FEATURES
Dual, 1024-Position Resolution
25 k, 250 k Full-Scale Resistance
Low Temperature Coefficient: 35 ppm/C
Nonvolatile Memory
1
Preset Maintains Wiper Settings
Permanent Memory Write-Protection
Wiper Settings Read Back
Actual Tolerance Stored in EEMEM
1
Linear Increment/Decrement
Log Taper Increment/Decrement
SPI Compatible Serial Interface
3 V to 5 V Single Supply or 2.5 V Dual Supply
26 Bytes User Nonvolatile Memory for Constant Storage
Current Monitoring Configurable Function
100-Year Typical Data Retention T
The ADN2850 provides dual-channel, digitally controlled programmable resistors
2
with resolution of 1024 positions. These devices
perform the same electronic adjustment function as a mechanical
rheostat with enhanced resolution, solid-state reliability, and
superior low temperature coefficient performance. The ADN2850’s
versatile programming via a standard serial interface allows
16 modes of operation and adjustment, including scratch pad programming, memory storing and retrieving, increment/decrement,
log taper adjustment, wiper setting readback, and extra user
defined EEMEM
1
.
Another key feature of the ADN2850 is that the actual tolerance
is stored in the EEMEM. The actual full-scale resistance can
therefore be known, which is valuable for tolerance matching
and calibration.
In the scratch pad programming mode, a specific setting can be
programmed directly to the RDAC
2
register, which sets the resistance between terminals W and B. The RDAC register can also
be loaded with a value previously stored in the EEMEM register.
The value in the EEMEM can be changed or protected. When
changes are made to the RDAC register, the value of the new
setting can be saved into the EEMEM. Thereafter, such value will
be transferred automatically to the RDAC register during system
power ON, which is enabled by the internal preset strobe.
EEMEM can also be retrieved through direct programming and
external preset pin control.
*Patent pending
NOTES
1
The term nonvolatile memory and EEMEM are used interchangeably.
2
The term programmable resistor and RDAC are used interchangeably.
REV. B
Information furnished by Analog Devices is believed to be accurate and
reliable. However, no responsibility is assumed by Analog Devices for its
use, nor for any infringements of patents or other rights of third parties that
may result from its use. No license is granted by implication or otherwise
under any patent or patent rights of Analog Devices.
ADN2850
*
FUNCTIONAL BLOCK DIAGRAM
Figure 1. RWB(D) vs. Decimal Code
The linear step increment and decrement commands enable the
setting in the RDAC register to be moved UP or DOWN, one step
at a time. For logarithmic changes in wiper setting, a left/right
bit shift command adjusts the level in ±6 dB steps.
The ADN2850 is available in the 5 mm 5 mm 16-lead frame chip
scale LFCSP and thin 16-lead TSSOP packages. All parts are
guaranteed to operate over the extended industrial temperature
range of –40°C to +85°C.
With respect to GND, VDD = 5 V2.4V
With respect to GND, VDD = 5 V0.8V
With respect to GND, VDD = 3 V2.1V
With respect to GND, VDD = 3 V0.6V
With respect to GND,
V
= +2.5 V, V
DD
= –2.5 V2.0V
SS
With respect to GND,
= +2.5 V, V
V
DD
R
= 2.2 kΩ to 5 V4.9V
PULL-UP
IOL = 1.6 mA, V
VIN = 0 V or V
= –2.5 V0.5V
SS
= 5 V0.4V
LOGIC
DD
±2.25µA
5pF
POWER SUPPLIES
Single-Supply Power RangeV
Dual-Supply Power RangeV
Positive Supply CurrentI
Positive Supply CurrentI
Programming Mode CurrentI
Read Mode Current
7
Negative Supply CurrentI
Power Dissipation
8
Power Supply SensitivityP
CURRENT MONITOR TERMINALS
Current Sink at V
Current Sink at V
DYNAMIC CHARACTERISTICS
9
1
2
5, 10
Resistor Noise Spectral Densitye
Analog Crosstalk (C
)CTVB1 = VB2 = 0 V, Measured VW1 with
W1/CW2
DD
DD/VSS
DD
DD
DD(PG)
I
DD(XFR)
SS
P
DISS
SS
I
1
I
2
N_WB
VSS = 0 V3.05.5V
±2.25± 2.75V
VIH = VDD or VIL = GND,
= 25oC24.5µA
T
A
VIH = VDD or VIL = GND3.56.0µA
VIH = VDD or VIL = GND35mA
VIH = VDD or VIL = GND0.339mA
VIH = VDD or VIL = GND,
= +2.5 V, V
V
DD
= –2.5 V3.56.0µA
SS
VIH = VDD or VIL = GND1850µW∆VDD = 5 V ± 10%0.0020.01%/%
0.000110mA
10mA
R
= 25 kΩ/250 kΩ, f = 1 kHz20/64nV/√Hz
WB_FS
= 100 mV p-p @ f = 100 kHz,
V
W2
Code 1 = Code 2 = 200
H
–65dB
REV. B–2–
ADN2850
ParameterSymbolConditionsMinTyp2MaxUnit
INTERFACE TIMING CHARACTERISTICS (apply to all parts)
Clock Cycle Time (t
CS Setup Timet
CLK Shutdown Time to CS Riset
Input Clock Pulsewidtht4 , t
Data Setup Timet
Data Hold Timet
CS to SDO – SPI Line Acquiret
CS to SDO – SPI Line Releaset
CLK to SDO Propagation Delay
CS High Pulsewidth
CS High to CS HighRDY Rise to CS Fallt
CS Rise to RDY Fall Timet
Read/Store to Nonvolatile EEMEM
CS Rise to Clock Edge Setupt
Preset Pulsewidth (Asynchronous)t
Preset Response Time to Wiper Settingt
)t
CYC
12
13
13
1
2
3
5
6
7
8
9
t
10
t
12
t
13
14
15
14
t
16
17
PRW
PRESP
Clock Level High or Low10ns
From Positive CLK Transition5ns
From Positive CLK Transition5ns
RP = 2.2 kΩ, CL < 20 pF50ns
Applies to Command 2H, 3H, 9
Not Shown in Timing Diagram50ns
PR Pulsed Low to Refresh140µs
Wiper Positions
FLASH/EE MEMORY RELIABILITY
Endurance
Data Retention
NOTES
1
Parts can be operated at 2.7 V single supply, except from 08C to –408C, where minimum 3 V is needed.
