DXT Plug-In Unit Descriptions: Mbif-C, Mbif-Cr
DXT Plug-In Unit Descriptions: Mbif-C, Mbif-Cr
DXT Plug-In Unit Descriptions: Mbif-C, Mbif-Cr
MBIF-C, MBIF-CR
DN02176011
Issue 1-2
MBIF-C, MBIF-CR
The information in this document applies solely to the hardware/software product (“Product”) specified
herein, and only as specified herein. Reference to “Nokia” later in this document shall mean the respective
company within Nokia Group of Companies with whom you have entered into the Agreement (as defined
below).
This document is intended for use by Nokia's customers (“You”) only, and it may not be used except for the
purposes defined in the agreement between You and Nokia (“Agreement”) under which this document is
distributed. No part of this document may be used, copied, reproduced, modified or transmitted in any form
or means without the prior written permission of Nokia. If You have not entered into an Agreement
applicable to the Product, or if that Agreement has expired or has been terminated, You may not use this
document in any manner and You are obliged to return it to Nokia and destroy or delete any copies thereof.
The document has been prepared to be used by professional and properly trained personnel, and You
assume full responsibility when using it. Nokia welcomes your comments as part of the process of
continuous development and improvement of the documentation.
This document and its contents are provided as a convenience to You. Any information or statements
concerning the suitability, capacity, fitness for purpose or performance of the Product are given solely on
an “as is” and “as available” basis in this document, and Nokia reserves the right to change any such
information and statements without notice. Nokia has made all reasonable efforts to ensure that the
content of this document is adequate and free of material errors and omissions, and Nokia will correct
errors that You identify in this document. Nokia's total liability for any errors in the document is strictly
limited to the correction of such error(s). Nokia does not warrant that the use of the software in the Product
will be uninterrupted or error-free.
NO WARRANTY OF ANY KIND, EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO
ANY WARRANTY OF AVAILABILITY, ACCURACY, RELIABILITY, TITLE, NON-INFRINGEMENT,
MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, IS MADE IN RELATION TO THE
CONTENT OF THIS DOCUMENT. IN NO EVENT WILL NOKIA BE LIABLE FOR ANY DAMAGES,
INCLUDING BUT NOT LIMITED TO SPECIAL, DIRECT, INDIRECT, INCIDENTAL OR CONSEQUENTIAL
OR ANY LOSSES, SUCH AS BUT NOT LIMITED TO LOSS OF PROFIT, REVENUE, BUSINESS
INTERRUPTION, BUSINESS OPPORTUNITY OR DATA THAT MAY ARISE FROM THE USE OF THIS
DOCUMENT OR THE INFORMATION IN IT, EVEN IN THE CASE OF ERRORS IN OR OMISSIONS
FROM THIS DOCUMENT OR ITS CONTENT.
This document is Nokia proprietary and confidential information, which may not be distributed or disclosed
to any third parties without the prior written consent of Nokia.
Nokia is a registered trademark of Nokia Corporation. Other product names mentioned in this document
may be trademarks of their respective owners.
Only trained and qualified personnel may install, operate, maintain or otherwise handle this
product and only after having carefully read the safety information applicable to this product.
The safety information is provided in the Safety Information section in the “Legal, Safety and
Environmental Information” part of this document or documentation set.
Nokia is continually striving to reduce the adverse environmental effects of its products and services. We
would like to encourage you as our customers and users to join us in working towards a cleaner, safer
environment. Please recycle product packaging and follow the recommendations for power use and proper
disposal of our products and their components.
If you should have questions regarding our Environmental Policy or any of the environmental services we
offer, please contact us at Nokia for any additional information.
Table of Contents
This document has 53 pages
1 MBIF-C/-CR overview.................................................................... 7
3 Structure of MBIF-C/-CR.............................................................. 11
3.1 Mechanical structure of MBIF-C/-CR............................................11
3.2 Logical structure of MBIF-C/-CR.................................................. 11
3.3 Operating principles of MBIF-C/-CR............................................ 21
3.4 Interfaces of MBIF-C/-CR.............................................................23
3.4.1 MBIF-C/-CR interfaces................................................................. 23
3.4.1.1 MBIF-C/-CR interface with clock equipment................................ 28
3.4.1.2 MBIF-C/-CR, interface with microcomputer unit...........................29
3.4.1.3 Message Bus interface.................................................................42
3.4.1.4 MBIF-C/-CR, other interfaces.......................................................44
3.4.2 Settings of MBIF-C/-CR............................................................... 45
4 Operation of MBIF-C/-CR.............................................................46
List of Figures
Figure 1 Operating environment of MBIF-C/-CR in microcomputer units........... 8
Figure 2 Segmented Message Bus. An MBRP and one MBIF-C in each rack
has been integrated to MBIF-CR..........................................................8
Figure 3 Implementation principle of Message Bus segment and Repeater Bus
for a single segment............................................................................. 9
Figure 4 Block diagram of the MBIF-C/-CR...................................................... 12
Figure 5 Message Bus receiver........................................................................ 13
Figure 6 Message Bus transmitter.................................................................... 15
Figure 7 Message Bus and Repeater interfaces.............................................. 19
Figure 8 Interfaces of the MBIF-C.................................................................... 24
Figure 9 Interfaces of the MBIF-CR..................................................................25
Figure 10 An asymmetrical current-controlled interface..................................... 28
Figure 11 Timing of the 8M and 8K basic timing signal...................................... 29
Figure 12 Interface of the MBIF-C with the Message Bus.................................. 44
Figure 13 Interfaces of the MBIF-CR with the Message Bus and Repeater Bus....
44
Figure 14 Front panel of MBIF-C/-CR.................................................................46
Figure 15 W6 switch, LED indicators and backplane connectors of MBIF-C and
MBIF-CR.............................................................................................48
List of Tables
Table 1 Timing of unit requesting reservation..................................................18
Table 2 Timing of free message bus segment's swapping period...................20
Table 3 Interface signals of the MBIF-C/-CR...................................................25
Table 4 I/O Read operations of the MBIF-C/-CR.............................................29
Table 5 Structure of address and command word...........................................34
Table 6 Selection of command message type.................................................35
Table 7 Structure of command messages.......................................................35
Table 8 I/O Write operations of MBIF-C/-CR plug-in units.............................. 36
Table 9 The principle of operation of the receive program.............................. 40
Table 10 Fill-up level of the receive-FIFO......................................................... 42
Table 11 Use of MB interface signals................................................................43
Table 12 Power Consumption of MBIF-C/-CR.................................................. 47
Table 13 Interchangeability code settings of MBIF-C/-CR (W6)........................48
Summary of changes
Changes between document issues are cumulative. Therefore, the latest document
issue contains all changes made to previous issues.
