RU2/MK3 ELECTRONICS

HARDWARE ARCHITECTURE

GKX-14-0042-SP_B2
2 July 2014

CONFIDENTIAL – between PA and BDR Thermea

 PA Regional Office:
PA Consulting Group
Cambridge Technology Centre
Melbourn
Hertfordshire, SG8 6DP
UK
Tel: +44 1763 261 222
Fax: +44 1763 260 023
www.paconsulting.com Version no: 1.0

GKX-14-
0042-SP_B1
Prepared by: Alastair Smith Document reference:

CONFIDENTIAL – between PA and BDR Thermea

CONFIDENTIAL – between PA and BDR Thermea
1 DOCUMENT CONTROL

1.1 Document Revision

Document Number Reason for Author Issue Date

Revision
A1 Ready for review Alastair Smith 16 Apr 2014

A2 Reviewed by Duncan Nick Warrington 24 Apr 2014

Evans
A Released for review Alastair Smith 25 Apr 2014

B1 Updated with proximity Alastair Smith 2 May 2014

sensor experiment
B2 Updated following Alastair Smith 2 July 2014
schematic design

1.2 Document Authorisation

Role Name Signature Date

Author Alastair Smith

PA Assignment Steve Carden

Director
PA Assignment Nick Warrington
Manager
BDR Project Manager Fedde De Waal

BDR Technical Lead Jasper Rijnsburger

CONFIDENTIAL – between PA and BDR Thermea

CONTENTS

1 DOCUMENT CONTROL 4
1.1 Document Revision 4
1.2 Document Authorisation 4

2 INTRODUCTION 3

3 HARDWARE DESIGN OVERVIEW 4

3.1 The two units 4
3.2 Architecture overview 4
3.3 Design rationale 5

4 KEY ELECTRONIC COMPONENTS 2

4.1 Main Processor 2
4.2 Display 2
4.3 Random Access Memory (RAM) 3
4.4 Nonvolatile storage (FLASH) 3
4.5 Wi-Fi Module 3
4.6 Optical encoder 3
4.7 PIC Boost Controller 3
4.8 Power supply architecture 4

5 MECHANICAL DESIGN CONSIDERATIONS 6

5.1 Introduction 6
5.2 Temperature sensing 6
5.3 Optical encoder 6
5.4 EMC Precautions 7

6 PROGRAMMING AND DEBUG ARCHITECTURE 8

6.1 Introduction 8
6.2 Programming and debug 8

7 TEST STRATEGY 9
7.1 Introduction 9
7.2 Manual test 9
7.3 Manufacturing tests 9
7.4 Equipment requirements 10

APPENDICES 11

WI-FI MODULE DECISION JUSTIFICATION 13

1
CONFIDENTIAL – between PA and BDR Thermea
A.1 Introduction 13
A.2 Options 13
A.3 Assessment of USB module approach 13
A.4 Low power module on main PCB 14
A.5 Low power module on mezzanine PCB 14
A.6 Custom low power module 14

ARM A5 BOOT AND DEBUG STRATEGIES 16

B.1 Introduction 16
B.2 ARM boot sequence 16
B.3 Debug peripherals 18

USB WI-FI MODULE POWER CHARACTERISTICS 19

C.1 Introduction 19
C.2 Experimentation 19

OPTICAL ENCODER EXPERIMENTATION 22

D.1 Introduction 22
D.2 Experimentation 22

PROXIMITY AND AMBIENT LIGHT SENSOR EXPERIMENTATION 26

E.1 Introduction 26
E.2 Experimentation 26

POWER CONSUMPTION CALCULATION 30

F.1 Atmel processor system 30
F.2 Power regulator efficiencies 30
F.3 RU2 device power budget 31
F.4 MK3 Device power budget 32
F.5 Supercapacitor calculation (RU-2 only) 32
F.6 Real-time clock considerations (MK-3 only) 33

MECHANICAL DESIGN CONSIDERATIONS 34

G.1 Introduction 34
G.2 Temperature sensing 34

2
CONFIDENTIAL – between PA and BDR Thermea
2 INTRODUCTION

This electronics hardware architecture document sets out the architectural design for
the RU2 and MK3 units. This describes
 The key components of the system and rationale for choosing them;
 How the electronic components are interfaced with one another;
 How the electronic components will integrate into the mechanical hardware;
 How the processor subsystem will be programmed and debugged;
 Detailed appendices of calculations and experimentation that has gone in to reaching the
conclusions presented.
The design assumes the ‘Best Choice design’ under consideration in Stage Gate 2 of the project. This
architecture should supply the perceived best user experience within hard constraints and decisions
made during the design process between the design authority (PA Consulting) and the client (BDR
Thermea).

3
CONFIDENTIAL – between PA and BDR Thermea
3 HARDWARE DESIGN OVERVIEW

This section provides an overview of the electronics design. This includes an

overview of the commonality and differences between platforms, selected key
components, component features, how they will connect and rationale for choosing
the component.