2
Typicals represent average readings at 258C and VDD = 5 V.
3
Resistor position nonlinearity error R-INL is the deviation from an ideal value measured between the maximum resistance and the minimum resistance wiper positions.
R-DNL measures the relative step change from ideal between successive tap positions. IW ~ 50 µA for V
4
Resistor terminals W and B have no limitations on polarity with respect to each other.
5
Guaranteed by design and not subject to production test.
6
Common-mode leakage current is a measure of the dc leakage from any terminal B and W to a common-mode bias level of V
7
Transfer (XFR) mode current is not continuous. Current consumed while EEMEM locations are read and transferred to the RDAC register. See TPC 9.
8
P
DISS
9Applies to photodiode of optical receiver.
10
All dynamic characteristics use VDD = +2.5 V and V
11
See timing diagram for location of measured values. All input control voltages are specified with tR = tF = 2.5 ns (10% to 90% of 3 V) and timed from a voltage level of 1.5 V.
Switching characteristics are measured using both VDD = 3 V and 5 V.
12
Propagation delay depends on value of VDD, R
13
Valid for commands that do not activate the RDY pin.
14
RDY pin low only for commands 2, 3, 8, 9, 10, and PR hardware pulse: CMD_8 ~ 1 ms; CMD_9, 10 ~ 0.1 ms; CMD_2, 3 ~ 20 ms. Device operation at TA= –40°C
and VDD < 3 V extends the save time to 35 ms.
15
Endurance is qualified to 100,000 cycles as per JEDEC Std. 22 method A117 and measured at –40 °C, +25°C, and +85°C; typical endurance at +25°C is 700,000 cycles.
16
Retention lifetime equivalent at junction temperature (TJ) = 55°C as per JEDEC Std. 22, Method A117. Retention lifetime based on an activation energy of 0.6 V will
derate with junction temperature.
Specifications subject to change without notice.
The ADN2850 contains 16,000 transistors. Die size: 93 mil 103 mil, 10,197 sq mil.
15
16
is calculated from (IDD VDD) + (ISS VSS).
PULL_UP
= –2.5 V.
SS
, and CL. See Applications section.
5, 11
20ns
10ns
1t
10ns
4t
0ns
0.150.3ms
H
35ms
10ns
100K Cycles
100Years
= 2.7 V and IW ~ 400 µA for V
DD
/2.
DD
40ns
50ns
= 5 V.
DD
CYC
CYC
REV. B
–3–
ADN2850
TIMING DIAGRAMS
CS
CLK
CPOL = 1
SDO
SDI
RDY
CPHA = 1
t
t
LSB
3
t
9
t
15
t
t
2
t
8
*
t
14
*NOT DEFINED, BUT NORMALLY LSB OF CHARACTER PREVIOUSLY TRANSMITTED.
THE CPOL = 1 MICROCONTROLLER COMMAND ALIGNS THE INCOMING DATA TO THE POSITIVE EDGE OF THE CLOCK.
1
t
5
t
4
t
10
MSBLSB OUT
t
7
t
6
MSB
t
11
Figure 2a. CPHA = 1 Timing Diagram
12
t
13
t
17
t
16
CS
CLK
CPOL = 0
SDO
SDI
RDY
CPHA = 0
t
12
t
1
t
2
t
8
MSB OUTLSB
MSB IN
t
14
*NOT DEFINED, BUT NORMALLY MSB OF CHARACTER JUST RECEIVED.
THE CPOL = 0 MICROCONTROLLER COMMAND ALIGNS THE INCOMING DATA TO THE POSITIVE EDGE OF THE CLOCK.
Stresses above those listed under Absolute Maximum Ratings may cause permanent damage to the device. This is a stress rating; functional operation of the device
at these or any other conditions above those listed in the operational sections of this
specification is not implied. Exposure to absolute maximum rating conditions for
extended periods may affect device reliability.
2
Maximum terminal current is bounded by the maximum current handling of the
switches, maximum power dissipation of the package, and maximum applied
voltage across any two of the B and W terminals at a given resistance.
3
Includes programming of nonvolatile memory.
4
Applicable to TSSOP-16 only. For LFCSP-16, please consult factory for details.
ADN2850BCP2525± 2±4–40 to +85LFCSP-16CP-1696BCP25
ADN2850BCP25-RL725± 2±4–40 to +85LFCSP-16CP-161,000BCP25
7" Reel
ADN2850BCP250250±2± 4–40 to +85LFCSP-16CP-1696BCP250
ADN2850BCP250-RL7250±2±4–40 to +85LFCSP-16CP-161,000BCP250
7" Reel
ADN2850BRU2525± 2±4–40 to +85TSSOP-16RU-16962850B25
ADN2850BRU25-RL725± 2±4–40 to +85TSSOP-16RU-161,0002850B25
7" Reel
*Line 1 contains product number, ADN2850, line 2 Top Mark branding contains differentiating detail by part type, line 3 contains lot number, line 4 contains product
date code YYWW.
CAUTION
ESD (electrostatic discharge) sensitive device. Electrostatic charges as high as 4000 V readily
accumulate on the human body and test equipment and can discharge without detection. Although
the ADN2850 features proprietary ESD protection circuitry, permanent damage may occur on
devices subjected to high energy electrostatic discharges. Therefore, proper ESD precautions are
recommended to avoid performance degradation or loss of functionality.
REV. B
–5–
WARNING!
ESD SENSITIVE DEVICE
ADN2850
TOP VIEW
(Not To Scale)
1
2
3
4
5
6
7
8
SDI
SDO
GND
V
1
V
SS
W1
ADN2850BRU
CLK
B1
16
15
14
13
12
11
10
9
CS
PR
WP
V
DD
V
2
W2
B2
RDY
SDO
GND
V
1
2
3
SS
4
V
1
CS
RDY
CLK
SDI
16
15 14 13
ADN2850BCP
CHIP SCALE
PACKAGE
8765
B2
B1
W1
W2
PIN CONFIGURATIONS
12
PR
11
WP
10
V
DD
9
V
2
ADN2850BCP PIN FUNCTION DESCRIPTIONS
Pin
No. Mnemonic Description
1SDOSerial Data Output Pin. Open-Drain output
requires external pull-up resistor. CMD_9 and
CMD_10 activate the SDO output. See
Instruction Operation Truth Table (Table II).
