FemtoClock® NG Octal Universal
Frequency Translator
8T49N285
Datasheet
Description
The 8T49N285 has a fractional-feedback PLL that can be used as a
jitter attenuator or frequency translator. It is equipped with six integer
and two fractional output dividers, allowing the generation of up to 8
different output frequencies, ranging from 8kHz to 1GHz. Three of
these frequencies are completely independent of each other and the
inputs. The other five are related frequencies. The eight outputs may
select among LVPECL, LVDS, HCSL or LVCMOS output levels.
This functionality makes it ideal to be used in any frequency
translation application, including 1G, 10G, 40G, and 100G
Synchronous Ethernet, OTN, and SONET/SDH, including ITU-T
G.709 (2009) FEC rates. The device may also behave as a frequency
synthesizer.
The 8T49N285 accepts up to two differential or single-ended input
clocks and a crystal input. The PLL can lock to either input clock, but
both input clocks must be related in frequency.
The device supports hitless reference switching between input
clocks. The device monitors both input clocks for Loss of Signal
(LOS). It generates an alarm when an input clock failure is detected.
Automatic and manual hitless reference switching options are
supported. LOS behavior can be set to support gapped or un-gapped
clocks.
The 8T49N285 supports holdover with an initial accuracy of ±50ppB
from the point where the loss of all applicable input reference(s) has
been detected. It maintains a historical average operating point that
may be returned to in holdover at a limited phase slope.
The device places no constraints on input to output frequency
conversion, supporting all FEC rates, including the new revision of
ITU-T Recommendation G.709 (2009), most with 0ppm conversion
error.
The PLL has a register-selectable loop bandwidth from 1.4Hz to
360Hz.
Each output supports individual phase delay settings to allow
output-output alignment.
The device supports Output Enable inputs and Lock, Holdover and
LOS status outputs.
2
The device is programmable through an I
2
C master capability to allow the register configuration to be read
I
from an external EEPROM.
C interface. It also supports
Typical Applications
• OTN or SONET / SDH equipment Line cards (up to OC-192, and
supporting FEC ratios)
• OTN de-mapping (Gapped Clock and DCO mode)
• Gigabit and Terabit IP switches / routers including support of
Synchronous Ethernet
• SyncE (G.8262) applications
• Wireless base station baseband
• Data communications
• 100G Ethernet
Features
• Supports SDH/SONET and Synchronous Ethernet clocks
including all FEC rate conversions
• <0.3ps RMS typical jitter (including spurs),12kHz to 20MHz
• Operating modes: locked to input signal, holdover and free-run
• Initial holdover accuracy of ±50ppb
• Accepts two LVPECL, LVDS, LVHSTL, HCSL or LVCMOS
input clocks
• Accepts frequencies ranging from 8kHz up to 875MHz
• Auto and manual input clock selection with hitless switching
• Clock input monitoring, including support for gapped clocks
• Phase-Slope Limiting and Fully Hitless Switching options to
control output phase transients
• Operates from a 10MHz to 40MHz fundamental-mode crystal
• Generates 8 LVPECL/LVDS/HCSL or 16 LVCMOS output clocks
• Output frequencies ranging from 8kHz up to 1.0GHz (diff)
• Output frequencies ranging from 8kHz to 250MHz (LVCMOS)
• Four General Purpose I/O pins with optional support for status &
control:
• Four Output Enable control inputs may be mapped to any of the
eight outputs
• Lock, Holdover & Loss-of-Signal status outputs
• Open-drain Interrupt pin
• Nine programmable PLL loop bandwidth settings from 1.4Hz to
360Hz.
• Optional Fast Lock function
• Programmable output phase delays in steps as small as 16ps
The 8T49N285 can be locked to either of the input clocks and
generate a wide range of synchronized output clocks.
It could be used for example in either the transmit or receive path of
Synchronous Ethernet equipment.
The 8T49N285 accepts up to two differential input clocks ranging
from 8kHz up to 875MHz. It generates up to 8 output clocks ranging
from 8kHz up to 1.0GHz.
The PLL path within the 8T49N285 supports three states: Lock,
Holdover and Free-run. Lock & holdover status may be monitored on
register bits and pins. The PLL also supports automatic and manual
hitless reference switching. In the locked state, the PLL locks to a
valid clock input and its output clocks have a frequency accuracy
equal to the frequency accuracy of the input clock. In the Holdover
state, the PLL will output a clock which is based on the selected
holdover behavior. The PLL within the 8T49N285 has an initial
holdover frequency offset of ±50ppb. In the Free-run state, the PLL
outputs a clock with the same frequency accuracy as the external
crystal.
Upon power up, the PLL will enter Free-run state, in this state it
generates output clocks with the same frequency accuracy as the
external crystal. The 8T49N285 continuously monitors each input for
activity (signal transitions).