3.1 The two units

The RU-2 unit is to be placed within a room, remote from a boiler control unit, and is intended as a
replacement to the thermostat control. Furthermore, it is intended to act as the hub for sending control
and diagnostic information back to BDR servers and receiving control information from the servers.
The MK3 unit, in contrast to the RU-2, is directly connected to the boiler and used by the installation
engineer to set up various aspects of the heating control system within a building, including control of
heating zones and connected devices.
Both RU-2 and MK3 units have very similar electronic designs in order for them to allow user
interaction and communicate with other devices. The key features are:
 Colour screen visual interface
 Rotating dial and buttons for user input
 Bus communication with controller device
Much of the functionality is common, and therefore development effort can be save by utilising much
of the same hardware in both designs.
The particular components that will be common are centred on the embedded system, which includes
the processor and memory subsystem. Economies of scale in pricing will be achieved, and particularly
software development effort will be shared across the two platforms, which also enables a consistent
user experience.
The key differences between the two devices are as follows:
 RU-2 device requires Wi-Fi IEEE 802.11n connectivity with a home router;
 RU-2 device requires connection to the controller using the BDRRoomBus;
 RU-2 device requires a temperature sensor device;
 RU-2 power budget is 1.5 Watts;
 MK-3 device requires connection to the controller using a CAN physical layer interface;
 MK-3 device requires a real-time clock; and
 Screen sizes are different (4.3 inch and 5.0 inch).

3.2 Architecture overview

3.2.1 Description
The RU-2 device provides the greater engineering challenge compared to the RU-2, due to the power
consumption constraints, inclusion a Wi-Fi module and thermal considerations in order to ensure the
temperature sensor measures the correct value.

4
CONFIDENTIAL – between PA and BDR Thermea
3.2.2 Diagrams
The overall architecture picture is shown below. Individual components are listed later in this
document.

3.3 Design rationale

The design has been made with the following guiding principles:
 Cost: The cost of the unit is a key requirement and the design has been developed to ensure low
unit cost.
 User Experience: The units are required to deliver a premium experience. The RU-2 and MK3
should be units you are proud to have in your home, components have therefore been chosen
which are able to deliver this experience.
 Time to development: The target product launch is Q1 in 2015 for which manufacturable
prototypes must be ready in October 2014.It is important therefore to ensure that the solution is
one which can be developed in this compressed time frame.
 Power consumption: For both units there is a maximum power budget due to power which can
be drawn from the BDR buses. The architecture must therefore remain within these budgets. In
addition there is a requirement to minimize power as much as possible to ensure a good energy
rating.

5
CONFIDENTIAL – between PA and BDR Thermea
 Thermally Isolated RU-2 Main PCB
region of PCB

Temperature
sensor
ADC

Proximity
CS_BOOT NAND Flash OSRAM SFH
push button MT29F1G08ABADAH4 4650-Z irLED

MMU
Room Bus Circuit SiLabs Si1141
DDR2 RAM Proximity sensor
MT47H32M16NF-25E I2C
500kHz

MCU UART Wi-Fi Module

BDR Roombus
modulator GPIO GPIO GPIO

Atmel ARM A5 CPU LCD panel assembly PCB

RGB DSN43600
display Ribbon cable
enable
interface LCD panel

1 x PWM LED driver

Power supply LED backlight
1.2V
circuit – see
supply
separate * 3.3V rail from voltage reg
1.8V
architecture * 12V rail from boost reg
24V supply supply
3.3V
Power supply
LED
Control Wheel functions

Quadrature Optical encoder

encoder AEDR-8300-1Wx
input
2X push buttons
Clock
DBGU UART GPIO
generator
COTS Component

Clocks: Fitted for development only

1 x push buttons
32.768 kHz 3 x indicator LEDs Separate PCB assembly
Programming UART 12 MHz
connector connector PCB region

Figure 1: The RU-2 electronics architecture

1
CONFIDENTIAL – between PA and BDR Thermea
 MK-3 Main PCB
Coin cell battery holder
and replaceable coin cell
battery

CAN PHY VDD

CAN
TJA1051T/3 BU

CS_BOOT NAND Flash

push button MT29F1G08ABADAH4
MMU
DDR2 RAM
MT47H32M16NF-25E

LCD panel assembly PCB

P.LPFO-TP2004DA050S-02
RGB display Ribbon
interface cable
Atmel ARM A5 CPU LCD panel
LED
1 x PWM
driver
LED backlight
Power supply
1.2V * 3.3V rail from voltage reg
circuit – see
supply * 9V rail from boost reg
separate
1.8V
architecture
supply
3.3V
Power Control Wheel PCB
supply
on LED
Quadrature Optical encoder
encoder AEDR-8300-1Wx
input
2X push buttons
Clock
DBGU UART GPIO
generator
COTS Component

Clocks: Fitted for development only

1 x push buttons
Micro USB 32.768 kHz Separate PCB assembly
UART 3 x indicator LEDs
debug 12 MHz
connector PCB region
connector

Figure 2: The MK3 electronics architecture

1
CONFIDENTIAL – between PA and BDR Thermea
 1
CONFIDENTIAL – between PA and BDR Thermea
4 KEY ELECTRONIC COMPONENTS

This section outlines the key functional features of each electronic component or function. Power
consumption issues are discussed later in this document.

