back to learn
back to power supplies

published: Jun-17-2015, last edit: Mar-16-2017

post separation
To snubber or not to snubber … and how..

Sometimes in power supply circuits there are (small value) capacitors mounted in parallel to the rectifier diodes accompanied with many claims of improved sound quality. Not many people know what they are for but still many people just solder them in parallel when they aren’t installed. Sometimes these capacitors are used to ‘short’ high frequency signals or are used to lower RFI emissions generated by diodes or the circuits behind it.
Below an explanation as to why and when and how snubber circuits are used.

A snubber circuit may be needed if the circuit is connected to mains and the whole circuit is not allowed to induce RF (Radio Frequencies) back into the mains, for instance when lacking a proper mains filter. The reason a rectifier can do this (generate RF signals) is explained below.

The schematic of a standard bridge rectifier. As explained above a rectifier is only working a short period of time. Namely only when the AC voltage on the input of the rectifier is about 0.6V higher than that on the reservoir capacitor behind the rectifier. This is shown in the picture below where in red the diode is ‘conducting‘ and thus ‘on‘ and during the pink time period the diode is ‘off‘ because the voltage of the reservoir capacitor is higher than that on the input of the rectifier.

rectbr2 on-offWhen the diode is not conducting it actually becomes a (somewhat voltage dependent) capacitor with a very small value. For the well known 1N4007 this capacitance lies in the order of 10pF. For Schottky diodes this usually is much smaller and can be below 1pF.  This means that during a small period the diode in the rectifier is a diode (a switch that conducts ONLY when the input voltage is higher than the output voltage) and the entire period it doesn’t conduct it becomes a capacitor. Below you see the substitute diagram for this.
What may be obvious here is that the reservoir capacitor and load has been replaced by a wire (short) which at first appearance seems strange but the reason for this is that a reservoir capacitor is very low impedance for higher frequencies (certainly when decoupled with a smaller ceramic capacitor) and is thus irrelevant for the explanation/occurring phenomenon. As can be seen there are 4 diodes of which 2 pairs of parallel capacitors that are effective in series. They thus form a single capacitor of the same value of a single diode. The secondary winding of the transformer is an inductance. In practice the transformer also has some capacitance and the wiring between the transformer and rectifier also form a capacitance. This means the capacitances are in parallel (and thus can be added) to the inductor = Cdiode + Ctrafo + Cwiring. This creates a resonance circuit, but as the frequency present on the mains is just 50 or 60 Hz nothing really will resonate as the resonance frequency of that circuit is in the + 1MHz range… so no problem it seems.

resonance circuitBut at the top of the voltage of that 50/60Hz frequency the capacitor is replaced by a ‘switch’ and shortly after it became a switch it goes off again and becomes a capacitor. At this switching point high currents/voltage and speeds are present and thus energy is involved.

Most people have pulled a mains plug with a device that was active quite a few times in their lives and sometimes, when the mains was at its peak, you can see a massive ‘flash’. Also visible in light switches e.t.c. The same can happen with a mains switch. The voltage peaks that are generated at that point can be very high and the energy as well, creating a spark as the circuit basically doesn’t like to be interrupted.
So when a circuit is switched off on a high peak voltage on the  inductor-side is the result. In the case of a bridge rectifier this happens 100x (or 120 times) per second. Because the inductor has capacitor(s) in parallel that peak becomes a resonance (slowly dying sinewave).