Changes between issues 1-2 and 1-1
Table 13 corrected.
Changes between Issues 1-1 and 1-0
Figure in Section MBIF-C C74990, MBIF-CR C104965 updated.
Issue 1-0
First issue.
1 MBIF-C/-CR overview
Main functions of MBIF-C and MBIF-CR
The Message Bus Interface (MBIF-C) acts as a bi-directional interface between a
microcomputer unit and a 16 bit parallel Message Bus (MB). The MBIF-C independently
controls the Message Bus (MB): all arbitration, request and release functions of the MB
have been decentralised to all MBIF-C plug-in units connected to the MB. The MBIF-CR
includes the same functionality as the MBIF-C, in addition to which it also provides the
repeater function. The MBIF-C/-CR plug-in unit also has built-in hardware and bus
diagnostics.
g Note: In this documentation, MBIF-C/-CR is used to refer to both MBIF-C and MBIF-
CR, as the basic functionality is the same in both plug-in units. MBIF-CR is used only
when repeater-specific information is presented.
When the CPU of the microcomputer unit writes a message to the MBIF-C/-CR, the
compressing logic of the plug-in unit compresses it simultaneously and moves it to the
transmit-FIFO. The transmit command reserves the Message Bus (MB) and sends the
message to the MB without any control from the CPU. The MBIF-C/-CR of the receiving
microcomputer unit receives and uncompresss the message under the control of its own
logic and gives an interrupt to the CPU for the incoming message. The MBIF-C/-CR can
receive messages at four different addresses, two of which are group addresses and can
be chosen as desired with the software.
The MBIF-C/-CR plug-in unit is an interface between the microcomputer and the
message bus segment. It is used in M98 mechanics, having the same basic functions as
the MBIF-B plug-in unit. The repeater function of the MBIF-CR connects the message
bus (MB) segments to the repeater bus (RB). The repeater bus connects MB segments
to form the full length message bus. The task of the MBIF-CR repeater function is to split
the message bus into several blocks so that signal quality is improved and higher
transmission rates can be used.
Operating environment of MBIF-C/-CR
The MBIF-C/-CR is connected to the CPU of the microcomputer unit via the PCI bus and
to the MBIF-C/-CR plug-in units of the other microcomputer units in the exchange via the
MB. The MBIF-C/-CR plug-in units receive their basic timing signals from the Clock
Equipment (CLO).
The figure below illustrates the operating environment of the MBIF-C/-CR plug-in unit in
the microcomputer units.
REPEATERFUNCTION
BLOCK(RPB)ONLY IN
MESSAGEBUS(MB)
MBIF-CR
OPERATION CGS
AND _INT0
MAINTENANCE ALARMTEST
MBLOAD
UNIT (OMU)
MBAL MBIF-C(R) MSGCOUNT CPU
CLOCK _INT1
EQUIPMENT 8M0,8K0
(CLS) 8M1,8K1
PCIBUS
DN0128737
The figure below shows the connection principle of the microcomputer units to the
duplicated Message Bus (MB) whose total length is approximately 150 m. Both MBs are
normally in use (active) simultaneously. One MB is chosen with the software as the
active bus along which all call control messages are sent to make sure that their order is
retained at the receive end. The other bus can be used to send e.g. various loading
messages. If either MB becomes faulty, the entire load is transferred to the intact MB. No
hardware-based changeover signal is associated with the selection of the MB in the
MBIF-C/-CR plug-in units.
Figure 2 Segmented Message Bus. An MBRP and one MBIF-C in each rack has
been integrated to MBIF-CR.
REPEATERBUS
M MBRP M M
MBIF-CR MBIF-CR
B MBIF-C B B
MBIF-CR
MBIF-C S S MBIF-C S
E MBIF-C E E
MBIF-C G G MBIF-C G
MBIF-C
The following figure illustrates the implementation of a message and repeater bus in a
single segment.
Figure 3 Implementation principle of Message Bus segment and Repeater Bus for a
single segment.
RearView
RepeaterBus0 RepeaterBus0
RepeaterBus1 RepeaterBus1
1A-1L
1N-1X
2A-2L
MB0
MB1
DN03310905
3 Structure of MBIF-C/-CR
PACKED
MAE
MAC
PCIBUS
INTERFACE D31:0 RECEIVER
_INT1 MD15...MD0,MDP
MWR
MESSAGEBUS(MB)
MWR
PCIBUS
MD15...MD0,MDP
TRANS-
MITTER MAC
MAE
_INT0 CONTROL PACKED
LOGIC
MRE,MRE2
ARBITER MSY
8M TIMING
8K
LOGIC
CGS
DN0128749
Generating of timing
The timing of the MBIF-C/-CR plug-in unit is generated by multiplying the 8.192 MHz
basic timing signal by four in a phase lock circuit. The resulting 32 MHz clock signal is
divided to obtain the arbitration frequency (2.048 MHz) and the transmission frequencies
chosen with settings.
The timing of all MBIF-C/-CR plug-in units in the exchange is synchronized to the same
phase with the 8 kHz and 8 MHz basic timing signal. The reserving and releasing of the
Message Bus, the incrementation and synchronization of the arbitration counters and the
entire sending and receiving of the command message take place fully synchronously in
all MBIF plug-in units of the exchange under the sole control of the timing signals with
the same nominal phase formed from the basic timing signals.
The writing of the messages into the receive-FIFO is asynchronous directly under the
control of the write signal sent at the same time as the transmit data.
The basic timing signals are supervised by examining the state of the carry output of a
four-bit synchronous counter incremented with the 8.192 MHz basic timing signal at 125
µs intervals. If the carry output of the counter is not in state 1 at the rising edge of the 8
kHz basic timing signal, or if the frequency of the 8 kHz signal is too low, a clock alarm is
activated.
Message Bus receiver
At the hardware level, the Message Bus receiver of the MBIF-C/-CR uncompresses
automatically messages compressed by the transmitter and saves the uncompressed
message into the FIFO from which the CPU interface can read the messages. The
receiver consists of the following parts:
• Message bus
• Control logic of the receiver
• Receive-FIFO
• Control and uncompressing logic
• Read-FIFOs
• (PCI interface).
