Binary to Ascii and back again

Question

I'm trying to interface with a hardware device via the serial port. When I use software like Portmon to see the messages they look like this:

42 21 21 21 21 41 45 21 26 21 29 21 26 59 5F 41 30 21 2B 21 27
42 21 21 21 21 41 47 21 27 21 28 21 27 59 5D 41 32 21 2A 21 28  

When I run them thru a hex to ascii converter the commands don't make sense. Are these messages in fact something different than hex? My hope was to see the messages the device is passing and emulate them using c#. What can I do to find out exactly what the messages are?

Answer 1

Does the hardware device specify a protocol? Just because it's a serial port connection it doesn't mean that it has to be ASCII/Readable english Text. It could as well be just a sequence of bytes where for example 42 is a command and 21212121 is data to that command. Could be an initialization sequence or whatever.

At the end of the day, all you work with is a series of bytes. The meaning of them can be found in a protocol specification or if you don't have one, you need to manually look at each command. Issue a command to the device, capture the input, issue another command.

Look for patterns. Common Initialization? What could be the commands? What data gets passed?

Yes, it's tedious, but reverse engineering is rarely easy.

The ASCII for the Hex is this:

B!!!!AE!&!)!&Y_A0!+!'
B!!!!AG!'!(!'Y]A2!*!(

That does look like some sort of protocol to me, with some Initialization Sequence (B!!!!) and commands (AE and AG), but that's just guessing.

Answer 2

The decive is sending data to the computer. All digital data has the form of ones and zeroes, such as 10101001010110010... . Most often one combines groups of eight such bits (binary digits) into bytes, so all data consists of bytes. One byte can thus represent any of the 2^8 values 0 to 2^8 - 1 = 255, or, in hexadecimal notation, any of the numbers 0x00 to 0xFF.

Sometimes the bytes represent a string of alphanumerical (and other) characters, often ASCII encoded. This data format assigns a character to each value from 0 to 127. But all data is not ASCII-encoded characters.

For instance, if the device is a light-intensity sensor, then each byte could give the light intensity as a number between 0 (pitch-black) and 255 (as bright as it gets). Or, the data could be a bitmap image. Then the data would start with a couple of well-defined structures (namely this and this) specifying the colour depth (number of bits per pixel, i.e. more or less the number of colours), the width, the height, and the compression of the bitmap. Then the pixel data would begin. Typically the bytes would go BBGGRRBBGGRRBBGGRR where the first BB is the blue intensity of the first pixel, the first GG is the green intensity of the first pixel, the first RR is the red intensity of the first pixel, the second BB is the blue intensity of the second pixel, and so on.

In fact the data could mean anything. Whay kind of device is it? Does it have an open specification?