Other commands shift out the previously
loaded SDI bit pattern delayed by 24 clock
pulses. This allows daisy-chain operation of
diode configured transistor
5W1Wiper terminal of RDAC1 ADDR
(RDAC1) = 0
.
H
6B1B terminal of RDAC1
7B2B terminal of RDAC2
8W2Wiper terminal of RDAC2. ADDR
9V
2
(RDAC2) = 1
Log Output Voltage 2 generated from internal
.
H
diode configured transistor
10V
DD
Positive Power Supply Pin
11WPWrite Protect Pin. When active low, WP
prevents any changes to the present register
contents, except PR and CMD_1 and CMD_8
will refresh the RDAC register from EEMEM.
Execute a NOP instruction before returning
to WP high.
12PRHardware Override Preset Pin. Refreshes the
scratch pad register with current contents of
the EEMEM register. Factory default loads
midscale 51210 until EEMEM loaded with
a new value by the user (PR is activated at
the logic high transition).
13CSSerial Register chip select active low.
Serial register operation takes place when
CS returns to logic high.
14RDYReady. Active high open-drain output. Identifies
completion of commands 2, 3, 8, 9, 10, and PR.
15CLKSerial Input Register Clock Pin. Shifts in
16SDISerial Data Input Pin. Shifts in one bit at a time
one bit at a time on positive clock edges.
on positive clock CLK edges. MSB loaded first.
ADN2850BRU PIN FUNCTION DESCRIPTIONS
Pin
No. Mnemonic Description
1CLKSerial Input Register Clock Pin. Shifts in
one bit at a time on positive clock edges.
2SDISerial Data Input Pin. Shifts in one bit at
a time on positive clock CLK edges.
MSB loaded first.
3SDOSerial Data Output Pin. Open-drain out put
requires external pull-up resistor. CMD_9
and CMD_10 activate the SDO output. See
Instruction Operation Truth Table (Table II).
Other commands shift out the previously
loaded SDI bit pattern delayed by 24 clock
pulses. This allows daisy-chain operation of
Command bits are C0 to C3. Address bits are A3–A0. Data bits D0 to D9 are applicable to RDAC wiper register whereas D0 to D15 are applicable to EEMEM
Register. Command instruction codes are defined in Table II.
00000XXX XX • • • • X XX • • • • • • XNOP: Do nothing. See Table XI for Programming
example.
10001000 A0X • • • • X XX • • • • • • XRetrieve contents of EEMEM(A0) to RDAC(A0)
Register. This command leaves device in the Read
Program power state. To return part to the idle state,
perform NOP instruction 0. See Table XI.
20010000 A0X • • • • X XX • • • • • • XSAVE WIPER SETTING: Write contents of RDAC(A0)
to EEMEM(A0). See Table X.
4
3
0011A3 A2 A1 A0D15 • • • • D8D7 • • • • • D0 Write contents of Serial Register Data Bytes 0 and
1 (total 16-bit) to EEMEM(ADDR). See Table XIII.
5
4
0100000 A0X • • • • X XX • • • • • • XDecrement 6 dB: Right shift contents of RDAC(A0)
Register, stops at all “Zeros.”
5
5
0101XXX XX • • • • X XX • • • • • • XDecrement All 6 dB: Right shift contents of all RDAC
Registers, stops at all “Zeros.”
5
6
0110000 A0X • • • • X XX • • • • • • XDecrement contents of RDAC(A0) by “One,” stops
at all “Zeros.”
5
7
0111XXX XX • • • • X XX • • • • • • XDecrement contents of all RDAC Registers by
“One,” stops at all “Zeros.”
81000XXX XX • • • • X XX • • • • • • XRESET: Load all RDACs with their corresponding
EEMEM previously saved values.
91001A3A2A1A0X • • • • X XX • • • • • • XTransfer contents of EEMEM (ADDR) to Serial
Register Data Bytes 0 and 1, and previously stored
data can be read out from the SDO pin. See Table XIV.
101010000A0X • • • • X XX • • • • • • XTransfer contents of RDAC (A0) to Serial Register
Data Bytes 0 and 1, and wiper setting can be read
from the SDO pin. See Table XV.
111011000A0X • • • • D9 D8D7 • • • • • D0 Write contents of Serial Register Data Bytes 0 and
1 (total 11-bit) to RDAC(A0). See Table IX.
12
5
1100000 A0X • • • • X XX • • • • • • X Increment 6 dB: Left shift contents of RDAC(A0),
stops at all “Ones.” See Table XII.
13
5
1101XXX XX • • • • X XX • • • • • • X Increment All 6 dB: Left shift contents of all RDAC
Registers, stops at all “Ones.”
14
5
1110000 A0X • • • • X XX • • • • • • XIncrement contents of RDAC(A0) by “One,” stops
at all “Ones.” See Table X.
15
5
1111XXX XX • • • • X XX • • • • • • XIncrement contents of all RDAC Registers by “One,”
stops at all “Ones.”
NOTES
1
The SDO output shifts out the last 24 bits of data clocked into the serial register for daisy-chain operation. Exception: for any instruction following Instruction 9 or 10,
the selected internal register data will be present in data byte 0 and 1. The instructions following 9 and 10 must also be a full 24-bit data-word to completely clock out
the contents of the serial register.
2
The RDAC register is a volatile scratch pad register that is refreshed at power ON from the corresponding nonvolatile EEMEM register.
3
Execution of the above operations takes place when the CS strobe returns to logic high.
4
Instruction 3 writes 2 data bytes (total 16-bit) to EEMEM. But in the cases of addresses 0 and 1, only the last 10 bits are valid for wiper position setting.
5
The increment, decrement, and shift commands ignore the contents of the shift register data bytes 0 and 1.
REV. B
–7–
ADN2850
OPERATIONAL OVERVIEW
The ADN2850 programmable resistor is designed to operate as
a true variable resistor. The resistor wiper position is determined
by the RDAC register contents. The RDAC register acts as a
scratch pad register which allows unlimited changes of resistance
settings. The scratch pad register can be programmed with any
position setting using the standard SPI serial interface by loading
the 24-bit data-word. The format of the data-word is that the first
4 bits are instructions, the following 4 bits are addresses, and the
last 16 bits are data. Once a specific value is set, this value can be
saved into a corresponding EEMEM register. During subsequent
power-ups, the wiper setting will automatically be loaded at that
value. Saving data to the EEMEM takes about 25 ms and consumes approximately 20 mA. During this time the shift register
is locked, preventing any changes from taking place. The RDY pin
indicates the completion of this EEMEM saving process. There
are also 13 two-bytes addresses, of user defined data that can be
stored in EEMEM.