Read
FIFO
0
MB
interface
Decompression
andcontrollogic
Receiver Read
Receive
controland FIFO
FIFO
paritylogic CRCcheckand 1
readlogic
PCI
interface
The control logic recognizes the addresses of the unit in the message (unit and group
addresses) and the type of the message (normal message or a command). In addition,
the control logic does the parity count. If a parity-error is detected in a word, receiving is
discontinued at once and the remaining part of the message is not written in the receive-
FIFO. A faulty message is recognized by its length: it is not long enough. If the control
logic recognizes the address designated to the unit, it performs the following:
1. The uncompressing algorithm waits in idle state until it notices that the receive-FIFO
is not empty, and it has received a permission from the control logic to start
uncompressing the message.
2. The uncompressing algorithm reads a word from the receive-FIFO and checks from
the PACKED signal, whether it is compressed. At the same time, the counter of the
uncompressing logic is reset. Had the word not been compressed, the
uncompressing logic writes the word into a read-FIFO number 0 or 1, pointed by the
control logic, and step 2 is repeated. Had the word been compressed, the
uncompressing algorithm moves to step 3.
3. The uncompressing logic writes the compressed word into read-FIFO 0 or 1 pointed
by the control logic, and increments its counter (by one). The counter is then
compared to the number of compressed words received from the upper byte of the
compressed word. If the values were equal, the logics return to step 2. If the values
were different, step 3 will be repeated.
The Control logic controls the writing of messages into the read-FIFOs assisted by
uncompressing logic.
After reseting, the control logic remains idle waiting for a message to arrive into the
receive-FIFO. When it notices that the receive-FIFO is not empty anymore, it gives the
uncompressing logic a command to start uncompressing the message. The first
message is always written into read-FIFO 0. While the uncompressing logic reads
receive-FIFO, the control logic looks for the start indication bit of a new message. When
the control logic notices the beginning of a new message and the uncompressing logic
has completed uncompressing of the previous word, the control logic starts writing into
read- FIFO 1 and, at the same time, gives an interrupt concerning the message waiting
in read-FIFO 0. After read-FIFO 1 has been written into, the following message will be
written into read-FIFO 0.
If receive-FIFO empties while receiving is in progress, the uncompressing logic will
remain waiting for the following word, and continues then in normal manner when there
is data in the FIFO again. If a receive-FIFO becomes empty and receiving is not in
progress anymore, the uncompressing logic will be reset, and the following message will
be written into the other FIFO.
The Control logic prevents the uncompressing of the messages when neither of the
read-FIFOs is not empty and the start indication bit following message is noticed. After
this, the uncompressing of the messages can start again when either of the FIFOs is
empty.
When a read-FIFO has been read empty, the control logic allows the writing of a new
message into it. In such case, the read-FIFO cannot be read again before the other FIFO
has been read first. This is to prevent the corruption of the message being written (see
Receiving of messages in Operating principles of MBIF-C/-CR).
The MBIF-C/-CR gives an interrupt only if the whole message has been received into
either read-FIFO, and an interrupt is given for each message. A new receiving interrupt
is given whenever the previous interrupt has been acknowledged and receiving interrupt
is allowed, and there is data in both read-FIFOs or there is data in the other read-FIFO
and the receive-FIFO is empty and receiving is not in progress.
The uncompressing and control logic writes a word (16 bits) into the read-FIFO at every
rising edge of the 66 MHz clock signal when there is data in the receive-FIFO and/or
uncompressing logic has not uncompressed the last word.
Read-FIFOs contain thus only one (uncompressed) message at a time. Therefore, their
size is the same as that of the write-FIFO, in other words, 32 x 2k. To optimize the PCI
interface, the read-FIFOs are 32-bit wide. Because they are written at the width f a word,
it must be possible to control the writing into them word for word. This can be achieved
with a signal which activates the writing signal of the upper and lower word in turns while
writing is in progress.
CRC-sum is calculated for every message read from read-FIFO. The CRC counter is
initialized by a read command of the first word. CRC-sum is not written to read-FIFO with
message, it is kept in register inside FPGA. CRC-error sets status bit but it has no other
impact on hardware. Checking the CRC-error (status bit) requires software support.
The uncompressing logic and read-FIFOs are reset with the same signal as the receive-
FIFO.
Control- and uncompressing logics are implemented by the FPGA circuit.
Message Bus transmitter
On hardware level, the Message Bus transmitter compresses messages automatically.
The Message Bus transmitter consists of the following (see also the figure below):
• (PCI interface)
• Write-FIFO
• Compressing logic and parity count
• Transmit-FIFO
• Interface circuits of the Message Bus
• Transmit logic and CRC-check.
PCI MB
interface interface
WriteFIFO, Transmit
Compression Transmit and
logic FIFO parity
andCRC logic
calculator
The master CPU writes messages via the PCI interface into write-FIFO. The write-FIFO
is integrated within the FPGA circuit and is three double words deep.
The CRC logic calculates a 16-bit CRC checksum of the message being written. A 16-bit
checksum is calculated from a 32-bit data by performing a XOR-operation to the bits next
to each other and by calculating the needed checksum from the resulting 16-bit word.
Transmit/Send command returns the CRC counter into initial state.
Compressing logic reads the 32-bit write-FIFO. It calculates parity (over 16 bits) from the
compressed data and saves the compressed part of the message and parity bit into
transmit-FIFO. The compressing algorithm is as follows:
1. The compressing algorithm remains in idle state until it notices that the write-FIFO is
not empty anymore. It moves to step 2.
2. Read the double word (32 bits) from the FIFO.
3. Compare bytes within the word. If the bytes are not equal, write two words (2 x 16-bit
addressing) with their parities into transmit-FIFO and return to step 1. Store also the
info that the word is not compressed (PACKED-signal) into the transmit-FIFO. If, on
the other hand, the bytes are equal, proceed to step 4.
4. Initialize value 2 into compressing counter.
5. Read the following double word and compare the bytes to the bytes read in step 3. If
the bytes are not the same or if the value of the compressing counter is FFH, write
the compressed part into the 16-bit transmit-FIFO in the following manner: save the
recurrent byte into the lower byte and the value of the compressing counter (how
many words the byte concerned recurs) into the upper counter together with parity
and compressing information. Return to step 3 and use as data the word you have
read last. If the bytes were the same, proceed to step 6.
