Various message length in serial port communication

[bikekowal]

Hi all,

I have a following problem. I am using LabView to communicate with a device over serial port. The device should not be asked more frequently than once a second. I am sending a question and should get an answer looking as follows:

FC 00 09 xxxx 4352

where xxxx should be 9 bytes long. Unfortunaltely the message has a various length - it always has a prefix and a tail but the length of part xxxx is various. I have already found this:

http://digital.ni.com/public.nsf/allkb/2D5F972CD550BBC386256F0B005E9A63

How can I cope with that problem? I enclosed a screenshot of a vi block diagram.

[grahamwebb]

I would remove the delay before the serial read, and set the bytes to read to a large number so you can get all available data back from the device. Then you might see what the device is trying to send.

[bikekowal]

Thank for a reply.

How to set the bytes to read? In which block function?

[johnsold]

Use a Property Node set to Bytes at Port.

Your first post has a contradiction: You say xxxx should be 9 bytes long and then you say the length of the xxxx portion varies.  Which is correct?  Is the 09 byte preceding xxxx actually a "length of data" byte?

Is the "tail" always 4352? Is there a termination character?

What kind of device sends this data? Do yuo have documentation of the communications protocol?

There may be easier ways to approach this.  Please give us as much information as you can.

Lynn

[bikekowal]

OK, i will add that property node.

There should be 9 bytes at xxxx but in fact the number of bytes is different every time I read a message. The tail is always 4352 and there is no termination character (or I don't know anything about that).

This data is sent by a temperature controller which I am using. It is not a mass product, and I have a documentation only in Polish. Should I send it?

[johnsold]

Most of us do not read Polish, so do not post it unless someone specifically requests it.

You can use Match Pattern to search for the 4352 and another one to search for the prefix string.  The string between the two search strings will be your data string.

I suggest that you look through the documentation carefully to see if it explains the variable number of bytes.  That variation could make it difficult to interpret the result.

Lynn

[bikekowal]

I have already used matched pantern and this is how i know this results. The doxumentation says clearly about nine bits.

[johnsold]

If it is always sending 9 bytes, I do not understand why you sometimes see more.  Can you post some data where you see varying numbers of bytes?

Lynn

[bikekowal]

That is the problem. I don't understand this either.

I attached few screenshots of buffer - you can see that the number of bytes is random.

[johnsold]

It looks like the extra bytes are always 00.  This suggests that default values are getting in somehow.

1. How are you configuring the port? Handshaking? Termination character enabled? Timeout?

2. Do you get any errors? What are they?

3. Try connecing an indicator to the "return count" output of the Read.  Does the count match the size of the strings you see?

Lynn