OPERATION DETAIL
There are 16 instructions that facilitate users’ programming
needs. Referring to Table II, the instructions are:
0. Do Nothing
1. Restore EEMEM setting to RDAC
2. Save RDAC setting to EEMEM
3. Save user data or RDAC setting to EEMEM
4. Decrement 6 dB
5. Decrement all 6 dB
6. Decrement one step
7. Decrement all one step
8. Reset all EEMEM settings to RDAC
9. Read EEMEM to SDO
10. Read Wiper Setting to SDO
11. Write data to RDAC
12. Increment 6 dB
13. Increment all 6 dB
14. Increment one step
15. Increment all one step
Tables VIII to XIV provide a few programming examples by using
some of these instructions.
Scratch Pad and EEMEM Programming
The basic mode of setting the programmable resistor wiper position
(programming the scratch pad register) is done by loading the
serial data input register with the instruction 11, the corresponding
address, and the data. Since the scratch pad register is a standard
logic register, there is no restriction on the number of changes
allowed. When the desired wiper position is determined, the user can
load the serial data input register with the instruction 2, which stores
the setting into the corresponding EEMEM register. The EEMEM
value can be changed at any time or permanently protected by
activating the WP command. Table III provides a programming
example listing the sequence of serial data input (SDI) words and
the corresponding serial data output (SDO) in hexadecimal format.
Table III. Set and Save RDAC with Independent Data
to EEMEM Registers
SDISDOAction
B00100
XXXXXXHLoads data 100H into RDAC1 register,
H
Wiper W1 moves to 1/4 full-scale
position.
20xxxx
H
B00100
Saves copy of RDAC1 register content
H
into corresponding EEMEM1 register.
B10200
H
20xxxx
Loads 200H data into RDAC2 register,
H
Wiper W2 moves to 1/2 full-scale
position.
21xxxx
H
B10200
Saves copy of RDAC2 register contents
H
into corresponding EEMEM2 register.
At system power ON, the scratch pad register is automatically
refreshed with the value previously saved in the corresponding
EEMEM register. The factory preset EEMEM value is midscale.
During operations, the scratch pad register can also be refreshed
with the current contents of the EEMEM registers in three different
ways. First, executing instruction 1 retrieves the corresponding
EEMEM value. Second, executing instruction 8 resets the EEMEM
values of both channels. Finally, pulsing the PR pin also refreshes
both EEMEM settings. Operating the hardware control PR
function, however, requires a complete pulse signal. When PR
goes low, the internal logic sets the wiper at midscale. The
EEMEM value will not be loaded until PR returns to high.
EEMEM Protection
The write-protect (WP) disables any changes of the scratch pad
register contents regardless of the software commands, except
that the EEMEM setting can be refreshed and can overwrite the
WP by using commands 1, 8, and PR pulse. To disable WP, it is
recommended to execute a NOP command before returning
WP to logic high.
Linear Increment and Decrement Commands
The increment and decrement commands (14, 15, 6, 7) are useful
for linear step adjustment applications. These commands simplify
microcontroller software coding by allowing the controller to
just send an increment or decrement command to the device. The
adjustment can be individually or gang controlled. For increment command, executing instruction 14 will automatically move the
wiper to the next resistance segment position. The master increment
instruction 15 will move all resistor wipers up by one position.
Logarithmic Taper Mode Adjustment (6 dB/step)
There are four programming instructions which provide the
logarithmic taper increment and decrement wiper position control by either individual or gang control. 6 dB increment is
activated by instructions 12 and 13 and 6 dB decrement is activated by instructions 4 and 5. For example, starting at zero
scale, executing 11 times the increment instruction 12 will move
the wiper in 6 dB per step from the 0% of the full-scale R
the full-scale R
. The 6 dB increment instruction doubles the
WB
WB
to
value of the RDAC register contents each time the command is
executed. When the wiper position is near the maximum setting,
the last 6 dB increment instruction will cause the wiper to go to
the full-scale 1023-code position. Further 6 dB per increment
instruction will no longer change the wiper position beyond its
full-scale, Table IV.
6 dB step increment and decrement are achieved by shifting the bit
internally to the left and right, respectively. The following information explains the nonideal ±6dB step adjustment at certain
REV. B–8–
ADN2850
conditions. Table IV illustrates the operation of the shifting
function on the individual RDAC register data bits. Each line
going down the table represents a successive shift operation. Note
that the left shift 12 and 13 commands were modified such that
if the data in the RDAC register is equal to zero, and the data is
left shifted, the RDAC register is then set to code 1. Similarly, if the
data in the RDAC register is greater than or equal to midscale,
and the data is left shifted, then the data in the RDAC register is
automatically set to full scale. This makes the left shift function
as ideal a logarithmic adjustment as possible.
The right shift 4 and 5 commands will be ideal only if the LSB is
zero (i.e., ideal logarithmic—no error). If the LSB is a one, then
the right shift function generates a linear half LSB error, which
translates to a number of bits-dependent logarithmic error as
shown in Figure 3. The plot shows the error of the odd numbers
of bits for ADN2850.
Table IV. Detail Left and Right Shift Functions for 6 dB
Step Increment and Decrement
Actual conformance to a logarithmic curve between the data contents in the RDAC register and the wiper position for each right
shift 4 and 5 command execution contains an error only for odd
numbers of bits. Even numbers of bits are ideal. The graph in
Figure 3 shows plots of Log_Error [i.e., 20 log
ADN2850. For example, code 3 Log_Error = 20 log
(error/code)]
10
(0.5/3)
10
= –15.56 dB, which is the worst case. The plot of Log_Error is
more significant at the lower codes.
0
–20
–40
dB
Using Additional Internal Nonvolatile EEMEM
The ADN2850 contains additional internal user storage registers
(EEMEM) for saving constants and other 16-bit data. Table V
provides an address map of the internal storage registers shown
in the functional block diagram as EEMEM1, EEMEM2, and
and 26 bytes (13 addresses 2 bytes each) of USER EEMEM.