In automatic reference switching, when an input clock has been
validated the PLL will transition to the locked state. If the selected
input clock fails and there are no other valid input clocks, the PLL will
quickly detect that and go into holdover. In the Holdover state, the
PLL will output a clock which is based on the selected holdover
behavior. If the selected input clock fails and another input clock is
available then the 8T49N285 will hitlessly switch to that input clock.
The reference switch can be either revertive or non-revertive.
The device supports conversion of any input frequencies to three
different, independent output frequencies on the Q0 and Q[2:3]
outputs. Additionally, a further five output frequencies may be
generated that are integer-related to the three independent
frequencies. These additional five frequencies are on the Q1 and
Q[4:7] outputs.
The 8T49N285 has a programmable loop bandwidth from 1.4Hz to
360Hz.
The device monitors all input clocks and generates an alarm when an
input clock failure is detected.
The device supports programmable individual output phase
adjustments in order to allow control of input to output phase
adjustments and output to output phase alignment.
The device is programmable through an I
autonomously read its register settings from an internal One-Time
Programmable (OTP) memory or an external serial I
Crystal Input
The crystal input on the 8T49N285 is capable of being driven by a
parallel-resonant, fundamental mode crystal with a frequency range
of 10MHz - 40MHz.
2
C and may also
2
C EEPROM.
8T49N285 Datasheet
The oscillator input also supports being driven by a single-ended
crystal oscillator or reference clock.
The initial holdover frequency offset is set by the device, but the long
term drift depends on the quality of the crystal or oscillator attached
to this port.
Bypass Path
For system test purposes, the PLL may be bypassed. When
PLL_BYP is asserted the CLK0 input reference will be presented to
the Q4 output dividers and the CLK1 Input reference will be
presented to the Q5 output dividers.
Additionally, CLK0 or CLK1 may be used as a clock source for the
output dividers of Q[4:7]. This may only be done for input frequencies
of 250MHz or less.
Input Clock Selection
The 8T49N285 accepts up to two input clocks with frequencies
ranging from 8kHz up to 875MHz. Each input can accept LVPECL,
LVDS, LVHSTL, HCSL or LVCMOS inputs using 1.8V, 2.5V or 3.3V
logic levels. To use LVCMOS inputs, refer to the Application Note,
Wiring the Differential Input to Accept Single-Ended Levels for
biasing instructions.
In Manual mode, only one of the inputs may be chosen and if that
input fails the PLL will enter holdover.
Manual mode may be operated by directly selecting the desired input
reference in the REFSEL register field. It may also operate via
pin-selection of the desired input clock by selecting that mode in the
RE FSE L register fie ld. In tha t cas e, GP IO[2 ] mus t be u sed as a Clock
Select input (CSEL). CSEL = 0 will select the CLK0 input and
CSEL = 1 will select the CLK1 input.
In addition, the crystal frequency may be passed directly to the output
dividers for Q[4:7] to use as a reference.
Inputs do not support transmission of spread-spectrum clocking
sources. Since this family is intended for high-performance
applications, it will assume input reference sources to have stabilities
100ppm or better, except where gapped clock inputs are used.
of +
If the PLL is working in automatic mode, then each of the input
reference sources is assigned a priority of 1-2. At power-up or if the
currently selected input reference fails, the PLL will switch to the
highest priority input reference that is valid at that time (seeInput
Clock Monitor for details).
Automatic mode has two sub-options: revertive or non-revertive. In
revertive mode, the PLL will switch to a reference with a higher
priority setting whenever one becomes valid. In non-revertive mode
the PLL remains with the currently selected source as long as it
remains valid.
The clock input selection is based on the input clock priority set by
the Clock Input Priority control registers. It is recommended that all
input references be given different priority settings in the Clock Input
Priority control register.
Each clock input is monitored for Loss of Signal (LOS). If no activity
has been detected on the clock input within a user-selectable time
period then the clock input is considered to be failed and an internal
Loss-of-Signal status flag is set, which may cause an input
switchover depending on other settings. The user-selectable time
period has sufficient range to allow a gapped clock missing many
consecutive edges to be considered a valid input.
User-selection of the clock monitor time-period is based on a counter
driven by a monitor clock. The monitor clock is fixed at the frequency
of the PLL’s VCO divided by 8. With a VCO range of 3GHz - 4GHz,
the monitor clock has a frequency range of 375MHz to 500MHz.
The monitor logic for each input reference will count the number of
monitor clock edges indicated in the appropriate Monitor Control
register. If an edge is received on the input reference being
monitored, then the count resets and begins again. If the target edge
count is reached before an input reference edge is received, then an
internal soft alarm is raised and the count re-starts. During the soft
alarm period, the PLL tracking will not be adjusted. If an input
reference edge is received before the count expires for the second
time, then the soft alarm status is cleared and the PLL will resume
adjustments. If the count expires again without any input reference
edge being received, then a Loss-of-Signal alarm is declared.
