back down to 0 volts, the circuit will draw current from these "temporary batteries" known as filter caps. This picks the voltage in the power supply back up to the peak value of the pulsed DC. However, as the circuit draws current from the filter caps, the filter caps discharge, which causes the supply voltage to drop at a rate determined by the filter cap value and the amount of current being drawn from it (again, supply and demand). The circuit continues to draw current from the filter caps until the DC pulse rises to a value above the remaining charge in the filter cap. The pulse then recharges the cap while the circuit now draws current from the main supply source, until the pulse reaches peak value and falls again, thus the cycle starts all over again. Below is a diagram that illustrates what pulsed DC looks like once it's been filtered by 1 stage of filtering. The dashed lines in the illustration represent the pulsed DC itself that has been filtered out by the filter circuit. As you can see, when the voltage from the rectifier wants to fall, the filter caps act like a buffer of sorts and take over as the primary power supply while the pulsed DC output is falling, then rising back up. The fall and rise of the rectified DC is still there...however, the amplifier circuitry never sees it because the capacitors keep the voltage up by supplying current to the amplifier during the fall and rise times, hence the "filtering" effect. As the circuit pulls electrons from the filter caps, the filter caps slowly discharge and drop voltage as they do so, indicated by the slope angle of the line in between the pulsed DC peaks. The slope angle of this line is indicative of and determined by the discharge rate of the filter caps. The faster the discharge rate of the filter caps, the steeper the angle of this slope, and vice versa. As long as there is an angle in that slope line, we're still not where we wanna be, but most of the filtering has been done and we are now much closer to a steady DC supply voltage. So what do we do? Add more capacitance! By adding more capacitance to the circuit, the filter circuit will be able to store more electrons, which gives them more electrons to supply for a longer period of time without dropping voltage. Once this happens, the output of the power supply looks more like this. The line representing voltage now stays constant between the peaks because of the higher capacity of the filter circuit...it takes longer to deplete the supply of electrons because there's simply more electrons available in the filter circuit to supply to the amplifier circuit. We can add capacitance in 1 of two ways - * Increase the value of the filter caps at the first filter stage or... * Add more filter caps across the first filter stage Most tube amplifier filter circuits use the latter approach. This allows them to use resistors between each stage that act as a multi-stage voltage divider to obtain different operating voltages for the preamp, phase inverter, and output stages. The resistors between each filter stage also provides isolation between the preamp, driver/phase inverter, and output stages so that each stage has its own "temporary battries" to pull current from and current is available on demand at each stage within the amp without one stage trying to steal current from another (i.e. "intermodulation distortion). Because there is a resistor that the current must flow through to get to the preceding amplifier stage, the current sees the stage hooked up to its capacitor source as the "easiest path" and instead flows through the stage that's connected to the filter cap itself instead of being "robbed" by the preceding stage. Tonal Effects of Filtering The amount of capacitance that a filter circuit has can have a great effect on tone as well as the dynamic response of the amplifier. Typically you won't notice it at low volume so much as you will at high volume, depending on how much capacitance the filter circuit has. What happens is that when the amplifier is driven hard, the plate resistance of the tubes is at its lowest, allowing the amp circuit to draw the maximum current that it can from the power supply. If the current demands of the amplifier exceed the amount of current the filtering circuit can supply (there it is again, supply and demand), you will experience a voltage drop at the supply, which some refer to as voltage "sag". This will have a great effect on how loose or tight the amp sounds. More sag = looser sounding/feeling amp whereas less sag = tighter sounding/feeling amp. Now this only comes into play if you have an amplifier that originally had a power supply that couldn't keep up with the amplifier's current draw demand and you bump the filtering up in the amplifier so that it can. If the filter circuit in your amplifier is already of high capacitance and can keep up with the current draw demand of the amplififer as it stands now, bumping the filtering up at this point won't make much of a difference, if any at all. You will eventually reach a point of diminishing returns. Summary The filter circuit is nothing more than a temporary power supply comprised of "temporary backup batteries" that supply the amplifier with power while the DC pulses from the rectifier are falling and rising. By adding more capacitance, our filter circuit can store more electrons and better supply the current demands of the powered circuit. The better the filter circuit can supply the current demands of the powered circuit, the less the filter circuit will discharge between the pulsed DC peaks and the straighter that slope will be. The trick is to get that slope between the peaks as straight as possible. The straighter that line between the peaks, the more steady the DC voltage at the power supply. It's all about supply and demand...the current supply of the filter circuit must be equal to or greater than the demand of the amplifier circuitry. The better it can do this, the more steady the supply voltage will be.