4.1 Main Processor

The main processor selected for this device is the ATMEL ARM Cortex A5 SAMA5D34 with the
following features that cover all required functionality for bot RU-2 and MK3 units.
 DDR2 controller
 NAND flash interface
 MMU
 Quadrature decoder logic
 CAN bus (x2)
 LCD controller
 GPIO for push buttons
 USB host
 USB device
 SPI
 USART x4 UART x2
 ADC
 RTC
 Standby mode for low power

4.2 Display
4.2.1 RU2: Densitron DSN43600 LCD TFT 4.3in display
The key features of this display are:
 4.3 inch display
 480x272 pixels
 16.7M pixels
 White LED backlight

4.2.2 MK3: New vision custom LCD TFT 5.0in display

The key features of this display are:
 5.0 inch display
 480x240 pixels
 16.7M pixels
 White LED backlight

2
CONFIDENTIAL – between PA and BDR Thermea
4.3 Random Access Memory (RAM)
Recommended part is manufactured by Micron Technology with part number MT47H32M16NF-25E.
Key features are:
 512 Mbit
 DDR2

4.4 Nonvolatile storage (FLASH)

NAND flash provides the lowest cost non-volatile storage mechanism that will hold the program and
any long term storage requirements such as screen brightness settings.
The part intended for use in this design is manufactured by Micron Technology with part number
MT29F2G08ABAEAH4-IT. Features are as follows:
 2Gbit
 8-bit width
 3.3V power supply

4.5 Wi-Fi Module

The selected module is the Gainspan GS2100, which offers the require power profile at a competitive
cost. This will be placed on the main PCB for the required medium and high-cost variants only. The
module allows the offload of the TCP/IP stack from the main processor, allowing the Wi-Fi connectivity
to be controlled entirely separately to the main processor. This allows the minimisation of power
consumption. Key features are therefore:
 Low power consumption in receive and transmit
 Can be woken up from ultra-low power consumption mode

4.6 Optical encoder

The selected part is the Avago AEDR-8300-1W0. This has the following features:
 3.3V-5V supply;
 Reflective encoder; and
 Quadrature outputs will interface with chosen processor.

4.7 Proximity sensor

The selected part is the Silicon Labs Si1141, in conjunction with the OSRAM SFH4650-Z infrared
LED. This has the following features:
 Reflective sensor with good user experience (see Appendix);
 Range of over 50cm;
 Sense rates suitable for human interaction;
 Will work with plastics and materials being used in this design; and
 Low average power consumption.

4.8 PIC Boost Controller

The selected microcontroller is the Atmel ATTiny9. This is one of the most simple, low cost micro
controllers on the market, with the following features:
 Low cost
 Low power

3
CONFIDENTIAL – between PA and BDR Thermea
 1KB ISP flash memory
 32B SRAM
 4 general purpose I/O lines

4.9 Power supply architecture

4.9.1 Description
The power supply in embedded systems has a number of considerations. The predominant factors is
that there tend to be many voltages required for the various connected components to operate. The
common components and voltage rails are:
 24V input
 1.2V processor core voltage
 1.8V memory components
 3.3V processor peripherals
 3.3V-5V optical encoder
 3.3V display driver
 LED display driver (9V and 12V), with pulse width modulation for low power dimming modes.
The RU-2 has the following additional requirements:
 1.8V-5.5V microcontroller
 3.3V Wi-Fi
 5.5V Capacitor charge voltage
The MK3 has the following additional power rails:
 5V CAN chip

4.9.2 RU-2 Device

Region of PCB
Super cap
24V to 5V 5V to 3.3V
3.3V Wi-Fi
regulator LDO
regulator

enable
Super capacitor

24V to 3.3V
uC LDO 3.3V supply for microcontroller
regulator

enable

24V - 3.6V
Switching
24V in
regulator 12V LCD
LM25010 LED Driver 12V LED
RT9293B

1.8V memory
ACT8865
1.2V core
Regulator
3.3V processor peripheral
and encoder

4
CONFIDENTIAL – between PA and BDR Thermea
Figure 3: Power supply architecture for the RU-2 device

4.9.3 MK-3 Device

The MK3 device has a simplified power architecture

Region of PCB
24V - 3.6V
Switching
24V in
regulator 12V LCD
LM25010 LED Driver 12V LED
RT9293B

1.8V memory
ACT8865
1.2V core
Regulator
3.3V processor peripheral
and encoder

24V - 5V
Switching
5V CAN controller chip
regulator
LM25010

Figure 4: Power supply for the MK3 device

5
CONFIDENTIAL – between PA and BDR Thermea
5 MECHANICAL DESIGN
CONSIDERATIONS

5.1 Introduction
This section describes the key considerations of the electronics design that are impacted by the
mechanical design. Specifically:
 Temperature sensing
 The optical encoder

5.2 Temperature sensing

This subsection identifies the approach to sensing temperature. Existing solutions to thermal
management are reviewed in the Appendices.

5.2.1 Approach to thermal management

The electronics design will employ the following features:
 Thermal isolation of the thermistor from the rest of the PCB through the use of PCB through-holes;
 Mechanical design that isolates the thermistor from the internal environment by encasing the
thermistor as in the nest device;
 PCB plated holes to allow a flying lead to be attached to the PCB, in case the arrangement
requires a thermistor arrangement with flying leads for a more remote sensor measurement.
 Laying out the thermistor away from the heat producing components (Wi-Fi, supercapacitor/battery,
LCD and processor).
In addition software and mechanical design will take thermal issues into consideration:
 Software compensation through an algorithm;
 Close work of the PCB and mechanical design team to ensure heat transfer is minimised;

5.3 Optical encoder

The optical encoder has significant mechanical design constraints that must be met in order for the
processor system to correctly detect magnitude and direction of the control wheel rotation.
Experimentation with materials and distances is provided in the appendices.
The most significant of these mechanical considerations that feeds into the electronics design is that
the distance from the optical encoder to the wheel on which the markings are must be less than
2.5mm and greater than 1mm. Coupled with the mechanical constraints that require the wheel to be
placed above the LCD screen, it is evident that it will not be possible to locate the optical encoder on
the main PCB.
The solution to this is as follows:
 Separate PCB for the optical encoder;
 Mechanical attachment on the main pcb for minimising manual labour in manufacturing;
 Mechanical fixing is required to ensure the optical encoder is placed at distances within the
tolerances of the optical encoder.

