1 |
2 |
btltz |
SpaceWire (SpW) for What?
|
2 |
|
|
"Spw grown organically from the needs of on-board processing applications.
|
3 |
|
|
(SpW is a network with Routers)
|
4 |
|
|
This Standard addresses the handling of payload data and control information on
|
5 |
|
|
board a spacecraft. It is a standard for a high speed data link, which is intended
|
6 |
|
|
to meet the needs of future, high capability, remote sensing instruments and
|
7 |
|
|
other space missions...
|
8 |
|
|
Processing units, mass-memory units and down-link telemetry systems devel-
|
9 |
|
|
oped for one mission can be readily used on another mission, reducing the cost of
|
10 |
|
|
development, improving reliability and most importantly increasing the amount
|
11 |
|
|
of scientific work that can be achieved within a limited budget.
|
12 |
|
|
(SpW is currently being installed on several NASA and European Space Agency (ESA)
|
13 |
|
|
spaceships to support onboard communications during space missions.---27 January 2003)
|
14 |
|
|
SpaceWire has taken into consideration two existing standards, IEEE 1355-1995
|
15 |
|
|
and ANSI/TIA/EIA-644.
|
16 |
|
|
"Point-to-point serial links with LVDS. The maximum data signalling rate that can be
|
17 |
|
|
achieved is different from one system to another, depending on several factors such as:
|
18 |
|
|
1)cable length, 2)driver-receiver technology, and 3)encoder-decoder design, and is limited by 4)skew and jitter."
|
19 |
|
|
|
20 |
|
|
|
21 |
|
|
|
22 |
|
|
|
23 |
|
|
|
24 |
|
|
|
25 |
|
|
Diff & Relation with IEE1355 device(eg. Atmel.Inc "TSS901E" which also called "SMCS332")
|
26 |
|
|
"The SMCS devices implement revision A of the standard, which is based on IEEE1355. They do not implement
|
27 |
|
|
the current version of SpaceWire. However, the current version of SpaceWire is backwards compatible with the earlier revision."
|
28 |
|
|
|
29 |
|
|
//////////////////////////////////////////////////////////////////////////////////////
|
30 |
|
|
|
31 |
|
|
|
32 |
|
|
|
33 |
|
|
|
34 |
|
|
|
35 |
|
|
This Core?
|
36 |
|
|
- CODEC
|
37 |
|
|
- A pure synchronous design(so it can works well in fpga) written in verilog, completely open.
|
38 |
|
|
- Annother high speed CODEC use PLL and clock recovery/lock circuit.
|
39 |
|
|
-SWR
|
40 |
|
|
- ...Synchronously Clocked Crosspoint...(Switch Latency...)
|
41 |
|
|
- ... buffered crossbar...distributed scheduling(effective and simple). 16x16 now may be enough.
|
42 |
|
|
|
43 |
|
|
* Due to the SpaceWire protocol,the file/module "SPW_CODEC.v" could be used as a core's top module (as a "channel")
|
44 |
|
|
to be integrated with a "routing matrix" to form a router which is a Routing Switch in the SpaceWire network.
|
45 |
|
|
Sometimes the CODEC is also called a "interface" or "input/output ports" of the router or a node device.
|
46 |
|
|
See the scheme.
|
47 |
|
|
* The file/module "SPW_I_vlogcore.v" is 3 CODEC with glue logic and Wishbone interface(COMI/HOCI) could be used as top module(a interface of the host to the SpW network) and be integrated into node devices.
|
48 |
|
|
("SpaceWire nodes are the sources and destinations of packets in a SpaceWire network.").
|
49 |
|
|
* The file/module "SWR_vlogcore.v" is a top module of Spacewire Router which integrates some interfaces and a
|
50 |
|
|
Routing Matrix "SwitchCore.v" and a little other logic. This core is so rough now.
|
51 |
|
|
* Had use some ip on the opencores.org such as a fifo(i have consulted the "plesiochronous_fifo.v" for the
|
52 |
|
|
Receiver. For the transmitter which only need a synchronous fifo for internal usage,so i write a synfifo which was similar to Xilinx ip for less complexity.) and a CRC.
|
53 |
|
|
* Note the LVDS driver and receiver is not contain in modules.
|
54 |
|
|
* Triple Modulo Redundant (TMR).
|
55 |
|
|
|
56 |
|
|
|
57 |
|
|
|
58 |
|
|
|
59 |
|
|
|
60 |
|
|
|
61 |
|
|
|
62 |
|
|
|
63 |
|
|
A little specs:
|
64 |
|
|
-INTERFACE
|
65 |
|
|
- The block "Timer" in the scheme of the draft-1 is merged into the "SPW_FSM".
