Hi, friends!
I'm designing a split power supply for a class AB amplifier. The transformer produces approximately 16-0-16V, 2.5A. What continuous current should rectifier diodes be rated for? Will 5A be enough or do I need more? Perhaps for a short time when playing low frequencies the current will be greater? How much greater?? Or this short current peaks are not a problem?
I'm designing a split power supply for a class AB amplifier. The transformer produces approximately 16-0-16V, 2.5A. What continuous current should rectifier diodes be rated for? Will 5A be enough or do I need more? Perhaps for a short time when playing low frequencies the current will be greater? How much greater?? Or this short current peaks are not a problem?
The worst case for diodes is wen you turn on the amplifier... capacitors are not charged and rectifiers will reach maximum current ... so the value of filter capacitor is important .. and also short circuit current of the transformer ...
But even if you dont have all these values it is possible to figure out what is needed ... I think your 5 amp rectifier will do the job .. since initial surge current will last only a few cycles ...
But even if you dont have all these values it is possible to figure out what is needed ... I think your 5 amp rectifier will do the job .. since initial surge current will last only a few cycles ...
The answer is more complex, since it will depend on the filtering capacity you mount, the mounting layout and the transformer parameters.
The filtering capacity will in turn depend on the power supply ripple rejection offered by the amplifier itself in the listening range and the power supply noise level that you accept as tolerable. The tolerable power supply noise may also be defined by the sensitivity of the speakers you are using towards your listening point.
The most appropriate thing would be to start from the need imposed on the output of the amplifier and go in reverse towards the power supply.
The filtering capacity will in turn depend on the power supply ripple rejection offered by the amplifier itself in the listening range and the power supply noise level that you accept as tolerable. The tolerable power supply noise may also be defined by the sensitivity of the speakers you are using towards your listening point.
The most appropriate thing would be to start from the need imposed on the output of the amplifier and go in reverse towards the power supply.
Last edited:
One thing you can do to reduce the peak current at startup is a soft-start circuit. There are several methods to do a soft start. Of course does nothing for the peak/average required for music.
See here for how to design a soft start circuit: https://neurochrome.com/pages/the-ultimate-guide-to-soft-start-design
The current through the diodes is actually not that easy to calculate. The diodes only conduct for a short amount of time so the peak current gets pretty high even though the average current remains reasonable. The best approach if you want an exact answer is to use a tool such as PSUD2. You'll want to find the repetitive peak current in steady-state along with the average current and choose a diode that can handle this (plus margin) as a minimum.
If you're using discrete diodes rather than a bridge rectifier keep in mind that each diode pair only conducts for half of the cycle, so the average current is lower than the average output current of the total supply. A 5 A bridge rectifier or maybe a quad of 1N540x (3 A rated) diodes should work just fine. An equally valid approach would be to just bolt a KBPC1010 (10 A) or KBPC2510 (25 A) bridge rectifier to the chassis and calling it good.
For more information about which type of rectifier diode to choose, have a look here: https://neurochrome.com/pages/rectification-snubbers
Tom
The current through the diodes is actually not that easy to calculate. The diodes only conduct for a short amount of time so the peak current gets pretty high even though the average current remains reasonable. The best approach if you want an exact answer is to use a tool such as PSUD2. You'll want to find the repetitive peak current in steady-state along with the average current and choose a diode that can handle this (plus margin) as a minimum.
If you're using discrete diodes rather than a bridge rectifier keep in mind that each diode pair only conducts for half of the cycle, so the average current is lower than the average output current of the total supply. A 5 A bridge rectifier or maybe a quad of 1N540x (3 A rated) diodes should work just fine. An equally valid approach would be to just bolt a KBPC1010 (10 A) or KBPC2510 (25 A) bridge rectifier to the chassis and calling it good.
For more information about which type of rectifier diode to choose, have a look here: https://neurochrome.com/pages/rectification-snubbers
Tom
Generally, the continuous current rating should be based on the average current rating: since the Vf is supposed to be independent from the current, no rms considerations should apply; it is a simplification of course, but it holds well in practical situations, thus 3A diodes should be OK in your case.
Another aspect to consider is the startup: older databook use to provide data about the maximum capacitance vs. source resistance to stay within the non-repetitive surge current limit (Ifsm).
Nowadays, you need to do the work by yourself, with the help of a simulator for example.
Note that with transformer-based supplies, the Ifsm of a diode capable of handling the average current required will generally be sufficient because of the resistance and leakage inductance of the transformer.
With direct, off-mains supplies the situation is (was) different.
Nowadays, such supplies require a PFC in most of the world.
Note that there might be exceptions: schottky diodes tend to have a smallish Ifsm, and also if you create an auxiliary supply from a large winding using small diodes, say 1N4148, their Ifsm will likely be exceeded, even though it is a beefy 4A