Debugging rs232

[RocketScienceSmurf]

I found a small, embedded machine that I wanted to turn in to a small firewall. It's an x86-based industrial computer with 3xNIC and a 9 pin D-sub connector for console access. The OS  and applications are stored on a CF card and with the software I intend to use I needed to get the serial access going so I could make the initial configuration but no matter what baud rate I tried I could not get it to work. 1200 baud was the "best" but still only gibberish so I had to figure out a way to examine the signaling on the serial port and determine the actual baud rate from that. I found some web sites that explained how an oscilloscope could be used so I tried it. It worked OK but I wanted to get more data and also the signal so I connected the transmit pin to an optocoupler (through a constant current limiting device) and fed the signal to my logic analyzer. At first I tried the "auto baud rate function" and a lot of different settings but then I looked at the signals and looked for the "shortest time" or "highest frequency" which was 680Hz. Half of that time equals 1360 which is the baud rate that was used. After I figured that out I could decipher the information sent on the serial port and start configuring my device.

I thought this information could be useful to others.

[profdc9]

It would be cool to take a USB protocol chip such as the FT232H and write some software for it so it would try to figure out what kind of protocol it is looking at and then you could connect to it using one of the many protocols it implements, UART, SPI, I2C, JTAG, etc.   Sort of a universal low speed protocol translator device.  A technique like this could be used to try to figure out what kind of signal and the data rate.

Dan

I found a small, embedded machine that I wanted to turn in to a small firewall. It's an x86-based industrial computer with 3xNIC and a 9 pin D-sub connector for console access. The OS  and applications are stored on a CF card and with the software I intend to use I needed to get the serial access going so I could make the initial configuration but no matter what baud rate I tried I could not get it to work. 1200 baud was the "best" but still only gibberish so I had to figure out a way to examine the signaling on the serial port and determine the actual baud rate from that. I found some web sites that explained how an oscilloscope could be used so I tried it. It worked OK but I wanted to get more data and also the signal so I connected the transmit pin to an optocoupler (through a constant current limiting device) and fed the signal to my logic analyzer. At first I tried the "auto baud rate function" and a lot of different settings but then I looked at the signals and looked for the "shortest time" or "highest frequency" which was 680Hz. Half of that time equals 1360 which is the baud rate that was used. After I figured that out I could decipher the information sent on the serial port and start configuring my device.

I thought this information could be useful to others.