STL Help again!!

STL??? · Jul 26, 2006

Hi,
I need to add some code to an existing program to provide error messages for all sensor faults. The address range of the sensors are M140.0 -> M159.7. The current method utilises pointers to load the address i.e LAR1 P#M140.0 into a Diagnostic block where it is broke down in to a number which decides which Textlist entry is displayed on the HMI.

Before this i want add my timeout block to generate an Error message if the sensor is active for more than a pre-determined time.

In the simulator the code below seems to work ok, However in actual operation the timer(#fault_Timer) doesn't timeout?

Could anybody give me some "Pointers"(no pun intended!!)

FUNCTION "Sensor Timeout" : VOID

TITLE =

VERSION : 0.1

 

VAR_INPUT

State : BOOL ; //Test for 1 ~ State=True / Test for 0 ~ State=False

Fault_Timer : TIMER ; 

Fault_time : S5TIME ; 

Reset : BOOL ; 

DB_No : BLOCK_DB ; 

Offset : DWORD ; //Value to Minus From Address Register

Station_in_front : BOOL ; 

END_VAR

VAR_OUTPUT

Alarm_State : BYTE ; //1 = On 2 = Off ~ 0 no processing

END_VAR

VAR_TEMP

SaveAR1 : DWORD ; 

Address : DWORD ; 

Byte_Value : WORD ; 

END_VAR

BEGIN

NETWORK

TITLE =Reset

 

A #Reset; 

R #Fault_Timer; 

NETWORK

TITLE =Run conditions

 

 

 

AN #Station_in_front; 

BEC ; 

 

NETWORK

TITLE =Zero Alarm status list

 

L 0; 

T #Alarm_State; 

 

NETWORK

TITLE =Save the contents of AR1 for later processing

 

TAR1 ; 

T #SaveAR1; 

T #Address; //Load from AR1

NETWORK

TITLE =Check Address range

 

L DW#16#FFFF; //Byte filter

AD ; 

SRW 3;

T #Byte_Value; 

 

L 140; //Min Address

>=I ; 

A( ; 

TAK ; 

L 159; //Max Address

<=I ; 

) ; 

JC OK; // Address range OK

JU END; // If out of range ~ Restore AR1 + End Block

NETWORK

TITLE =Sensor should be off

 

OK: A #State; 

JC NW5; 

 

AN M [AR1,P#0.0]; 

L #Fault_time; 

SD #Fault_Timer; 

 

A #Fault_Timer; 

JC NW6; 

JU END; 

NETWORK

TITLE =sensor should be on

 

NW5: A M [AR1,P#0.0]; 

L #Fault_time; 

SD #Fault_Timer; 

 

A #Fault_Timer; 

JC NW6; 

JU END; 

NETWORK

TITLE =Send Error Message state ON/OFF

// 

// 

//

NW6: AN #State; 

JC OFF; 

L 1; 

JU WRT; 

OFF: L 2; 

WRT: T #Alarm_State; 

NETWORK

TITLE =Send Sensor Number Message 

//Message 1 = 40.0 for m140.0 

//

L #Address; 

L DW#16#FFFF; // mask off all except Byte/Bit area

AD ; // Apply 32 bit Mask

T #Address; 

L #Offset; // Minus Offset to Point to first message address 

SLD 3; // Move to Byte Area of the Pointer

-D ; // Subtract from Original Value

LAR1 ; 

 

 

OPN #DB_No; 

A #Fault_Timer; 

= DBX [AR1,P#0.0]; // Write Message

 

 

NETWORK

TITLE =Block End restore address register 1

 

END: L #SaveAR1; 

LAR1 ; 

 

 

END_FUNCTION

Thanks again
STL

STL??? · Jul 26, 2006

Online:

SimonGoldsworthy · Jul 26, 2006

How many times are you calling this block ? I presume you are using a different timer for each call.

Just an immediate comment, you should not execute a BEC in a FC before you have evaluated the logic for an output. The FC will always write a value to the output, so if it has not been evaluated the output will be a random value (the value will depend on the use of local data by other blocks, so sometimes it appears to work ok, sometimes it doesn't).

Gambrinus · Jul 26, 2006

Hm, I don't like to use AR1 for parameter passing... I amn't sure do you access correct timer or not. Using of MD defined as area-internal pointer is more easy and safe. IMHO, something wrong with it. But if someone else before you did it through AR1... Try the following:

FUNCTION "Sensor Timeout" : VOID
TITLE =
VERSION : 0.1

VAR_INPUT
State : BOOL ; //Test for 1 ~ State=True / Test for 0 ~ State=False
Fault_Timer : TIMER ;
Fault_time : S5TIME ;
Reset : BOOL ;
DB_No : BLOCK_DB ;
Offset : DWORD ; //Value to Minus From Address Register
Station_in_front : BOOL ;
END_VAR
VAR_OUTPUT
Alarm_State : BYTE ; //1 = On 2 = Off ~ 0 no processing
END_VAR
VAR_TEMP
SaveAR1 : DWORD ;
Address : DWORD ;
Byte_Value : WORD ;
END_VAR
BEGIN
NETWORK
TITLE =Reset

