Download Final Design Report - University of Idaho

2009-2010
K10 Inductive Charger
Dan Mathewson
Brett Russell
Kyle Ryan
Brady Coyle
3
Anna Camery
[email protected]
Executive Summary
Team Induct-Us has designed an inductive charging system for the NASA Ames K10
Lunar Rover.
The inductive charging system must be able to charge the K10’s batteries, Ocean Server
Intelligent Battery System (IBPS), fully. The system needs to be integrated in the current K10
rover at Ames, autonomously run, and have desired 60% power transfer efficiency.
This project is split into two separate sections; mechanical and electrical. The mechanical
consists of an arm design, charging station structure and docking aids. The electrical consists of
appropriate components for the charging circuit, proper feedback, and power circuit design. The
arm designs and station are tested for ease of use, complexity and price. The charging circuit has
been tested for power requirements and optimum power transfer efficiency.
The project began by researching known inductive charging examples. From there,
technical specifications and calculations can be made to determine circuit components. Using the
data and calculations gained and specifications given by the client, the testing can begin. After
iterative testing, a final product is then constructed. This process is recounted in detail within the
following report.
ii
Table of Contents
Executive Summary ……………………………………………………………….ii
Background…………………………………………………………………………1
Problem Definition………………………………………………………………....2
Conceptual Development…………………………………………………………..3
Product Description………………………………………………………………...8
Design Evaluation…………………………………………………………………13
Recommendations and
Future Work.............................................................................................................16
Appendix A: Mechanical Models
Appendix B: Electrical Schematics
Appendix C: Datasheets
iii
Background
The K10 lunar rover is a remote, extraterrestrial surveillance robot under development at
NASA Ames in Moffett Field, California. Tasked with increasing the efficiency of future lunar
missions, the K10 will deploy to the moon before manned missions. Once there, it will map the
terrain and look for resources. NASA uses this data to plan manned missions, astronaut
excursions, and choose sites for future infrastructure. This frees up the astronauts for tasks not
completed by the rovers and eliminates unproductive astronaut outings. The K10 packs several
instruments to carry out this task including, a 360°, panoramic, 3D scanning laser; a ground
penetrating radar; a color panoramic camera, and a high resolution terrain imager.
Once on the moon, all lunar rovers have power requirements that have to be dealt with.
Most extraterrestrial rovers rely on direct solar power to charge batteries. This method can
compromise the rover by taking up precious equipment space and making them far more fragile.
On the moon, large portions of the surface receive little light while blocked by the Earth making
solar power impractical. The K10 rover avoids these pitfalls by charging its batteries through 3
sets of manually operated plugs. This presents a problem since while the rover is remotely
exploring the lunar surface, it cannot plug itself in. Before deployment, a new method of
charging needs to be developed.
Charging stations are a proposed idea for charging extraterrestrial rovers. Even with a
charging station, a power transfer still must occur between the station and the rover. Conductive
charging plugs, like the current physical connections, are hard to connect remotely and are
susceptible to debris and other blockages. The University of Idaho RLEP Fellows proposed
making this power transfer with induction.
1
Problem Definition
The goal for this design project is to develop an inductive charging station that will be
able to fully charge the K10 lunar surveillance rover provided at NASA Ames. Currently during
testing, the K10 rover must be stopped to allow human intervention to replace the
batteries. There are three existing charge ports on the rover that must be connected to allow
recharging. When the K10 is on a mission, this task is time consuming for the operators. That is
why the ability to dock and charge itself is desirable.
Table 1: Specifications
General
Specific
Acceptable
Preference
Charge batteries
wireless charging
efficiency
power requirements
No interaction
60% efficiency
300 watts for batteries
Requirement
Desired
Requirement
Ocean Server boards
power requirements
max 16 A, 120V for ocean
server boards
Requirement
Docking station
power requirements
emergency stop
switch
120VAC from wall
shuts down within .05 sec
power on/off
Requirement
Requirement
Requirement
Ease of use
easily dock remotely
dock within 1 minute
Desired
Integration
movable primary core
plus or minus 6 inches (mag. of
x y plane)
Desired
Arm
fits in primary core
plus or minus .25 inch
Desired
By the end of this project, the team has a working prototype of a charging arm that could
be easily integrated into the current K10 system.