Table V. EEMEM Address Map
EEMEM
NumberAddressEEMEM Content For
10000RDAC1
20001RDAC2
30010USER1
1, 2
3
40011USER2
:::
151110USER13
161111% Tolerance
NOTES
1
RDAC data stored in EEMEM locations are transferred to their corresponding
RDAC REGISTER at power-on, or when instructions 1, 8, and PR are executed.
2
Execution of instruction 1 leaves the device in the read mode power consumption
state. After the last instruction 1 is executed, the user should perform a NOP,
instruction 0 to return the device to the low power idling state.
3
USER <data> are internal nonvolatile EEMEM registers available to store and
retrieve constants and other 16-bit information using instructions 3 and 9 respectively.
4
Read only.
4
Calculating Actual Full-Scale Resistance
The actual tolerance of the rated full-scale resistance R
WB1
is
stored in EEMEM register 15 during factory testing. The actual
full-scale resistance can therefore be calculated, which will be
valuable for tolerance matching or calibration. Notice this value
is read only, and the full-scale resistance of R
R
of typically 0.1%.
WB1_FS,
WB2_FS
matches
The tolerance in % is stored in the last 16 bits of data in EEMEM
register 15. The format is sign magnitude binary format with the
MSB designates for sign (0 = positive and 1 = negative), the next
7 MSB designate for the integer number, and the 8 LSB designate
for the decimal number. See Table VI.
Table VI. Tolerance in % from Rated Full-Scale Resistance
Figure 3. Plot of Log_Error Conformance for Odd
Numbers of Bits Only (Even Numbers of Bits Are Ideal)
REV. B
–9–
ADN2850
Daisy-Chain Operation
The serial data output pin (SDO) serves two purposes. It can be
used to read out the contents of the wiper settings or EEMEM
values using instructions 10 and 9 respectively. If these instructions are not used, SDO can be used for daisy-chaining multiple
devices in simultaneous operations (see Figure 4). The SDO pin
contains an open-drain N-Ch FET and requires a pull-up resistor if SDO function is used. Users need to tie the SDO pin of
one package to the SDI pin of the next package. Users may need
to increase the clock period because the pull-up resistor and the
capacitive loading at the SDO-SDI interface may induce time
delay to the subsequent devices (see Figure 4). If two ADN2850s
are daisy-chained, a total 48 bits of data is required. The first
24 bits (formatted 4-bit instruction, 4-bit address, and 16-bit
data) go to U2 and the second 24 bits with the same format go
to U1. The CS should be kept low until all 48 bits are clocked into
their respective serial registers. The CS is then pulled high to
complete the operation.
V
DD
SCLK
C
MOSI
SS
ADN2850
U1
SDISDO
CLK
R
2.2k
P
SDISDO
ADN2850
U2
CLKCSCS
PRWP
CLK
CS
SDI
VA LI D
COMMAND
COUNTER
COMMAND
PROCESSOR
AND ADDRESS
DECODE
SERIAL
REGISTER
ADN2850
5V
SDO
GND
R
PULLUP
Figure 5. Equivalent Digital Input-Output Logic
V
DD
INPUTS
LOGIC
PINS
300
GND
Figure 6a. Equivalent ESD Digital Input Protection
Figure 4. Daisy-Chain Configuration
DIGITAL INPUT/OUTPUT CONFIGURATION
All digital inputs are ESD protected. Digital inputs are high
impedance and can be driven directly from most digital sources.
Active at logic low, PR and WP should be biased to V
if they
DD
are not used. There are no internal pull-up resistors present on
any digital input pins. To avoid floating digital pins that may
cause false triggering in a noisy environment, pull-up resistors
should be added to these pins. However, this only applies to the
case where the device will be detached from the driving source
once it is programmed.
The SDO and RDY pins are open-drain digital outputs. Similarly,
pull-up resistors are needed if these functions are used. To optimize
the speed and power trade-off, use 2.2 kΩ pull-up resistors.
The equivalent serial data input and output logic is shown in
Figure 5. The open-drain output SDO is disabled whenever
chip select CS is logic high. ESD protection of the digital inputs
is shown in Figures 6a and 6b.
V
DD
INPUT
WP
300
GND
Figure 6b. Equivalent WP Input Protection
SERIAL DATA INTERFACE
The ADN2850 contains a 4-wire, SPI compatible, digital interface (SDI, SDO, CS, and CLK). The 24-bit serial word must be
loaded with MSB first, and the format of the word is shown in
Table I. The Command Bits (C0 to C3) control the operation of
the programmable resistor according to the instruction shown
in Table II. A0 to A3 are assigned for address bits. A0 is used to
address RDAC1 or RDAC2. Addresses 2 to 14 are accessible by
users. Address 15 is reserved for the factory. Table V provides an
address map of the EEMEM locations. The data bits (D0 to D9) are
the values that are loaded into the RDAC registers at instruction 11. The data bits (D0 to D15) are the values that are loaded
into the EEMEM registers at instruction 3.
The last instruction prior to a period of no programming activity
should be applied with the No Operation (NOP), instruction 0. It
is recommended to do so to ensure minimum power consumption
in the internal logic circuitry
The SPI interface can be used in two slave modes, CPHA = 1,
CPOL = 1 and CPHA = 0, CPOL = 0. CPHA and CPOL refer to
the control bits that dictate SPI timing in these microconverters
and microprocessors: ADuC812/ADuC824, M68HC11,
and MC68HC16R1/916R1.
REV. B–10–
ADN2850
SW(1)
SW(0)
SWB
B
R
S
R
S
SW(2
N
–
1)
W
SW(2
N
–
2)
RDAC
WIPER
REGISTER
AND
DECODER
R
S = RWB
/2
N
R
S
DIGITAL
CIRCUITRY
OMITTED FOR
CLARITY
TERMINAL VOLTAGE OPERATING RANGE
The ADN2850 positive VDD and negative VSS power supply
defines the boundary conditions for proper two-terminal programmable resistance operation. Supply signals present on terminals W
and B that exceed V
or VSS will be clamped by the internal
DD
forward biased diodes (see Figure 7).