6
CONFIDENTIAL – between PA and BDR Thermea
5.4 EMC Precautions
Processor subsystems and wireless telecommunications devices typically run at high operating
frequencies with many clock speeds. A consequence of this is that electromagnetic interference
issues may prevent the product from being certified.
There are two main methods that will be employed to minimise EMC issues:
 A ‘Screening can’ will be placed above the processor and memory subsystem to act as a Faraday
cage and prevent transmission of EM waves. A screening tin has already been used for the
Gainspan module, which will be soldered directly onto the main PCB.
 Placement of high frequency signals between power and ground planes, which will prevent
resonant effects across of through the PCB.

5.5 Proximity and ambient light sensing

The proximity sensor has a number of mechanical design considerations. These are due to the active
nature of the sensor, and due to the requirement for it to sense outside the optical environment of the
RU-2 device. As a consequence the following are required:
 Internal optical isolation of the sensor from the infrared LED source with a black rubber seal;
 Separation of the sensor from the source by a distance of greater than 10mm;
 Placing the infrared source and sensor as close as possible to the covering panel; and
 Use of a transparent plastic above both sensor and emitter.

An arrangement recommended by Silicon Labs is shown below.

5.6 Mating with the electrical contacts

TBC

7
CONFIDENTIAL – between PA and BDR Thermea
6 PROGRAMMING AND DEBUG
ARCHITECTURE

6.1 Introduction
Appendix B provides an overview of the ARM A5 boot sequence. The boot and debug architecture
takes advantage of ARM A5 features, which include an advanced boot ROM that enables debug to be
performed using a USB interface.

6.2 Programming and debug

6.2.1 Non-volatile code storage
The design includes NAND flash as the primary non-volatile program code storage. In production this
device needs to be programmed prior to their assembly onto the PCB. Standard equipment shall be
used to do this, as recommended by the volume manufacturer.

6.2.2 Debugging environment

During debug the flash will be programmed by forcing the processor into debug mode. This will be
done by using a pushbutton on the PCB, which temporarily disables the Flash, allowing the processor
to boot into debug mode. The processor and Flash can then be programmed through this interface,
using a standard SAM-BA environment.

8
CONFIDENTIAL – between PA and BDR Thermea
7 TEST STRATEGY

7.1 Introduction
The test strategy considers the various stages of the project and of the manufacturing process. This
process will be developed into a separate manufacturing test document in later stages of this project.

7.2 Manual test

The first PCB will be manufactured and populated during stage gate 3. In order to validate the
electronics design and develop the detailed manufacturing test plan, a manual process will be
developed comprising of:
 Test power rails under minimum and maximum loads;
 Test voltage drops across diodes;
 Test oscillator frequencies;
 Power consumption tests (including a tailored Wi-Fi test for the for RU-2 unit);
 Test manual flash programming;
 Full software functionality.
The process of conducting the manual test will enable and better inform the manufacturing test.

7.3 Manufacturing tests

The manufacturing tests are summarised in the following figure.

PCB mounted PCB ICT Testing Operational test

components •Bare PCB assumed tested for •The device should be booted to
•NAND flash programmed short and open circuits ensure that all subsystems are
•In circuit test of components working correctly
•Wi-Fi module assumed tested
and tolerances
•Processor chip and other
discretes assumed tested •Visual inspection

Before the DUT is powered up, FULL in-circuit component testing must take place. Where possible
the following should be tested:
 All passive components where possible must be tested for tolerances;

9
CONFIDENTIAL – between PA and BDR Thermea
 All tantalum capacitors must be tested for correct insertion;
 All active components must be tested for correct insertion using visual or diode testing techniques;
 All diodes and transistors must be tested for correct forward and reverse voltage drops.

7.4 Equipment requirements

It has been assumed that a bed of nails method of connection between the device and ICT equipment.
Test points will be located on the back (assumed to be layer 6) of the PCB.
The ICT equipment used at this stage must be capable of measuring component tolerances from 1%
for resistive and 5% for capacitive and inductive components. It is acknowledged that circuit topology
may reduce the effectiveness of these tests, and so this will need to be taken into account.

10
CONFIDENTIAL – between PA and BDR Thermea
APPENDICES

WI-FI MODULE DECISION JUSTIFICATION 13

A.1 Introduction 13
A.2 Options 13
A.3 Assessment of USB module approach 13
A.4 Low power module on main PCB 14
A.5 Low power module on mezzanine PCB 14
A.6 Custom low power module 14

ARM A5 BOOT AND DEBUG STRATEGIES 16

B.1 Introduction 16
B.2 ARM boot sequence 16
B.3 Debug peripherals 18

USB WI-FI MODULE POWER CHARACTERISTICS 19

C.1 Introduction 19
C.2 Experimentation 19

OPTICAL ENCODER EXPERIMENTATION 22

D.1 Introduction 22
D.2 Experimentation 22

POWER CONSUMPTION CALCULATION 26

E.1 Atmel processor system 30
E.2 Power regulator efficiencies 30
E.1 RU2 device 31
E.2 MK3 Device 32
E.3 Supercapacitor calculation (RU-2 only) 32
E.4 Real-time clock considerations (MK-3 only) 33

MECHANICAL DESIGN CONSIDERATIONS 34

F.1 Introduction 34
F.2 Temperature sensing 34

RISK LOG ERROR! BOOKMARK NOT DEFINED.