6. Increment the compressing counter by one and return to step 5.
When the message is read from the transmit-FIFO the parity is calculated over 17 bits
over data and compressing information signal so that the receiver will be able to identify
faulty compressing information.
The compressing algorithm is implemented with a programmable logic so that the
compressing speed is 30 ns (33 MHz clock frequency) per double word. In other words,
the compressing logic can also function at the full speed of the PCI bus. Because the
transmit-FIFO is a 16-bit device, it must be written with a 66 MHz clock signal. This
signal is formed by the phase-locked loop from the backend clock signal of the PCI
connection.
The compressing algorithm is activated whenever the write-FIFO is not empty. If the
write-FIFO is emptied while the compressing logic as at step 5 and the part of the
message that has been compressed but has not been written into the transmit-FIFO, the
logic remains in idle state waiting for new data in the write-FIFO. Thus, the transmit logic
has to give a transmit command to the compressing logic, after which the compressing
logic writes, if need be, the last word into the transmit-FIFO and returns to initial state. In
addition, the compressing logic always finishes off by reading the CRC-sum and saving it
with its parity to the transmit-FIFO unless the message to be transmitted is a command.
The transmit logic decodes the send command from the master CPU and activates
transmitting as soon as the write-FIFO is empty (compressing is completed).
Command messages are handled with the compressing logic in a normal way because
they cannot be compressed.
The transmit-FIFO's size is 8 kilobytes. Its memory is 18 x 4k (including data,
compressing information and parity).
Transmit logic, CRC-calculation, compressing logic, parity count and write-FIFO are
implemented with a FPGA circuit. Write-FIFO (FPGA's internal) is written by using a
signal coming from the 33 MHz PCI bus. The transmit-FIFO is written by using a 66 MHz
clock signal that has been formed by the phase-locked loop from the 33 MHz clock
signal mentioned above.
The CRC-check requires support from the MBDRIV driver program.
Message Bus arbiter
The arbitration algorithm developed for the MBIF-C/-CR improves the performance of the
plug-in unit from that of the MBIF-A when the load factor of the Message Bus (MB) is
relatively low and one or more microcomputer units are waiting for transmission turn.
With the MBIF-C/-CR, the MB arbiter is integrated into the FPGA circuit and only the
2.048 MHz arbitration frequency is available. This means that the arbitration cycle lasts
488 ns.
The MBIF-C/-CR plug-in units of the active side of the duplicated microcomputer unit
which always belongs to the standard equipping of the exchange feed the
synchronization signal MSY to themselves and to all other microcomputer units on the
MB. The MSY signal loads the same initial value to the counters of all arbiters at the
beginning of every arbitration round.
The principle of operation of the new arbitration system is as follows:
• The Message Bus (MB) is divided into eight (16, if four bits is used) groups
(message bus segments) according to the addresses set by the MBRP. Each group
includes always microcomputer units included in the same cabinet (segment). The
first group includes the microcomputer units of the first cabinet (cabinet address 0),
the second group the microcomputers of the second cabinet etc. The arbitration
counter typically has only three bits, that is eight steps. If none of the microcomputer
units waits for transmission turn, one entire arbitration round is 8 × 488 ns = 3.904
µs. If a configuration has more than eight cabinets 4 bit address can be taken into
use with the _CA3EN signal (through the back plane connector).
• When the arbiter is incremented to a group, every microcomputer unit in the group
waiting for transmission turn stops the arbitration with the MRE2 signal. In addition,
each unit in the group controls the bus data bit that corresponds to the D7, D6, D1
AND D0 (UA) signals of their MB address for 180 ns. For example, address 40H
controls MB data bit MD4, and address 04H controls data bit MD0. Each unit
inspects which data bits are active, after which the microcomputer units that control
the lowest data bit have the right to set the data bits corresponding to the four middle
address bits (D5…D2) as active data bit MD1 for address 04H). The unit waiting for
request sets the data bit it controls into state 1 and the unit not waiting for request
sets the bit into state 0. MD signals need only be controlled for 180 ns, since all units
possibly requesting transmission are in the same cabinet and all units can identify
the reservation signal in this time.
• The microcomputer unit that has set the lowest data bit active starts transmitting by
activating the MRE signal. In addition to the unit transmitting a message, the other
units in the group waiting for transmission also hold the MRE2 signal active.
• Once the microcomputer unit has transmitted the entire message, it deactivates MRE
and MRE2 signals and sets the data bits into high impedance state. The deactivation
of the MRE signal is an indication for the other units in the group waiting for their turn
that the next unit in the queue can reserve the bus. This operation takes place within
the next two arbitration cycles. During the first arbitration cycle, all units notice the
bus being free. During the second cycle, they run the data line corresponding to their
address into state they wish (reserve request, that is state 1 or tristate). After these
two cycles, the unit that has activated the lowermost data line starts transmitting.
When even the last unit in the group has sent a message, it deactivates both MRE
and MRE2 signal, after which arbitration counters step forward as usual. In other
words, the arbiter steps whenever the MRE2 signal is passive.
• Only the 2,048 MHz arbitration frequency is available for the MBIF-C/-CR, and thus
the arbitration cycle lasts 488 ns. The termination resistors keep the signals in
deactive state. During the request cycle the timing of the MBIF-C/-CR follows the
values listed in the table below, provided that the microcomputer unit wishes to
reserve the message bus and the arbitration counter receives cabinet addresses. If
the unit does not wish to reserve the message bus, it naturally does not control any
signal.
Table 1 Timing of unit requesting reservation.
274,5 MDx data bit selected with MB addresses D7, D6, D1 and D0 (UA) set as active. The
microcomputer unit controls the data bit at this stage only if it notices that it has set a
data bit of higher priority status (MD0 highest, MD11 lowest) at the start of the
request cycle.
457,5 The winner activates MRE, MRE2 is read, and the arbitration logic stops the
artbitration counter.
518,5 The first word of the message is transmitted to the message bus (30,5 ns after the
start of the arbitration cycle).