resonance hitAs the diodes are in reverse mode (so not conducting) they do NOT form a load and the resonance circuit will ‘swing’ out as shown above. The frequency of this resonance is thus determined by the inductor (trafo) and all the capacitances in parallel to it (Cdiode + Ctrafo + Cwiring) and lies in the MHz range. When the capacitance of the wiring is very low (short wires wide apart) and the trafo has little capacitance the influence of the diodes is bigger. The higher powered the diode the bigger it is and the larger the capacitance. When using Schottky diodes this means the resonance frequency will be higher than  when ‘normal’ diodes are used. Higher frequencies usually means more potential trouble and higher ’emission’ by the way.
For this reason you sometimes see capacitors in parallel to diodes which effectively enlarges the capacitance and thus moves the resonance to lower frequencies… BUT they still are there. below a picture of a diode with a capacitor in parallel. On the right you see what it essentially becomes in reverse voltage. The capacitors used in parallel should be around 1nF in practice. The diode capacitance of 1pF or 10pF then has no influence any more. So when using a Schottky diode instead of a ‘normal’ (bipolar) diode they are equally ‘fast’ and have the same ‘capacitance’. The only difference being the voltage across the ‘switch’ which is between 0.5V and 1V for bipolar ones and between 0.2V and 0.8V for Schottky diodes. Using Schottky diodes thus only creates higher RF (when no counter measures are taken) and ensure a slightly (0.4 to 0.8V higher DC voltage on the reservoir capacitor.

diode capAs stated the resonance still exists at the inductor side (NOT at the DC side, the reservoir capacitor) and isn’t damped as there is no ‘load’ to damp it. With added caps that frequency is lower and thus less harmful in most cases. What is needed is a load on the input side of the rectifier. A load that can dissipate the energy of that peak/resonance and turn it into heat.

snubber circuitFor this we need a damping resistor that is either in series (above on the left) or in parallel (2nd from the left). A resistor in series is NOT possible as the resistor would be in series with the diodes severely limiting the current and thus power. We will need a resistor in parallel to transformer coil. To dissipate energy it would have to be a low value (between 10Ω and 100Ω). Needless to say that when using a 15V trafo and 10Ω resistor  in parallel this resistor would dissipate 22 Watt and become steaming hot. Not very practical….
Fortunately the resonance frequency is very high. This means that when the damping resistor has a capacitor in series that is much higher in value than that of the total capacitance of the circuit that capacitor in series with the damping resistor will act as a ‘short‘ (impedance will be close to 0 Ohm) for those frequencies. For 50/60Hz the impedance of that series capacitor should be high so at 50/60Hz there will be practically no AC current flowing. For this the capacitance needs to be small.  In essence the snubber circuit on the right needs to have a capacitor (C1) with a value much higher than that of the present capacitance and small enough to limit the current for 50/60Hz AC currents. R1 needs to dissipate the power of the resonance + the power from the 50/60Hz AC. When R1 = there will be no damping only the resonance frequency will be lowered. When this resistance value is too high there will also be no damping as there is no load.

Practical values of C1 will be between 47nF and 1μF and practical values of R1 will be between 4.7Ω and 47Ω depending on the situation.
In general it is better to use a 220nF and 10Ω resistor than not use anything at all. If you do not care about a little bit of RF common mode energy fed back into the mains you can leave the snubber circuit out.

This results in the circuit below and shows the proper way to prevent RF emissions back onto the mains via the transformer (common mode currents)

snubbered rectifierIn short, capacitors in parallel to the diodes only shifts the problems in frequency (which makes it less problematic) but mounting a snubber circuit in parallel to the transformer/rectifier input ensures the resonating energy is removed quickly and eliminates the possible RF garbage source.

Does this have any influence on the ‘sound’ off the amplifier ? Most likely not… Does it reduce radiated emission ? Yes it will.

post separation

back to learn
back to power supplies


  1. gello says:

    sir, i have a problem with my circuit im using cd4011 ic… what the problem is that when i plug in another device ex: cellphone charger my circuit will be interupted thus will result unpredictable output in my circuit. the circuit will change stage ex: the output is low it would become high and vice versa. so im thinking that the problem is in the power supply… so you have an idea on how to solve it sir. please i will wate for your answer. thanks much… 🙂

    • Solderdude says:

      Looks like a power supply dip issue. Possible solutions are increasing the reservoir capacitor or feed the circuit’s power supply through a Shottky diode followed by a large reservoir capacitor.
      Also make sure the 4011 has a 100nF cap directly between pin 7 and pin 14 and always connect the inputs of unused ports to either the power line or the 0V point.

  2. saads says:

    good article keep it up. but to calculate RC Snubber read this ref page.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.