It is expected that for normal (non-gapped) clock operation, users will
set the monitor clock count for each input reference to be slightly
longer than the nominal period of that input reference. A margin of
2-3 monitor clock periods should give a reasonably quick reaction
time and yet prevent false alarms.
For gapped clock operation, the user will set the monitor clock count
to a few monitor clock periods longer than the longest expected clock
gap period. The monitor count registers support 17-bit count values,
which will support at least a gap length of two clock periods for any
supported input reference frequency, with longer gaps being
supported for faster input reference frequencies. Since gapped
clocks usually occur on input reference frequencies above 100MHz,
gap lengths of thousands of periods can be supported.
Using this configuration for a gapped clock, the PLL will continue to
adjust while the normally expected gap is present, but will freeze
once the expected gap length has been exceeded and alarm after
twice the normal gap length has passed.
Once a LOS on any of the input clocks is detected, the appropriate
internal LOS alarm will be asserted and it will remain asserted until
that input clock returns and is validated once 8 rising edges have
been received on that input reference. If another error condition on
the same input clock is detected during the validation time then the
alarm remains asserted and the validation time starts over.
Each LOS flag may also be reflected on one of the GPIO[3:0]
outputs. Changes in status of any reference can also generate an
interrupt if not masked.
Holdover
8T49N285 supports a small initial holdover frequency offset in
non-gapped clock mode. When the input clock monitor is set to
support gapped clock operation, this initial holdover frequency offset
is indeterminate since the desired behavior with gapped clocks is for
the PLL to continue to adjust itself even if clock edges are missing. In
gapped clock mode, the PLL will not enter holdover until the input is
missing for two LOS monitor periods.
The holdover performance characteristics of a clock are referred as
its accuracy and stability, and are characterized in terms of the
fractional frequency offset. The 8T49N285 can only control the initial
frequency accuracy. Longer-term accuracy and stability are
determined by the accuracy and stability of the external oscillator.
When the PLL loses all valid input references, it will enter the
holdover state. In fast average mode, the PLL will initially maintain its
most recent frequency offset setting and then transition at a rate
dictated by its selected phase-slope limit setting to a frequency offset
setting that is based on historical settings. This behavior is intended
to compensate for any frequency drift that may have occurred on the
input reference before it was detected to be lost.
The historical holdover value will have three options:
•Return to center of tuning range within the VCO band
•Instantaneous mode - the holdover frequency will use the
DPLL current frequency 100msec before it entered
holdover. The accuracy is shown in the AC Characteristics
Table, Table 11A.
•Fast average mode - an internal IIR (Infinite Impulse
Response) filter is employed to get the frequency offset.
The IIR filter gives a 3 dB attenuation point corresponding
to nominal a period of 20 minutes. The accuracy is shown
in the AC Characteristics Table, Table 11A.
When entering holdover, the PLL will set a separate internal HOLD
alarm internally. This alarm may be read from internal status register,
appear on the appropriate GPIO pin and/or assert the nINT output.
While the PLL is in holdover, its frequency offset is now relative to the
crystal input and so the output clocks will be tracing their accuracy to
the local oscillator or crystal. At some point in time, depending on the
stability & accuracy of that source, the clock(s) will have drifted
outside of the limits of the holdover state and the system will be
considered to be in a free-run state. Since this borderline is defined
outside the PLL and dictated by the accuracy and stability of the
external local crystal or oscillator, the 8T49N285 cannot know or
influence when that transition occurs.
remain in the Holdover state internally.
As a result, the 8T49N285 will
Input to Output Clock Frequency
The 8T49N285 is designed to accept any frequency within its input
range and generate eight different output frequencies that are
independent from the input frequencies. The internal architecture of
the device ensures that most translations will result in the exact
output frequency specified. Where exact frequency translation is not
possible, the frequency translation error will be minimized. Please
contact Renesas for configuration software or other assistance in
determining if a desired configuration will be supported exactly.
Synthesizer Mode Operation
The device may also act as a frequency synthesizer with the PLL
generating its operating frequency from just the crystal input. By
setting the SYN_MODE register bit and setting the STATE[1:0] field
to Freerun, no input clock references are required to generate the
desired output frequencies.
When operating in Synthesizer Mode as described above, the
8T49N285 has a fixed loop bandwidth of approximately 200kHz.
When Operating in all other modes, the following information applies:
The 8T49N285 uses no external components to support a range of
loop bandwidths:1.40625Hz, 2.8125Hz, 5.625Hz, 11.25Hz, 22.5Hz,
45Hz, 90Hz, 180Hz or 360Hz.
The device supports two different loop bandwidth settings:
acquisition and locked. These loop bandwidths are selected from the
list of options described above. If enabled, the acquisition bandwidth
is used while lock is being acquired to allow the PLL to “fast-lock”.
