2011 Microchip Technology Inc. DS22067J-page 9
11XX
3.6 Device Standby
The 11XX fe atures a low-po wer Stand by mode du ring
which the device is waiting to begin a new command.
A high-to-low transition on SCIO will exit low-power
mode and prepare the device for receiving the start
header.
Standby mode will be entered upon the following
conditions:
• A NoMAK followed by a SAK (i.e., valid termina-
tion of a command)
• Reception of a standby pulse
3.7 Device Idle
The 11XX features an Idle m ode du ring w hi ch all s eria l
data is ignored until a standby pulse occ urs . Idle mod e
will be entered upon the following conditions:
• Invalid device address
• Invalid command byte, including Read, CRRD,
Write, WRSR, SETAL and ERAL during a write
cycle.
• Missed edge transition
• Reception of a MAK following a WREN, WRDI,
SETAL, or ERAL command byte
• Reception of a MAK following the data byte of a
WRSR command
An invalid start header will indirectly cause the device
to enter Idle mode. Whether or not the start header is
invalid cannot be detected by the slave, but will
prevent the slave from synchronizing properly with the
master. If the slave is not synchronized with the
master, a n e dge t r ans iti on will be mi ss ed , th us c au sin g
the device to enter Idle mode.
3.8 Synchronization
At the beginning of every command, the 11XX utilizes
the start header to determine the master’s bus clock
period. This period is then used as a reference for all
subsequent communication within that command.
The 11XX features re-synchronization circuitry which
will monitor the position of the middle data edge during
each MAK bi t and subs equently adju st the internal tim e
reference in order to remain synchronized with the
master.
There are two variables which can cause the 11XX to
lose synchronization. The first is frequency drift,
defined as a ch ange i n the bit period, TE. The second is
edge jitter, which is a single occurrence change in the
position of an edge within a bit period, while the bit
period itself remains constant.
3.8.1 FREQUENCY DRIFT
With in a syst em, th ere is a p ossibi lity that fr equen cies
can drift due to changes in voltage, temperature, etc.
The re-synchronization circuitry provides some toler-
ance for such frequency drift. The tolerance range is
specified by two parameters, FDRIFT and FDEV. FDRIFT
specifies the maximum tolerable change in bus fre-
quency per byte. FDEV specifies the overall limit in fre-
quency deviation within an operati on (i.e., from the en d
of the start header until communication is terminated
for that o pera tion). T he sta rt head er at th e begi nning of
the next operation will reset the re-synchronization cir-
cuitr y an d a l low fo r an o t her FDEV amount of frequency
drift.
3.8.2 EDGE JITTER
Ensuring that edge transitions from the master always
occur exactly in the middle or end of the bi t period is not
alwa ys po s si bl e. T h er ef or e, t he re - sy nc hro ni z at i on cir -
cuitry is designed to provide some tolerance for edge
jitter.
The 11XX adjusts its phase every MAK bit, so TIJIT
specifies the maximum allowable peak-to-peak jitter
relative to the previous MAK bit. Since the position of
the previous MAK bit would be difficult to measure by
the maste r , the m inimum and maximum jitte r values for
a system should be considered the worst-case. These
values will be based on the execution time for different
branch paths in software, jitter due to thermal noise,
etc.
The difference between the minimum and maximum
values , as a percent age of the b it period, should be cal-
culated and then compared against TIJIT to determine
jitter compliance.
Note: In the case of the WRITE, WRSR, SETAL,
or ERAL commands, the write cycle is initi-
ated upon receipt of the NoMAK, assuming
all other w rite requi rem en ts have been met.
Note: Because the 11XX only re-synchronizes
during the MAK bit, the overall ability to
remain synchronized depends on a combi-
nation o f frequency d r if t and ed ge j itter (i.e.,
if the MAK bit e dge is experien cing the max-
imum allowable edge jitter, then there is no
room for frequency drift). Conversely, if the
frequency has drifted to the maximum
amount tolerable within a byte, then no
edge jitter can be present.