2
Concept Development
Mechanical Design
Arm
During the course of the semester we came up with three distinctive arm designs; a static
arm, a planer XY arm, and a multi-axis robotic arm. The multi-axis robotic arm would remind
someone of a fully functional human arm in which the core would be mounted in place of the
wrist. The arm would be mounted on top of the base station. This design was originally pursued
due its ability to adapt to almost any environment and position. However, do to the complexity
and need of additional controls this design was dismissed.
The planer XY arm is an arm that functions parallel to the rear planner face of the K10
rover. This arm would compensate for any misalignment in the horizontal or vertical directions.
However, this would require additional controls, thus this design was dismissed as well.
The static arm is the simplest design, requiring no additional controls. This design is
simply a rigid arm that would extend from the back of the K10. This arm mounts to the
secondary side circuit enclosure. The enclosure is mounted in such a way that it can shift in order
to compensate for different rover placement. Do to the simplicity of this arm it was the option
selected for the final design.
Base Design
In association with the three arm designs, two station designs where initially created; one
design each for the XYZ arm and one for the static and XY arms. Both designs fulfill the same
basic tasks: housing all electronics, the E-Stop, and the primary coil. Both designs sit on three
main points so the stations are not over-constrained and do not rock. The primary coil for the
3
static and XY arm is attached to a vertical screw drive that helps compensate for large
misalignments.
Based on suggestions received at a design review with NASA Ames, the base station
designs where changed to include two tire guide tracks. These ramps will insure that the K10
docks the same every time. These ramps will raise the K10 off the ground ensuring the rover is
always perpendicular to base.
Electrical Design
The components used in the power flow diagram include: AC/DC full bridge rectifier,
DC/AC inverter, step down transformer, optical feedback loop, battery server board and lithium
ion batteries. The following options were pursued throughout the project design.
AC/DC Full Bridge Rectifier
The full wave bridge rectifier used in the circuit is made by Center Semiconductor
Corporation. Alternating current from the local utility supplying 120 V and 15 A at 60 Hz is the
input for the rectifier. The voltage is then dropped across two diodes on the positive and negative
period of the sine wave. The negative period is inverted and made positive allowing for
maximum power transfer. A capacitor is placed in parallel to the rectifier output and controls DC
ripple.
DC/AC Inverter
The DC/AC inverter used in the circuit is made by Maxim Integrated Products
Incorporated. There are two proposed selections for the inverter. Firstly, a flyback PWM inverter
type was selected because it utilized a less complex parameter circuit and only requires that three
windings be used on the transformer. The second option is a forward inverter PWM type selected
4
because it able to handle the needed supplied power, however, the package does not have a built
in optical sensor and it requires four windings on the transformer.
Core
The decision of what magnetic core to use is an important part of the inductive charger.
Not only does it affect the size and shape of the mechanical arm and station designs, but also
provides the required power transfer efficiency that must be obtained from our given
specifications. Of course, as for any design, there are many aspects that must be considered. The
most important element to the core is how well the magnetic field lines are concentrated. The
goal of the transformer is to get the magnetic field lines to pass through as many turns as possible
to obtain maximum power transfer efficiency. The material of the core also is an important
characteristic to explore. Ferrite is the ideal material to use for this given application. This allows
for a higher frequency range (2 kHz – 3 MHz) and lowers core loss, as well as having a high
permeability.
The three specific core geometry’s that were explored for this design were, a cylindrical
plain core, an E-core, and a pot core.
Table 2: Ferrite core comparative geometry considerations
Core Bobbin Winding Winding Assembly Mounting Heat
Shielding
cost cost
cost
flexibility
flexibility dissipation
Cylindrical High Low
core
Low
Good
Simple
Good
Good
Good
E core
Pot core
Low
Low
Excellent
Good
Simple
Simple
Good
Good
Excellent
Poor
Poor
Excellent
Low Low
High Low
A cylindrical plain core is a piece of ferromagnetic material that resembles a small
circular disk in shape. A coil of wire could be wrapped around the core while having the
5
component attached to the end of the mechanical arm. This is a technique that is taken directly
from electric vehicle chargers. While the design is fairly simple and straight forward, the field
losses would be greater than desired and would not allow for maximum efficiency.
Figure 1: Cylindrical core insertion
The E-core was another core geometry that was investigated. This geometry design
contains the magnetic flux very well as opposed to the cylindrical plain core. It would also allow
room for wire leads. Wire can be wrapped in a plastic bobbin that would be placed over the
center leg of the E. Among many other advantages, the E-core is a good heat dissipater. A
disadvantage to using this design, however, would be the higher probability for misalignment
during the docking process. Two E-cores, one being a primary and the other the secondary,
would need to be brought together every time for a direct, repeatable connection.
