Start
• System
Overview
• LCD
• Analog
•
Software
•
Implementation
•
Hand
that Feeds
•
Where
to Now?
•
Sources
and PDF
LCD
The
displays were surprisingly easy to drive with
software. The Epson E2202CKR LCD with a SED1565
controller can be driven in Parallel mode, which
allows the status and busy bits to be read, or
SPI mode, which allows only writing. However,
writing to the display at 2.5 MHz in SPI mode
appeared to have no negative effect. The datasheets
suggest 6 MHz, but this appeared to present a
problem. It might be feasible in Parallel mode
because you’d be able to read the busy bit. But
in SPI mode, I had no choice but to limit the
speed. However, a full-screen refresh at 2.5 MHz
takes roughly 7.2 ms, which includes the indexing
of arrays overhead. That’s no problem because
2.5-MHz update rates of moving objects prove to
be too fast for the LCD anyway. Moving objects
become blurred because the pixel elements take
too long to change state.
The
display runs at 3.3 V, which makes it perfectly
compatible with the LPC2138 microcontroller (see
Figure 2, p. 50). It draws a maximum of 260 µA,
which makes it easy to appreciate where the displays
came from. The screens retain their information
without having to be physically refreshed. The
SED1565 LCD controller takes care of this. Therefore,
it’s feasible that you could capitalize on the
low-power consumption of the ARM device in Power
Down mode at 500 µA to get long life from a battery.
|
(Click
here to enlarge)
|
Figure
2—Three elements are mounted on the board:
the E2202CKR LCD, the DX-045D-90 sensor, and
the buttons for calibration functions. The
other two elements are the LPC2138 and the
speaker on the MCB2130 evaluation board. |
A
first look at the register set will put you in
a comfort zone. Some of the functions and the
way they’re selected aren’t too dissimilar from
a simple alphanumeric display with an HD44780
controller. The E2202CKR LCD doesn’t have a character
ROM or display control commands, but there is
enough there to write information to the screen.
The biggest difference is that you can control
the LCD bias voltages with the register set instead
of a simple preset potentiometer.
The
only problem I experienced involved driving the
screen’s contrast. There are a number of parameters
for setting up the onboard switcher that controls
the bias to the LCD. The one that matters the
most, v5_volume, appears to be a lot more critical
than any of the others. With a value of 0x20,
you have a black screen. With a value of 0x05,
you wouldn’t know the screen was on. It is a balancing
act between v5_volume and v5_res_ratio to get
the screen to be useful.
The
LCD seems to have an infrared sensor in it to
adjust contrast. No, it isn’t about the rise in
temperature changing the switching characteristics
of the pixels, because a simple infrared LED changes
the contrast of the display. I stumbled onto it
when I tried to take a picture of the project
in sunlight. It makes sense—not that the controller
part of the screen would ever see sunlight, but
the fact that the unit is sensitive to IR. This
means the contrast can be automatically adjusted
in the intended application of mobile phones.
It’s a useful little undocumented feature (or
a pitfall if you didn’t find out about it).