• Since the arbitration counter compares cabinet addresses instead of unit addresses,
restrictions for address use in the MBIF-B no longer apply. The equipment enables
any address for a unit within the scope of 00H…BFH. The addresses C0H…FFH are
not available as unit addresses, since these addresses would drive the data bits
MD15…12 at the start of the request cycle and would thus collide with the cabinet
address driven by the MBRP. However, these addresses are available as group
addresses. Cabinet addresses can be read from the MBIF-C status register. The
reservation is directed at the start by with bits D7 and D6, in order to avoid the
aforementioned collision with MBRP (D7 and D6 can not have the status 1 for any
unit). D1 and D0 distribute effectively units in the same cabinet, as one and the same
cabinet often has several sequential addresses.
• If a microcomputer unit requires transmitting turn when the MRE2-signal is active, it
cannot take the turn before the MRE2-signal has been released and the arbiter has
been incremented, even if the group the unit belongs to had a transmission turn. This
is to secure all transmitting groups transmitting turn in due time.
• In estimating the power handling capacity of the terminator resistor. It has to be taken
into account that more than one unit at a time may drive the MRE2 signal.
The MBIF-CR plug-in unit contains an asynchronic repeater function, the task of which is
to repeat message bus signals that it receives to the opposite segment. In other words, it
repeats messages received from the Repeater Bus (RB) to the Message Bus (MB), and
messages from the RB to the MB. The task of the repeater is to ensure the quality of
signals on a long segment and high transfer frequency.
In addition to bus interfaces, the repeater also includes a control logic, which sets the
interfaces into the desired states in the following way:
MD15...MD0 MD15...MD0
MBS_MDP RB_MDP
MBS_MRE RB_MRE
MBS_MRE2 RB_MRE2
MESSAGEBUSSEGMENT MBIF-C/-CR REPEATERBUS
MBS_MWR RB_MWR
MBS_MAC RB_MAC
MBS_MAE RB_MAE
MBS_MSY RB_MSY
MBS_PACKED RB_PACKED
DN03310929
The connection direction is recognized from the arbiter counter and cabinet address
during the arbitration cycle. When the arbiter counter corresponds to its own cabinet
address, the transmission direction is from the segment to the Repeater Bus (RB). If the
arbiter counter does not correspond to its own cabinet address, it is possibly the turn of
some other segment to transmit and the connection direction is from the RB to the
segment.
Due to the fault monitor of segments, the signal connection directions are graded during
the arbitration cycle.
Due to transit time delays, address signals driven to the MB segment and fault control
signals all transmit and receive segments are connected separately to the FPGA. The
Data, MWR, PACKED and MDP signals are controlled with the internal logic of the FPGA
depending on the situation at hand. The entry points of the signals are always active,
and only the output ports are controlled when necessary. The repeater function also has
to recognize the message type (the MD8 state of the data bit at the start of the message
transmission).
When the repeater notices a command message it has to change its direction in the
middle of transmitting the message so that the status and receipt bits can be sent from
the receiver to its sender. In other words, the repeater connected to the MB of the sender
of the command message activates its RB connection whereas the repeater of the MB of
the receiver closes its MB connection for data signals and activates its RB connection.
The segment that has received the command message is recognized by the MAC receipt
signal received from it. Sending the command signal takes three arbitration cycles, and
the status and receipt bits are transmitted during the last arbitration cycle. In this case
the change of direction of the repeater has to be timed in the following manner:
• To prevent simultaneous running of the RB, the deactivation of the MRE signal has to
set RB signals to a state of high impedance.
The above is applicable also when the transmitter and receiver of a command message
are both in the same segment.
The data signals MD12…15 function in the direction of the MB segment differently from
other data signals, as the repeater sets a cabinet address for these when the bus is in
deactivated state. Thus, MD12…15 are active towards the segment when the bus is in
deactive state while the cabinet address delivered from the back connector is directed to
them. When the bus is activated from either direction, cabinet addresses are set first in
state 1 and, approximately 50 ns later, to a state of high impedance. Data bits remain in
state of high impedance if the reservation has come from the MB, and they are
reactivated if the reservation comes from the RB. A cabinet address can be directed to
the segment whenever the MRE2 signal is in deactive state.
Fault control of the MB segment is implemented during the arbitration cycle. Each MBIF-
CR whose arbiter does not correspond to its own cabinet address sets the data bits
MD0…MD11 into state 1 at the start of an arbitration cycle. After 457,5 ns from the
beginning of the arbitration cycle each MBIF-C in the cabinet reads the data bits. Data
bits MD0…MD11 should be '1' for all MBIF-Cs. If this is not the case, the MBIF-C that
noticed the error sets the fault control status register bit D2 and implements an interrupt.
The precise functioning and timing are displayed in table 1 below. Interrupt is edge
sensitive, and a new interrupt occurs only once the previous one has been
acknowledged and the status register bit is off.
Time (ns)
457,5 If MRE2 is active, RD0…RD15, MDP, PACKED and MWR are connected from the RB
to the MB.
The units read the bites MD0...MD11.
Data signals are set into deactive state whenever it is noticed that the MRE signal
(MRE_MB or MRE_RB signals) is in deactive state. Data signals are run for 80…150 ns,
after which they are directed into state of high impedance.
The sender of the synchronization signal MSY is always the same unit that synchronizes
the bus. This unit remains constant practically all the time. The unit can change only due
to a change of sides in the synchronizing unit, but the back-up units must belong to the
same MB segment. In this case the MSY signal scan be connected directly from the
receiver of the MB connector to the RB transmitter, and vice versa. The transmit ports
are closed in initial state. The transmitter of the RB is opened with rising edge of the
MSY signal coming from the MB, and vice versa. Activation of the other transmitter is
prevented simultaneously with opening the transmitter, that is, the MSY signal coming
from the MB prevents the activation of the MB transmitter. Once the transmitter has been
activated or closed, it remains in the same state until the unit is zeroed with the reset
signal. In this case the transmitters are reset in both directions in tristate, waiting for the
new synchronization signal. The entry ports of the MSY signal are always active.
The entry ports of the MAC and MAE signals are always active. As with MRE signals, the
signal data of output ports are constantly active (in state 0). The output ports of the MAC
and MAE signals are controlled with the MRE signal and the received MAC and MAE
signals. The output ports can be activated only in the direction from which the MRE
signals have been received. In other words, the MAC and MAE signals can only be
active towards the RB if the MRE signal has been received from the RB direction, and
towards the MB if the MRE signal has been received from the MB direction. This is
because only the sender of a message can read its receipt signals. However, output
ports are opened only if a MAC or MAE signal is received from the opposite segment
than the MRE signal. Output ports controlling the signals are kept active for as long as
the MAC and MAE signals are active. In other words, the deactivation of the MRE signal
does not close the output ports.