The final core design that was explored was the pot core. Pot cores are self-shielding
cores that provide complete enclosure of the magnetic field lines. This allows for shielding from
6
outside electromagnetic interference (EMI) which could have a direct effect on the field lines.
They provide all the major advantages that E-cores offer, while allowing for an ease of docking.
Most of the requirements needed for this project seem to be met by the pot core; however there
still is one pressing need that must be explored. The pot core does a poor job at heat dissipation.
Ways to combat this would be to incorporate a heat sink.
Figure 2: Pot core insertion
The core recommendation that was decided on was the pot core. This design allows for
the best possible maximum power transfer efficiency. The major loss in the electrical power
system will come from the transformer, so having the best materials and highest efficient design
is essential.
Optical Feedback Loop
The optical feedback system is composed of an infrared emitter and phototransistor
detector couple with a voltage regulator. Both the emitter and detector are made by RadioShack
7
and a selection for the voltage regulator has been made. Voltage information is sent from the DC
side of the second rectifier back to the inverter to regulate the power going into the battery server
boards through the transformer.
Battery Server Board
The battery server boards are made by Ocean Server Technology Incorporated. The
output from the rectifier is the input to the server boards. The two main boards are battery
management modules, a low and high current version. The high current board feeds a DC-DC
converter to power the computer and the low current board charges the batteries.
Lithium Ion Batteries
14.4 V, 6.6 AHr Lithium ion battery packs that was also fabricated by Ocean Server
Technology Incorporated. The number of battery packs used in full scale design is 24 packs but 8
are intended to be used for testing purposes.
Product Description
Mechanical Design
Primary Side (Base Station)
The final design for the base station, pictured below in Fig. 3, has three major features.
The ramps, the spring loaded arm, and the basket.
8
Figure 3: Solid model of the final base station.
The purpose of the ramps is to allow the rover to dock in the same position every time,
enhancing alignment. They also prevent variations and changes in the terrain under the base
station from causing misalignment. Each ramp features extended outside ramp walls to prevent
K-10 from climbing off the station and damaging itself and the ramps. The entrances are also
flared to help guide the rover’s wheels onto the ramps. Each ramp was constructed out of 1/8th
inch aluminum sheet for ease of manufacturing and to save weight. The fabrication of the ramps
was completed by the University of Idaho’s facilities machine shop.
To lift the primary side core up to the rover’s level, a vertical arm is attached to rear of
the station. The arm is connected to the base station via a spring-loaded hinge. This allows for
flexibility in the horizontal direction in the event that the rover overruns the end of the ramps,
9
keeping the primary core safe from damage. The vertical arm is constructed out of slotted
aluminum stock allowing the height of the primary side core to be adjusted for different size
rovers.
The primary side core is housed in a tapered basket. The purpose of the basket is twofold. First it assures a proper connection by guiding the secondary arm in during docking,
secondly it acts as a heat sink to help cool the cores during charging. The basket was turned out
of round aluminum stock.
Secondary Side (Rover)
The rover side of the mechanical design consists of a housing to hold circuitry and a
secondary core, spaced from the housing. The secondary side features a standoff for the
secondary core to ensure proper spacing from the rover to the vertical arm. It also features an
externally mounted enclose, this is to protect the circuitry as well as to enable use on multiple
rovers. The enclosure was purchased from an electronics store and is made from injectionmolded plastic, while the standoff was made from steel pipe.
Electrical Design
Power Supply
It was decided that the simplest and most effective way to transfer energy was to use an
existing power supply instead of attempting to build a complete circuit from scratch. The idea
was that it already meets all of our needs with it being able to take the 120VAC power from the
wall and outputting 15VDC for the batteries. It became more feasible and less time consuming to
purchase a product that already was capable of providing the necessary functions that was
required for the project.
10
The power supply that was chosen was the Mean Well SP-320-15. This power supply
can take input voltage from the wall (120 VAC) and output the required voltage (15 VDC) to
power the Ocean Server boards. The Ocean Server boards will then provide the voltage to the
battery packs.
