DS28EC20: 20Kb 1-Wire EEPROM
22 of 27
Master-to-Slave
For a write-one time slot, the voltage on the data line must have crossed the V
TH
threshold before the write-one low
time t
W1LMAX
is expired. For a write-zero time slot, the voltage on the data line must stay below the V
TH
threshold
until the write-zero low time t
W0LMIN
is expired. For the most reliable communication, the voltage on the data line
should not exceed V
ILMAX
during the entire t
W0L
or t
W1L
window. After the V
TH
threshold has been crossed, the
DS28EC20 needs a recovery time t
REC
before it is ready for the next time slot.
Figure 11. Read/Write Timing Diagram
Write-One Time Slot
RESISTOR MASTER
V
PUP
V
IHMASTER
V
TH
V
TL
V
ILMAX
0V
t
F
t
SLOT
t
W1L
Write-Zero Time Slot
RESISTOR MASTER
t
REC
V
PUP
V
IHMASTER
V
TH
V
TL
V
ILMA
X
0V
t
F
t
SLOT
t
W0L
Read-Data Time Slot
RESISTOR MASTER DS28EC20
t
REC
V
PUP
V
IHMASTER
V
TH
V
TL
V
ILMA
X
0V
Master
Sampling
Window
t
F
t
SLOT
t
RL
t
MSR
DS28EC20: 20Kb 1-Wire EEPROM
23 of 27
Slave-to-Master
A read-data time slot begins like a write-one time slot. The voltage on the data line must remain below V
TL
until the
read low time t
RL
is expired. During the t
RL
window, when responding with a 0, the DS28EC20 starts pulling the data
line low; its internal timing generator determines when this pulldown ends and the voltage starts rising again. When
responding with a 1, the DS28EC20 does not hold the data line low at all, and the voltage starts rising as soon as
t
RL
is over.
The sum of t
RL
+ δ (rise time) on one side and the internal timing generator of the DS28EC20 on the other side
define the master sampling window (t
MSRMIN
to t
MSRMAX
) in which the master must perform a read from the data line.
For the most reliable communication, t
RL
should be as short as permissible, and the master should read close to
but no later than t
MSRMAX
. After reading from the data line, the master must wait until t
SLOT
is expired. This
guarantees sufficient recovery time t
REC
for the DS28EC20 to get ready for the next time slot. Note that t
REC
specified herein applies only to a single DS28EC20 attached to a 1-Wire line. For multidevice configurations, t
REC
needs to be extended to accommodate the additional 1-Wire device input capacitance. Alternatively, an interface
that performs active pullup during the 1-Wire recovery time such as the DS2482-x00 or DS2480B 1-Wire line
drivers can be used.
IMPROVED NETWORK BEHAVIOR (SWITCHPOINT HYSTERESIS)
In a 1-Wire environment, line termination is possible only during transients controlled by the bus master (1-Wire
driver). 1-Wire networks, therefore, are susceptible to noise of various origins. Depending on the physical size and
topology of the network, reflections from end points and branch points can add up or cancel each other to some
extent. Such reflections are visible as glitches or ringing on the 1-Wire communication line. Noise coupled onto the
1-Wire line from external sources can also result in signal glitching. A glitch during the rising edge of a time slot can
cause a slave device to lose synchronization with the master and, consequently, result in a Search ROM command
coming to a dead end or cause a device-specific function command to abort. For better performance in network
applications, the DS28EC20 uses a new 1-Wire front-end, which makes it less sensitive to noise.
The 1-Wire front-end of the DS28EC20 differs from traditional slave devices in three characteristics:
1) There is additional low-pass filtering in the circuit that detects the falling edge at the beginning of a time slot.
This reduces the sensitivity to high-frequency noise. This additional filtering does not apply at overdrive speed.
2) There is a hysteresis at the low-to-high switching threshold V
TH
. If a negative glitch crosses V
TH
but does not go
below V
TH
- V
HY
, it is not recognized (Figure 12, Case A). The hysteresis is effective at any 1-Wire speed.