G.1 Uknowns Error! Bookmark not defined.
G.2 Decisions pending Error! Bookmark not defined.

11
CONFIDENTIAL – between PA and BDR Thermea
 12
CONFIDENTIAL – between PA and BDR Thermea
 WI-FI MODULE DECISION JUSTIFICATION

A.1 Introduction
This section describes the options and trade-offs involved in the options for 802.11n connectivity.

A.2 Options
 USB Module
– This option involved using a standard USB connector to attach a low-cost COTS USB Wi-Fi
dongle as an expansion option.
 Low power module on main PCB
– This option involves employing a COTS Wi-Fi module and soldering it to the main board of the
RU2 unit.
 Low power module on mezzanine PCB
– This option involves employing a COTS Wi-Fi module and soldering it to the main board of the
RU2 unit.
 Custom low-power module
– This option involves designing a custom low power Wi-Fi module, purchasing a Wi-Fi chipset
from an OEM company, developing the analogue antenna circuit, and ensuring Wi-Fi
certification of the product.

A.3 Assessment of USB module approach

A.3.1 Summary of approach
Advantages
Lowest cost option: Wi-Fi modules can be sourced for around the €3 mark.
Low risk: hardware exists.
Disadvantages
High power.
Software development effort would be required.
Show stopper
Power consumption is ~420mW idle, which is significantly more than the power budget.
Steps to make option more viable
A custom TCP/IP stack and USB stack could be considered and evaluated. Estimated time would be a
one week study. If viable following further analysis, the custom stack would involve a 50% software
effort to the custom module approach.

A.3.2 Recommendation
This approach should not be used without significant investment into assessing a low power dongle
mode. This would carry risk of it not working.

13
CONFIDENTIAL – between PA and BDR Thermea
A.4 Low power module on main PCB
A.4.1 Summary of approach
Advantages
Low power: A super capacitor arrangement will provide for peak current scenarios.
Low risk: hardware and software exists, minimal integration required.
Disadvantages
High cost: assuming there is a module on every unit, this approach has the highest BOM cost of over
€7.
Show stopper
Cost is prohibitive.
Steps to make option more viable
See mezzanine PCB option.

A.4.2 Recommendation
This approach should not be used without significant investment into assessing a low power dongle
mode. This would carry risk of it not working.

A.5 Low power module on mezzanine PCB

A.5.1 Summary of approach
Advantages
Low power: A super capacitor arrangement will provide for peak current scenarios.
Low risk: hardware and software exists, minimal integration required.
Lower cost than main PCB scenario.
Disadvantages
High cost: assuming there is a module on every unit, this approach has the highest BOM cost of over
€8. However, this reduces to €1 on the low cost units without Wi-Fi.
Show stopper
Cost is prohibitive.
Steps to make option more viable
See custom Wi-Fi module.

A.5.2 Recommendation
This approach should not be used without significant investment into assessing a low power dongle
mode. This would carry risk of it not working.

A.6 Custom low power module

A.6.1 Summary of approach
Advantages
Low power: A super capacitor arrangement will provide for peak current scenarios.
Lower cost than other modular approaches.
Disadvantages

14
CONFIDENTIAL – between PA and BDR Thermea
Long development time.
Risk is high due to:
 potential infringement and licensing. For example one module supplier has 9 patent relating to low
power Wi-Fi.
 Cost of existing Wi-Fi solutions may reduce by time of high volumes, eating away at original
business case
 Must sell over 400k units to see ROI, based on estimates, this will be in Y3/Y4

High development cost: of €1M, based on:

 IP search: €50,000
 Systems Eng: €50,000
 HW D&T: €350,000
 SW D&T: €300,000
 Manufacture setup: €50,000
 Equipment and hire: €50,000
 Wi-Fi certification: €100,000
 Other approvals: €50,000.
Show stopper
Cost outlay is prohibitive.
Steps to make option more viable
Licensing software stack or design from Gainspan corporation.

A.6.2 Recommendation
This approach should not be used without significant investment into assessing a low power dongle
mode. This would carry risk of it not working.

15
CONFIDENTIAL – between PA and BDR Thermea
 ARM A5 BOOT AND DEBUG STRATEGIES

B.1 Introduction
This section describes the available and recommended strategies for booting and debugging the
candidate ARM 5 processor.

B.2 ARM boot sequence

The ARM processor has an embedded ROM used to control the boot sequence. The system always
boots from ROM address 0x0 in the first stage bootloader process. It will initialise the chip and then
attempt second stage boot from Non-Volatile Memory locations.
Failure to locate valid entry code will result in the boot sequence entering SAM-BA mode, which
initialises USB peripherals and waits for relevant control sequences on these interfaces to enable
upload of code to memory and boot of the processor.
The Atmel A5 development board stores program memory on a 1GB flash memory chip, and therefore
boots from this location in the second stage. A push button is provided as an input chip select port on
the NAND flash allowing the user to disable automatic boot from NAND flash at power up and entry to
SAM-BA mode for debug mode, or to enable write access to the NAND flash.
The same method of boot is proposed for the RU2 and MK3 units.

16
CONFIDENTIAL – between PA and BDR Thermea
Figure 5: The boot sequence for the Atmel ARM A5 processor

17
CONFIDENTIAL – between PA and BDR Thermea
Figure 6: The debug mode state diagram for the Atmel ARM A5 processor

B.3 Debug peripherals

The Atmel ARM A5 has a debug

18
CONFIDENTIAL – between PA and BDR Thermea
 USB WI-FI MODULE POWER
CHARACTERISTICS

C.1 Introduction
This appendix describes the experimental procedure used in assessing the viability of the Wi-Fi USB
module approach to connect the RU2 unit to the public internet.

