| |
|
April 24, 1997
Circuit tricks
increase LCD contrast
Steve
Hageman, Hewlett-Packard, Santa Rosa, CA
A popular combination in
embedded designs is the Microchip Technology PIC µC
mated to one of the many available LCD modules. These
combinations provide a low-cost, simple way to add
intelligence to any design. The LCD modules are specified
to operate from one 5V supply, but the lack of a negative
voltage to bias the LCD backplane severely limits the
available LCD contrast. The true spirit of embedded PIC
designs lies in their simplicity and the minimum number
of parts they use. Hence, most designs usually compromise
with less-than-optimum contrast. However, by adding a few
extra parts, you can obtain suitable bias voltages that
give you your money's worth in LCD contrast.
 The usual connection of the LCD
bias pin is to connect a trim pot between VDD
and ground (Figure 1). However, the best contrast
occurs with the bias pin at ground if only a single
supply is available. The contrast is optimal with a 0.1
to 0.9V bias on the LCD. Under these conditions, the
LCD draws 100 to 600 µA from the bias supply (depending
on the display's size). The LCD data sheets specify a
temperature coefficient of 16 mV/ºC to compensate the
display for contrast changes with temperature. In
practice, most PIC designs are meant for human use, which
limits the temperature range from about 15 to 35ºC.
Testing indicates that over this temperature range,
compensating the LCD bias is unnecessary, because the
contrast stays within acceptable limits when you use any
of the bias circuits shown.
 The circuits in Figures 2 and 3 draw an extra 2 mA or so from the 5V
supply--a drain that could limit the circuits' use in
designs that require extremely long-life battery
operation. For most designs, however, another 2 mA or so
won't be noticeable. You have a choice of several bias
circuits, depending on your system's exact configuration.
You can use the first circuit in nearly any design (Figure
2a). It
operates as a charge pump that obtains its drive from the
µC's master clock. You can build this design with a
couple of SOT-23s and chip capacitors; hence, the design
takes up little space. You can adjust the contrast
trimpot to provide a slightly negative bias voltage to
the LCD, thereby improving contrast. The design works for
ceramic-resonator oscillator configurations that operate
to about 5 MHz. Above this frequency, the FET's input
capacitance starts to load the oscillator excessively.
However, 80% of all PIC designs use 4 MHz as the
oscillator frequency.
The second design (Figure
2b) is useful
for limiting oscillator loading; it extends the possible
operating frequency by using an extra CMOS gate. You can
use nearly any uncommitted gate type to drive the charge
pump in a manner similar to that of the first design.  Many PIC circuits incorporate
RS-232C communications. If yours does, you might be able
to steal some negative bias from the RS-232C circuits. Figure 3 shows how to tap 8V from an LT1383
single-supply RS-232C transceiver IC. Circuits such as
the Linear Technology LT1383 contain an internal +5 to
±8V charge pump, and if such an IC is already in the
circuit, you might as well use the voltage as the
LCD-bias supply. Loading is minimal with the resistive
divider shown.
 A common trick when using a PIC is
to drive the PIC's input pin directly from the RS-232C
line through a 22-kilohms resistor (Figure 4). The PIC's internal protection diodes
clamp the voltage at the input to the PIC between the
supply and ground. Software can then decode the RS-232C
communications. If your circuit has frequent or
continuous RS-232C communications, you may be able to
steal the LCD bias by directly clipping the negative
peaks from the RS-232C input. This approach assumes that
the RS-232C driver is using the true specified bipolar
signal levels, not just the 0 to 5V levels that it
sometimes uses. You can increase the value of C1
to approximately 2.2 µF if you need more holdup time.
You can also increase the time between communications by
making sure that your RS-232C transmitter
"marks," or drives, the transmit line low,
between transmissions. This method is usually the way
communications end by default, but some small software
changes can force this state at the end of a
transmission. (DI #2012)
|