3) There is a time window specified by the rising edge hold-off time t
REH
during which glitches are ignored, even if
they extend below V
TH
- V
HY
threshold (Figure 12, Case B, t
GL
< t
REH
). Deep voltage droops or glitches that
appear late after crossing the V
TH
threshold and extend beyond the t
REH
window cannot be filtered out and are
taken as the beginning of a new time slot (Figure 12, Case C, t
GL
t
REH
).
Devices that have the parameters V
HY
and t
REH
specified in their electrical characteristics use the improved 1-Wire
front-end.
Figure 12. Noise Suppression Scheme
V
PUP
V
TH
V
HY
0V
t
REH
t
GL
t
REH
t
GL
Case A
Case C
Case B
DS28EC20: 20Kb 1-Wire EEPROM
24 of 27
CRC GENERATION
The DS28EC20 uses two different types of CRCs. One CRC is an 8-bit type and is stored in the most significant
byte of the 64-bit ROM. The bus master can compute a CRC value from the first 56 bits of the 64-bit ROM and
compare it to the value stored within the DS28EC20 to determine if the ROM data has been received error-free.
The equivalent polynomial function of this CRC is X
8
+ X
5
+ X
4
+ 1. This 8-bit CRC is received in the true
(noninverted) form. It is computed at the factory and programmed into the ROM.
The other CRC is a 16-bit type, generated according to the standardized CRC16 polynomial function
X
16
+ X
15
+ X
2
+ 1. This CRC is used for fast verification of a data transfer when writing to or reading from the
scratchpad and with the Extended Read Memory command. In contrast to the 8-bit CRC, the 16-bit CRC is always
communicated in the inverted form. A CRC generator inside the DS28EC20 (Figure 13) calculates a new 16-bit
CRC, as shown in the command flowchart (Figure 7). The bus master compares the CRC value read from the
device to the one it calculates from the data, and decides whether to continue with an operation or to reread the
portion of the data with the CRC error.
With the Write Scratchpad command, the CRC is generated by first clearing the CRC generator and then shifting in
the command code, the target addresses TA1 and TA2, and all the data bytes as they were sent by the bus
master. The DS28EC20 transmits this CRC only if the data bytes written to the scratchpad include scratchpad
ending offset 11111b. The data can start at any location within the scratchpad.
With the Read Scratchpad command, the CRC is generated by first clearing the CRC generator and then shifting in
the command code, the target addresses TA1 and TA2, the E/S byte, and the scratchpad data as they were sent
by the DS28EC20 starting at the target address. The DS28EC20 transmits this CRC only if the reading continues
through the end of the scratchpad, regardless of the actual ending offset.
With the initial pass through the extended read memory flow, the 16-bit CRC value is the result of shifting the
command byte into the cleared CRC generator, followed by the two address bytes and the data bytes. Subsequent
passes through the extended read memory flow generate a 16-bit CRC that is the result of clearing the CRC
generator and then shifting in the data bytes. For more information on generating CRC values refer to Application
Note 27: Understanding and Using Cyclic Redundancy Checks with Maxim i
Button Products (www.maxim-
ic.com/AN27).
Figure 13. CRC16 Hardware Description and Polynomial
Polynomial =
X
16
+ X
15
+ X
2
+ 1
X
0
X
1
X
2
X
3
X
4
X
5
X
6
X
7
X
8
X
9
X
10
X
11
X
12
X
13
X
14
X
15
X
16
1
st
STAGE
2
nd
STAGE
3
rd
STAGE
4
th
STAGE
6
th
STAGE
5
th
STAGE
7
th
STAGE
8
th
STAGE
9
th
STAGE
10
th
STAGE
11
th
STAGE
12
th
STAGE
13
th
STAGE
14
th
STAGE
15
th
STAGE
16
th
STAGE
INPUT DATA
CRC
OUTPUT

DS28EC20P+

Mfr. #:
Manufacturer:
Maxim Integrated
Description:
EEPROM 20Kb 1-Wire EEPROM
Lifecycle:
New from this manufacturer.
Delivery:
DHL FedEx Ups TNT EMS
Payment:
T/T Paypal Visa MoneyGram Western Union

Products related to this Datasheet