One Cycle Control vs current mode control

[Circlotron]

Keyue Ma Smedley is well known in her circles for inventing the concept of one cycle control. It basically works like this: If, for example, you have a buck regulator, put an integrator that is driven from the switching stage raw output. Begin the integration at the start of the switching cycle and terminate the switching cycle when the integrator output reaches a reference level corresponding to the desired output voltage. Then reset the integrator ready for the next cycle. It is said to have very fast response time when maintaining a desired DC output.

So far so good. But then it struck me - this doesn't seem all that different to current mode control. Instead of having an integrator you either directly or indirectly sense the inductor current and terminate the switching cycle when the inductor current reaches a reference value. So then, what's the big deal about OCC by comparison?

https://ieeexplore.ieee.org/abstract/document/471281

https://www.researchgate.net/figure/One-cycle-control-schematic_fig2_277333522

[xavier60]

I first saw an idea like that with the TDA4601 and TDA4605 controllers. Switching transistor current is simulated by a capacitor being charged via a resistor from the HVDC while the transistor is on and fully discharged when the transistor is off. The controllers operated in transition mode, never CCM.
I guess the idea is to avoid the need for a CS resistor which would be great advantage in high current designs.

[David Hess]

It reminds me of slope compensation when used with constant off-time and current mode control.