V
DD
W
B
V
SS
Figure 7. Maximum Terminal Voltages Set by VDD and V
The ground pin of the ADN2850 device is primarily used as a digital
ground reference that needs to be tied to the PCB’s common
ground. The digital input control signals to the ADN2850 must
be referenced to the device ground pin (GND), and satisfy the
logic level defined in the Specifications table of this data sheet.
An internal level shift circuit ensures that the common-mode
voltage range of the two terminals extends from V
SS
to V
regardless of the digital input level. In addition, there is no
polarity constraint on voltage across terminals W and B. The
magnitude of |V
| is bounded by V
WB
DD
– VSS.
Power-Up Sequence
Since diodes limit the voltage compliance at terminals B and W
(see Figure 7) it is important to power V
first before apply-
DD/VSS
ing any voltage to terminals B and W. Otherwise, the diode will be
forward biased such that V
For example, applying 5 V across V
will be powered unintentionally.
DD/VSS
will cause the VDD terminal
DD
to exhibit 4.3 V. Although it is not destructive to the device, it may
affect the rest of the user’s system. As a result, the ideal power-up
sequence is in the following order: GND, V
and V
. The order of powering VB, VW, and Digital Inputs is not
B/W
important as long as they are powered after V
, VSS, Digital Inputs,
DD
.
DD/VSS
Regardless of the power-up sequence and the ramp rates of the
power supplies, once V
are powered, the power-on reset
DD/VSS
remains effective, which retrieves EEMEM saved values to the
RDAC registers (see TPC 7).
Layout and Power Supply Bypassing
It is a good practice to employ compact, minimum-lead length
layout design. The leads to the input should be as direct as possible with a minimum of conductor length. Ground paths should
have low resistance and low inductance. To minimize the digital
ground bounce, the digital signal ground reference can be joined
remotely to the analog ground terminal of the ADN2850.
Similarly, it is also a good practice to bypass the power supplies
with quality capacitors for optimum stability. Supply leads to the
device should be bypassed with 0.01 µFto0.1 µF disc or chip
ceramics capacitors. Low ESR 1 µF to 10 µF tantalum or electro-
lytic capacitors should also be applied at the supplies to minimize
any transient disturbance (see Figure 8).
REV. B
DD
ADN2850
V
DD
V
SS
10F
10F
+
C3
C1
0.1F
+
C4
C2
0.1F
V
DD
V
SS
GND
Figure 8. Power Supply Bypassing
RDAC STRUCTURE
The patent-pending RDAC contains a string of equal resistor
segments, with an array of analog switches, that act as the wiper
connection. The number of positions is the resolution of the
device. The ADN2850 has 1024 connection points, allowing it to
provide better than 0.1% setability resolution. Figure 9 shows an
equivalent structure of the connections between the two terminals
that make up one channel of the RDAC. The S
ON, while one of the switches SW(0) to SW(2
SS
one at a time depending on the resistance position decoded from
will always be
WB
N
– 1) will be ON
the data bits. Since the switch is not ideal, there is a 50 Ω wiper
resistance, R
. Wiper resistance is a function of supply voltage
W
and temperature. The lower the supply voltage or the higher the
temperature, the higher the resulting wiper resistance. Users
should be aware of the wiper resistance dynamics if accurate
prediction of the output resistance is needed.
Figure 9. Equivalent RDAC Structure
Table VII. Nominal Individual Segment Resistor Values
Device Resolution25 kΩ250 kΩ
1024-Step24.4244
CALCULATING THE PROGRAMMABLE RESISTANCE
The nominal full-scale resistance of the RDAC between terminals
W and B, R
, is available with 25 kΩ and 250 kΩ with 1024
WB_FS
positions (10-bit resolution). The final digits of the part number
determine the nominal resistance value, e.g., 25 kΩ = 25 and
250 kΩ = 250.
The 10-bit data-word in the RDAC latch is decoded to select one
of the 1024 possible settings. The following discussion describes
the calculation of resistance R
(D) at different codes of a 25 kΩ
WB
part. The wiper’s first connection starts at the B terminal for
data 000
. RWB(0) is 50 Ω because of the wiper resistance and it
H
is independent of the full-scale resistance. The second connection
is the first tap point where R
(1) becomes 24.4 Ω + 50 = 74.4 Ω
WB
–11–
ADN2850
for data 001H. The third connection is the next tap point representing R
(2) = 48.8 + 50 = 98.8 Ω for data 002H and so on. Each
WB
LSB data value increase moves the wiper up the resistor ladder
until the last tap point is reached at R
(1023) = 25026 Ω. See
WB
Figure 9 for a simplified diagram of the equivalent RDAC circuit.
25
20
15
(D) – k
WB
10
R
R
WB_FS
5
0
01023256
= 25k
512768
CODE – Decimal
Figure 10. RWB(D) vs. Code
The general equation that determines the programmed output
resistance between Wx and Bx is:
RD
()
WBWB FSW
D
=× +
1024
RR
_
(1)
where D is the decimal equivalent of the data contained in the
RDAC register, R
W and B, and R
is the full-scale resistance between terminals
WB_FS
is the wiper resistance.
W
For example, the following output resistance values will be set for
the following RDAC latch codes with V
Note that in the zero-scale condition a finite wiper resistance of
50 Ω is present. In this state, care should be taken to limit the
current flow between W and B to no more than 20 mA to avoid
degradation or possible destruction of the internal switches.
Channel-to-channel R
scale. The change in R
matching is well within 1% at full-
WB
with temperature has a 35 ppm/°C
WB
temperature coefficient.