The repeater is always functional, even if the MBIF-CR is separated from the bus.
timing signals. If the arbitration counter in one MBIF-C/-CR has been incorrectly
synchronized, its transmission turn may overlap with the transmission turn of another
MBIF-C/-CR. This will lead to an error condition only when both MBIF-C/-CRs are
attempting to send a message at the same time.
Sending of messages
The 16–bit words of the message to be sent, the PACKED information is written into the
transmit-FIFO with a memory addressing, which prevents the activation of the interrupt
signal during the writing into the FIFO.
When the message in the transmit-FIFO is written to MB, the transmit logic releases
MRE and drives receive signal to zero for a period of three clock cycles.
Sending of command messages
When sending command messages, the sender deactivates the line transmitter circuits
corresponding to the low byte after the address word has been sent. After one arbitration
cycle, the recipient of the command message activates the line transmitter circuits
corresponding to the low byte for the transmission of the status and acknowledgement
byte. The received status and acknowledgement byte is written at the end of the
command message cycle of FPGA. The MB is reserved for command messages for
three arbitration cycles.
Receiving of messages
The low byte of the received address word is compared with the wired address, with the
programmed group addresses and with the selected 7EH or 7FH general group address
by the FPGA circuit IC26. If one of the four comparators identifies the address, the
receiving of the message begins. The parity of all words to be received are checked.
All the write signals of receive-FIFO circuits are formed by the FPGA circuit from the
MWR signal of the Message Bus.
If a parity error is detected in connection with any byte in the rest of the message, writing
to receive-FIFO is prevented. If acknowledgements are enabled, the enable signal
controlling the MAC signal of the MB is deactivated and the enable signal controlling the
MAE signal of the MB is activated.
Receiving of command messages
All received command messages are decoded by decoding logic which identifies the
different command message types. For a reset command message, output pin 5 of the
monostable multivibrator controlling the _RSTOUT signal is triggered into state 1 for a
duration of approx. 400 µs.
Generating clock alarm
Supervision of basic clock signals is implemented in FPGA. 10 bit counter circuit is
clocked by inverted 8 MHz clock. The counter is always loaded with value 0 when 8 KHz
signal is in state 1. If counter is not 3FF hex when 8 kHz pulse arives or counter reaches
3FF hex and there is no 8 kHz pulse a alarm is given.
When the basic timing signals contain imterference, _ICLOCKALARM signal is set to 0
for approx. 50 ms. When zero, _ICLOCKALARM sets MBAL signal also to zero. MBAL
signal can be made active also by activating the ALTST signal into state 1.
If the alarm is continuous , the _ICLOCKALARM signal will remain continuously in state
0.
The operation of all line transceiver circuits of the MB interface is disabled for 0.5 us if
clock interference occurs.
PCI bus interface
The MBIF-C/-CR is connected to the master computer via a PCI bus. The connection
uses DMA burst transmission between the central memory of the CPU and the receive-
FIFO of the MBIF-C/-CR. Other operations to be applied to the MBIF-C/-CRs (writing
commands and reading status registers) are single addressings. The CPU takes care of
initialising the DMA controller on the MBIF-C/-CR by writing to it the transmission start
address in the central memory, the length of the block to be transmitted and the
transmission direction (from the memory to the transmit-FIFO or from the read-FIFO to
the memory). Subsequently, the CPU starts the DMA controller, which transmits
independently the data as a burst. During the DMA transmission, the MBIF-C/-CR
operates as PCI bus master. When the DMA controller has performed the transmission,
it gives an interrupt to the CPU, after which it can be initialised to perform a new
transmission. The DMA transmission is always performed until the end, even if a receive
error is detected while reading the receive-FIFO. The error is identified from error bits
D17, D21 and D22 of the receive status register.
_CA3EN
SETTINGOF UA7...UA0 REPEATER
MESSAGEBUS CA3...CA0 _REPEN FUNCTION
ANDCABINET "0" "0" ENABLE
ADDRESSES "1"
_INT0
SELECTIONOFMB'S CGS
_INT1
SYNCHRONIZATION SYNC0
MSGCOUNT CPU
SYNC1 MBLOAD
_RSTOUT
TDI
TDO
JTAG
TCK
INTERFACE
TMS
DN03493938 _TRST
AL OPERATION AND
CLOCKEQUIPMENT 8M ALTEST MAINTENANCE
(CLO) 8K
_CA3EN
SETTINGOF UA7...UA0 REPEATER
MESSAGEBUS CA3...CA0 _REPEN FUNCTION
ANDCABINET "0" "0" ENABLE
ADDRESSES "1"
_INT0
SELECTIONOFMB'S CGS
_INT1
SYNCHRONIZATION SYNC0
MSGCOUNT CPU
SYNC1 MBLOAD
_RSTOUT
TDI
TDO
JTAG
TCK
INTERFACE
TMS
DN03493926 _TRST
Signal Description
Signal Description
GND Ground
Signal Description
_RSTOUT Activation signal of the initial reset request coming from the
Message Bus as command message
Signal Description
The PCI bus interface signals of the MBIF-C/-CR meets the PCI specification. The
interface signals with the Clock Equipment use asymmetrical current-controlled
interfaces, whereas MB signals comply with the MLVDS stantard. All other interfaces are
TTL or CMOS level. The figure below shows the principle of an asymmetrical current-
controlled interface.
TRM TRM
MLVDS-B
TTL TTL
The MBIF-C/-CR receives the necessary external basic timing signals 8M and 8K from
the Clock Equipment. The figure below shows the timing of the signals.
8M
8k
DN9987087
A pull-up resistor (10k) is added to the A-signal of the input counters to prevent noise in
case of blockout.
D13 In state 0, if the clock alarm output test signal ALARMTEST is active
(state 1).
10H or 18H D31…D0 Reading of interrupt and receive status register and
acknowledgement of receive errors. The reading of the register from
address 18H does not acknowledge receive errors.
10H (18H) D9 In state 0, if the interrupts of the MBIF-C/-CR have been masked.