|
66 |
|
|
|
67 |
|
|
-[Clock]
|
68 |
|
|
-Clock : Synchronous method
|
69 |
|
|
- The block "Receiver's clock recovery" in the scheme of the draft-1 is merged into the "Receiver".
|
70 |
|
|
- The block "Transmitter clock generator" in the scheme of the draft-1 is merged into the "Transmitter".
|
71 |
|
|
-Clock : High speed method
|
72 |
|
|
-The wide -> narrow process unit could be found in Transmitter and PPU.
|
73 |
|
|
-
|
74 |
|
|
|
75 |
|
|
- The credit count keeped in "crd_cnt" which indicates the buffer space of the opposite Rx bufer is in the transmitter; at the same time the outstanding count keeped in "osd_cnt" which indicates the buffer data of the opposite Tx is in the receiver.
|
76 |
|
|
- Notice that according to the custom of circuit operation,some name near the interface between the Tx/Rx and the fifo may be a little different from the standard document.(eg."TX_READY, TX_WRITE" may only present as a output pulse signal "rdbuf_o" to just read that mem.
|
77 |
|
|
- Note that the Receiver do not generate local EEP to the host. It receive remote EEP, and if a error occurs when PPU is in Run state, a Link error output then will be asserted by the PPU. Then the EEP may be written to the transmitter by the host(the user's application software).
|
78 |
|
|
- "crc32_lib.v" on OC has been tailored to fit this pjt. It is optional.
|
79 |
|
|
- Not like the Atmel's "TSS901E"(or called "SMCS332"),this ip core dosen't use a
|
80 |
|
|
"clk10" but only a global clock "gclk".The "clk10" is not indispensable according
|
81 |
|
|
to the SpaceWire standard and this may also be convenient.
|
82 |
|
|
Note that not like TSS901E,the node channels doesn't support wormhole routing.
|
83 |
|
|
All routing function is performed by Routers.
|
84 |
|
|
|
85 |
|
|
|
86 |
|
|
|
87 |
|
|
|
88 |
|
|
|
89 |
|
|
|
90 |
|
|
-ROUTER
|
91 |
|
|
- bufferd switch
|
92 |
|
|
- ... minimum switching latency and lowest propagation delay ...
|
93 |
|
|
- ... ...
|
94 |
|
|
- ...operate directly on variable-size packets,...no packet
|
95 |
|
|
segmentation and no scheduler(distributed scheduling decisions)
|
96 |
|
|
- "sop_ack_o" is a sychrnous acknowledge signal indicate that the cell which a package want to be loaded has space to load. If it's true, after 1 clock, the controller will load data to that cell.
|
97 |
|
|
|
98 |
|
|
|
99 |
|
|
*********************************************************
|
100 |
|
|
- +Conformity Limitations:
|
101 |
|
|
|-- Link Interfaces --+
|
102 |
|
|
| |-
|
103 |
|
|
+-- Router --+
|
104 |
|
|
|-
|
105 |
|
|
|
106 |
|
|
|
107 |
|
|
|
108 |
|
|
TODO: the "Router" and various interface to other bus.
|
109 |
|
|
*************************************************************
|
110 |
|
|
|
111 |
|
|
/////////////////////////////////////////////////////////////////////////////////
|
112 |
|
|
|
113 |
|
|
|
114 |
|
|
|
115 |
|
|
|
116 |
|
|
|
117 |
|
|
|
118 |
|
|
|
119 |
|
|
Some interesting topics:
|
120 |
|
|
1)Open or close?
|
121 |
|
|
In the "SpaceWire Router Requirements Specification(Issue:1 Rev: 5 Date:28.02.2003)" clause 4.1 has some words about
|
122 |
|
|
this question:
|
123 |
|
|
"Open VHDL Core
|
124 |
|
|
The SpaceWire input/output ports shall be implemented with the Open VHDL core developed in the ESM006 contract for ESA [AD2].
|
125 |
|
|
Rationale: It will minimise design effort and help to validate the VHDL core."
|
126 |
|
|
It did not refer to open the Router's IP,besides i have not found free "IO ports" core available on internet.
|
127 |
|
|
|
128 |
|
|
2)Active things:(observe at 03-26-2005)
|
129 |
|
|
If someone has deep interest,you could pay attention to a conference "DASIA(DAta Systems In Aerospace) 2005" in
|
130 |
|
|
Sheraton Grand Hotel, Edinburgh, Scotland during 30th May and 2nd June 2005.