Once locked the PLL will use the locked bandwidth setting. If the
acquisition bandwidth setting is not used, the PLL will use the locked
bandwidth setting at all times.
Output Dividers
The 8T49N285 supports eight output dividers. Six of the output
dividers will have IntN capability only (see Table 3) and the other two
Each integer output divider block consists of two divider stages in a
series to achieve the desired total output divider ratio. The first stage
divider may be set to divide by 4, 5 or 6. The second stage of the
divider may be bypassed (i.e. ÷1) or programmed to any even divider
ratio from 2 to 131,070. The total divide ratios, settings and possible
output frequencies are shown in Table 3.
In addition, the first divider stage for the Q[4:7] outputs support a
bypass (i.e. ÷1) operation for some clock sources.
For the FracN output dividers Q2, Q3, the output divide ratio is given
by:
Output Divide Ratio = (N.F)x2
N = Integer Part: 4, 5, ...(2
F = Fractional Part: [0, 1, 2, ...(2
For integer operation of these output dividers, N = 3 is also
supported.
18
-1)
28
-1)]/(228)
Table 3. Q[0:1], Q[4:7] Output Divide Ratios
1st-Stage
Divide
4147501000
515600800
616500666.7
428375500
5210300400
6212250333.3
4416187.5250
5420150200
6424125166.7
4131,070524,2800.00570.0076
5131,070655,3500.00460.0061
6131,070786,4200.00380.0051
NOTE: Above frequency ranges for Q[4:7] apply when driven directly
2nd-Stage
Divide
from the PLL.
Total
Divide
...
Minimum
MHz
F
OUT
Maximum
F
MHz
OUT
Output Divider Frequency Sources
Output dividers associated with the Q[0:3] outputs take their input
frequency directly from the PLL.
Output dividers associated with the Q[4:7] outputs can take their
input frequencies from the PLL, Q2 or Q3 output dividers, CLK0 or
CLK1 input reference frequency or the crystal frequency.
Output Banks
Outputs of the 8T49N285 are divided into three banks for purposes
of output skew measurement:
When the 8T49N285 switches between input references or enters or
leaves the holdover state, there are two options on how the output
phase can be controlled in these events: phase-slope limiting or fully
hitless switching (sometimes called phase build-out) may be
selected. The SWMODE bit selects which behavior is to be followed.
If fully hitless switching is selected, then the output phase will remain
unchanged under any of these conditions. Note that fully hitless
switching is not supported when external loopback is being used.
Fully hitless switching should not be used unless all input references
are in the same clock domain. Note that use of this mode may
prevent an output frequency and phase from being able to trace its
alignment back to a primary reference source.
If phase-slope limiting is selected, then the output phase will adjust
from its previous value until it is tracking the new condition at a rate
dictated by the SLEW[1:0] bits. Phase-slope limiting should be used
if all input references are not in the same clock domain or users wish
to retain traceability to a primary reference source.
Input-Output Delay Control
When using the 8T49N285 in external loopback or in a situation
where input-output delay needs to be known and controlled, it is
necessary to examine the exact signal path through the device. Due
to the flexibility of the device, there are a large number of potential
signal paths from input to output through it that depend on the desired
configuration. Each of those potential paths may include or exclude
logic blocks from the path and change the absolute value of the delay
(Static Phase Offset or SPO) through the device. Considering the
range of SPO values to cover all those potential paths would not be
useful in achieving the target delays for any specific user
configuration. Please contact Renesas for the specific SPO value
associated with a desired input-output path. Note that events such as
switchovers, entering or leaving holdover or re-configuring the signal
path can result in one-time changes to the SPO due to that path
re-configuration. The AC Characteristics table (Table 11A) indicates
the maximum variation in SPO that could be expected for a particular
path through the device.
Output Phase Alignment
The device has a programmable output to output phase alignment for
each of the eight output dividers. After power-up and the PLL has
achieved lock, the device will be in a state where the outputs are
synchronized with a deterministic offset relative to each other. After
synchronization, the output alignment will depend on the particular
configuration of each output according to the following rules. The
step size is defined as the period of the clock to that divider:
1) Only outputs derived from the same source will be aligned with
each other. 'Source' means the reference selected to drive the output
divider as controlled by the CLK_SELn bit for each output.
2) For integer dividers (Q[0:1], Q[4:7]) when both divider stages are
active, edges are aligned. This case is used as a baseline to compare
the other cases here.
3) For integer dividers where the 1st-stage divider is bypassed (only
Q[4:7] support this), coarse delay adjustments can't be performed.
The output phase will be one step earlier than in Case 2.
4) Fractional output dividers (Q2 or Q3) do not guarantee any specific
phase on power-up or after a synchronization event.
5) Integer dividers using Q2 or Q3 as a source (Q[4:7] support this
option) will be aligned to their source divider's output (Q2 or Q3).
