On 9/2/2012 7:52 PM, Jamie wrote:
I am looking at a chip which only uses 1.8 volt power and has diode
clamps to the power and ground rails. I want to drive a 3 volt LCD
display. The chip maker tells me, "Swinging 3.3v open drain is pretty
easy". Maybe I am having a brain cramp, but I can't think of how to do
this in a way that doesn't require a transistor or buffer. That was
the context, I stated I could use other devices that don't require the
use of buffers for 3.3 volt drive.
Obviously the "easy" way is to drive an SPI or I2C port device with
enough outputs to control the LCD, or even to use an LCD driver chip.
That's fine if that is what I have to do. I just don't want to add
another chip or two or three driving up the cost by another several
bucks. At that point I would just use a chip with LCD compatible
drivers built in.
Am I missing something? Is this easy to do without buffers or external
It's rather easy to use a low voltage uC to signal/drive higher voltage
If the IO line needs to be converted then a simple NPN can be used with
the emitter to common and the base being driving from the IO line via a
small bias R. This NPN on the collector side will then be the sink
for the 3.3 volt device.
If the IO output can not be inverted then you can use a NPN and connect
the emitter to the IO output. The base will then be biased with a
dropping R from the 1.8V supply.. With this, it'll only conduct when the
Emitter gets pulled low via the IO line and the collector will act as
the pull down.. The only problem with this is, the current limit of what
you pull down is limited by what the uC IO output can handle, but, it'll
give you the level shift you desire..
In any case, I like serial LCD displays, it limits the required IO
lines. The only problem there, is response time and it could be
important if you are doing graphics on the LCD display in real time.
This maybe of some interest to you.
Thanks for the reply, but I am aware that you can use a buffer or
transistor to buffer the output signal. I have no interest in that
since I can just select an MCU that will directly drive an LCD if that
is what the best approach is.
I was told by the maker of a 1.8 volt I/O MCU, as I stated above,
"Swinging 3.3v open drain is pretty easy". I am assuming by this he
means it is pretty easy without adding a small truck load of parts and
none of them costing over a penny. The LCD I am looking at driving has
over 30 segments if I use a static device. If I bail on the 1.8 volt
part and use a standard MCU with LCD drive I can even use a
multiplexed drive bringing the pin count below 20 for an even more
If he were talking about switching a 3.3 volt TTL level input, then I
could agree since the input only needs to swing from below 0.8 volts
to 2 volts and I might be able to do that with a combination of
resistors (I haven't actually checked to see how much noise margin is
left, if any). But an LCD needs to swing close to 3 volts minimum.
I think this company is a bit unusual... to put it nicely. They think
it is better for you to add external devices than it is to
"complicate" their chip with real world I/O voltages and hardware I/O
peripherals. That may work for async serial ports and SPI, but so much
of today's hardware needs more performance or just plain needs to
simplify the bill of materials.
He meant as you understood it, you need to use a truck load of parts if
you plan on implementing many IO lines up to 3.3 volt levels. Or, use a
dense level shifter logic chip.
I think it would be cheaper to use a 3.3 capable processor. There are
DC DC converters that will swing a single cell source up to 3.3V if
the power source is a concern.
Yes, I have come to the same realization. No, the power rail is not the
issue. The issue is with the MCU with lots of internal horsepower, but
very, very poor I/O. I can live a rich full life with a cheaper MCU
with LCD drivers built in. But it won't be as much fun I guess...