Figure 4: Mean Well power supply
The reasoning for choosing this supply was it provides the correct voltage values on the
input and output sides. Also, the 320 W is enough to power the battery packs on-board the K10
rover. Since NASA Ames is currently using a similar product, it can be easily integrated into the
current setup they use for testing. However, the only problem with this decision is it does not
provide for the isolation necessary to allow for the rover to charge its batteries through the means
of induction. So the next step was to alter the board of the power supply to allow for separation
of the primary and secondary sides.
PCB Separation
Investigation of the power supply board showed the primary and secondary circuitries
were conveniently separated. The only components that bridged the two sides were the
11
transformer and the optocouplers. Since we had planned to create our own transformer and
feedback system anyways, it was decided to pull off these components and physically cut the
PCB along the primary and secondary sides to allow for two divided parts.
Figure 5: Milling of PCB Power Supply Board
The primary side of the circuit board is located in the housing on the base station. The
secondary is inside the exterior box situated on K10.
Transformer Redesign
In order to duplicate the existing transformers characteristics, we measured the
inductance values of each winding using an RLC meter. Since our system has a fixed 100 kHz
switching frequency, we needed to measure the inductance for this same frequency.
12
Using printed ABS plastic bobbins, we wrapped our own windings to match the
inductance values from the existing transformer. Following each winding, we used a layer of
specialized transformer tape to enclose each winding. If two windings come into contact with
each other, voltage arching can occur and cause the transformer to fail. The final design took
120VAC at its input on the primary station side, transferred power across the transformer to the
secondary side rover and stepped the voltage down to 15VDC.
Design Evaluation
Mechanical
Looking at the specifications, Table 1, the mechanical evaluation can be broken down
into four parts: the on/off switch, ease of docking, the movable basket, and docking tolerance.
Table 3: Mechanical design evaluation
Part
Specification
On/off switch
Power on and off
Ease of docking
Docks in 1 min.
Movable basket 12 in total vert. displacement
Docking tolerance
within .25 in.
Ramp strength
500 lbs total
Actual
Spec Met?
Power on/off
Yes
<1 min
Yes
17 in.
Yes
.2
Yes
supports rover
Yes (?)
max unknown
When looking at the actual results of the mechanical side of the project it is a little hard to
predict the actual results with the real K-10. For example docking time; we don’t have a working
rover so we don’t know exactly how fast it can climb the ramps, just that it can and it does not
take overly long for us to dock our rover by hand.
13
Electrical
A summarized list of the electrical results from this project from the Design Failure Mode
and Effects Analysis can be found in Table 4.
Table 4: DFMEA summary
Item
Potential Failure
Recommended Action
Power transfer across
transformer
Isolated regulation of
output voltage
Emergency switch
integration
No charge, dangerous
environment
Circuit component
overload
Will not kill power to
batteries
Repeated design calculations
and testing
Have feedback, regulation
circuit
Insert in charging circuit and
test results
Action
Results
Success
pending
Pending
Testing Procedure
Our design had three main specifications that could possibly fail. We have conducted a
series of tests to evaluate our design and they are listed below.
Test 1. The first test that we did was to test the original circuit operations. For this test, it
was important to measure different voltage ratings from different parts of the circuit. We left the
circuit completely as it would come out of the box and opened up the lid to have access to
reference points. Once the charger was powered on, we took voltage readings at the input, across
the windings of the transformer, over the optocouplers, and at the output terminals. A full list of
the test results can be found in the appendix, but the main results showed 122.6VAC at the input
and a regulated 15.15VDC output.
Test 2. The next step for our testing was to take out the filter capacitor that bridged the
primary and secondary side. The reason for removing this was that, in our circuit, there would be
no way to have that filter capacitor bridge our station to rover side. It would make our circuitry
14
no longer isolated. The results from this test showed that both the input and output voltage stayed
consistent with the first test.
Test 3. The next few tests deal with the circuit being altered. This test was to test out our
new, redesigned transformer. We already pre wound both the primary and secondary cores. It
was important to be able to match the inductances of the original transformer. The redesigned
transformer had similar inductance values and acceptable leakage inductance values. A full list
of inductances can be found in the Appendix. We then soldered on wires to both the primary and
secondary circuit that would be the approximate real length that the windings and core would
have to be placed on. The only change from the last test was the switched out transformer. The
input value stayed the same, but the output value was approximately 10 volts too high. It was
also discovered that the only way to make this circuit work, was to remove one of the
optocouplers that controlled the voltage regulation. We believed that this was due to the increase
of output voltage that would cause the circuit to fail. This meant that the output was unregulated
and would vary from 24VDC-35VDC. This test did, however, prove effective power transfer
across two completely isolated sides. See Figure 6 for the set-up of this final test
Figure 6. Test 3 Circuit Set-Up
15
This is currently where we are. We have been able to get effective induced power through
our circuit. The Recommendations and Future Work section explains the remaining tests that still
need to be completed to finalize our project.