Note that the output skews described above are not included in any
of the phase adjustments described here.
Once the device is in operation, the outputs may have their phase
adjustments re-synced in one of two ways:
1) If the PLL becomes unlocked, the coarse phase adjustments will
be reset and the fine phase adjustments will be re-loaded once it
becomes locked again.
2) Toggling of PLL_SYN bit may also be used to force a re-sync /
re-load for the outputs.
The user may apply adjustments that are proportional to the period
of the clock source each output divider is operating from. For
example, if the divider associated with Output Q3 is running off the
PLL, which has a VCO frequency of 4GHz, then the appropriate
period would be 250ps. The output phase may be adjusted in these
steps across the full period of the output.
• Coarse Adjustment: all Output Dividers may have their phase
adjusted in steps of the source clock period. For example a
4GHz VCO gives a step size of 250ps. The user may request an
adjustment of phase of up to 31 steps using a single register
write. The phase will be adjusted by lengthening the period of
the output by 250ps at a time. This process will be repeated
every four output clock periods until the full requested
adjustment has been achieved. A busy signal will remain
asserted in the phase delay register until the requested
adjustment is complete. Then a further adjustment may be
setup and triggered by toggling the trigger bit.
• Fine Adjustment: For the Fractional Output Dividers associated
with the Q2 and Q3 outputs, the phase of those outputs may be
further adjusted with a granularity of 1/16th of the VCO period.
For example a 4GHz VCO frequency gives a granularity of
16ps. This is performed by directly writing the required offset
(from the nominal rising edge position) in units of 1/16th of the
output period into a register. Then the PLL_SYN bit must be
toggled to load the new value. Note that toggling this bit will
clear all Coarse Delays for all outputs associated with the PLL,
so Fine Delays should be set first. The output will then jump
directly to that new offset value. For this reason, this adjustment
should be made as the input is initially programmed or in
High-Impedance.
Each output has the capability of being inverted (180° phase shift).
Jitter and Wander Tolerance
The 8T49N285 can be used as a line card device and therefore is
expected to tolerate the jitter and wander output of a timing card PLL
(e.g. 82P33714).
The Q[0:7] clock outputs are provided with register-controlled output
drivers. By selecting the output drive type in the appropriate register,
any of these outputs can support LVCMOS, LVPECL, HCSL or LVDS
logic levels.
The operating voltage ranges of each output is determined by its
independent output power pin (V
) and thus each can have
CCO
different output voltage levels. Output voltage levels of 2.5V or 3.3V
are supported for differential operation and LVCMOS operation. In
addition, the LVCMOS output operation supports 1.8V (V
CCO
).
Each output may be enabled or disabled by register bits and/or GPIO
pins configured as Output Enables. The outputs will be enabled if the
register bit and the associated OE pin are both asserted (high). When
disabled an output will be in a high impedance state.
LVCMOS O p eratio n
When a given output is configured to provide LVCMOS levels, then
both the Q and nQ outputs will toggle at the selected output
frequency. All the previously described configuration and control
apply equally to both outputs. Frequency, phase alignment, voltage
levels and enable / disable status apply to both the Q and nQ pins.
When LVCMOS levels are selected, the Q and nQ outputs can be
selected to be phase-aligned with each other or inverted relative to
one another. Phase-aligned outputs will have increased
simultaneous switching currents which can negatively affect phase
noise performance and power consumption. It is recommended that
use of this selection be kept to a minimum.
Power-Saving Modes
To allow the device to consume the least power possible for a given
application, the following functions are included under register
control:
• Any unused output, including all output divider and phase
adjustment logic, can be individually powered-off.
• Clock gating on logic that is not being used.
Status / Control Signals and Interrupts
General-Purpose I/Os & Interrupts
The 8T49N285 provides four General Purpose Input / Output (GPIO)
pins for miscellaneous status & control functions. Each GPIO may be
configured as an input or an output. Each GPIO may be directly
controlled from register bits or be used as a predefined function as
shown in Table 4. Note that the default state prior to configuration
being loaded from internal OTP or external EEPROM will be to set
each GPIO to function as an Output Enable.
If used in the Fixed Function mode of operation, the GPIO bits will
reflect the real-time status of their respective status bits as shown in
Table 4. Note that the LOL signal represents the lock status of the
PLL. It does not account for the process of synchronization of the
output dividers associated with that PLL. The output dividers
programmed to operate from that PLL will automatically go through a
re-synchronization process when the PLL locks or re-locks or if the
user triggers a re-sync manually via register bit PLL_SYN. This
synchronization process may result in a period of instability on the
affected outputs for a duration of up to 350ns after the re-lock (LOL
de-asserts) or the PLL_SYN bit is de-asserted.
