Power analysis was completed on the circuit below in spice. This 2×2 MIMO converter was designed by a heuristic process: using KVL analysis and properties of DC-DC switched capacitor converters to meet the desired system specifications below. Theory of operation is covered in the equation and derivation section.
In the above experimented scenario the two inputs were modeled as ideal 1V sources. The capacitors are all 100 micro farad components and the transistors are power mosfet PSMN0R9-25YLC, data sheet provided in the extras section tab. In this circuit distribution of power is realized by switching between two phases in which capacitors are either in parallel or series with one another and the sources.
First the converter was tested by applying constant 10 ohm loads on the output capacitors C5 and C3. Under these conditions the converter maintained an efficiency of roughly 77%. A screenshot of the output power vs the combined input power is shown below.
MIMO POWER RESULTS
The second experiment was to examine how the converter performed when faced with dynamic loads, namely a pulsing 10 to 1k ohm load. The figure below shows the output voltage waveforms along with the associated instantaneous power output. The period of the pulse loads in this case is 10 micro seconds. Since both outputs are inherently dependent on one another there is a clear ripple in the voltage outputs no matter whether the just a single output faces a pulsing load. The converter still manages to reach steady stage outputs efficiency though.
The second part of this simulation process was to compare the MIMO converter against traditional converters. For comparison I designed the two buck converters below to meet the desired system specifications. Buck converters have traditionally been used for power regulation because they are highly efficient, provide accurate regulation and can handle high output power. In this scenario I sized the components used in the two converters similarly to the MIMO converter. Both units take 3V input sources but step down the voltage to 1.5V and 2V. This is simply accomplished by changing the duty cycle for each respective converter. The switching frequency of the MOSFET is operating at the same 1Mhz frequency as the MIMO. Results are shown below.
The first apparent difference between the two converters is the time required to meet the desired target output voltage. The MIMO converter managed to reach steady state in a few tens of micro seconds. Comparatively the bucks require roughly 1.4ms.
The input power for each converter is shown below. These converters manage to reach a combined efficiency of 82%. With further optimizations of the duty cycle and the components it’s not difficult for bucks to reach greater than 95% efficiency ratios. However, as no optimization has been completed on the MIMO converter, the optimization of the buck was left off as to create a level playing field. In this regard, the MIMO converter, without any control applied, can obtain comparable efficiency to traditional methods.