Hi, am I allowed to ask similar question here?
Essentially I have this same problem with FC6 status: W#16#80B1

This is using Step7 V5.5 w/ S7300 PLC and CP 343-1

Communications are set up for passive TCP/IP with CP enabled for SEND/RECV

My application data has different packet lengths depending on the purpose of the data.
Therefore when idle, my program calls FC6 with a length of 44 bytes (my application header). Inside of that header contains a field for attached data length. So I always process FC6 for 44 bytes, when I receive those 44 bytes my program pulls out the "attached data length" and puts that new length in the next call of FC6.

For example:
"Packet Type 1" comes across network as 712 bytes. (44 bytes header + 668 bytes data)
   - FC6 receives 44 bytes.

   - User program processes bytes to find data length (in this case 668), puts this length in the next FC6 call.

   - FC6 receives 668 bytes.

   - User program now has full telegram.

   - User program sets FC6 length back to 44 bytes in anticipation of next telegram

   - FC6 status = W#16#8081 until next telegram arrives

This procedure I've outlined above WORKS the first time, and only that one time.
The next time data is sent across network, FC6 errors with status = W#16#80B1 when in reality it should have received 44 bytes.

I see in the help document: "Remedy: Run another receive call with maximum receive buffer size"
Could someone explain a little further what that means? Are they suggesting to read 8192 bytes? If this is the case I would have to wait until 8192 bytes collect on the card before I get "NDR". This cannot be, but I am not understanding what else they might mean.

Any help is greatly appreciated! Thanks in advance everyone-

Just one extra detail. When I mention "set FC6 length", I really mean assign length to a common ANY pointer. I am using the same any pointer to call FC6 with, just changing "count" and "address" fields depending on where received data should go.

Examples:

"header"
   - count = 44

   - address = P#myHeader

"packet 1 data"

   - count = 668

   - address = P#myPacket1

----------------------------------------------------------------------------------------------------
Split form FC6(AG_RECV) ERROR 16#80B1.

Thanks for your response, I have spent much time just now verifying the pointers are correct.

I have gone so far as to hardcode static structures with multiple calls of FC6, but it still returns #80B1.
So even when I don't manually build any pointers, it still does not work properly.

BUT, I think I am onto something. I notice that the FC6 AG_RECV "len" output parameter always returns a value of "240".

My application data requests "684" bytes.

Could this be the problem? Maybe I need to issue multiple calls of FC6 where my data exceeds "684" bytes.
Considering an "FC" is not containing static data, I would imagine it would loop through internally until hit application data or 8k. Since I don't see how it would have any memory of previous call.
BUT, it is possible that the CP card retains memory from my last call, so it sees this transaction "User equest 684 byte, CP card deliver 240 bytes, User request 44 bytes, CP card deliver error".

I've done much speculating which I hate to do, but I am trying to think through what may be happening.

I am going to test calling (240 bytes, 240 bytes, 204 bytes - 3 calls to achieve 684 bytes) tomorrow morning (just realized this is possible, but must head to class).

I will continue to look back tonight incase someone has some feedback on my idea.

Thanks-

Kaulquappe

Hello cmgrass,
I still imagine your pointer is wrong. You need an ANY pointer as static variable. Can you post the HEX value for the pointer which causes error 80B1?

Regards,
Kaulquappe

 Hi Kaulquappe, thanks for your interest. I have attached some screen shots which might help.


EDIT: Sorry I should have proofread my comments in the attached picture. Specifically in the green box. Should read D#16#85000490. But I'm sure you'll see what I mean.
Thanks again-

Hi guys, I have this working now! 

I'm not sure the technical reason, nor can I say for sure that this will fix other peoples similar problems.
But what I have done is not allow FC6 to be called every scan. I am essentially letting FC6 run until I receive data. Then I change parameters and call the function again in separate program scan. The way my state machine works, there can't be two consecutive calls of FC6.

So my transaction is as follows (assuming already connected)

Scan 1) logic is ready to receive new data, populate ANY pointer, call FC6

Scan 2) if data received, process this data

Scan 3) if logic is ready to receive new data, populate ANY pointer, call FC6

Scan 4) if data received, process this data

.

.

.

If someone can share technical reason why this is working I would be thankful.

All I know is that it functions similar to "FB64 TRCV", where the difference is that "FC6 AG_RECV" has NO ".EN" input, while "FB64 TRCV" has the enable input. My take away is that I need to treat the "FC6" rung/block enable in similar method to the "FB64" enable input pin/parameter.

It is always the user program causing these issues, as much as I want to blame hardware sometimes I must remember this . It is the fun of programming, you learn a lot!