REV. B–12–
Typical Performance Characteristics–ADN2850
CODE
36
16
012002004006008001000
34
28
22
20
18
32
30
26
24
OHMS
TEMPERATURE – C
4
–40–20020100
CURRENT – A
40
3
2
1
0
–1
6080
ISS @ VDD/VSS = 2.7V/0V
I
DD
@ VDD/VSS = 2.7V/0V
I
SS
@ VDD/VSS = 5V/0V
I
DD
@ VDD/VSS = 5V/0V
0.25
FREQUENCY – Hz
0.0E+00
I
DD
– mA
0
0.20
0.15
0.10
0.05
2.0E+06 4.0E+06 6.0E+06 8.0E+06 1.0E+07 1.2E+07
MIDSCALE
FULL SCALE
ZERO SCALE
VDD/VSS = 5V/0V
R
AR
= 25k
1.0
0.8
0.6
0.4
0.2
0
R-INL ERROR – LSB
–0.2
–0.4
–0.6
0200400600800
DIGITAL CODE
TPC 1. R-INL vs. Code, TA = 40C, 25C,
85
C Overlay, RAB = 25 k
0.4
0.2
0
Ω
+25C
–40C
+85C
1000
TPC 4. Wiper On-Resistance vs. Code
–0.2
–0.4
R-DNL ERROR – LSB
–0.6
–0.8
02004006001000
DIGITAL CODE
TPC 2. R-DNL vs. Code, TA = 40C, 25C,
85
C Overlay, RAB = 25 k
120
100
C
80
60
40
20
0
REV. B
–20
–40
RHEOSTAT MODE TEMPCO – ppm/
–60
–80
TPC 3.∆RWB/∆T Rheostat Mode Tempco
25k VERSION
250k VERSION
3842561280
CODE – Decimal
Ω
VDD/VSS = 5.0V/0V
T
640512
= 25C
A
768
800
896
1023
–13–
TPC 5. IDD vs. Temperature, RAB = 25 k
Ω
TPC 6. IDD vs. Clock Frequency, RAB = 25 k
Ω
ADN2850
W
B
NC
I
W
DUT
V
MS
NC = NO CONNECT
A
DUT
V
SS
ICM
W
B
V
DD
NC
NC
VCM
GND
A
NC = NO CONNECT
IW = IA
= 25C
T
A
0.5V/DIV
RWB(D)
EXPECTED
MIDSCALE
NORMALIZED RESISTANCE
VA LU E
50S/DIV
TPC 7. Memory Restore During Power-On Reset
5V/DIV
CS
5V/DIV
CLK
V
5V/DIV
SDI
I
DD
20mA/DIV
4ms/DIV
TPC 8. IDD vs. Time (Save) Program Mode
100
TA = 25C
10
– mA
WB_MAX
1
0.1
THEORETICAL – I
0.01
TPC 10. I
R
= 25k
WB_FS
R
WB_FS
CODE – Decimal
vs. Code
WB_MAX
= 250k
8967686405123841282560
1024
TEST CIRCUITS
Test Circuits 1 to 3 show some of the test conditions used in the
Specifications table.
5V/DIV
CS
5V/DIV
CLK
SDI
5V/DIV
I
DD
2mA/DIV
SUPPLY CURRENT RETURNS TO MINIMUM POWER
CONSUMPTION IF INSTRUCTION 0 (NOP) IS
EXECUTED IMMEDIATELY AFTER INSTRUCTION 1
(READ EEMEM)
4ms/DIV
TPC 9. IDD vs. Time (Read) Program Mode
Test Circuit 1. Resistor Position Nonlinearity
Error (Rheostat Operation; R-INL, R-DNL)
0.1V
R
=
SW
DUT
B
W
ISW
CODE = 00
TO V
V
SS
I
SW
H
+
0.1V
_
DD
Test Circuit 2. Incremental ON Resistance
Test Circuit 3. Common-Mode Leakage Current
REV. B–14–
ADN2850
PROGRAMMING EXAMPLES
The following programming examples illustrate the typical sequence
of events for various features of the ADN2850. Users should refer
to Table II for the instructions and data-word format. The instruction numbers, addresses, and data appearing at SDI and SDO pins
are displayed in hexadecimal format in the following examples.
Table IX. Scratch Pad Programming
SDISDOAction
B00100
XXXXXXHLoads data 100H into RDAC1 register,
H
Wiper W1 moves to 1/4 full-scale
position.
B10200
H
B00100
Loads data 200H into RDAC2 register,
H
Wiper 2 moves to 1/2 full-scale position.
Table X. Incrementing RDAC Followed by Storing
the Wiper Setting to EEMEM
SDISDOAction
B00100
XXXXXXHLoads data 100H into RDAC1 register,
H
Wiper W1 moves to 1/4 full-scale position.
E0XXXX
E0XXXX
B00100HIncrements RDAC1 register by one to 101H.
H
E0XXXXHIncrements RDAC1 register by one to 102H.
H
Repeat the increment command –
(E0XXXX
until desired wiper
H)
position is reached
20XXXX
XXXXXXHSaves RDAC1 data into EEMEM1
H
Optionally tie WP to GND to protect
EEMEM values
Table XII. Using Left Shift by One to Increment 6 dB Steps
SDISDOAction
C0XXXX
XXXXXXHMoves wiper 1 to double the present
H
data contained in RDAC1 register.
C1XXXX
C0XXXXHMoves wiper 2 to double the present
H
data contained in RDAC2 register.
Table XIII. Storing Additional User Data in EEMEM
SDISDOAction
32AAAA
XXXXXXHStores data AAAAH into spare EEMEM
H
location USER1. (Allowable to address
in 13 locations with maximum 16 bits
of data).
335555
32AAAAHStores data 5555H into spare EEMEM
H
location USER2. (Allowable to address
in 13 locations with maximum 16 bits
of data).
Table XIV. Reading Back Data From Various Memory Locations
SDISDOAction
92XXXX
XXXXXXHPrepares data read from USER1
H
location.
00XXXX
92AAAAHNOP instruction 0 sends 24-bit word
H
out of SDO where the last 16 bits
contain the contents of USER1 location.
NOP command ensures device returns
to idle power dissipation state.
Table XI. Restoring EEMEM Values to RDAC Registers
EEMEM values for RDACs can be restored by: Power-On,
Strobing PR pin or Programming shown below.
SDISDOAction
10XXXX
XXXXXXHRestores EEMEM1 value to RDAC1
H
register.
00XXXX
H
100100
NOP. Recommended step to minimize
H
power consumption.
8XXXXX
00XXXXHReset EEMEM1 and EEMEM2
H
values to RDAC1 and RDAC2 registers
respectively.
Table XV. Reading Back Wiper Setting
SDISDOAction
B00200
C0XXXX
XXXXXXHSets RDAC1 to midscale.
H
B00200
H
Doubles RDAC1 from midscale to
H
full-scale.
A0XXXX
C0XXXXHPrepares reading wiper setting from
H
RDAC1 register.
XXXXXX
A003FFHReadback full-scale value from RDAC1
H
register.
