Folie 1 - RWTH Aachen University

Download Report

Transcript Folie 1 - RWTH Aachen University

Compilation of Dis-/Advantages
of DC-DC Conversion Schemes
Katja Klein
1. Physikalisches Institut B
RWTH Aachen University
Power Task Force Meeting
December 16th, 2008
Advantages: Grounding
• Standard grounding scheme
 Module ground potentials are all the same
 Common ground reference for bias, analogue and digital voltage for
whole substructure (rod, petal)
 Bias voltage ground reference is the same for all modules
 Easier for slow controls (difficult in SP to sense voltages)
Katja Klein
Discussion of DC-DC Conversion
2
Advantages: Communication
• Readout and control scheme is very standard
 AC-coupling of communication not needed
 Control chips can be supplied independently of modules
Katja Klein
Discussion of DC-DC Conversion
3
Advantages: Start-Up & Selective Powering
• Easy start-up
 Control chips can be powered on first
 If one converter per module, single modules can be powered on/off
 In scenario w/ charge pump per chip, single chips can be powered on/off
Katja Klein
Discussion of DC-DC Conversion
4
Advantages: Different Voltages
• Different voltages can be provided
 Buck-type converters: the same converter chip can be configured for
different output voltages
 Via a resistive bridge
 Two conversion steps can be combined
 No efficiency loss (in contrast to linear regulation in SP)
 Can cope with Vopto > Vchip
 Can cope with Vana ≠ Vdig
 Charge pumps: only integer conversion ratios, defined by configuration
Katja Klein
Discussion of DC-DC Conversion
5
Advantages: Flexibility
• Great flexibility with respect to
 combination of modules with different load
 Different numbers of readout chips
 Trigger modules vs. standard modules
 power groups with different number of modules
 TEC vs. barrel
• In contrast, with SP current is fixed to highest current needed by any
chain member  chains must be uniform to avoid burning power in regulators
Katja Klein
Discussion of DC-DC Conversion
6
Advantages: Changing Loads
• Compatibility with changing loads, relevant for
 pixel detector
 load is driven by occupancy
 trigger modules
• SP: the highest current potentially needed must always be provided
 inefficiency
Katja Klein
Discussion of DC-DC Conversion
7
Disadvantages: Chip Technology
• Need for a “high voltage“ tolerant process (10-12V)
 Good candidate identified, radiation hardness still to be proven
 Strong dependency on foundry: support of process over years?
 Any changes in process must be followed closely and irradiation
tests be repeated
Katja Klein
Discussion of DC-DC Conversion
8
Disadvantages: Converter Efficiency
• Converter efficiency will be around 80%
(ESR of passive components, Ron of transistors, switching losses)
 Local generation of heat cooling of DC-DC converters needed
 Local efficiency decreases with lower conversion factor (Uout/Uin)
 Local efficiency decreases with higher switching frequency
 In two-step schemes efficiencies multiply
Katja Klein
Discussion of DC-DC Conversion
9
Disadvantages: Currents in Cables
• Cannot compete with Serial Powering
 Currents in power group with DC-DC conversion = I0nr
 I0 = current of a single module
 n = number of parallely powered modules in the power group
 r = conversion ratio = Uout/Uin
 Current in Serial Powering chain = I0, independent of n
 E.g. for 20 modules in power group need r = 20 to compensate
 Higher efficiency in SP (at least up to FE)  less cooling needed
 Cables inside tracker volume can be thinner with SP
Katja Klein
Discussion of DC-DC Conversion
10
Disadvantages: Risks
• We have to stick with parallel powering
 Multiplicity (modules per cable) as today or higher
 Open connections (e.g. at PP1, PP0) lead to loss of power group
 Short on module leads to loss of power group
 Protection needed? Use DC-DC converter to switch off module?
 Converter can break: can imagine isolated failures (loss of
regulation...) and failures that lead to loss of power group (short)
 More risky if one converter powers several modules
 Do we need redundancy?
 This adds mass
Katja Klein
Discussion of DC-DC Conversion
11
Disadvantages: Material & Space
• Material budget and space considerations
 Amount of copper in cables scales with current = I0nr
 I0 = current of a single module
 n = number of modules in the power group
 r = conversion ratio = Uout/Uin
 Air-core inductor (even if integrated into PCB, it needs a lot of copper)
 Filter capacitors, maybe other filter components
 With regulation (buck etc.), PCB traces can be narrow
 Without regulation (charge pumps), input voltage must be exact
 Linear regulator or rather solid input traces
 Is shielding needed? How to design good low mass shielding?
Katja Klein
Discussion of DC-DC Conversion
12
Disadvantages: Material Budget
Simulated components:
 Kapton substrate with 4 copper layers
 Copper wire toroid
 Resistors & capacitors
 Chip
TEC
1 converter / module
Motherboards
Analog
OptoHybrids
Kapton
circuits
FE-hybrids
Katja Klein
Discussion of DC-DC Conversion
13
Disadvantages: Material Budget
Total gain for strip tracker with 1 converter per module and a conversion ratio of 8;
with power cables and motherboards modified accordingly:
Strip tracker
Katja Klein
Discussion of DC-DC Conversion
14
Disadvantages: Noise
• DC-DC converters are undoubtedly noise sources (by design)
 Conductive noise through cables
 Ripple on output voltage: switching frequency (1-5MHz) + higher harmonics
are in the bandpath of the amplifier
 Switching leads to high frequency noise (tens of MHz, not so critical)
 Both CM and DM contributions
 Radiated noise
 From inductor near field via inductive (and capacitive?) coupling
 From cables
 Has to be taken into account for all aspects of electronics system design:
readout chip, FE-hybrid, grounding & shielding, motherboard, layout ...
 Not clear what to prepare for: noise depends on implementation
 For same chip, noise emission can be rather different depending on PCB etc.
 Scalability from lab system to complete detector not obvious
 PS noise requirements to be understood
Katja Klein
Discussion of DC-DC Conversion
15