Table 4. GPIO Configuration
Configured as Input
Fixed Function
Output
GPIO Pin
Enable
(default)
3OE[3]OE[7]-GPI[3]LOS[1]GPO[3]
2OE[2]OE[6]CSELGPI[2]LOS[0]GPO[2]
1OE[1]OE[5]-GPI[1]HOLDGPO[1]
0OE[0]OE[4]-GPI[0]LOLGPO[0]
Output
Enable
Clock
Select
General
Purpose
Configured as
Output
Fixed
Function
General
Purpose
Interrupt Functionality
Interrupt functionality includes an interrupt status flag for the PLL
Loss-of-Lock Status (LOL), PLL Holdover Status (HOLD) and Input
Reference Status (LOS[1:0]) that is set whenever there is an alarm
on any of those signals. The Status Flag will remain set until the
alarm has been cleared and a ‘1’ has been written to the Status
Flag’s register location or if a reset occurs. Each Status Flag will also
have an Interrupt Enable bit that will determine if that Status Flag is
allowed to cause the Interrupt Status to be affected (enabled) or not
(disabled). All Interrupt Enable bits will be in the disabled state after
reset. The Device Interrupt Status flag and nINT output pin are
asserted if any of the enabled Interrupt Status flags are set.
Device Hardware Configuration
The 8T49N285 supports an internal One-Time Programmable (OTP)
memory that can be pre-programmed at the factory with one
complete device configuration. If the device is set to read a
configuration from an external, serial EEPROM, then the values read
will overwrite the OTP-defined values.
This configuration can be over-written using the serial interface once
reset is complete. Any configuration written via the programming
interface needs to be re-written after any power cycle or reset. Please
contact Renesas if a specific factory-programmed configuration is
desired.
AADa ta 0AAData nPAOffset Addr MSBOffset Addr LSB AData 1A
Device Start-up and Reset Behavior
The 8T49N285 has an internal power-up reset (POR) circuit and a
Master Reset input pin nRST. If either is asserted, the device will be
in the Reset State.
For highly programmable devices, it is common practice to reset the
device immediately after the initial power-on sequence. Renesas
recommends connecting the nRST input pin to a programmable logic
source for optimal functionality. It is recommended that a minimum
pulse width of 10ns be used to drive the nRST input pin.
While in the reset state (nRST input asserted or POR active), the
device will operate as follows:
• All registers will return to & be held in their default states as
indicated in the applicable register description.
• All internal state machines will be in their reset conditions.
• The serial interface will not respond to read or write cycles.
• The GPIO signals will be configured as OE[3:0] inputs.
• All clock outputs will be disabled.
• All interrupt status and Interrupt Enable bits will be cleared,
negating the nINT signal.
Upon the latter of the internal POR circuit expiring or the nRST input
negating, the device will exit reset and begin self-configuration.
The device will load an initial block of its internal registers using the
configuration stored in the internal One-Time Programmable (OTP)
memory. Once this step is complete, the 8T49N285 will check the
register settings to see if it should load the remainder of its
configuration from an external I
continue loading from OTP. See the section on I2C Boot-up
Initialization Mode for details on how this is performed.
Once the full configuration has been loaded, the device will respond
to accesses on the serial port and will attempt to lock the PLL to the
selected source and begin operation. Once the PLL is locked, all the
outputs derived from it will be synchronized and output phase
adjustments can then be applied if desired.
2
C EEPROM at a defined address or
Serial Control Port Description
Serial Control Port Configuration Description
The device has a serial control port capable of responding as a slave
2
C compatible configuration to allow access to any of the
in an I
internal registers for device programming or examination of internal
status. All registers are configured to have default values. See the
specifics for each register for details.
The device has the additional capability of becoming a master on the
2
C bus only for the purpose of reading its initial register
I
configurations from a serial EEPROM on the I
configuration to the serial EEPROM must be performed by another
device on the same I2C bus or pre-programmed into the device prior
to assembly.
I2C Mode Operation
The I2C interface is designed to fully support v2.1 of the I2C
Specification for Normal and Fast mode operations. The device acts
as a slave device on the I
address defined in the Serial Interface Control register (0006h), as
modified by the S_A0 input pin setting. The interface accepts
byte-oriented block write and block read operations. Two address
bytes specify the register address of the byte position of the first
register to write or read. Data bytes (registers) are accessed in
sequential order from the lowest to the highest byte (most significant
bit first). Read and write block transfers can be stopped after any
complete byte transfer. During a write operation, data will not be
moved into the registers until the STOP bit is received, at which point,
all data received in the block write will be written simultaneously.
For full electrical I
2
pull-up resistors for SDATA and SCLK. The internal pull-up resistors
have a size of 51k typical.
2
C bus at 100kHz or 400kHz using the
C compliance, it is recommended to use external
2
C bus. Writing of the
Figure 3. I2C Slave Read and Write Cycle Sequencing
When operating in I2C mode, the 8T49N285 has the capability to
become a bus master on the I2C bus for the purposes of reading its
configuration from an external I
will be supported.
