Decode this protocol

Joined: 11 Aug 2005 Posts: 282

Can someone decode this? I'm trying to write my own data acquisition software for a multimeter. But first, I have to reverse engineer it!

Here are some hex dumps
60Hz:
11 27 3d 47 5e 67 7d 8f 9d a0 b0 c0 d2 e8
0Hz:
11 27 3d 4f 5d 67 7d 87 9d a0 b0 c0 d2 e8
1Mohm:
13 20 35 4f 5d 67 7d 87 9d a0 b2 c4 d0 e8
0.988kohm:
13 27 3d 4b 5f 67 7f 83 9f a2 b0 c4 d0 e8
98.2ohm:
13 27 3d 43 5f 67 7f 89 9f a0 b0 c4 d0 e8
0v:
17 27 3d 4f 5d 67 7d 87 9d a0 b0 c0 d4 e8
5.27v:
17 27 3d 43 5e 6d 7b 81 95 a0 b0 c0 d4 e8
-5.27v:
17 2f 3d 43 5e 6d 7b 81 95 a0 b0 c0 d4 e8

Joined: 05 Aug 2005 Posts: 697 Location: Hard to Explain

I like pointing out the obvious:

It looks like the first byte is the mode the multimeter is in, and the second byte the scale it is set to.

Could the last byte be part of a xsum since they are the same through all of the dumps?

Joined: 11 Aug 2005 Posts: 282

I think the "e8" is a stop byte.
http://www.lsc-group.phys.uwm.edu/beowulf/medusa/construction/multimeter_instructions.html
My multimeter uses a different protocol. It's still 14 bytes per packet, however.

Here's a link to the product page (including the provided program):
http://www.p-mastech.com/products/04_dm/ms8226t.html

Posted: Thu Sep 13, 2007 9:26 pm

I tend to agree with ry.david's guess that the 1st byte is the measurement mode (freq, resistance, voltage, etc.)

I don't agree that the last byte is a checksum. Based on the +/- 5.27V readings, there isn't a separate checksum byte, since those just differ in an early byte and are, otherwise, identical.

I think that the final E8 byte is a termination/sample-separator byte.

Beyond that, you'll need more data to fully decode stuff. Another approach - seen in the 1st link you posted - is to look at each sample as ASCII. Perhaps that shows you something?

Joined: 11 Aug 2005 Posts: 282

It just gives junk when read as ASCII.

Joined: 05 Aug 2005 Posts: 697 Location: Hard to Explain

Could you hook it up to a variable power supply and get many different samples to compare with? That would make the voltage side a little easier, and probably once you figure that out, then it will probably carry over to resistance and freq.

Joined: 11 Aug 2005 Posts: 282

I got an idea: how about download the original program and try inputting your own data to see how it works? To start, it uses 2400 8N1 from what a serial sniffer could tell.