Transcript Overview of the Interdisciplinary Microsystems Group

OPS - Energy Harvesting
•
•
Name: Alex Phipps
Goals:
– Simulate the direct charge circuit to determine why the efficiency is so low.
– Assist with presentation preparation.
– Completely understand Shengwen’s circuits.
•
Key accomplishments.
– Simulated data for the direct charge circuit.
• Measured the input and output power with an optimal load.
• Measured input and output power with varying R1 values.
• Compared simulated efficiencies with measured ones.
– Nearly completed the setup of the charge pump circuit.
•
Key issues
– Simulations still do not match up with the measurements.
• The sims are very inefficient.
– Rusty with Cadence and buffer design.
Direct charger efficiency
Beam Efficiency vs. Acceleration
16
14
Efficiency (%)
12
Simulated
10
Ropt
8
R1 = 2.41M
R1 = 2.00M
6
R1 = 1.5M
4
2
0
0
2
4
6
8
Acceleration (m/s^2)
10
12
Model used for simulations
L1 
I   U
F
V 

Mm
2
Rm
R1  2

C1   2Cm
Ceb
Re
Explanation of direct charger inefficiency
• The cantilever beam is very non-stable in its parameters.
– There needs to be a calibration done to find the R1, C1, etc,
parameters every time that tests are performed with the beam.
Charge Pump
•
•
Name: Alex Phipps
Goals:
– Work on first draft of test plan.
– Conclude the individual simulations of the ring oscillator and charge pump
circuit.
– Combine the two individual components together (oscillator and pump) and
simulate the entire circuit.
– Begin to optimize/size the design so that everything will fit on the chip.
•
Key accomplishments.
– Individual simulations of the oscillator and charge pump completed.
• Waveforms behavior is as expected.
– The simulations of the entire circuit work properly.
•
Key issues
– New design goal.
• Small reservoir for quicker charge.
– Optimization of the circuit is slow due to complexity and inexperience.
Old Design
•
In the original circuit, the charge pump is used to bring the battery
voltage up to 1.2V. At this threshold, the controller is turned on, and
Shengwen’s power processor takes over.
– The charge pump allows the voltage driving the circuit to be less than
1.2V.
– The charge pump is very inefficient, and is turned off once the battery
voltage is high enough to power the controller.
Vbat>1.2V
Vbat<1.2V
Source
Charge
Pump
Source
Converter
Battery
Battery
Controller
New Design Goals
•
The battery acts as a giant capacitor, and the time required to charge it to
1.2V will reflect the initial charge on the battery.
– When the battery voltage is very low, the charge pump will have to be on for
longer periods of time, which lowers the efficiency of operation.
•
The solution is to use a secondary reservoir/capacitor.
– A smaller capacitor is can be charged up to 1.2V more quickly than the large
battery.
Vbat > 1.2V
Charge
pump
Seconary
Cap
(1.2V)
Controller
Battery
Source
Converter
New Design Issues
• There are several control issues that arise due to the second
reservoir.
– We want to use the charge pump as little as possible due to its
inefficiency.
• When the battery voltage is close to 1.2V, the capacitor will only
have to power the controller for a short time while the converter
charges the battery.
– Once this occurs the charge pump can be switched off.
• However, when the battery voltage is closer to 0V, the converter
will take longer to charge the battery, and the charge pump will
have to maintain the capacitor voltage for an extended period.
– These issues will have to be addressed with Shengwen, as I am not
yet an expert with his control scheme.
Test Plan
•
Buffer
(Inverter
Chain)
Ring
Oscillator
Vout pump
Charge
Pump
Vout pump


– Charge pump, ring oscillator
(clock), and reservoir.
– The first draft test plan is seen
below:
 • Explanations for these test
points follow.

Voutcap
Charge Reservoir
(Capacitor)
There are 3 main components of
this design:
Ring Oscillator / Buffer
• The ring oscillator is very sensitive to variations in the load
capacitances of each inverter. A buffer is used to minimally
load the oscillator and the two clock signals can be taken
from the output of the buffer.
– Measurement directly from the oscillator could cause excessive
loading from probe capacitance.
• A small resistance will be introduced between the buffer and
the charge pump because the signal is taken off of the chip.
– The resistance is from bond pads and from the wire needed to
connect the two pins.
– The buffer design will have to take this parasitic effect into account.
Voutpump
• Voutpump will allow for the output of the charge pump circuit
to be tested independently of the reservoir.
– The capacitance that is being charged can be varied to examine and
debug charge pump circuitry.
• The possibility exists that the charge reservoir will need to
be too large to put on chip. There is a finite capacitance
that can be achieved, and this may not be large enough to
hold the charge that Shengwen’s circuit needs.
Voutcap
• Voutcap will be used to measure the amount of energy that
can be stored on the reservoir. Connecting it to different
loads will allow us to see how the well the charge pump is
able to maintain the voltage when energy is being removed.
Simulations
•
Simulations of the entire circuit have been successful.
– Generic values transistor and capacitor values were used without concern for
practical sizing.
•
Design optimization will be needed to determine transistor and capacitor sizing.
DC
Buffer
Chain