An estimated cost of making a replicated prototype is outlined in Table 5.
Table 5: Breakdown of fabrication costs
Raw
Materials
$104.00
Labor
$160.00
Electrical
Components
$180.00
Total
$444.00
Sheet
Stock
Hardware
Machining
Fabrication
Power Supply
Cores
Wire
Recommendations and Future Work
As the product must be finalized and delivered to our client, there is still some work to be
completed. First, we need to continue our testing of the circuit with our new optocouplers and the
circuit board separated. This would allow our circuit to be completely isolated while regulating
the voltage. Another means for regulation would be to apply a regulation circuit to the output of
our circuit. Further testing would have to prove which method would be most advantageous.
From there, all that would be left to do would be to test the entire circuit with the station and arm
design. A full systems integration test would allow the product to be finished and ready to be
shipped to Ames.
The purpose of our project was to design and build an inductive charging circuit. This
idea is something that is completely new to the K10 project and is something that is required for
actual use of the rover on the moon. Since Ames currently has to manually plug in the batteries,
an isolated system is important toward having an autonomous rover. There also, currently is no
16
other system that is designed to do what ours does, so having this new system will be used until a
new version is created.
Since this system is in its first version, there is still future work that can be done to
improve the design. We are currently outputting an unregulated voltage to the battery
management system. This variation is not one that is ideal, but one that will be able to be handled
by the management system.
17
Appendix A: Mechanical Models
A1
A2
A3
A4
A5
Appendix B: Electrical Schematics
B1
Power Loss
Note: Optimal design losses, taking into account minimal losses for each component
Appendix C: Datasheets
C1
C2
Matched Infrared Emitter and Phototransistor Detector
Model: 276-142 | Catalog #: 276-142
What's in the box


1 x infrared emitter - 2V 40mA
1 x phototransistor detector - 20V 25mA
Tech Specs
Dimensions:
Product Height 0.34 INCHES
Product Width 0.2 INCHES
General Features:
Model 276-142
Product Type Phototransistor
Body Material Multi
Fits What:
Model 276-142
Miscellaneous Features:
Min Operating Temperature
-40 Fahrenheit
Max Operating Temperature
176 Fahrenheit
Supported Languages
English
C3
C4
C5
DC-DC Up Converter with efficiencies of up to 96% depending on input voltage. Converts
from battery power or charge power to a higher output voltage for operating user devices.
The device can operate at about 100 Watts with no ariflow, and 240 Watts with a small
amount of airflow.
The DC2U-1V provides four input power connectors to attach any of the IBPS battery
management modules (BB-xx,MP-xx and XP-xx). The input connection is by the standard
8 pin Molex Mini-Fit Jr connector. Cable 19-00026-20 is an example cable for connecting
the input voltage. The output connector is a 10 pin Molex Mini-Fit Jr connector.
The input Voltage range is 9V - 24V (Input Voltage must be lower than the output
voltage)
If you need additional output power these converters can be combined in parallel to scale
the output power. The modules have dynamic current sharing to distribute the load.
The standard modules support 19V, 24V, 28V and 48V outputs, other voltages may be
special ordered (19-48V range).
RoHS Compliant
C6
C7
Rechargeable "smart" lithium ion battery pack with integrated
SMB fuel gauge. Integrated safety circuit monitors and
reports pack, Temp, Voltage, Current, Time to Empty, Time
to Full and other key parameters via the IBPS BB-xx, MP-xx
and XP-xx controllers. The battery controller module
continously monitors the status of the battery pack and can
report this information to the user via the serial port on the
battery controller.
The BA95HC-FL battery pack has a cable and latching
connector for attaching the battery pack to the battery
controller. This is different from a standard Li-ion battery pack
as users are not required to implement a backplane with a
connector as used on Notebook computers. This battery
pack connects directly to OceanServer's IBPS Battery
Management Modules. In applications where batteries are
located more than 9" from the IBPS Modules, a Battery
Extension Cable, 19-00035-06, can be used.