Operation Overview
The device has a number of different displays or modes, each summarizing different
information about the serial link. The device is operated completely via two of
the three buttons S1, S2 and S3. These are located from left to right across the
bottom of the SKP16C62P demo board.
The buttons S2 and S3 are used to navigate the display system. S3 acts as a NEXT
button and moves the device through the modes of analysis. S2 acts as an ENTER
button performing an appropriate action on the current menu displayed.
Operation Modes/Displays
In the initial display, the device shows the status of the two serial connections
DB9F and DB9M, the female and male connectors respectively. If a cross-wired cable
is detected on a port, then "Xvr" will be displayed next to the name of the port.
If the port is unconnected or not detected, then it will not be shown on the
display. If neither port is in use then "No Cable Detected" will be displayed.
When at least one port has been detected then the user can move onto other displays.
Note, if both serial ports are being used then baud-rate detection and analysis
occurs only on the DB9F connector. To detect on the DB9M port simply remove the
DB9F cable.
The voltage display shows the RS232 voltage levels. These voltages are measured
on the receive pin of each input. S2 changes the display format to include two
decimal places. It is this voltage that is used to detect the presence and
polarity of a cable. This information is a useful check to ensure that the
hardware layer of the serial link is present and operational.
| |
|
Photo 3 The cable detect mode with two cables connected.
CLICK TO ZOOM
|
|
Photo 4 The Voltage display mode with two cables connected.
CLICK TO ZOOM
|
Monitor mode is used to initiate communication on the USB connection to allow
third party monitoring of the data sent and received on the DB9 connectors.
The USB connection uses the same baud-rate and settings as those detected on
the DB9 connectors. S2 is used to toggle the USB connection on or off.
The number of "events" detected on the RS232 lines is available as a display
mode. The device has a maximum number of 128 events that can be analysed. The
event memory resets automatically if there is a one second pause in data on the
serial line. Pushing S2 can also clear the event memory. When an event is
detected the green LED will light momentarily.
The period display is used to show the width of the shortest pulse in the event
memory. This is used to calculate the baud-rate. Again, S2 changes the display
format to include decimal places. The display will automatically change units
between microseconds and milliseconds where appropriate.
In addition to this, the exact baud-rate calculated from the bit width is
available. As with other displays, S2 changes the display format to include a
decimal place. The baud is displayed in bits per second. This information is
useful when a micro-controller has been set up wrong, or has a clock that has
gone haywire and is no longer using a standard baud-rate.
| |
|
Photo 5 The period display mode.
CLICK TO ZOOM
|
|
Photo 6 The baudrate guess display mode.
CLICK TO ZOOM
|
Although the bits per second value given is useful, it is often slightly wrong
due to timing inaccuracies and doesn't always give a usable value. The
baud-rate guess mode calculates the closest commonly used baud-rate from an
internal table of standard baud-rates. The top line of the display shows how
close the measured data is to the guess. A value less than 90% usually means the
guess is wrong. This information can be a very valuable ballpark figure when
beginning to debug a serial link.
If the baud-rate guess looks good enough, you can then go on with some analysis
of the data in the event memory using the guessed baud-rate. If no reasonable
data can be found then "NoMATCH" is displayed. If the parity bits cannot be
determined then the display will read "xx bit" where xx is the number of bits
per word including start, stop and parity bits. The top line displays the number
of complete words or characters used in the analysis. If you have a good
baud-rate guess (greater than 98%) and a decent number of events (more than 32
or so) then this is usually bang on the money.
|