2
As an I
C bus master, the 8T49N285 will support the following
functions:
• 7-bit addressing mode
• Base address register for EEPROM
• Validation of the read block via CCITT-8 CRC check against value
stored in last byte (E0h) of EEPROM
• Support for 100kHz and 400kHz operation with speed negotiation.
If bit d0 is set at Byte address 05h in the EEPROM, this will shift
from 100kHz operation to 400kHz operation.
• Support for 1- or 2-byte addressing mode
• Master arbitration with programmable number of retries
2
C EEPROM. Only a block read cycle
8T49N285 Datasheet
• Fixed-period cycle response timer to prevent permanently hanging
2
the I
C bus.
• Read will abort with an alarm (BOOTFAIL) if any of the following
conditions occur: Slave NACK, Arbitration Fail, Collision during
Address Phase, CRC failure, Slave Response time-out
The 8T49N285 will not support the following functions:
2
•I
C General Call
• Slave clock stretching
2
C Start Byte protocol
•I
• EEPROM Chaining
• CBUS compatibility
• Responding to its own slave address when acting as a master
If enabled (via the BOOT_EEP bit in the Startup register), once the
nRST input has been de-asserted (high) and its internal power-up
reset sequence has completed, the device will contend for ownership
2
of the I
location on the I2C bus. The address of that memory location is kept
in non-volatile memory in the Startup register. During the boot-up
process, the device will not respond to serial control port accesses.
Once the initialization process is complete, the contents of any of the
device’s registers can be altered. It is the responsibility of the user to
make any desired adjustments in initial values directly in the serial
bus memory.
C bus to read its initial register settings from a memory
If a NACK is received to any of the read cycles performed by the
device during the initialization process, or if the CRC does not match
the one stored in address E0h of the EEPROM the process will be
aborted and any uninitialized registers will remain with their default
values. The BOOTFAIL bit (021Eh) in the Global Interrupt Status
register will also be set in this event.
If the BOOTFAIL bit is set, then the LOL indicator will be set.
Contents of the EEPROM should be as shown in Table 5.
Table 5. External Serial EEPROM Contents
EEPROM Offset
(Hex)
00 11111111
01 11111111
02 11111111
03 11111111
04 11111111
05 1111111
0618T49N285 Device I
07 00000000
08 - DFDesired contents of Device Registers 08h - DFh
RsvdR/W-Reserved. Always write 0 to this bit location. Read values are not defined.
Select number of times arbitration for the I
retried before being aborted. Note that this number does not include the original try.
Internal One-Time Programmable (OTP) memory usage on power-up:
0 = Load power-up configuration from OTP
1 = Only load 1st eight bytes from OTP
External EEPROM usage on power-up:
0 = Load power-up configuration from external serial EEPROM (overwrites OTP
values)
1 = Don’t use external EEPROM
2
C base address for serial EEPROM.
C bus to read the serial EEPROM will be
NOTE 1: These values are specific to the device configuration and can be customized when ordering. Please refer to the FemtoClock NG
Universal Frequency Translator Ordering Product Information guide or custom datasheet addendum for more details.
Table 6C. Device ID Control Register Bit Field Locations and Descriptions
Device ID Register Control Block Field Locations
Address (Hex)D7D6D5D4D3D2D1D0
0002REV_ID[3:0]DEV_ID[15:12]
0003DEV_ID[11:4]
0004DEV_ID[3:0]DASH_CODE[10:7]
0005DASH_CODE[6:0]1
Device ID Control Register Block Field Descriptions
Bit Field NameField TypeDefault Value Description
REV_ID[3:0]R/W0000bDevice revision.
DEV_ID[15:0]R/W0603hDevice ID code.
Device Dash Code:
DASH_CODE
[10:0]
NOTE 1: These values are specific to the device configuration and can be customized when ordering. Please refer to the FemtoClock NG
Universal Frequency Translator Ordering Product Information guide or custom datasheet addendum for more details.
R/WNOTE 1
Decimal value assigned by Renesas to identify the configuration loaded at the
factory. May be over-written by users at any time. Refer to FemtoClock NG Universal Frequency Translator Ordering Product Information guide to identify major
configuration parameters associated with this Dash Code value.
