Created Date: April 16,2007

Created by: FACTSTech

>So the question is, does STOP prevent asynchronous things like writing to the CPU memory from happening?

No. The BASIC Interpreter executes each statement to completion before going to the next statement.

I don't see any reason why line 19800 would not execute. If there was a problem with the F2-CP128 backplane interface and the S205_ instruction could not execute then the interpreter would print an error out of the command port and the program would stop.

What is the value of variable AD1 when line 19800 executes (add PRINT statement to see the value) and what value do you see in V1476 when the BASIC program stops?

In order to use calculated addresses in the S205_ or BMOVE statement see page 2.3 and 2.13 in the F2-CP-M manual (http://www.facts-eng.com/manuals/f2cpm.pdf) under the 'Advanced ' heading.

Created Date: April 17,2007

Created by: AZRoger

FACTSTech,

Thanks for the pointer to the pointer stuff. Page 2.13 is exactly what I needed. Odd about Hex 7 refering to VMem 6 ... but it works. The corrected lines are below.

19300 S205_(AD1+1)=MINUTEOFDAY : REM save the minute of day

19400 S205_(AD1+2)=AD2+I : REM point to the correct force bit word

19500 S205_(AD1+3)=K2 : REM save the bit(s) that is(are) to be forced

The other problem was in regular ladder logic and DS5. Aw I watched data view, the address sometimes changed to the expected new address, but was then reset to the first Vmem address. However, when the CoPro was stopped, the resets didn't happen.

I check the BASIC program and there were only the two references to that address. One to pick up the current value and one to stuff the new one. The XREF view in the ladder logic had NO references to those addresses. Ugh.

I think that problem was the result of my recovery from a DS5 crash. I had some initialization logic (LDA, OUT(box) sequences) as part of a long rung. I had deleted those rows from the rung. DS5 crashed during the SAVE to disk right after that. I didn't go through the "guaranteed " proceedure to recover. I just fired it up. It said the disk and PLC didn't match. I selected use disk and thought I was OK. However, it behaved like the graphical ladder and the functional ladder didn't match. The initialization logic was still running. I added a row to that long rung, put in a LDA, deleted the LDA and saved again. The problem went away.

Moral of THAT story: Take backups often and be prepared to use them. http://forum1.automationdirect.com/board/smile.gif

Thanks again for you help.

Roger

Created Date: April 16,2007

Created by: AZRoger

I am having two separate problems with the BASIC CoProcessor. The first has to do with passing addresses into the processor from ladder logic. The second has to do with the STOP command.

Let's cover STOP first. Line 19900 has a stop command so I can look at the variables and see what's happening. The line above it 19800 appears not to work. The target word is not filled in (data view does not see the change). If I copy that program line to the command line and run it, it does execute. So the question is, does STOP prevent asynchronous things like writing to the CPU memory from happening? This would be an odd behaviour, IMO, so it probably isn't what's causing the problem. But something is.

The second question has to do with getting an address from the CPU and using it in a basic statement to set a specified word in the CPU memory. Line 19000 gets an address into variable AD1. Lines 19300, 19400 and 19500 try to use that address to put values in the PLC memory. These lines don't work. Similar line with math in the address part (Lines 20300-20600) work fine. I put a 0 (zero) at the front of the VH(expression) to see if the math would work then, it didn't. I need to have the Ladder tell the Basic CoPro where to put data. And I 've run out of ideas.

17900 REM Load all the Stages in one Pass

18000 BMOVE R,SG(1000),SG(1577)

18100 K5=0 : REM counter for words changed

18200 FOR I=0 TO 23

18300 K1=OLDSTATES(I).XOR.SHARED(I*2)

18400 IF K1=0 THEN GOTO 20000 : REM no change

18500 K2=K1.AND.SHARED(I*2) : REM See is changed bit is ON

18600 IF K2=0 THEN GOTO 20000 : REM it went off so don't count it

18650 OLDSTATES(I)=SHARED(I*2) : REM copy the new value

18700 K5=K5+1 : REM count the mismatch

18900 REM: K2 has the bit(s) that need to be set

19000 AD1=S205_VH(1476) : REM get address for first word to hold data

19100 AD2=S205_VH(1473) : REM get address of first force bit

19200 PRINT2 AD1; " ";AD2

19300 S205_VH(0+AD1)=MINUTEOFDAY : REM save the minute of day

19400 S205_VH(0+AD1+1)=AD2+I : REM point to the correct force bit word

19500 S205_VH(0+AD1+2)=K2 : REM save the bit(s) that is(are) to be forced

19600 AD1=AD1+3 : REM advance for next data

19700 IF AD1=S205_VH(1475) THEN AD1=S205_VH(1474):REM if last, start over

19800 S205_VH(1476)=AD1 : REM set address for next sample

19900 STOP

20000 NEXT I

20100 IF K5=0 THEN GOTO 20800 : REM skip the next part if no changes

20200 REM need to add to words changed and minutes that had changes

20300 K6=S205_VB(1460+THEDAY) : REM get current count

20400 S205_VB(1460+THEDAY)=K5+K6

20500 K7=S205_VB(1450+THEDAY) : REM get current count

20600 S205_VB(1450+THEDAY)=K7+1

20700 PRINT2 "At ";MINUTEOFDAY; " ";K5; " words changed K6,K7 ";K6; ", ";K7

20800 REM end of counter updating section

Any help or pointers would be appreciated. Thanks.

Roger