C.2 Experimentation
C.2.1 Experimental setup
The experimental setup uses the following:
 Raspberry pi computer running raspbian OS
 USB male to female extension cable, with wires exposed to allow current probe on power rail
 Standard 802.11n Wi-Fi USB dongle (unbranded)
 Lecroy wavejet 324 oscilloscope
 Tektronix TCPA300 Current probe
The Wi-Fi module was chosen due to its reported low power characteristics. Various raspberry pi
forums reported that other solutions were too power hungry to work effectively with the raspberry pi
chip.
The figures below show the experimental setup.

19
CONFIDENTIAL – between PA and BDR Thermea
C.2.2 Results
The figures below demonstrate the results from the experiment. The average current consumed by the
Wi-Fi chip us 83mA, which translates to a power of 415mW. The peak current was 248mA, translating
to 1240mW (1.24W).
The settings for the measurements below are:
 Oscilloscope scope 20mV per division (vertical)
 Current probe: 1A/V
 Higher resolution photos are available in the electronics working directory.

20
CONFIDENTIAL – between PA and BDR Thermea
C.2.3 Conclusions
The USB Wi-Fi modular approach in this form consumes too much power. In order to develop this
solution further, a custom TCP/IP and USB driver combination should be considered.

21
CONFIDENTIAL – between PA and BDR Thermea
 OPTICAL ENCODER EXPERIMENTATION

D.1 Introduction
This section describes the experimentation performed in order to develop a proof of concept
electronics design for an accurate rotating dial that will be used to control the menu on the RU-2 and
MK-3 devices. This should prove the concept and then feed into development for the Stage 3 full
device prototypes.

D.2 Experimentation
D.2.1 Experimental setup
The experimental setup contains four key components:
 Avago optical encoder AEDR-8300-1W0;
 Prototype dials (with vertically mounted and horizontally designed ‘ribs’ used to trigger the optical
encoder output);
 Two vices, used to experiment with the positioning of the encoder and dial; and
 Oscilloscope for testing the clarity of the optical encoder output.

22
CONFIDENTIAL – between PA and BDR Thermea
Figure 7: The experimental setup

23
CONFIDENTIAL – between PA and BDR Thermea
Figure 8: The dial with horizontally mounted ribs

Figure 9: The optical encoder.

Figure 10: The optical encoder, control wheel arrangement with clean output observed on the
oscilloscope

D.2.2 Results and design considerations

Some experimentation was required to ensure the optical encoder could be demonstrated to work
correctly:
 The SLA material has unique properties that would not be exhibited from a high-volume
manufactured plastic. In order to make the encoder work correctly, the ribs were lightly sprayed
with a metallic paint. This increases the reflectivity of the surface, allowing the optical encoder to
pick up a cleaner difference.
 The length of the ribs needed to be increased in order to provide a clean output from the optical
encoder. This is the length relative to the radius of the cylinder the length of the ribs had to be
extended towards the centre if the radius of the cylinder. Again, this was to increase the reflectivity
in order for the encoder to pick up the signal.

24
CONFIDENTIAL – between PA and BDR Thermea
 There is a relatively tight tolerance on the distance of the wheel ribs from the encoder. The
tolerance is as stated in the Avago datasheet, and is achievable with mechanical fixings.
Some mechanical and electronic debug of the design is expected in stage 3. However, the proof of
concept is demonstrated in the figure below. This figure shows the electronic signals output from a fine
rotation of the dial (approximately 3 degrees).

Figure 11: The optical encoder output with a 3 degree movement of the dial.

25
CONFIDENTIAL – between PA and BDR Thermea
 PROXIMITY AND AMBIENT LIGHT
SENSOR EXPERIMENTATION

E.1 Introduction
This section describes the experimentation done to test the recommended proximity and ambient light
sensors. The purpose is to prove the concept, uncover any considerations for completing the
electronics design and identify any mechanical and other considerations.

E.2 Experimentation
The experimentation employs a Silicon Labs Infrared Multi-Function Board Evaluation Kit IRMFB-EK.

E.2.1 Proximity sensing

The development board has a mode that allows the proximity sensors to be tested. As proximity to the
sensor is detected, LEDs progressively lught up as the object comes c loser to the sensor.
The board is preloaded with the demo on the “bottom” side of the board. The switch selecting “bot
side” was enabled for the demo and the proximity sensor was tested by placing a hand at distances
and observing the distances at which the light came on. This is demonstrated in the figures below.
The sensor is accurate and provides an excellent user experience. It is highly responsive, with no
noticeable lag in response of the LEDs turning on.
The sensor was tested with a variety of transparent plastic materials, including polycarbonate and
acrylic. The existing plastic screen was removed in order for this to be tested. This confirmed that our
chosen screen material, polycarbonate, will allow the use of this sensor.

E.2.2 Considerations for other parts of the design

Rubber washer to prevent spill from emitter.
Space the sensor from the emitter a short distance.
Include a calibration parameter in software.