Table 6E. Digital PLL Input Control Register Bit Field Locations and Descriptions
Digital PLL Input Control Register Block Field Locations
Address (Hex)D7D6D5D4D3D2D1D0
0008REFSEL[2:0]FBSEL[2:0]RVRTSWMODE
00091110PRI_1[1:0]PRI_0[1:0]
000A11REFDIS_1REFDIS_0RsvdRsvdSTATE0[1:0]
000BRsvdPRE_0[20:16]
000CPRE_0[15:8]
000DPRE_0[7:0]
000ERsvdPRE_1[20:16]
000FPRE_1[15:8]
0010PRE_1[7:0]
0011RsvdRsvd
0012Rsvd
0013Rsvd
0014RsvdRsvd
0015Rsvd
0016Rsvd
Digital PLL Input Control Register Block Field Descriptions
Bit Field NameField TypeDefault Value Description
Input reference selection for Digital PLL:
000 = Automatic selection
001 = Manual selection by GPIO inputs
REFSEL[2:0]R/W000b
FBSEL[2:0]R/W000b
RVRTR/W1b
SWMODER/W1b
PRI_0[1:0]R/W00b
010 through 011 = Reserved
100 = Force selection of Input Reference 0
101 = Force selection of Input Reference 1
110 = Do not use
111 = Do not use
Feedback mode selection for Digital PLL:
000 through 011 = internal feedback divider
100 = external feedback from Input Reference 0
101 = external feedback from Input Reference 1
110 = do not use
111 = do not use
Automatic switching mode for Digital PLL:
0 = non-revertive switching
1 = revertive switching
Controls how Digital PLL adjusts output phase when switching between input
references:
0 = Absorb any phase differences between old and new input references at the PLL
output. Recommended for use when both input references are in the same clock
domain.
1 = Limit the maximum rate of phase change at the PLL output when adjusting to a
new input reference’s phase/frequency using phase-slope limiting as set in the
SLEWn bits. Recommended for use when the input references are not in the same
clock domain.
Switchover priority for Input Reference 0 when used by Digital PLL:
00 = 1st priority
01 = 2nd priority
10 = do not use
11 = do not use
Digital PLL Input Control Register Block Field Descriptions
Bit Field NameField TypeDefault Value Description
Switchover priority for Input Reference 1 when used by Digital PLL:
00 = 1st priority
PRI_1[1:0]R/W01b
REFDIS_0R/W0b
REFDIS_1R/W0b
STATE0[1:0]R/W00b
PRE_0[20:0]R/W000000hPre-divider ratio for Input Reference 0 when used by Digital PLL.
PRE_1[20:0]R/W000000hPre-divider ratio for Input Reference 1 when used by Digital PLL.
RsvdR/W-Reserved. Always write 0 to this bit location. Read values are not defined.
01 = 2nd priority
10 = do not use
11 = do not use
Input Reference 0 Switching Selection Disable for Digital PLL:
0 = Input Reference 0 is included in the switchover sequence for Digital PLL
1 = Input Reference 0 is not included in the switchover sequence for Digital PLL
Input Reference 1 Switching Selection Disable for Digital PLL:
0 = Input Reference 1 is included in the switchover sequence for Digital PLL
1 = Input Reference 1 is not included in the switchover sequence for Digital PLL
Digital PLL State Machine Control:
00 = Run automatically
01 = Force FREERUN state - set this if in Synthesizer Mode for PLL.
10 = Force NORMAL state
11 = Force HOLDOVER state
Holdover Averaging mode selection for Digital PLL:
00 = Instantaneous mode - uses historical value 100ms prior to entering holdover
HOLD[1:0]R/W00b
01 = Fast Average Mode
10 = Reserved
11 = Set VCO control voltage to VCC/2
Holdover Averaging Enable for Digital PLL:
HOLDAVGR/W0b
0 = Holdover averaging disabled
1 = Holdover averaging enabled as defined in HOLD[1:0]
Enables Fast Lock operation for Digital PLL:
FASTLCKR/W0b
0 = Normal locking using LCKBW & LCKDAMP fields in all cases
1 = Fast Lock mode using ACQBW & ACQDAMP when not phase locked and LCKBW
& LCKDAMP once phase locked
LOCK[7:0]R/W3Fh
Lock window size for Digital PLL. Unsigned 2’s complement binary number in steps of
2.5ns, giving a total range of 640ns. Do not program to 0.
Integer portion of the Delta-Sigma Modulator value. Do not set higher than FFh. This
DSM_INT[8:0]R/W02Dh
implies that for crystal frequencies lower than 16MHz, the doubler circuit must be
enabled.
DSMFRAC[20:0]R/W000000h
Fractional portion of Delta-Sigma Modulator value. Divide this number by 2
determine the actual fraction.
Delta-Sigma Modulator Order for Digital PLL:
00 = Delta-Sigma Modulator disabled
DSM_ORD[1:0]R/W11b
01 = 1st order modulation
10 = 2nd order modulation
11 = 3rd order modulation
Multiplier applied to instantaneous frequency error before it is applied to the Digitally
Controlled Oscillator in Digital PLL:
DCXOGAIN[1:0]R/W01b
00 = 0.5
01 = 1
10 = 2
11 = 4
Dither Gain setting for Digital PLL:
000 = no dither
001 = Least Significant Bit (LSB) only
010 = 2 LSBs