|
131 |
|
|
|
132 |
|
|
Tuesday 31st May 2005(SESSION 4B ON BOARD BUSES AND COMMUNICATION)
|
133 |
|
|
[Chairperson: P. Plancke, ESA/ESTEC, The Netherlands]
|
134 |
|
|
09:00 Quality from SpaceWire and Quality into SpaceWire Systems B.M. Cook, P.H. Walker - 4Links, UK
|
135 |
|
|
09:30 SpaceWire Remote Memory Access Protocol S. Parkes, C. McClements - University of Dundee, UK
|
136 |
|
|
10:00 SpaceWire Internet Tunnel S. Mills, S. Parkes - University of Dundee, UK
|
137 |
|
|
|
138 |
|
|
|
139 |
|
|
|
140 |
|
|
|
141 |
|
|
|
142 |
|
|
|
143 |
|
|
|
144 |
|
|
|
145 |
|
|
|
146 |
|
|
//////////////////////////////////////////////////////////////////////////////
|
147 |
|
|
Annex A:
|
148 |
|
|
|
149 |
|
|
Industry info and some Web site could be accessed for a further reference::
|
150 |
|
|
|
151 |
|
|
1) "SpaceWire UK are specialist providers of VHDL Intellectual Property & Design Services for the Space Industry".
|
152 |
|
|
The SpaceWire UK has developed a CODEC and SpaceWire Switch Core (Router) in VHDL for business. (I am sure i
|
153 |
|
|
don't want to disturb this new ip vendor and I also had consulted their ip schemes.My try is just a hobby and
|
154 |
|
|
have a heart to contribute to promoting the SpW standard.)
|
155 |
|
|
www.spacewire.co.uk
|
156 |
|
|
|
157 |
|
|
2)The EADS Astrium GmbH (formed in May 2000 by the merger of Matra Marconi Space (France/UK) and the space division
|
158 |
|
|
of DaimlerChrysler Aerospace (Germany)) is well known as the author of the "Router requirements spec"
|
159 |
|
|
(Steve Parkes, University of Dundee).
|
160 |
|
|
" ETD-031 'SpaceWire Router Development, Coordination and Validation' was Kicked-Off with Astrium the 2nd
|
161 |
|
|
January 2002. " And the issue 1 rev 5 had been published on 28.02.2003.
|
162 |
|
|
They described a router ASIC with internal configuration port,external pins for status/error monitoring
|
163 |
|
|
(with status/error registers and control registers) in Feb,2004.It had been implemented in an Atmel MH1RT
|
164 |
|
|
gate array.
|
165 |
|
|
www.astrium.eads.net
|
166 |
|
|
|
167 |
|
|
3)Austrian Aerospace (AAE) has a switch ASIC, the Dynamic Switch Matrix (DSM), built in Atmel.
|
168 |
|
|
" ESM-006 'SpaceWire Router ASIC development' is an ETD-031 sub-contract with Austrian Aerospace concerning
|
169 |
|
|
the SpaceWire CODEC VHDL, router FPGA and router ASIC. "
|
170 |
|
|
www.space.at
|
171 |
|
|
|
172 |
|
|
4)4links derives from University of Dundee is a company that specializes in SpaceWire / IEEE1355 &Ethernet.
|
173 |
|
|
They have some products such as EtherSpaceLink, SpaceWire-cPCI,SpaceWire-PCI,SpaceWire-Cables,and
|
174 |
|
|
IPs which are relative to these products. They use Xilinx fpga implement routers with a performance
|
175 |
|
|
well above 200 Mbit/s.
|
176 |
|
|
www.4links.co.uk/index.htm
|
177 |
|
|
|
178 |
|
|
5)Webs of Chinese laboratory of avionics(for space use) say they are developing a SpW network system.
|
179 |
|
|
www.spacee.net
|
180 |
|
|
|
181 |
|
|
6)Atmel offer SMCS332,SMCS Lite,and said their "potential future Standard ASICs & IPs include:SMCS Lite,
|
182 |
|
|
SpaceWire Router(2004),SpaceWire Rx/Tx" in 2002.
|
183 |
|
|
www.atmel.com
|
184 |
|
|
|
185 |
|
|
7)STAR-Dundee Ltd is a spin-out company of the University of Dundee set up specifically to support users
|
186 |
|
|
of SpaceWire(fund from the Scottish Executive). Dr Steve Parkes was the Managing Director of STAR-Dundee,
|
187 |
|
|
also the author of the SpaceWire standard document. "STAR-Dundee Ltd aims to fill the emerging market
|
188 |
|
|
niche for research and development, debug and support tools for SpaceWire."
|
189 |
|
|
|
190 |
|
|
Dr Steve Parkes: sparkes@computing.dundee.ac.uk
|
191 |
|
|
Press Officer: j.m.marra@dundee.ac.uk
|
192 |
|
|
|
193 |
|
|
www.dundee.ac.uk
|
194 |
|
|
|
195 |
|
|
8) :--) Ask Google for more and fresh news.
|
196 |
|
|
|
197 |
|
|
|
198 |
|
|
|
199 |
|
|
|
200 |
|
|
|