26
CONFIDENTIAL – between PA and BDR Thermea
 27
CONFIDENTIAL – between PA and BDR Thermea
Additional comments:

E.2.3 Ambient light

The same sensor package includes a visible light spectrum sensor. To test this we used the same
development board and enabled the development kit software “Silicon Laboratories Performance
Analysis Tool Version 2.35”, which allows raw outputs from the sensor to be observed. The
experimentation is show below, which displays three scenarios:
1. light shined directly onto the sensor from a smartphone camera flash-light;
2. ambient light in an office environment; and
3. darkened conditions where a hand has been placed over the sensor.
The sensor performs well in all conditions, with the following readings taken, with tests enumerated as
above.
1. Observed value around 300
2. Value around 40
3. Value less than 10

28
CONFIDENTIAL – between PA and BDR Thermea
Given this dynamic range, it should be feasible to design a simple software algorithm to detect the
lighting conditions and alter the LCD screen brightness.

The diagrams of each scenario are presented below. Higher resolution diagrams that display the
graphed screen outputs in more detail are stored in the project directory.

29
CONFIDENTIAL – between PA and BDR Thermea
 POWER CONSUMPTION CALCULATION

F.1 Atmel processor system

Nominal values have been taken for the purpose of this document. Further design corners are
presented in the Atmel A5 datasheet.
 1.2V, 90mA VDD Core power consumption in Active mode. Whilst datasheets report 108mA, Atmel
have verbally confirmed that in normal operation with graphics active, we should expect the
processor to consume.
 Backup mode for RTC: 1.2uA @ 3.3V. For the RTC, a battery with a minimum specification of
158mAh for a 15 year lifetime (131400 hours). A 3V CR2032 with 250mAh would satisfy this
requirement.

F.2 Power regulator efficiencies

There are a number of switched mode voltage regulators in the power supply architecture.

F.2.1 24V regulator

Runs at approximately 80% efficiency with the given input and output voltages.

F.2.2 Active Semi Processor regulator

Efficiency curves for the Active semiconductor ACT8865 power regulation chip are shown below. The
regulator is most efficient with a 3.6V input, supplying around 200mA of current. At this point the
efficiency is 90% for the 1.2V rails and 95% for the 3.3V rail.

F.2.3 Overall efficiency estimate

The overall efficiency estimate used in our calculations and is therefore 80%.

30
CONFIDENTIAL – between PA and BDR Thermea
Figure 12: Efficiency of the multiple rail voltage regulator

F.3 RU2 device power budget

Item Power Notes

Processor: ARM A5 108mW From datasheet, 25C 90mA @ 1.2V

Densitron Screen 528mW 40mA forward Current

WiFi Module 330mW (continuous receive) Based on GainSpan data.

726mW (transmit)

Encoder 50mW For optical encoder

Passive and Power 300mW Assuming 20% loss

Management

SDRAM 288mW Operating burst write current of 160mA,

configuration x16

Flash 115mW Based on program current of 35mA @ 3.3V

31
CONFIDENTIAL – between PA and BDR Thermea
 Total Power without 1306mW This includes a 20% inefficiency overhead
Wi-Fi

Total Maximum Power 2178mW This includes a 20% inefficiency overhead

F.4 MK3 Device power budget

Item Power Notes

Processor: ARM A5 108mW From datasheet, 25C 90mA @ 1.2V

New Vision Screen 432mW 45mA forward Current

Encoder 50mW For optical encoder

Passive and Power 300mW Assuming 20% loss

Management

SDRAM 288mW Operating burst write current of 160mA,

configuration x16

Flash 115mW Based on program current of 35mA @

3.3V

Total Maximum Power 1192mW This includes a 20% inefficiency overhead

F.5 Supercapacitor calculation (RU-2 only)

The power profile for the RU-2 device shown above requires the use of a supercapacitor in order to
supply peak power when the Wi-Fi is transmitting or receiving.
Assuming the use of the Panasonic 0.22F capacitor. The following characteristics must hold true:
 The voltage at the output of the capacitor must remain at 3.3V + V_do, the dropout voltage of the
regulator controlling power to the Wi-Fi module.
Wi-Fi module transmit and receive charge profile:
 Radio power up: 43mA 1.56 ms
 CCA 126mA, 1.8ms
 Tx Data 194mA 1.4ms
 Rx ACK 134mA 280 us
 Save state 22.5mA, 1.4ms
In pessimistic terms, the circuit must be able to supply 200mA for around 9 milliseconds
The RU-2 power supply architecture shows the arrangement of supercapacitor and voltage regulation.
The supercapacitor is charges at 5.5 volts and fed into a low dropout 3.3V regulator. Assuming the
power consumed is constant, the Wi-Fi module consumes 600mW (3.3V*200mA).
The voltage regulator cannot operate when the input drops below 3.3V plus the dropout voltage.
Therefore there is only ~1V of voltage drop before the circuit will cease to operate.

32
CONFIDENTIAL – between PA and BDR Thermea
The standard equation that governs this behaviour is t=0.5C(Vo^2-V1^2)/P. Therefore the
arrangement we have designed should operate for approximately 1 second before the circuit fails.
This 1 second figure is 100 times greater than the requirement, and should therefore provide ample
room to transmit all necessary data when all components are in operation. Furthermore, it will allow for
significant supercapacitor degradation over time.
This calculation is outlined in the table below.

Value Unit Description

3.3 V Module power rail

0.2 A Transmit current

0.66 W Power = Current of 200mA times voltage of 3.3V

5.5 V V0

V1 = 3.3V plus 1.2V dropout voltage (this is a

4.5 V pessimistic dropout voltage)

0.22 F Capacitance

Theoretical time that the capacitor can provide

1.666666667 s current for

Time that the capacitor can transmit for, including

1 s safety margin.

F.6 Real-time clock considerations (MK-3 only)

The calculation below provides a calculation outlining the required battery for the MK-3 Device. In
addition to the Mathematics there are the following considerations:
 The battery should be replaceable
 A mechanical part will be required in the case of the replaceable battery, contributing cost,
manufacturing time and potentially an additional installation step.

