Continuous Conduction Mode, Discontinuous Conduction Mode and Critical (Boundary) Conduction Mode in boost derived circuits

So we’re talking CCM, DCM and CrCM (BCM)

Continuous conduction CCM and discontinuous conduction modes DCM refer to whether the current in the energy storage element (inductor) in buck or boost derived circuits goes to zero each switching cycle or not. In CCM it does not reach zero so at the end of every switching cycle there is some energy left which is the ‘topped up’ in the next cycle to provide enough energy for the output. In DCM, all the energy each cycle is used up and the inductor sits with no current and no stored energy for part of the time. Energy stored relates to output load so whether you are in CCM or DCM depends on load. If you design for CCM (and you do have a choice) there will always be a lighter load when the circuit swaps to DCM. If you design for DCM at max load and minimum input voltage, it will stay in DCM for all lighter loads and higher input voltages. All assuming fixed frequency operation. The value of inductance, which relates to energy stored sets which mode you’re in - lower inductance forces DCM. So which is better? In DCM, the switch current always starts from zero (no stored energy) so this is a low dissipation switching transition which is good for efficiency and EMI. However, the current then has to peak at a high value to store enough energy for the output. In CCM, the switch current starts from some finite value so is not a zero dissipation transition and power can be lost hurting efficiency. The peak current is not so high though as there is some energy already stored to use. At high power in DCM, the high peak currents can be a big problem and cause more power loss than saving so CCM is used. At low power, DCM is usually used as the peak currents are manageable. Also the loop transfer function of DCM is simpler and easier to optimise. In CCM peak currents are lower for the same power but as I said, it will cross to DCM at some lower power out. This means the transfer function changes and it’s not possible to have a control loop that works optimally in both modes (without a lot of extra complexity).

Having said all that, this applies when the output rectifier is a diode in a buck or boost. If synchronous rectifier MOSFETs are used, at least in a buck converter, they can be arranged to sink and source current so the inductor current can stay continuous all the time so CCM is maintained whatever the load.

You have to do the calculations to see what works best as it’s a compromise between EMI, efficiency, component cost and loop stability. As an example, up to around 30W I might design for CCM at full load dropping into DCM at about 50% load.

Often simple converters are designed to always work on the boundary between CCM and DCM by varying their frequency with load variations. This is termed boundary or critical conduction mode. CRM. This can be an extremely simple circuit that doesn’t even need a control chip.