THE DESIGN CYCLE
VDD
RX
TX
R1 470
R2
10K
C2
R3
470
R4
470
19
18
5
4
3
2
1
11
12
13
14
15
16
17
10
9
6
20
7
8
D+
D-
GP7/TxLED
RST
OSC2
OSC1
VDD
RTS
RX
CTS
GP2
GP1/USBCFG
GP0/SSPND
Vusb
TX
GP3
GP6/RxLED
VSS
GP5
GP4
C3 470nF
VDD
J1
Y1
12MHz
MCP2200
■ SCHEMATIC 1. No rocket science here. The
MCP2200 circuitry is like taking a walk in the park.
MINI-B USB RECPT
6
3
2
5
4
1
SHIELD
D+
D-GND
ID
VBUS
A high-level block diagram of the MCP2200 is
depicted in Figure 1. The MCP2200 sure looks like a
“specialized” microcontroller. However, I can’t yet prove
that it is. So, let’s take a walk around its pins.
The MCP2200 can operate with voltages between 3.0
volts and 5. 5 volts. In that most low power embedded
applications that will employ the MCP2200 draw their
power directly from the USB portal, the MCP2200 will
find itself powered by the USB portal’s 5.0 volt VBUS line
in most instances. This mode of operation is termed Bus
Power Only mode in the MCP2200 datasheet. The
MCP2200 can also be configured in self-powered mode.
I have this microcontroller feel about the MCP2200.
One thing that really makes me think PIC is the MCP2200’s
need for an external 12 MHz quartz crystal or ceramic
resonator. However, my “It’s really a PIC” theory could be
debunked by getting up on my donkey and declaring the
clock signal is needed for the
USB interface and/or the
MCP2200’s internal controller.
Judging from the oscillator block
in Figure 1, I’m willing to bet
that the 12 MHz clock signal is
processed through a 4x PLL to
synthesize a 48 MHz USB clock.
I’m also betting the farm on the
control block really being a
microcontroller block as it
supports the configuration
calls out a muRata CSTCE12M0G15L; the MCP2200 Demo
Board User’s Guide specifies a muRata CSTCE12M0G15L99-R0.
I couldn’t find either of those ceramic oscillators that I
didn’t have to buy 3,000 of. So, I finally came up with
a muRata CSTCE12M0G55-R0 (Digi-Key part number
490-1197-1-ND) that we will use in our design.
■ SCREENSHOT 1. We could
shrink this design considerably
by replacing the SOIC with a
QFN and reducing the SMT
component size from 0603 to
0402. However, I think it’s small
enough right now, thank you.
July 2010 59
