SIPEX SP9842

SP9841/42
8-Bit Octal, 2-Quadrant Multiplying, BiCMOS DAC
■ Replaces 8 Potentiometers and 8 Op Amps
■ Operates from Single +5V Supply
■ 6.3 MHz 2-Quadrant Multiplying Gain Bandwidth
■ No Signal Inversion
■ Eight Reference Inputs, Eight Voltage
Outputs (SP9841)
■ Four Reference Inputs, Eight Voltage
Outputs (SP9842)
■ 3-Wire Serial Input
■ 0.8MHz Data Update Rate
■ +3.25 Volt Output Swing
■ Midscale Preset
■ Low 65 mW Power Dissipation (8mW/DAC)
DESCRIPTION…
The SP9841 and SP9842 are general purpose octal DACs in a single package. The SP9841
features eight individual reference inputs, while the SP9842 provides four pair of voltage
reference inputs. Both parts feature 6.3MHz bandwidth, two–quadrant multiplication, and a
three–wire serial interface. Other features include midscale preset, no signal inversion and low
power dissipation from a single +5V supply. Devices are available in commercial and industrial
temperature ranges.
DAC A
+
8
–
VINA
VOUTA
Data
Clock
Serial Data Input
Serial Data Output
SP9842 Block Diagram
SERIAL
REGISTER
8
8x8
DAC
REGISTER
Preset
Load
4
DAC A
+
8
VOUTA
–
LOGIC
VINH
8
DAC H
8
Data
Clock
Serial Data Input
Serial Data Output
SERIAL
REGISTER
8
Preset
Load
DAC B
8x8
DAC
REGISTER
+
8
–
VINA/B
VOUTB
Decoded
Address
SP9841
VOUTH
VREF Low
SP9841 Block Diagram
4
LOGIC
8
DAC G
+
8
–
VOUTG
Decoded
Address
DAC H
SP9842
+
8
–
VING/H
VOUTH
VREF Low
267
ABSOLUTE MAXIMUM RATINGS
These are stress ratings only and functional operation of the device
at these or any other above those indicated in the operation
sections of the specifications below is not implied. Exposure to
absolute maximum rating conditions for extended periods of time
may affect reliability.
CAUTION:
While all input and output pins have internal protection networks, these parts should
be considered ESD (ElectroStatic Discharge) sensitive devices. Permanent damage may occur on unconnected devices subject to high energy electrostatic fields. Unused devices must be stored in conductive
foam or shunts. Personnel should be properly grounded prior to handling this device.
The protective foam should be discharged to
the destination socket before devices are removed.
VDD to GND ...................................................................... -0.3V, +7V
VINX to GND ............................................................................... VDD
VREFL to GND ............................................................................. VDD
VOUTX to GND ............................................................................ VDD
Short Circuit IOUTX to GND ............................................ Continuous
Digital Input & Output Voltage to GND ....................................... VDD
Operating Temperature Range
Commercial: SP9841K/SP9842K .............................. 0°C to +70°C
Extended Industrial: SP9841B/SP9842B ................ -40°C to +85°C
Maximum Junction Temperature (TJ max) .......................... +150°C
Storage Temperature ................................................. -65° to 150°C
Lead Temperature (Soldering, 10 sec) ............................... +300°C
Package Power Dissipation ................................. (TJ max - TA)/∅JA
Thermal Resistance ∅JA
P-DIP .................................................................................. 57°C/W
SOIC-24 .............................................................................. 70°C/W
SPECIFICATIONS
(VDD = +5V, All VINX= +1.625V, VREFL = 0V, TA = 25° C for commercial–grade parts; TMIN ≤ TA = TMAX for industrial–grade parts; specifications apply to
all DAC's unless noted otherwise.)
PARAMETER
SIGNAL INPUTS
Input Voltage Range
Input Resistance
SP9841
SP9842
Input Capacitance
SP9841
SP9842
VREFL Resistance
VREFL Capacitance
DIGITAL INPUTS
Logic High
Logic Low
Input Current
Input Capacitance
Input Coding
STATIC ACCURACY
Resolution
Integral Nonlinearity
Differential Nonlinearity
Half-Scale Output Voltage
Zero-Scale Output Voltage
Output Voltage Drift
DYNAMIC PERFORMANCE
Multiplying Gain Bandwidth
Slew Rate
Positive
Negative
Total Harmonic Distortion
MIN.
0
5
2.5
268
MAX.
1.625
10
5
UNITS
V
kΩ
kΩ
CONDITIONS
VREFL = GND, VDD = 4.75V
D = 55H; Code Dependent
Code Dependent
0.375
19
38
0.75
190
30
60
250
pF
pF
kΩ
pF
0.8
±10
8
V
V
µA
pF
2.4
All D = ABH; Code Dependent
Code Dependent
Binary
1.600
8
±0.25
±0.2
1.625
20
25
±1.0
±1.0
1.650
100
Bits
LSB
LSB
V
mV
µV/°C
4
6.3
MHz
3.0
–3.0
7.9
–8.3
0.005
V/µs
V/µs
%
Output Settling Time
Crosstalk
Digital Feedthrough
Wideband Noise
TYP.
60
0.7
µs
70
6
42.5
dB
nVs
µV rms
Note 1
Note 1
PR = LOW, Sets D = 80H
D = 00H
PR = LOW, Sets D = 80H
VINX = 100 mV p-p+ 1.0V dc
Measured 10% to 90%
VOUTX = 100mV to +3.1V
VOUTX = +3.1V to 100mV
VINX = 0.8VDC + 1.4V p-p
D= FFH; 1kHz, fLP = 80 kHz
±1 LSB Error Band, 8H to
255H
Note 2
VREFL = +1.625V, D = 0 to FFH
VOUT = 3.25V; 400Hz to 80kHz
SPECIFICATIONS (continued)
(VDD = +5V, All VINX= +1.625V, VREFL = 0V, TA = 25° C for commercial–grade parts; TMIN ≤ TA = TMAX for industrial–grade parts; specifications apply to
all DAC's unless noted otherwise.)
PARAMETER
DYNAMIC PERFORMANCE
SINAD
MIN.
Digital Crosstalk
DAC OUTPUTS
Voltage Range
Output Current
0
±10
TYP.
MAX.
85
dB
6
nVs
±15
VDD–1.5
Capacitive Load
47,000
DIGITAL OUTPUT
Logic High
3.5
Logic Low
0.4
POWER REQUIREMENTS
Power Supply Range
4.75
5.00
5.25
Positive Supply Current
13
Power Dissipation
65
ENVIRONMENTAL AND MECHANICAL
Operating Temperature Range
Commercial
0
+70
Industrial
–40
+85
Storage Temperature Range
–65
+150
Package
SP9841N
24–pin Plastic DIP
SP9841S
24–pin SOIC
SP9842S
20–pin SOIC
Notes:
1
2
3
UNIT
V
mA
pF
V
V
V
mA
mW
CONDITIONS
VINX = 0.8VDC + 1.4V p-p
D= FFH; 1kHz, fLP = 80 kHz
SP9842 only; measured
between adjacent channels of
same pair; D = 7FH to 80H
RL = 5kΩ; VDD = 4.75V
∆VOUT < 10mV, VINX=1.625V,
PR = LOW
No Oscillation
IOH = -0.4mA
IOL = 1.6mA
To rated specifications
PR = LOW
°C
°C
°C
Note 3
The op amp limits the linearity for VOUT ≤ 100mV. When VREFL is driven above ground such that the
output voltage remains above 100mV, then the linearity specifications apply to all codes. For VREFL =
GND, VIN = 1.5V, codes 0 through 7 are not included in differential or integral linearity tests. Integral
and differential linearity are computed with respect to the best fit straight line through codes 8
through 255.
SP9841 is measured between adjacent channels, f = 100kHz; SP9842 is measured between
adjacent pairs, f = 100kHz.
For plastic DIP packaging of SP9842, please consult factory.
269
Plot 1. Integral Linearity Error versus Code.
Plot 2. Differential Non–linearity Error versus Code.
Plot 3. Integral Linearity Matching; VOUT A through VOUT D.
270
Plot 4. Integral Linearity Matching; VOUT E through VOUT H.
Plot 5. THD versus Frequency.
Plot 6. PSRR versus Frequency.
271
Plot 7. Small Signal Gain versus Frequency.
VDD = 5V
VIN = 0.05V to 1.55V
VOUT = 0.1V to 3.1V
Plot 8. Full Scale Pulse Response.
VDD = 5V
VIN = 0.05V to 1.55V
VOUT = 0.1V to 3.1V
Plot 9. Positive Full Scale Settling.
272
VDD = 5V
VIN = 0.05V to 1.55V
VOUT = 0.1V to 3.1V
Plot 10. Negative Full Scale Settling.
Plot 11. VIN(X) Current versus Code.
Plot 12. IREFL Current Input Current versus Code.
273
Plot 13. Typical Midscale Output versus Temperature.
Plot 14. Supply Current versus Temperature.
Plot 15. Output Short Circuit Current versus VOUT(X).
274
Plot 16. Sink Current at Zero Scale.
Plot 17. Typical VOUTmax versus VDD.
Plot 18. Typical VOUT min versus VDD versus ISINK.
275
Plot 19. Integral Error versus Code versus VDD; VIN(X) = 0.5V.
Plot 20. Integral Error versus Code versus VDD; VIN(X) = 0.6V.
276
a)
b)
c)
d)
Plot 21. Pulse Response — a) CLOAD = 470pF, RLOAD = 10MOhm; b) CLOAD = 470pF, RLOAD = 1kOhm; c) 50Ohms in series
with CLOAD = 470pF; d) RLOAD = 1kOhm, 50Ohms in series with CLOAD = 470pF.
277
a)
b)
c)
d)
Plot 22. Pulse Response — a) CLOAD = 4,700pF; b) CLOAD = 4,700pF, RLOAD = 1kOhm; c) 30 Ohms in series with CLOAD =
4,700pF; d) RLOAD = 1kOhm, 30Ohms in series with CLOAD = 4,700pF.
278
a)
b)
c)
d)
Plot 23. Pulse Response — a) CLOAD = 47,000pF; b) CLOAD = 47,000pF, RLOAD = 1kOhm; c) 15 Ohms in series with CLOAD =
47,000pF; d) RLOAD = 1kOhm, 15 Ohms in series with CLOAD = 47,000pF.
279
a)
b)
c)
d)
Plot 24. Pulse Response — a) CLOAD = 0.47µF; b) CLOAD = 0.47µF, RLOAD = 1kOhm; c) 8.2 Ohms in series with CLOAD =
0.47µF; d) RLOAD = 1kOhm, 8.2 Ohms in series with CLOAD = 0.47µF.
280
PINOUT
VOUTC 1
VOUTB 2
VOUTA 3
VINB 4
VINA 5
VREFL 6
PRESETL 7
VINE 8
VINF 9
VOUTE 10
VOUTF 11
VOUTG 12
24 VOUTD
23 VINC
22 VIND
21 VDD
SP9841
20 SDI
19
18
17
16
15
14
13
GND
SDO
CLOCK
LOADH
VINH
VING
VOUTH
VOUTC 1
VOUTB 2
VOUTA 3
VINA/B 4
VREFL 5
PRESETL
6
20 VOUTD
19 VINC/D
18 VDD
17 SDI
SP9842
VINE/F 7
VOUTE 8
VOUTF 9
VOUTG 10
16 GND
15 SDO
14
13
12
11
CLOCK
LOADH
VING/H
VOUTH
SP9841 PINOUT
Pin 1 — VOUTC — DAC C Voltage Output.
Pin 18 — SDO — Serial Data Output; active totem–
pole output.
Pin 2 — VOUTB — DAC B Voltage Output.
Pin 19 — GND — Ground.
Pin 3 — VOUTA — DAC A Voltage Output.
Pin 20 — SDI — Serial Data Input.
Pin 4 — VINB — DAC B Reference Voltage Input.
Pin 21 — VDD — Positive 5V Power Supply.
Pin 5 — VINA — DAC A Reference Voltage Input.
Pin 22 — VIND — DAC D Reference Voltage Input.
Pin 6 — VREFL — DAC Reference Voltage Input
Low, common to all DACs.
Pin 23 — VINC — DAC C Reference Voltage Input.
Pin 7 — PRESETL — Preset Input; active low; all
DAC registers forced to 80H.
Pin 8 — VINE — DAC E Reference Voltage Input.
Pin 9 — VINF — DAC F Reference Voltage Input.
Pin 10 — VOUTE — DAC E Voltage Output.
Pin 11 — VOUTF — DAC F Voltage Output.
Pin 12 — VOUTG — DAC G Voltage Output.
Pin 13 — VOUTH — DAC H Voltage Output.
Pin 14 — VING — DAC G Reference Voltage Input.
Pin 15 — VINH — DAC H Reference Voltage Input.
Pin 16 — LOADH — Load DAC Register Strobe;
active high input that transfers the data bits from the
Serial Input Register into the decoded DAC Register.
Refer to Table 1.
Pin 17 — CLOCK — Serial Clock Input; positive–
edge triggered.
Pin 24 — VOUTD — DAC D Voltage Output.
SP9842 PINOUT
Pin 1 — VOUTC — DAC C Voltage Output.
Pin 2 — VOUTB — DAC B Voltage Output.
Pin 3 — VOUTA — DAC A Voltage Output.
Pin 4 — VINA/B — DAC A and B Reference Voltage
Input.
Pin 5 — VREFL — DAC Reference Voltage Input
Low, common to all DACs.
Pin 6 — PRESETL — Preset Input; active low; all
DAC registers forced to 80H.
Pin 7 — VINE/F — DAC E and F Reference Voltage
Input.
Pin 8 — VOUTE — DAC E Voltage Output.
Pin 9 — VOUTF — DAC F Voltage Output.
281
Pin 10 — VOUTG — DACG Voltage Output.
FEATURES…
The SP9841 and SP9842 include eight separate op
amp–buffered eight–bit DACs. These can be used to
replace up to eight trimpots with eight low–impedance programmable sources. The SP9841 uses eight
separate multiplying reference inputs, while the
SP9842 provides four pair of multiplying inputs. All
of the reference inputs, in either case, are returned to
a common voltage reference low pin. The inherent 2X
gain from the two–quadrant multiplying reference
inputs to the outputs allows the use of AC or DC
multiplying reference inputs generated from a single,
low supply voltage.
Pin 11 — VOUTH — DACH Voltage Output.
Pin 12 — VING/H — DACG and H Reference
Voltage Input.
Pin 13 — LOADH — Load DAC Register Strobe;
active high input that transfers the data bits from the
Serial Input Register into the decoded DAC Register.
Refer to Table 1.
Pin 14 — CLOCK — Serial Clock Input; positive–
edge triggered.
Pin 15 — SDO — Serial Data Output; active totem–
pole output.
Each DAC has its own data register which holds its
output state. These data registers are updated from an
internal serial-to-parallel shift register which is loaded
from a standard 3-wire serial input digital interface.
Twelve data bits make up the data word clocked into
the serial input register. This data word is decoded
such that the first 4 bits determine the address of the
DAC register to be loaded and the last 8 bits are the
data. A serial data output pin at the opposite end of the
serial register allows simple daisy-chaining in mul-
Pin 16 — GND — Ground.
Pin 17 — SDI — Serial Data Input.
Pin 18 — VDD — Positive 5V Power Supply.
Pin 19 — VINC/D — DACC and D Reference
Voltage Input.
Pin 20 — VOUTD — DACD Voltage Output.
FIRST
LAST
D0
LSB
D1
D2
D3
D4
DATA
D5
D6
D7
MSB LSB
Table 1. Serial Input Decoded Truth Table
282
A0
A1
A2
ADDRESS
A3
MSB
A3
A2
A1
A0
0
0
0
0
0
0
0
0
1
1
.
.
1
0
0
0
0
1
1
1
1
0
0
.
.
1
0
0
1
1
0
0
1
1
0
0
.
.
1
0
1
0
1
0
1
0
1
0
1
.
.
1
D7
D6
D5
D4
D3
D2
D1
D0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
1
1
1
0
0
1
0
0
1
0
0
1
0
0
1
0
0
1
0
1
1
1
1
1
1
1
1
1
0
0
.
.
.
1
0
0
.
.
.
1
1
1
1
1
1
0
1
DAC Updated
No Operation
DACA
DACB
DACC
DACD
DACE
DACF
DACG
DACH
No Operation
.
.
No operation
DAC Output Voltage
VOUT = D/128 (VIN – VREFL) + VREF L
VREFL
1/128 (VIN – VREFL) + VREF L
.
.
.
127/128 (VIN – VREFL) + VREF L
VIN (Preset Value)
129/128 (VIN – VREFL) + VREF L
.
.
.
254/128 (VIN – VREFL) + VREF L
255/128 (VIN – VREFL) + VREF L
tiple DAC applications without additional external
decoding logic.
VIN
The SP9841/9842 consume only 65 mW from a
single +5V power supply. The SP9841 is available in
24-pin plastic DIP and SOIC packages. The SP9842
is available in a space–saving 20–pin SOIC package.
VDAC
DAC 1
8
+
VOUT
–
R
VREFL
R
VOUT = 2 X VDAC when VREFL = 0V
= 2(D/256) X VIN
= D/128 X VIN
VOUT = D/128 X (VIN – VREFL) +VREFL
For applications requiring code–controlled output
polarity reversal regardless of the reference input level
(i.e. four–quadrant multiplication), please see the
SP9840/SP9843 product data sheet.
Figure 1. DAC and Output Amplifier Circuit
USING THE SP9841/9842
Theory of Operation
Each of the eight channels of the SP9841/SP9842 can
be used for signal reconstruction, as a programmable
dc source, or as a programmable gain/attenuation
block, multiplying an ac reference input by factors of
0 to 1.992. The rugged, wideband output amplifiers
provide both current sink and source capability for dc
applications, even those driving difficult loads. The dc
source mode mimics the functionality of a programmable trimpot with the added benefit of a low impedance buffered output. The amplifier's bandwidth and
high open–loop gain allow its use in programmable
gain applications where even a low distortion, high
resolution signal (such as audio) must be gated on and
off or gain–controlled over a –42 to +6dB range.
configured for a non–inverting gain of 2 using equal
value thin–film feedback and gain–setting resistors.
Signal ground is the VREFL pin, the common reference
input return for the 8 DAC–op amp channels. As
shown in Figure 1, the DAC section can be thought of
as a potentiometer across VIN(X) to VREFL. When this
potentiometer reaches its maximum output value of
255/256 times VIN, the output will be 1+(RFB/RGAIN) or
2 times the value of VDAC (actually up to 1.9921875
times the input voltage, with VREFL tied to ground).
When the potentiometer is at its minimum value of 0/
256, the output will try to be 0V, again assuming VREFL
is tied to ground.
The true relation between the dc levels at the VIN pin,
VREFL and the output can be described as:
Each channel consists of a voltage–output DAC,
implemented using CMOS switches and thin–film
resistors in a inverted R–2R ladder configuration.
Each DAC drives the positive terminal of an op amp
2.5
VREFL= 0V
3.0
D = FFH
2.0
1.5
where Data is programmable from 0 to 255, and RFB = RG.
D = 80H
1.0
VOUT(X) (Volts)
VOUT(X) (Volts)
3.0
VOUT = ((1 + RFB/RG) * (Data/256) * (VIN – VREFL)) + VREFL
0.5
0
0
a)
D = 00H
0.5
1.0
VIN(X) (Volts)
1.5
b)
VREFL= 1.5V
2.5
2.0
D = 00H
80 H
1.0 D =
H
FF
0.5
=
D
0
0.75
1.5
VIN(X) (Volts)
1.5
2.25
Figure 2. a) Single–Quadrant, and b) Two–Quadrant Operation
283
When VREFL is tied to ground, this expression
reduces to:
VOUT = (Data/128) *VIN
Multiplication of Input Voltages
While both the SP9841 and SP9842 are capable of
two–quadrant multiplication, this terminology is not
very precise when describing a system which runs
from a single positive supply. Traditionally, the quadrants have been defined with respect to 0V. A two–
quadrant multiplying DAC could produce negative
output voltages only if a negative voltage reference
were applied. A four–quadrant device could also
produce a code–controlled negative output from a
positive reference, or a code–controlled positive output from a negative reference. If ground is used to
delineate the quadrants, then the SP9841/SP9842
should be considered single–quadrant multiplying
devices, as their output op amps cannot produce
voltages below ground.
In reality, it is possible to define a dc voltage as a signal
ground in a single supply system. If the DAC's VREFL
pin is driven to the voltage chosen as pseudoground,
then each voltage output will exhibit 2–quadrant
behavior with respect to pseudoground; that is the
output voltage will enter the quadrant below the
pseudoground only when the reference input voltage
goes below pseudoground. This mode of operation is
useful when implementing programmable gain/attenuator sections, especially when the input signal is
bipolar with respect to pseudoground, or is ac–
coupled into the VIN(X) pin. When VREFL is tied to
power supply ground, only output voltages greater
than VREFL are possible, and the device performs
single–quadrant multiplication, much like a buffered
programmable trimpot across a single supply. Figures 2a and 2b show single–quadrant and 2–quadrant
performance of the SP9841/SP9842. Applications
which require 4–quadrant operation with respect to
pseudoground should use the SIPEX SP9840 or
SP9843 4–quadrant multiplying DACs.
The choice of voltage to use for the pseudoground is
limited by the legal voltage swing at the op amp
output. The op amp exhibits excellent linearity for
output voltages between, conservatively, 100mV and
VDD – 1.5V. The op amp BiCMOS output stage
consists of an npn follower loaded by an NMOS
common sourced to ground. This circuit exhibits
wide bandwidth and can source large currents, while
retaining the capability of driving the output to voltages close to ground.
At output voltages below 25mV, feedback forces
some op amp internal nodes toward the supply rails.
The NMOS pull–down device gets driven hard and
the NMOS device enters the linear range — it begins
to function in the same manner as a 50 ohm resistor.
In reality, the wideband amplifier output stage sinks
some internal quiescent current even when driving the
output towards ground. This sunk current drops
across the output stage NMOS transistor ON–resistance and internal routing resistance to provide a
minimum output voltage, below which the amplifier
cannot drive. This minimum voltage is in the 15
to 25mV range. It varies within a package with
each op amp's offset voltage and biasing variations. If an input voltage lower than this minimum, such as code 0 or 1, when VREFL is ground,
VDD = 4.75V minimum; VOUTMAX >100mV
3.0
2.5
2.5
VIN(X) (Volts)
VIN(X) (Volts)
VDD = 4.75V minimum; VOUTMAX < +3.25V
3.0
2.0
1.5
1.0
a)
1.5
1.0
0.5
0.5
0
0
2.0
0.5
1.0
1.5
2.0
VREFL (Volts)
2.5
3.0
0
0
0.5
1.0
1.5
b)
Figure 3. Reference Voltages a) Normal Operation; b) Maximum Linearity Near Code 1
284
2.0
VREFL (Volts)
2.5
3.0
is requested, feedback within the op amp circuit
will force internal nodes to the rails, while the
output will remain saturated near this minimum
value. Non–saturated monotonic behavior returns
between 25mV and 100mV at the output, but full
open loop gain and linearity are not apparent until
the output voltage is nearly 100mV above the
negative supply. Applications which require good
linearity for codes near zero should drive the VREFL
input at least 100mV above the ground pin, as this
insures that the output voltage will not go below
100mV for any legal input voltage. Two–quadrant
applications (programmable gain/attenuator) usually
bias VREFL up at system pseudoground, well above this
saturation region, and therefore maintain linearity
even at high attenuations (i.e. at code 1).
The allowable, useful values of VIN(X) and VREFL are
limited if a legal output value is to be expected for all
input codes. At maximum gain (DAC code 255) VOUT
is approximately equal to 2VIN(X) – VREFL. By solving
this equation twice, once with VOUT set to 0V, and then
again with Vout set to VDD–1.5V, the chart of Figure
3a results. This chart can be used to find the maximal
VIN(X) voltage excursions for any given voltage
driven into VREFL. The upper line plots the maximum
voltage at VIN(X) and the lower line plots the minimum voltage at VIN(X) at each value of VREFL drive.
Normal operation would be for VIN(X) anywhere
between the two lines. For example, assume a 4.75V
supply voltage, and that the DAC code is set to 255. If
VREFL is driven to 1.6V, VIN(X) below 0.8V would
require the output amplifier to swing below ground.
VIN(X) above 2.425V would require output voltages
greater than VDD – 1.5V, or 3.25V.
Figure 3b shows the limits on VIN when the minimum VOUT is constrained to be greater than 100mV,
for extremely linear operation, even at DAC code
1. In this case, the lower line is 50mV above its
position in Figure 3a, except that below VREFL =
100mV, the minimum input voltage stays at
100mV. It should be noted that VIN(X) can always
be driven to or slightly beyond the supply rails
without harm. Under such circumstances, the DAC
code can always be set to provide sufficient attenuation to get an undistorted output.
Driving the Reference Inputs
The VIN inputs exhibit a code–dependent input resistance, as shown in the specifications. In general, these
inputs should be driven by an amplifier capable of
handling the specified load resistance and capacitance. The reference inputs are useful for both ac and
dc input sources. However, series resistance into these
pins will degrade the linearity of the DAC. A series
resistance of 50 Ohms can cause up to 0.5LSB of
additional integral linearity degradation for codes
near full scale, due to the code–dependent input
current dropping across this error resistance. AC–
coupled applications should use the largest capacitor
value (lowest series resistance) which is practical, or,
use an external buffer to drive the inputs.
The DAC switches function in a break–before–make
manner in order to minimize current spikes at the
reference inputs. As previously noted, the reference
inputs can withstand driving voltages slightly beyond
the power supply rails without harm. The gain of 2 at
the op amps limits the choice of VIN/VREFL combinations if clipping is to be avoided at the higher codes.
Output Considerations
Each DAC output amplifier can easily drive 1Kohm
loads in parallel with 15pF at its rated slew rate. The
unique BiCMOS amplifier design also ensures stability into heavily capacitive loads — up to 47,000pF.
Under these conditions, the slew rate will be limited by
the instantaneous current available for charging the
capacitance — the slew rate will be severely degraded,
and some damped ringing will occur. Especially
under heavy capacitive loading, a large, low impedance local bypass capacitor will be required. A
0.047µF ceramic in parallel with a low–ESR 2.2 to
10µFtantalumarerecommendedforworst–caseloads.
The amplifier outputs can withstand momentary
shorts to VDD or ground. Continuous short circuit
operation can result in thermally induced damage,
and should be avoided.
If the input reference voltage is reduced to 0.6V, then
both the amplifier and DAC are functional at room
temperature at supply voltages as low as 2.5V. At VDD
= 2.7V, power dissipation is 9.3mW typical, with the
serial clock at 4MHz, or 7.0mW typical with the serial
clock gated off.
Interfacing to the SP9841/SP9842
A simple serial interface, similar to that used in a
74HC594 shift–register with output latch, has been
implemented in these products. A serial clock is used
285
tPR
1
0
PRESET
tS
VOUT (FFH)
(08H)
SDI
1
0
CLOCK
1
0
LOAD
1
0
±1 LSB
ERROR BAND
A3
A2
A1
A0
D7
D6 D 5 D 4 D 3 D2 D1 D0
VOUTFS
0
SERIAL DATA INPUT TIMING DETAIL (PRESET = Logic "1"; VIN(X) = 1.5V; VREFL = 0V)
SERIAL 1
DATA IN 0
AX or DX
tDS
tDH
SERIAL 1
DATA OUT 0
CLOCK
1
0
LOAD
1
0
tPD
tCH
tCLKD
tCL
tLD
tLDCK
tS
VOUT (FFH)
(08H)
±1 LSB
ERROR BAND
CHARACTERISTICS
(Typical @ 25°C with VDD = +5V unless otherwise noted.)
PARAMETER
Input Clock Pulse Width (tCH, tCL)
Data Setup Time (tDS)
Data Hold Time (tDH)
CLK to SDO Propagation Delay (tPD)
DAC Register Load Pulse Width (tLD)
Preset Pulse Width (tPR)
Clock Edge to Load Time (tCKLD)
Load Edge to Next Clock Edge(tLDCK)
MIN.
50
30
20
TYP.
MAX.
100
50
50
30
60
UNIT
ns
ns
ns
ns
ns
ns
ns
ns
CONDITIONS
Figure 4. Timing.
SDI
X
CLK
L
Data
LOADH
PRESETL
LOGIC OPERATION
L
H
No Change
L
H
Shift In One Bit from SDI
Shift Out 12–clock delayed data at SDO
X
X
X
L
All DAC Registers Preset to 80H (Note 1)
X
L
H
H
Load Serial Register Data into DAC(X) Register
Note 1: "Preset" may not persist at all DACs if LOADH is high when PRESETL returns high.
Table 2. Logic Control Input Truth Table.
286
to strobe serial data into a 12–stage shift–register at
each rising clock edge. The first four serial bits contain
the address of the DAC to be updated, MSB first. The
next 8 bits contain the binary value to be loaded into
the desired DAC, again MSB first. After the 12th serial
bit is clocked in, the LOADH line can be strobed to
latch the 8 bits of data into the data holding register for
the desired DAC. The address bits feed a decoding
network which steers the LOADH pulse to the clock
input of the desired DAC data holding register. The
output of the 12th shift–register is also buffered and
brought out as the SERIAL DATA OUT (SDO),
which can be used to cascade multiple devices, or for
data verification purposes.
The address field is set up such that DAC A is
addressed at 0001 (binary). Address 0000(binary)
will not affect the operation of any channel, as this
combination is easily generated inadvertently at
power–up. Other no–operation addresses exist at
1001(binary) through 1111(binary). Another use for
no–operation addresses is to mask off updates of any
DAC channel in a multiple–part system with cascaded serial inputs and outputs. By sending a valid
address and data only to the desired channel, it is
possible to simplify the system hardware by driving
the LOADH pin at each part in parallel from a single
source. Table 1 shows a register–level diagram of the
addresses, data, and the resulting operation.
A fourth control pin, PRESETL, can be used to
simultaneously preset all DAC data holding registers
to their mid–scale (80H) values. This will asynchronously force all DAC outputs to buffer the voltages at
their respective inputs to their outputs with unity gain.
This feature is useful at power–up, as a simple resistor
to the supply and capacitor to ground can insure that
all DAC outputs start at a known voltage. It can also
be used to implement stand-alone (non–programmed)
applications, such as a unity gain octal cable driver.
Table 2 summarizes the operation of the four digital
control inputs.
The four digital control input pins have been
designed to accept TTL (0.8V to 2.0V minimum)
or full 5V CMOS input levels. Timing information
is shown in Figure 4. Serial data is fully clocked
into the shift–register after 12 clock rising edges,
subject to the described setup and hold times. After
the shift–register data is valid, the LOADH line
can be pulsed high to load data into the desired
DAC data register, which switches the DAC to the
new input code. The serial clock input should not
see a rising edge while the LOADH pulse is high
in order to prevent shift–register data from corruption during data register loading.
The serial clock and data input pins are designed to be
compatibleasslavesunder NationalSemiconductor's
Microwire™ and MicrowirePlus™ protocols and
under Motorola's SPI™ and QSPI™ protocols. In
some micro–controllers, the interface is completed by
programming a bit in a general–purpose I/O port as a
level, used to strobe the LOADH line at the DACs.
This is done in a manner similar to that used for
generating a CS signal, which is necessary when
driving some other Microwire™ peripherals.
Low Voltage Operation
At nominal VDD, the CMOS switches used in the
DAC obtain sufficient drive to maintain an ONresistance much lower than the thin–film resistors.
This keeps the non–linear voltage–dependent portion
of their ON-resistances low, and guarantees both
excellent DAC linearity versus code, and low–distortion multiplication of large–swinging AC inputs. The
devices in the op amp also receive sufficient drive to
guarantee the specified bandwidth and output drive
current. However, all circuits within the DACs are
quite "functional" at very low values of VDD. By
reducing the reference voltages such that the maximum VOUT is near the target of VDD-1.5V, the DACs
will provide better than 0.5LSB typical integral performance for DC output voltages between 100mV
and VDD-1.5V. Reducing the reference voltage actually aids the linearity of the DACs, even at nominal
VDD. This occurs because the NMOS half of the
CMOS switches are more fully utilized at reference
voltages closer to ground, thus further reducing the
ON–resistance of the switches. Reference input currents are proportional to the reference voltages and
will also decrease with the reference voltages.
Plot 19 shows typical DC output linearity for VIN(X)
set to 0.5V, with VDD at 2.5, and then 3.5V. Note that
at 3.5V, the linearity is actually much better than the
±0.25LSB typical performance at VIN(X) = 1.625V
and VDD = 5V. Similarly, Plot 20 shows that this
performance level persists for VDD = 4.5V and 5.5V,
with VIN(X) set to 0.6V. The price paid for low voltage
operation is in op amp gain, bandwidth and es–
pecially current sinking at the DAC output. Plots 17
287
+5V
RPULLUP
1.5kOhms
1/8 of SP9841
1.22V
ICL8069
45µF*
+
–
(*Optional Noise Reduction)
DCOUT
3
DCOUT = 20mV to +2.44V for DATA = 00H to FFH.
At PRESET, DCOUT = +1.22V
Figure 5. Inexpensive DC Source.
through 19 show that for lower output current values
less than 1 mA, the SP9841/9842 can be used effectively even with VDD in the range of 2.7 to 3.3V.
operate at 4.75V, 1.75kOhms is required; the
1.5kOhms shown will suffice even if its value is
5% high. To drive a single input, a 10kOhm
value could be used. In order to reduce reference
and supply generated noise, an optional capacitor of 1 to 100µF bypasses the reference.
Application Circuits
Figure 5 shows an inexpensive single–quadrant
DC source for generating voltages from near
ground to near 2.44V. When using a two–terminal reference, the pull–up resistor should be
chosen so that the minimum input resistance of
5kOhms at each VIN(X) can be driven at the
lowest expected VDD. At VDD = 4.75V, and VREFL
= 1.25V, each input to be driven needs 0.248mA,
and the regulator needs 0.1mA to stay well
regulated. Thus, to drive all eight inputs, RPULLUP
should be chosen to supply at least 2mA. To
Figure 6 shows a circuit which generates DC
voltages roughly symmetric with respect to
2.446V. Two bandgap references are stacked to
first drive VREFL to 1.223V, and the input to 2.446V.
The pull–up resistor value should again be scaled
for worst–case loading — in order to drive all eight
inputs at 4.75V, a value of 620 Ohms is required.
At fullscale, the DAC output is near 3.65V. While
typical units will source 5mA at an output voltage
+5V
2kOhm
+2.44V
45µF*
45µF*
1/8 of SP9841
1.22V
ICL8069
+1.22V
1.22V
ICL8069
+
–
DCOUT
3
DCOUT = +1.22V to +3.66V for DATA = 00H to FFH.
At PRESET, DCOUT = +2.44V
(*Optional Noise Reduction)
Figure 6. Pseudo Bipolar Source Generates Voltages Above and Below 2.44V.
288
+1.8V
1/8 of SP9841
R1
10KΩ
VI
45µF*
OP–90
+
–
+
+
–
DCOUT
R2
4.75KΩ
–
3
ICL8069
R3
1.2KΩ
1.22V
DCOUT = 20mV to +3.6V for DATA = 00H to FFH.
At PRESET, DCOUT = +1.8V
(*Optional Noise Reduction)
Figure 7. Generating Programmable DC Voltages.
of 3.75V while running from a 4.75V supply, this
behavior is not tested in device production. If
maximum linearity is required near the 3.66V
fullscale voltage, then output loading should be
kept under 1mA.
Figure 7 uses a 1.8V reference to provide an output
voltage range from near ground to almost 3.6V.
The external micropower reference uses a bandgap
in a bootstrapped configuration, which guarantees
excellent supply rejection. Voltage at V1 is set by
1.223 ∗
 R2 
 (R1 + R2) 
VOUT is 1.223V +V1. R3 is used to set the quiescent current through the bandgap, I=V1/R3. The
op amp will easily drive one to all eight inputs.
Figure 8 shows a non–programmed standalone application. By tying PRESETL to ground, all channels are
permanently set to unity gain. While preset, the input
impedance at each input is set to 40kOhms nominal
(20kOhms minimum), which minimizes required
input current drive. The TL431 reference is programmed by resistors R1 and R2 for 3.3V. R3 is chosen
to provide at least 165µA for each input driven, plus
0.5mA for the reference at the minimum supply value
to be considered. In the Figure, the 560 Ohms shown
will drive all eight inputs. The excellent capacitive
load capability of the output amplifiers handles any
value of capacitive bypass loads without oscillation;
however, to minimize ringing at powerup, load capacitance can be chosen to be greater than 0.1µF or
less than 1,000pF.
+5V
560Ω
DIGIN
AC04
1/8 of SP9841
32KΩ
+
2.5V
100KΩ
+
–
–
DCOUT
TL431
PRESET
DIGIN
DCOUT
0
1
3.3V
0V
Eight independent 3.3V @ 10mA supplies with
logic–level controlled shutdown.
Figure 8. Generating Up to Eight (8) 3.3V @ 10mA DC Supplies with Logic–Level Controlled Shutdown (Non–Programmed).
289
+12V
In
10µF
LM317
Out
Adjust
R1
499Ohms
10kOhms
VDD
1/8 of SP9841
ICL8069
1.22V
R4
681Ohms
+
–
VA
VOUT
R2
1kOhms
R3
511Ohms
3
VOUT = 4.5V TO 5.5V at 1Amp for DATA = 00H to FFH.
At PRESET, VOUT = 5.0V
Figure 9. Programmable 1Amp Power Source.
Figure 9 shows a DAC channel controlling the output
voltage of an LM317 voltage regulator. By programming the code, the DAC changes its own supply
voltage. This circuit can be modified for wider output
voltage ranges by reducing the value of R4. However,
the circuit as shown requires the DAC to sink 1mA to
the negative rail at code 0 at its lowest VDD, at which
point the output voltage is 62mV. Thus, programming
codes 0 through 5 will do little to influence the output.
If R4 is replaced with a short circuit, useful operation
would be between 3.9V and 6.15V output; however,
the DAC output must then sink 2.5mA at VDD = 3.9V,
which results in a minimum DAC output voltage of
around 150mV. Codes above 17 will then provide
equally spaced output voltage increments.
Figure 10 shows how the gain of an external non–
inverting op amp can be programmed. RF and RG are
chosen for nominal gain. RTRIMRANGE is then ratioed to
RF to provide the desired range of gain trim. A wide
gain grange is achievable — for example, with RF =
11kOhms,RG =1kOhmsandRTRIMRANGE =2.74kOhms,
gain would be programmed linearly from 8 to just
under 16.
The OP–491 shown in Figure 10 is capable for
+5V
2.2µF
1VP-P
1/4 of OP–491
+5V
VREFL
+
–
VDD
1/8 of SP9841
RIBIAS
50kOhms
+
–
VREFL
= Up to VDD/2
(2.5V nominal for
rail–to–rail output
at SIG OUT)
3
SIG OUT
RGAIN
2.5kOhms
RF
10kOhms
RTRIMRANGE
20kOhms
AV = 1 +
RF  D  RF , D = 0 to 255
+ 1 − 
RG  128  RTRIMRANGE
Set RTRIMRANGE for desired gain-trim range.
For RTRIMRANGE = 20kOhms:
Code 0, AV = +5.5
Code 128, AV = +5.0
Code 255, AV = +4.51
Figure 10. Adjustable Gain of External Non–Inverting Opamp Circuit; VOUT = Rail–to–rail.
290
rail–to–rail output swing. In order to obtain this
performance, VREFL must be externally driven to
VDD/2, perhaps by use of the circuit of Figure 11.
At VREFL near 2.5V and VDD = 4.75V, the typical
positive output headroom at the DAC is limited
to 1.15V above 2.5V, so that this circuit is useful
for rail–to–rail outputs for gains higher than
4.35 (i.e. 1.15VPP maximum input). Note that
VINH
while an AC–coupled input is shown, this circuit is just as useful for DC–coupled inputs
which are generated with respect to the VREFL
pseudoground voltage. RIBIAS is used for the
AC–coupled circuit for opamp bias current return when the DAC is programmed to code 0, as
no current flows into VIN(X) at code 0.
Figure 11 shows a minimal parts count method
1/8 of SP9841
+
–
VOUTH
At PRESET, VOUT = 1.3V
(not well–regulated)
3
VINB
1/8 of SP9841
+
–
Load code 112 for 1.4VOUT = VREFL
96 for 1.6VOUT = VREFL
70 for 2.2VOUT = VREFL
VOUTB
3
VINA
1/8 of SP9841
+
–
VREFL
RISRC
2kOhm
3
VOUTA
ICL8069
+
–
301Ohm
45µF*
1.22V
*(Optional Noise Reduction)
+5V
The usable range for the bootstrapped VREFL circuit is 1.4 to 2.4VOUT at VDD = 5V, RISRC = 2kOhms.
To increase the upper usable limit, decrease value of RISRC.
Figure 11. Programmable, Bootstrapped, 1.4V to 2.2V VREFL Drive.
291
of generating a programmable pseudoground
voltage at the VREFL terminal. A pseudoground is
very useful if any channels are to be used in AC–
multiplying applications. In such applications,
the pseudoground will set the DC offset of the
output signal. The voltage output of this circuit
as the code is decreased is non–linear because
the DAC bootstraps the increased output voltage by a larger fraction at each code. It is really
meant to be programmed only over a range of
codes between 104 and perhaps 60. It does
exhibit a fairly well–defined output, even if
non–intentional codes are programmed. For
codes above 112, the output stage resembles a
50 Ohm resistor to ground, and the VREFL output
will be near 1.3V, depending upon the loading at
the other VIN(X) inputs. For codes below 60, the
VINB
±0.75V
output voltage will continue to rise until limited
by available current through RISRC. Note that
RISRC supplies the actual current into VREFL, and
must be chosen in order to supply enough current for all channels, especially if any of the
other eight inputs are to be grounded. A plot of
VREFL versus code is shown with the Figure, for
all other inputs either grounded or tied to the
supply.
Figure 12 shows a programmable gain/attenuator section using the programmable VREFL
drive. The VREFL of each DAC is actually
internally connected. When the optional 45µF
noise reduction capacitor is included, this circuit is capable of 86dB of SNR and 74 to 84dB
of SINAD at 1kHz, depending on the pro-
1/8 of SP9841
2.2µF
+
–
VREFL
VINA
VOUTB
1/8 of SP9841
+
–
VREFL
VOUTA
3
RISRC
2kOhm
ICL8069
+
45µF*
+5V
–
301Ω
1.22V
*(Optional Noise Reduction)
Load DACA with code 90; sets VREFL =1.7V.
Then, load DACB with desired gain:
code 255 = +6dB
code 128 = 0dB
code 64 = –6dB
code 1 = –42dB
code 0 = –70dB
Figure 12. AC–Coupled, Programmable Gain/Attenuator with Bootstrapped Programmable Output DC Offset (VREFL Drive).
292
grammed gain. Please refer to the THD versus
Frequency plot, which was generated by terminating a 600 Ohm source with 150 Ohms to
ground, then into this circuit. For the best gain
linearity versus code, use the largest (lowest
series impedance) coupling capacitor available,
or externally buffer the input.
channel can be set up with a programmable DC
offset adjustment with its output summed through
a large resistor into the OP–491 inverting terminal. Note that when RTRIMRANGE is set up for
only unity gain change range as in Figure 12,
only –0.5 times the DAC output offset will
appear at SIG OUT.
Figure 13 shows an external op amp with an
inverting programmable gain. In this circuit the
maximum output swing at the DAC occurs at the
maximum circuit gain. Thus, the headroom restriction at the DAC output applies at the maximum gain, which, for rail–to–rail outputs (VREFL
= 2.5V or VDD/2) should be greater than 4.3. By
making the programmable gain range large, this
circuit can be used to provide rail–to–rail outputs even at the lower gains. This circuit has
been ratioed to provide exact integer gain increments for every increase in 25 codes, over the
range of –1 to –11. This large range of gain
comes at a slight cost — the output offset of the
DAC amplifier will be gained up by –5.12 times
at SIG OUT. If this is a problem, a second DAC
Another application for the circuits of both
Figure 10 and 13 could be to force precise gains
from circuits made from imprecise resistors. By
restricting the programmable gain range to ±2%
(by setting RTRIMRANGE to be 100 times RF), the
resistors could be 1% values and the programmable gain resolution would increase to better
than 12–bits (0.0156%). In this case, only 1% of
the DAC output offset voltage would appear at
SIG OUT.
Figure 14 shows a window comparator and two
channels of programmable-gain input. While
the input signal is shown as AC–coupled, DC
signals of up to rail–to–rail amplitude could be
measured by setting the attenuation at the signal
VREFL +5V
2.2µF
1.25VP-P
+
–
1/4 of OP–491
SIG OUT
RGAIN
7.68kOhms
+5V
VDD
1/8 of SP9841
+
–
VREFL
= Up to VDD/2
3
RF
7.68kOhms
RTRIMRANGE
1.5kOhms
AV = −
RF  D   RF 
−  
 , D = 0 to 255
RG  128   RTRIMRANGE 
Set RTRIMRANGE for desired gain-trim range.
For RTRIMRANGE = 1.5kOhms:
Code 0, AV = -1
Code 25, AV = -2
Code 75, AV = -4
Code 175, AV = -8
Code 225, AV = -10
Code 250, AV = -11
Gain resolution = 4%
Figure 13. Adjustable Gain of External Inverting Opamp Circuit, VOUT = Rail–to–Rail.
293
Figure 14. Two–Channel Multiplexed Window Comparator with Programmable Gain and Limits.
294
+5V
2.5V
10KΩ
10KΩ
10KΩ
+
–
VREF LOW =
1.67V
Any DC
voltage
SIG 2
±0.75V
Any DC
voltage
SIG 1
±0.75V
+
–
1.0µF
1.0µF
0.83V
OP-290 V
LOW =
VHI = 2.5V
3
SIG2OUT
SIG1OUT
499Ω
499Ω
VLO Compare
MUXED (or mixed!)
signal
1/2 of LM339
+
–
+
–
"Error"
LED
360Ω
+5V
SIG1OUT : Gain = 0 to 2x
for ±0.75VIN
0.17V < VOUT < 3.17V at
maximum gain for
DATA = 0H to FFH.
At PRESET, Gain = 1x.
VLOW Compare: 1.67V < VOUT < 0V
for DATA = 0H to FFH.
At PRESET, VOUT = 0.83V
1/2 of SP9841
+
–
+
–
+
–
+
–
VHI Compare
VHI Compare: 1.67V < VOUT < 3.33V
for DATA = 0H to FFH.
At PRESET, VOUT = 2.5V
input DACs to the proper code. The LM339
does not really drive the LED to full illumination, due to limited output current, but a pull–up
resistor alone will yield a functional TTL error
signal. External op amps could use the VREFL
voltage as pseudoground. The outputs of the two
signal DACs must be isolated with resistors if
the two signals are to be multiplexed. This will
reduce the signal gain to 255/256 maximum,
due to the resistive divider created at the comparator input. If only a single channel was to be
window–compared, then the maximum gain to
the comparator would be the usual 255/128.
Figure 15 shows the schematic of an evaluation
board, which can be used with an IBM–compatible (XT or AT) computer and the simple
QuickBasic routine of Figure 16 to load each
DAC channel with its desired code. A straight–
through 25-pin cable can be used, or the board
can be plugged directly into the back of the PC.
Data is first latched into each 'HC165 parallel–
to–serial converter. Then a small state machine
is initiated by strobing INI. It clocks the latched
data into the serial data input and strobes the
LOADH input at the DAC. A pair of banana
jacks is used for applying VDD from an external
supply. A trimpot–adjustable voltage reference
is tied to all eight DAC inputs. On the evaluation
board, jumpers will allow this reference to drive
any VIN(X) input or the VREFL pin. The other
three op amps in the quad OP–491 are available
for breadboarding circuits, such as in Figures 1
through 14. If the reference voltage is adjusted
down to 0.5V, the DAC and the board should
function with VDD as low as 2.5V.
Driving Capacitive Loads
Unlike many other products, the SP9841/9842
will not oscillate under purely capacitive loading. However, fullscale step outputs will show
overshoot and ringing of up to 40% at worst–
case purely capacitive loading (between 1,000
and 10,000pF). Figures 17 through 20 show
near fullscale steps under capacitive loads of
between 470pF and 0.47µF. For capacitance up
to 10,000pF, the addition of a resistive load to
ground at the op amp output will decrease settling times without adversely affecting the positive–going slew rate. For higher capacitances,
this settling time enhancement comes at the
expense of positive slew rate, as not all instantaneous current can be used to charge the capacitor. For all values of capacitive load, settling
time can be dramatically reduced by adding a
small resistor in series with the DAC outputs.
Such series resistors will degrade the current
sinking ability at the DAC outputs for voltages
near ground; while the DACs typically sink
2mA at VDD =5V at VOUT = 110mV, the addition
of a 50Ohm resistor would require 210mV after
the resistor to sink 2mA. Large capacitances
require lower values of series resistance in order
to obtain critical damping.
295
Figure 15. Evaluation Board — Loads SP9841/9842 from IBM PC Parallel Port
296
J1
18
16
14
9
8
7
6
5
4
3
2
1
GND
INI
ALF
D7
D6
D5
D4
D3
D2
P5V
10µF
C11
CLK
CTRL0
P5V
1
2
3
4
Q
5
Q
6
12
9
8
U2
GND
B 12
A 11
5 G
6
H
7 SON
13
12
11
10
P5V
SI 10
SO 9
D 14
C 13
3 E
4 F
SH/LD VDD 16
CLK CLK-IH 15
SO
A 11
SI 10
B
D 14
C 13
U3B
CLR
Q
Q
8
9
ADDR LSB
ADDR1
ADDR2
ADDR MSB
CLK
D
VDD
VDD
ADDRESS
PR
J1 — BANANA JACK
J2 -J4 — 0.5 INCH JUMPERS
S1 — 25-PIN D CONNECTOR
(AMP #747469–4)
R1 — 49.9KΟΗΜ
R2 — 2KOHM TRIMPOT
R3 — 1.5KOHM
R4 — 1.5KOHM
R5 — 3KOHM
R6, R7, R8 — 10KOHM
U3A
CLR
D
CLK
PR
CLKN
2
1
U1
GND
D7 (MSB)
D0
8
5 G
6
H
7 SON
D6
STR
D1
3 E
4 F
D5
D4
U1, U2 — 74HC165
U3, U6 — 74FC74
U4 — 74HC02
U5 — DALE X0–438–8.0, 8MHZ OSCILLATOR
U7 — 74HC161
U8 — ICL8069
A1 — OP–491
C1 – C10 — 0.1µF CERAMIC
C11, C12 — 10µF/30V
C13 — 3PF
S1
2
D3
SH/LD VDD 16
CLK CLK-IH 15
1
D2
PC PRINTER
PORT
D1
D0
DATA
12
11
C5
0.1µF
U4A
13
13
12
11
10
GND
7
U5
OUT
14
VDD
P5V
U6A
CLR
D
CLK
PR
Q
Q
8
8
9
CLK
C8
0.1µF
6
5
PWRUPL
R1
49.9KOHM
P5V
6
5
4
3
2
7
10
4
Q
Q
3
2
P3
P2
P1
P0
Q3
Q2
Q1
Q0
U4D
U7
CLK
PE
TE
TC
CLR
D
CLK
PR
U6B
MR
9 SPE
1
DACLD
U4B
P5V
LOADL
6
5
CLKN
P5V
8
9
1
11
12
13
14
U4C
15
DACLD
2
1
3
4
10
U8
ICL8069
GCLK
CLKN
CLK
P5V
R6
10KOHM
C10
0.1µF
4 V+
5
INB+
6
INB7
OUTB
1 OUTA
2
INA3 INA+
A1
VOUTH 13
VING 14
VINH 15
LD 16
CLK 17
SDO 18
GND 19
SDI 20
VDD 21
VIND 22
VINC 23
VOUTD 24
OUTC
INC-
INC+
V-
IND+
IND-
OUTD
C9
0.1µF
8
9
10
11
12
13
R8
10KOHM
R7
14 10KOHM
GCLK
P5V
NOTE:
1. ALL DIGITAL IC'S BYPASSED WITH 0.1µFTO GROUND
2. THREE UNCOMMITTED OP AMPS IN THE OP–491 PACKAGE
ARE AVAILABLE FOR USER APPLICATIONS.
C12
10µF
C13
3PF
3KOHM
R5
SP9841
R4
1.5KOHM
12 VOUTG
11 VOUTF
10 VOUTE
9 VINF
8 VINE
7 PRE
6 VREFL
5 VINB
4 VINA
3 VOUTA
2 VOUTB
R2
2KOHM
R3
1.5KOHM
P5V
PWRUPL
TEST
POINTS
1 VOUTC
SP9841.BAS
'This program accepts an address (1 through 8) and data (0 through 255)
'in decimal and sends them to the DAC. Addresses 1 through 8 will
'correspond to converters A through H respectively. The appropriate
'output will be: Vout-(data/128)*VREF volts.
'We found that for our IBM PC/AT the LPT1 port address was 378H (Data
'Register 378H and control register 37AH) while for our IBM PC/XT the
'LPT1 port address was 3BCH (Data Register #BCH and control register 3BEH).
DIM
DIM
DIM
DIM
DIM
CLS
lsb AS INTEGER
msb AS INTEGER
datareg AS INTEGER
contrlreg AS INTEGER
n AS INTEGER
DO
INPUT "Enter type of PC, AT or XT: ", type$
IN UCASE$(type$) = "AT" OR UCASE$(type$) = "XT" THEN
EXIT DO
ELSE PRINT "Please enter either AT or XT.": PRINT
END IF
LOOP
IF UCASE$(type$) = "AT" THEN datareg = &H378: cntrlreg = &H37A
IF UCASE$(type$) = "XT" THEN datareg = &H3BC: cntrlreg = &H3BE
CLS
n=0
DO
WHILE n=0
DO
test$ =""
INPUT "Enter Address (1 through 8): ", lsb
IF lsb < 1 or lsb> 8 THEN test$ = "false"
IF test$ = "false" THEN PRINT "Please enter a valid address.": PRINT
LOOP UNTIL test$ <> "false"
DO
test$ = ""
PRINT
INPUT
IF msb < 0
IF test$ =
LOOP UNTIL
"Enter Data (0 through 255 in decimal): ", msb
or msb > 255 THEN test$ = "false"
"false" THEN PRINT "Please enter valid data.": PRINT
test$ <> "false"
OUT cntrlreg, $H3
OUT datareg, &H0 + msb
OUT cntrlreg, &H2
'set both latch clocks low
'send most significant byte to port
'clock U1
OUT datareg, &H0 + lsb
OUT cntrlreg, &H0
OUT cntrlreg, &H4
'send least significant byte to port
'clock U2
'enable U7, set U1 & U2 to serial out mode
PRINT :
PRINT "Strike spacebar to enter new data or Q to quit."
DO
X$ = INKEY$
IF UCASE$(X$) = "Q" THEN n=1
LOOP UNTIL X$ = " " OR UCASE$(X$) = "Q"
LOOP
END
Figure 16. Microsoft qbasic Program to Load Evaluation Board with Desired Codes.
297
ORDERING INFORMATION
Model
SP9841KN
SP9841KS
SP9842KS
SP9841BN
SP9841BS
SP9842BS
298
Reference Inputs
Temperature Range
Package
................................... Eight, independent ................................................... 0° to + 70°C .............................. 24–pin, 0.3" Plastic DIP
................................... Eight, independent ................................................... 0° to + 70°C ........................................ 24–pin, 0.3" SOIC
................................... Four pair ................................................................... 0° to + 70°C ........................................ 20–pin, 0.3" SOIC
................................... Eight, independent ............................................... –40° to + 85°C .............................. 24–pin Plastic, 0.3" DIP
................................... Eight, independent ............................................... –40° to + 85°C ........................................ 24–pin, 0.3" SOIC
................................... Four pair ............................................................... –40° to + 85°C ........................................ 20–pin, 0.3" SOIC