Component Value Unit Notes

RTC current draw 0.0000012 Amps from Atmel ARM A5 datasheet

Lifetime requirement 131400 hours 24 hours*365 days*15 years

Required charge 0.15768 Ah

Required charge 157.68 mAh

Storage and leakage margin 236.52 mAh suggested battery 250mAh

33
CONFIDENTIAL – between PA and BDR Thermea
 MECHANICAL DESIGN
CONSIDERATIONS

G.1 Introduction
This appendix outlines existing solutions and design considerations that feed into the electronic and
mechanical design of the RU-2 and MK3 devices. This includes the arrangement of the temperature
sensor.

G.2 Temperature sensing

G.2.1 Existing IP and solutions
The current qSense device manufactured by Honeywell for BDR Thermea includes a number of
considerations for measuring the temperature:
 A thermistor that is thermally isolated from the rest of the PCB using cutouts to prevent through-
PCB conduction;
 Plastic casing that fits over the thermistor and prevents convective currents through the air gap in
the device;
 Thermally conductive material to transfer heat from the front of the casing to the thermistor.
These features are presented in the photograph below, which shows the PCB next to the thermally
conducting material. This thermally conducting material typically sits directly on top of the thermistor
PCB: in the figure it is shown facing the camera.

PCB holes for thermal isolation

Thermally conductive material

Holes provide Flying lead

mitigation strategy

Plastic casing that sits

around thermistor

Thermistor

34
CONFIDENTIAL – between PA and BDR Thermea
Figure 13: Honeywell device temperature sensor and mechanical fixing

The Nest device employs a similar mechanical design with four additional features:
4. The device has a hole to the external environment;
5. The casework is more carefully designed to fit on top and in contact with the PCB. This is
achieved more accurately by placing the thermistor next to the fixing hole that attaches the unit to
the wall. Therefore allowing the screw to tighten the casework onto the PCB.
6. A thermally insulating coating has been applied to the surface of the PCB.
7. The high-end processor subsystem and screen have been place on a separate PCB with
insulating casework between the two printed circuit boards.

Figure 14: Nest device temperature sensor and mechanical fixing

35
CONFIDENTIAL – between PA and BDR Thermea
 WI-FI MODULE TESTING

H.1 Introduction
This appendix discusses the testing conducted with the candidate Wi-Fi module. The key aspects
required of the Wi-Fi module are:
 Low power consumption, in line with numbers quoted by the manufacturer’s marketing material;
and
 Sufficient communications range from the module, allowing connection to the internet from within a
house or flat environment.

H.2 Experimentation
H.2.1 Setup
The module tested was the gainspan GS2100MIE, which is fitted to a Gainspan development board:
EVB3, revision 2.0. This setup employs an external antenna. Despite presence of a PCB antenna, a
separate module must be used to test this functionality (the GS2100MIP), but no development
environment is available for this configuration.
The development environment allows a PC to communicate with the board using a serial terminal. We
have used teraterm via a standard COM port. The development board uses FTDI USB to serial
converter chips to allow the PC to communicate via a USB to serial emulator/driver.

H.2.2 Power consumption

H.2.3 Range
Some basic tests were performed with the GS2100MIE. This was done by querying the RSSI
measured by the Gainspan module at various points around the PA office environment and by
comparing to mobile phone RSSI measurements. The teste were as follows:
1. Test within 2 metres of a Wi-FI router.
2. Test 60 metres from the Wi-Fi router (not line-of-sight), this is at far end of range.
3. Comparative test with smartphone Samsung Note 3.
4. Comparative test with Wi-Fi dongle Ralink 5370.

The results were as follows:

1. Module reported RSSI of -30dBm.
2. Module reported RSSI of -78dBm.
3. The range for the module and for a standard cellular phone were equivalent.
4. The range for the low-cost Wi-Fi dongle was insufficient, with the connection being unstable at
2/3rds of the distance of the module and smartphone.

36
CONFIDENTIAL – between PA and BDR Thermea
 Corporate headquarters
123 Buckingham Palace Road
London SW1W 9SR
United Kingdom
Tel: +44 20 7730 9000

paconsulting.com

This document has been prepared by PA

on the basis of information supplied by the
client and that which is available in the public
domain. No representation or warranty is given
as to the achievement or reasonableness
of future projections or the assumptions
underlying them, management targets,
valuation, opinions, prospects or returns,
if any. Except where otherwise indicated,
the document speaks as at the date hereof.

© PA Knowledge Limited 2014.

We are an employee-owned firm of over 2,500 people, All rights reserved.
operating globally from offices across North America,
This document is confidential to the
Europe, the Nordics, the Gulf and Asia Pacific.
organisation named herein and may not
be reproduced, stored in a retrieval system,
We are experts in energy, financial services, life sciences
or transmitted in any form or by any means,
and healthcare, manufacturing, government and public electronic, mechanical, photocopying or
services, defence and security, telecommunications, otherwise without the written permission
transport and logistics. of PA Consulting Group. In the event that
you receive this document in error, you
Our deep industry knowledge together with skills in should return it to PA Consulting Group,
123 Buckingham Palace Road, London
management consulting, technology and innovation SW1W 9SR. PA accepts no liability
allows us to challenge conventional thinking and whatsoever should an unauthorised recipient
deliver exceptional results with lasting impact. of this document act on its contents.

37
CONFIDENTIAL – between PA and BDR Thermea