D17 In state 1, if the identification bit of the first word of a message has
been erroneusly in 0 state in the read command
In the MBIF-C/-CR this means that the read command of the first
word in the message (change-over command of the read-FIFO) has
been given before the FIFO currently being read is emptied.
D20 In state 1, if an attempt has been made to write into a full receive-
FIFO.
0CH D23...D0 Reading of the status register on the transmit side and reading of the
received status and acknowledgement byte.
D7 Always in state 1.
D9 Always in state 1.
D17 In state 1, if the recipient has detected a parity error on the MB.
D28...31 Cabinet address bytes CA0...3. The MBRP runs these bytes on
Message Bus signals when the MD12...15 segments are free.
D15...D0 Contents of the Message Bus load counter. The counter counts
MBLOAD signal pulses. The counter turns around and does not have
to be initialised.
D31…D16 Contents of the Message Bus message counter. The counter counts
MSGCOUNT signal pulses. The counter turns around and does not
have to be initialised.
D1 In state 0, if error has been found in the line circuit of the repeater
(only in MBIF-CR).
Always in state 1 in MBIF-C.
Write operations
A message is written into the write-FIFO. The write-FIFO is in the memory space and the
writing takes place as memory addressing. Every write is decoded as a double word
wide addressing into the write-FIFO. Compressing logic is activated whenever there is
data in the write-FIFO, and it transfers/forwards the compressed message into the
transmit-FIFO.
The total length of the message must be divisible by four. Also the receive-FIFO must
always be read in double words (the MBIF-C/-CR ignores byte selection signals of the
bus). For error conditions, an additional word (content is irrelevant) must therefore be
written into normal messages (not into command messages) after the address and
command word and only after this the message header. The reason for this is that
reading a corrupted message in the FIFO the first complete double word of the next
intact message is lost (the reading will be interrupted when the beginning of a new
message is detected too early). The part of the message getting lost must not contain
header data (e.g. length of the message).
The table below describes the structure of the address and command word to be written.
Signal Description
D7...D0 MB address
Signal Description
The table below describes the selection of the command message type (chosen with
address and command word bits D14...D9)
Signal Description
D9 = 0 Separation from MB
D9 = 1 Connection of MBIF-C/-CR to MB
D10 = 1 Restart; the receiver does not receive an interrupt from the
command message.
X = MB address
The transmit-FIFO of the MBIF-C/-CR plug-in unit is in the memory space and writing
into the memory space is done by memory addressing. Each write operation of the
memory area into the address area of the MBIF-C/-CR is decoded as double word
addressing into the transmit-FIFO. Other write commands of the MBIF-C/-CR are in the
I/O space. The MBIF-C/-CR has all the same I/O-write operations as the MBIF-A. The
reset command of the transmit-FIFO resets the write-FIFO as well as the compressing
logic. Respectively, the resetting of the receive-FIFO resets also the read-FIFOs and
release logic. The enabling of receive interrupt is also in use with I/O-write (OUT base
address + 3CH).
The table below describes the write operations of the MBIF-C/-CR plug-in units.
D0 Not used
MBLOADEN
7E_7FSEL
ALLMSGRECEN
• The status register on the transmit side is examined to find out if everything is ready.
• The message is written into the transmit-FIFO. The total length of the message to be
transmitted must be N x 4 bytes.The transmission must be carried out in double word
mode, because the MBIF-C/-CR does not check the byte selection signals, but
presumes the addressings being always in double word mode.
• The message transmission is started with the _TRANSMIT command. The message
is written into the transmit-FIFO (previous phase) and always sent completely, also in
error conditions. The recipient detects the corrupted message and rejects it.
The transmit-FIFO must be reset after every failed message transmission. It is never
allowed to write a new message into the transmit-FIFO until the transmit-FIFO is empty.
If the message transmission fails, an interrupt is given to the CPU. The cause of the
interrupt is found out by reading the interrupt register. A transmission error interrupt must
be acknowledged with the _TRAERRQUIT command which also resets the transmit-
FIFO.
When writing into the transmit-FIFO, the parity bit of the 16-bit word is not written at the
same time. The parity bit is calculated and added to each data word during transmission
Messages can be sent only when the MBIF-C/-CR is connected to the bus. If a message
transmission is started when the MBIF-C/-CR is separated from the bus (the
_SEPARATED signal in state 0), the transmission will immediately end with a
transmission error interrupt, and the Message Bus is not reserved at all.
All line transmitters of the MBIF-C/-CR are separated from the MB for the duration of a
clock disturbance. The clock disturbance does not in any way affect the actual
_SEPARATED signal which handles the separation. When the clock disturbance ends,
the transmitters remain separated about 0.5 ms longer. Any transmission attempts made
during this time will immediately end with a transmission error interrupt. The receiving of
normal messages and their writing into the FIFO is not disabled during the clock
disturbance. The receiving of command messages is completely disabled during a clock
disturbance.
Receiving of messages
With the MBIF-C/-CR, the reading of the first word is separated from the reading of the
remaining part of the message with one bit address decoding. MBIF-C/-CR accesses 1
MB from the PCI bus, and therefore the MSB it uses is A19. The MBDRIV reads the first
word of the message from address 80000H (A19 is 1) and the remaining part of the
message (with DMA) from address 00000H (A19 is 0).
The principle of operation of the receive program with regard to faultless messages is
presented below:
Operation Bits
Reading the first word from address 80000H. With this 32 bits
command, the receive interrupt is prevented and
acknowledged. Also, with this command, the control
logic changes the read-FIFO to be addressed.
Checking the first word. The MBDRIV copies from the 2 × 32 bits
bus header of the message the information needed
for initializing the DMA controller (the length and the
type of the message) for the first word. The DMA
controller is then initialized and activated with a
memory addressing.
If a parity error has been detected in the reading of the receive-FIFO, the read message
is corrupted, but the next message can be read normally.
Sending of command messages
When writing command messages, bit D8 in the recipient's address and command
message must be set into state 1. The command messages are sent with the
_CMDMSGSEND command.
A command message can be received only at the wired address of the unit. The
command messages affecting the state of the microcomputer unit must also have a
certain exact format for them to be accepted.
The error acknowledgement signal MAE is never activated when receiving command
messages, because if a parity error occurs in connection with the transmission, the
recipient will not identify its address.
The command message is never compressed because the bytes of the doubleword are
not identical. Thus the compressing algorithm does not have to individually recognize the
command message and uncompress it at the receiving end.