Analog Devices offers a user-friendly ADN2850EVAL evaluation
kit that can be controlled by a personal computer through the printer
port. The driving program is self-contained, so no programming
languages or skills are needed.
REV. B
–15–
ADN2850
APPLICATIONS
Optical Transmitter Calibration with ADN2841
Together with the multirate 2.7 Gbps Laser Diode Driver ADN2841,
the ADN2850 forms an optical supervisory system where the dual
programmable resistors are used to set the laser average optical
power and extinction ratio (see Figure 11). The ADN2850 is
particularly ideal for the optical parameter settings because of its
high resolution, compact footprint, and superior temperature
coefficient characteristics.
The ADN2841 is a 2.7 Gbps laser diode driver that uses a unique
control algorithm to manage both the laser average power and
extinction ratio after the laser initial factory calibration. It stabilizes
the laser data transmission by continuously monitoring its optical
power, and correcting the variations caused by temperature and
the laser degradation over time. In the ADN2841, the I
the laser diode current. Through its dual-loop power and extinction
ratio control, calibrated by the ADN2850, the internal driver
controls the bias current I
It also regulates the modulation current I
and consequently the average power.
BIAS
by changing the
MODP
modulation current linearly with slope efficiency. Any changes in
the laser threshold current or slope efficiency are therefore compensated. As a result, this optical supervisory system minimizes the
laser characterization efforts and enables designers to apply comparable lasers from multiple sources.
Incoming Optical Power Monitoring
The ADN2850 comes with a pair of matched diode connected
PNPs, Q
and Q2, that can be used to configure an incoming optical
1
power monitoring function. With a reference current source, an
instrumentation amplifier, and a logarithmic amplifier, this feature
can be used to monitor the optical power by knowing the dc
average photodiode current from the following relationships:
I
V=V =V
1 BE1T
V=V =VI
2 BE2 T
In
CS1
I
1
I
CS2
n
I
2
monitors
MPD
(2)
(3)
V
CC
I
ADN2841
DIN
DINQ
MPD
I
MODP
IDTONE
I
BIAS
CLK
SDI
CS
ADN2850
CONTROL
EEMEM
EEMEM
RDAC1
RDAC2
W1
B1
W2
B2
DIN
DINQ
IDTONE
PSET
ERSET
Figure 11. Optical Supervisory System
Knowing IC1 = a1 I
therefore a and I
are matched. Combining Equations 2 and 3
S
PD, IC2
= a2 I
and Q1–Q2 are matched,
REF,
theoretically yields:
I
V–V=VIn
REF
T21
I
PD
Where IS1 and IS2 are saturation current
, V2 are VBE, base-emitted voltages of the diode connector
V
1
transistors
is the thermal voltage, which is equal to k × T/q.
V
T
= 26 mV at 25°C
V
T
k = Boltzmann’s constant = 1.38E–23 Joules/Kelvin
q = electron charge = 1.6E–19 coulomb
T = temperature in Kelvin
= photodiode current
I
PD
= reference current
I
REF
Figure 12 shows such a conceptual circuit.
V
CC
(4)
POST
V
DD
V
SS
ADN2850
W
1
B
1
AMP
0.75 BIT RATE
AD623
G
IN AMP
LOG AMP
LPF
(1 + 100k/RG) (V2 – V1)
VT COMPENSATION
C
PRC
THERMISTOR
TIA
10nF
I
V
W
1
2
Q
1
B
2
GND
–5V
I
PD
REF
R
V
2
Q
2
CDR
DATA
CLOCK
LOG
AVERAGE
POWER
Figure 12. Conceptual Incoming Optical Power Monitoring Circuit
REV. B–16–
The output voltage represents the average incoming optical power.
The output voltage of the log stage does not have to be accurate
from device to device, as the responsivity of the photodiode will
change between devices. An op amp stage is shown after the log
amp stage, which compensates for V
variation over temperature.
T
Equation 4 is ideal. If the reference current is 1 mA at room
temperature, characterization shows that there is an additional
30 mV offset between V
VV=0.026In
2 —1
and V1. A curve fit approximation yields
2
0.001
×
+
003.
I
PD
(5)
Such offset is believed to be caused by the transistors self-heating
and the thermal gradient effect. As seen in Figure 13, the error
between an approximation and the actual performance ranges is
less than 0% to –4% from 0.1 mA to 0.1 A.
0.30
0.25
0.20
– V
1
0.15
– V
2
V
0.10
0.05
DEVICE 1
DEVICE 2
DEVICE 3
CURVE FIT
I
T
REF
A
= 1mA
= 25C
ERROR
12
9
6
3
0
APPROXIMATING ERROR – %
–3
ADN2850
B2
51200
W2
R
(6)
W1
B1
Figure 14. Reduce Resistance by Half with Linear
Adjustment Characteristics
Much lower resistance can also be achieved by paralleling a
discrete resistor as shown in Figure 15.
W1
B1
Figure 15. Resistor Scaling with Pseudo-Log Taper
Adjustment Characteristics
The equivalent resistance at a given setting is approximated as:
DR
×+
R=
eq
×++×
WB FS_
WB_FS
51200 1024DRR
In this approach, the adjustment is not linear but pseudologarithmic. Users should be aware of the need for tolerance matching
as well as temperature coefficient matching of the components.
BASIC RDAC SPICE MODEL
0
1.E-071.E-061.E-051.E-041.E-03
IPD – A
Figure 13. Typical V2 – V1 vs. I
PD
at I
REF
–6
= 1 mA
and TA = 25°C
Resistance Scaling
The ADN2850 offers either 25 kΩ or 250 kΩ full-scale resistance.
Users who need lower resistance and still maintain the numbers
of step adjustment can parallel two or more devices. Figure 14
shows a simple scheme of paralleling both channels of the programmable resistors. In order to adjust half of the resistance
linearly per step, users need to program both devices coherently
with the same settings. Note that since the devices will be programmed one after another, an intermediate state will occur, and
this method may not be suitable for certain applications.
RDAC
25k
CW = 80pF
W
C
B
B
= 11pF
Figure 16. RDAC Circuit Simulation Model (RDAC = 25 kΩ)
The internal parasitic capacitances and the external capacitive
loads dominate the ac characteristics of the RADCs. A general
parasitic simulation model is shown in Figure 16.
Listing I provides a macro model net list for the 25 kΩ RDAC: