LED bar-graph display represents two digits
Novel circuit uses a reference ladder signal based on the most-significant-digit display.
Ajoy Raman, Bangalore, India; Edited by Paul Rako and Fran Granville -- EDN, September 22, 2011
This circuit uses two National Semiconductor LM3914 dot/bar-display-driver ICs to implement a two-digit, 0 to 5V LED voltmeter that mimics a subranging flash ADC. An LED bar graph comprising five LEDs, each representing 1V of input signal, represents the MSD (most-significant digit). Nine LEDs in dot mode, in which only one LED lights, represent the LSD (least-significant digit). The circuit senses the operation of the MSD LEDs and uses them to change the input reference ladder of the chip that drives the LSD. The input signal ranges from 0 to 5V, and accuracy is better than ±50 mV. The circuit operates over a supply voltage range of 5 to 8V.
R1 and R2 divide the input voltage
in half, such that a 5V maximum input
is 2.5V at the LM3914s, IC1 and IC2
(Figure 1). You strap the mode pin of IC1
high, so it operates as a bar graph, and use
VR1 to adjust the REFOUT pin of IC1 to
2.5V. Thus, each of the IC1 output pins
lights successively in 0.5V increments.
Because this IC makes the MSD, you
wire in only five LEDs on every other
output, starting at output D2, meaning
that the five LEDs will light at 1V intervals
from 1 to 5V. The LM3914’s data
sheet explains how you can use R3 to set a
constant-current output on the LED pins
(Reference 1). The current in each LED
is approximately 10 times the current
that you draw from the REFOUT output
pin. The part maintains 1.25V between
the REFADJ and REFOUT pins. The
VR2/R10/R13 voltage divider causes a load,
which, along with the 1.5-kΩ value of
R3, sets a fixed output current in LEDs
D1 through D5. You should select these
LEDs from the same batch so that their
forward voltage drops match.
You then wire a resistor and a transistor
around each of the four LEDs.
The voltage across the LED also presses
across the resistors, so these LEDs
form four constant-current sources that
operate in conjunction with the LEDs.
Adjust VR3 such that each LED when on adds 500 mV to their summed output.
You send this signal to RLO, the bottom
of the internal resistor string in the second
LM3914 (Figure 2). You then send
the 50%-divided input signal to the SIG
Pin of IC2. Use an op amp, IC3, to add a
fixed 500-mV offset plus the summed-current
signal from the outputs of IC1.
R1 and R2 reduce the input signal to the
circuit by 50%, so a 500-mV excursion
at IC2’s SIG Pin input represents 1V of
the input excursion.As the input to the circuit goes from 0 to 1V, the SIG inputs to both bargraph ICs go from 0 to 0.5V. No LEDs light on IC1, meaning that IC2 has RLO at 0V and RHI at the 500-mV offset you adjusted with VR2. The LED outputs of IC2 now light in sequence as the input to the chip goes from 0 to 0.45V, corresponding to a 0 to 0.9V input at the Signal-in Port. When the input signal is high enough to light LED D1, the value at IC2’s RLO jumps to 500 mV, and the input at RHI jumps to just 500 mV higher than RLO, or 1V. Because IC2’s internal resistor ladder is now biased between 0.5 and 1V, IC2 indicates 0.1V steps between 1 and 2V at the Signal-in Port. Leave the Mode Pin on IC2 floating so that the part operates in dot mode instead of bar-graph mode.
At a 4.9V input to the Signal-in Port, LEDs D1 through D4 illuminate, resulting in 2V at the RLO input of IC2. The op amp adds 500 mV to that value and presents it to the RHI input of IC2 for a total of 2.5V. The input to IC2 is 2.45V, so the D9 output of IC2 lights D14, correctly indicating the LSB (least-significant bit) of the measurement as nine-tenths.
|
Reference |
|
Talkback
-
Thank you for responding to my inquiry.
So the 4 collectors are only tied together with R11 and the other R's who's designation I can't read in this print and IC2 pin4 , right? That makes sense.
Any way to make that type larger so it will show in this bit map graphic?
Would it be asking too much to have EDN fix the schematic and publish the correction out of respect to the author and the readers? I agree that that many errors in a professional publication is less than acceptable.
It is a clever circuit indeed. A long time ago, in the 70's when these IC's came out I built a 10X10 led scope out of these chips. I still have it. Not cutting edge but was fun. If I recall the data sheet you can make it up to 100 X 100 leds. I thought it would make a nice wall hanging with a couple of mics as inputs or connected to the stereo as a L-R X-Y display.
If anyone's interested I'd publish it.
George Kaye - 2011-5-10 11:50:15 PDT -
Yet another reason why schematic lines should not cross and connect at the same point. "Pluses" do not connect, "Ts" do.
Don Crafton - 2011-5-10 11:01:34 PDT -
On Figure 1, delete the four intersection dots that are shown connecting the 'T2 500mV/Digit' signal to the IC1 outputs on pins 12, 14, 16, 18.
Also on Figure 1, op-amp IC3A is missing pin information. Pin 2 should be marked as the (-) input and pin 3 should be marked as the (+) input.
Also on Figure 1, there is no reason to show the terminated wires on unused outputs such as LED 9 (pin 11). As drawn, this implies that some sort of connection should be made to these IC pins.
On Figure 2, the circle terminator shown on the left side of the 'Mode-Select Amplifier' is incorrect. It is supposed to be an arrow pointing up at the ten LED comparators to show that this sub-circuit's output influences the LED signals (overriding the comparator outputs).
After study of the circuit, these are the ONLY drafting errors I noticed. And so many drafting errors are unacceptable for a professional publication in my opinion.
Paul Schmidt - 2011-4-10 09:46:25 PDT -
This is drawn wrong. The used outputs of IC1 are all tied together putting all the LEDS and transistors/resistor strings in parallel. All LEDS will light when any led lights. Would you please correct this? Thank you
George Kaye - 2011-23-9 09:52:57 PDT