No CRC sum is to be calculated of the command message.
Databits are to be directed into 1-state also after sending a command message.
Interrupts
The MBIF-C/-CR units use two interrupt outputs (_INT0 and _INT1), which are
connected to the CPU interrupt inputs _INT0.._INT3. The interrupt signals are run with
the LVT circuit. The circuit is closed, when the output is actively run up (to 3 V).
The interrupt signal _INT1 is used to give the DMAREADY interrupt with which the DMA
controller announces to the CPU that it has dispatched the transmission initialised for it
and is ready to start a new transmission.
Five interrupts have been connected to the interrupt signal _INT0:
When any one of the above interrupts is on, the outgoing interrupt signal is activated into
state 0. Every interrupt can be acknowledged separately.
A new receive interrupt is always activated whenever the previous interrupt has been
acknowledged and receive interupt is enabled and there is data in both read-FIFOs, or
there is data in one read-FIFO and the receive-FIFO is empty and receiving is not in
progress. This means that a receive interrupt is given separately for each message. The
interrupt is edge sensitive. In this case, there is always a complete message ready in the
receive-FIFO when the interrupt is given.
When reading the first word, the receive interrupt is deactivated and a new receive
interrupt is prevented from passing through. The receive interrupt is enabled with a
separate command or by reading from the interrupt and receive register address 00H.
Receive errors can be acknowledged in connection with the reading of the interrupt and
receive register as it is done when reading a received message (see section ). In other
cases, the interrupt register must be read without acknowledging receive errors.
Bits of
the
receive
status
registe
r
D24 D26 D28 D16 D25 D27 D29 D18 Fill-up level
0 0 1 1 0 0 1 1 = 0%
1 0 1 1 0 0 1 1 0...6.25%
1 1 1 1 0 0 1 1 6.25...25%
1 1 0 1 0 0 1 1 25...43.75%
1 1 0 0 0 0 1 1 43.75...50%
1 1 0 0 1 0 1 1 50...56.25%
1 1 0 0 1 1 1 1 56.25...93.7%
1 1 0 0 1 1 0 1 93.7...100%
1 1 0 0 1 1 0 0 = 100%
approximately 120 kohm. The maximum number of units connected to the Message Bus
is 20 units for the Message Bus segments and 10 units for the repeater bus. The line
transceiver circuits do not either cause an adverse load on the bus even if the operating
voltages are missing in the MBIF-C/-CR so the failure of the power supply of any
microcomputer unit, or its intentional switching off, will not prevent the operation of the
MB.
The table below describes the use of the MB interface signals.
Signal Use
MDP The parity of the parity signal is even (e.g. with the 0000H data,
the state of the parity signal is 0).
MRE The message sender reserves the bus by activating the request
signal MRE into state 0 for the entire duration of the message
transmission.
MRE2 MRE2 is set active by all units that are in the same arbitration
group as the unit currently sending and that want to request the
bus.
MWR The transmission clock signal MWR sent at the same time as the
data to be transmitted controls the writing into the receive-FIFO.
The figures below illustrate the interface of the MBIF-C with the MB and the interfaces of
the MBIF-CR with the MB and RB. The MB is terminated at both ends with terminator
units.
MBRP _RX_EN
RX_DATA
TX_EN
TX_DATA
MESSAGEBUS
SEGMENT
MBIF-C
DN03403653
Figure 13 Interfaces of the MBIF-CR with the Message Bus and Repeater Bus.
_RX_EN
RX_DATA
MBRP TX_EN
TX_DATA
REPEATERBUS
_RX_EN
RX_DATA
TX_EN 100R
MBIF-C TX_DATA
MESSAGEBUS
SEGMENT
DN03403665
An interrupt is given to the microcomputer unit for a clock alarm generated by the MBIF-
C/-CR so that every microcomputer unit will be able to supervise the basic timing signals.
To supervise the operation of the cabling of the alarm signals and the alarm input of the
alarm interface unit, the unit handling the operation management can activate the clock
alarm signal MBAL by setting the ALARMTEST signal into active state (state 1). The
activation of the ALARMTEST signal does not turn on the LED indicator for clock alarm
or otherwise affect the operation of the MBIF-C/-CR. The inverted state of the
ALARMTEST signal (_ALARMTEST) can be read by the software from the status
register.
The operation of the entire clock alarm logic can also be tested by the software by using
the _MBALTEST command.
4 Operation of MBIF-C/-CR
Front panel
The front panel of the MBIF-C/-CR contains two LED indicators, an upper green one and
a lower red one. The red LED is illuminated when a clock alarm is active and the green
LED is illuminated when the plug-in unit is connected to the MB.
DN0283653
Front panel dimensions
The front panel dimensions of the MBIF-C/-CR are 262 x 20 mm.
Backplane connectors
The MBIF-C/-CR is connected to the motherboard of the cartridge via four five-row
connectors. The connector types are as follows:
+5 V 250 mA 1.25 W
-5 V 11 mA 55 mW
J5
J4
LD1
J3
LD3
W6
MBIF-
1 8
C/-CR Interchangeability 2 7
J1 3 6
W6 4 5
DN04108947
The W6 pin header is used for setting the interchangeability code of the plug-in unit,
when required, see the table below.
There are no standard nor alternative settings available on the plug-in unit.
The other jumper groups are used only for tests or other special purposes.
Interchangeability code settings (W6)
Interchangeability is coded with four bits using the W6 pin header. Interchangeability
codes are described in the table below.
D ON ON OFF OFF
F ON OFF ON OFF
G OFF ON ON OFF
H ON ON ON OFF
K ON OFF OFF ON
L OFF ON OFF ON
M ON ON OFF ON
N OFF OFF ON ON
P ON OFF ON ON
R OFF ON ON ON
Not defined ON ON ON ON
6 GND GND
7 GND GND
9 GND GND
Connector J4
A B C D E F G
3 GND MD 5 B MD 5 A GND MD 4 B MD 4 A
12 GND GND
14 GND GND
Connector J3
1 GND MDP B MDP A GND MWR B MWR A GND
2 GND GND 2N
4 GND GND 2P
5 GND GND
8 GND GND 2T
9 GND GND
14 GND GND
15 GND GND
16 GND GND
17 GND GND
18 GND GND
19 GND GND
Connector J1
1 GND 5V 3.3V 5V GND
12 GND GND
14 GND GND
23 GND 5V GND