A #Reset;
R #Fault_Timer;
NETWORK
TITLE =Run conditions

AN #Station_in_front;
BEC ;

NETWORK
TITLE =Zero Alarm status list

L 0;
T #Alarm_State;

NETWORK
TITLE =Save the contents of AR1 for later processing

TAR1 ;
T #SaveAR1;
L DW#16#FFFF; //Byte filter
AD ;
T #Address; //Load from AR1
NETWORK
TITLE =Check Address range

SRW 3;
T #Byte_Value;

L 140; //Min Address
>=I ;
A( ;
TAK ;
L 159; //Max Address
<=I ;
) ;
JC OK; // Address range OK
JU END; // If out of range ~ Restore AR1 + End Block
NETWORK
TITLE =Sensor should be off

OK: A #State;
JC NW5;

AN M [#Address];
L #Fault_time;
SD #Fault_Timer;

A #Fault_Timer;
JC NW6;
JU END;
NETWORK
TITLE =sensor should be on

NW5: A M [#Address];
L #Fault_time;
SD #Fault_Timer;

A #Fault_Timer;
JC NW6;
JU END;
NETWORK
TITLE =Send Error Message state ON/OFF
//
//
//
NW6: AN #State;
JC OFF;
L 1;
JU WRT;
OFF: L 2;
WRT: T #Alarm_State;
NETWORK
TITLE =Send Sensor Number Message
//Message 1 = 40.0 for m140.0
//
L #Address;
L #Offset; // Minus Offset to Point to first message address
SLD 3; // Move to Byte Area of the Pointer
-D ; // Subtract from Original Value
T #Address

OPN #DB_No;
A #Fault_Timer;
= DBX [#Address]; // Write Message

NETWORK
TITLE =Block End restore address register 1

END: L #SaveAR1;
LAR1 ;

END_FUNCTION

But better to transfer data offset to your function not through AR1, but through MD...

Gambrinus · Jul 26, 2006

STL, what you see online in AR1 in network 1?

STL??? · Jul 26, 2006

Hi Simon,
Currently there are 2 FC's which handle the sequencing/Error display. the syntax for the original setup is as follows:

AN "Part Control"
JC M042
A "VKE=1"
S "START-TUE"
S "START-TWA"
L S5T#10MS
T DBW 12
L S5T#500MS
T DBW 14
LAR1 P#M 240.6
AN "REQUEST START CYCLE"
JCN M008

LAR1 P#M 146.5
A "ST1".SENSOR._8S6
JCN M004

The Labels M008 & M004 denote the Jump destination depandant on if a High or Low signal is required to display. this is where my block call will be x2 for each block (4 in total)

Both FC101 and FC102 share the same pointer addresses, which i think is risky ~ But it works!

Just an immediate comment, you should not execute a BEC in a FC before you have evaluated the logic for an output

Point taken, i forgot about that

I think i see the problem, although the pointer is accurate to the program stage,it isnt set to 1 anywhere and if viewed outside the block call is not visable as it is used in the second call?

Need to have a rethink!!

Thanks again
STL

Gambrinus · Jul 26, 2006

Another question like Simon question: how do you define TIMER parameter in FC call?

Gambrinus · Jul 26, 2006

Using the following higher-language constructions may cause the contents of the DB register and address register AR1 to be modified:

Fully-qualified DB access (for example, DB20.DBW10) as an actual parameter for a function (FC)
Function block (FB) and multiple-instance calls
Structure components of a formal parameter as an address within an FC or FB
Structure components of a formal parameter as an actual parameter for an FC or FB

What is A "ST1".SENSOR._8S6? Access to structure element...

STL??? · Jul 26, 2006

Another question like Simon question: how do you define TIMER parameter in FC call?



      CALL  "Sensor Timeout"

       State           :=FALSE

       Fault_Timer     :=T111

       Fault_time      :=S5T#30S

       Reset           :="RESET ALARM"

       DB_No           :="HMI~ALARM"

       Offset          :=DW#16#64

       Station_in_front:="ST1 IN LOAD POSITION"

       Alarm_State     :="HMI~ALARM".SIGNAL_STATUS_ST1

Thanks for all the reply's, You have got my thought process started again.

I'm going to completely rewrite this code, i wasnt happy with it anyway!
Unfortunately timewise i have to work with whats already there.

Gambrinus - thanks for the alternative addressing method.

regards STL

STL??? · Jul 26, 2006

What is A "ST1".SENSOR._8S6? Access to structure element...

This is why i think the code wont work in my timeout block. It is a datablock bit address mapped from an ASI PID area.

A M [AR1,P#0.0];

As the Memory bit value is never written i'm assuming its zero, the actual decision based logic is via A "ST1".SENSOR._8S6

So i think the timer can never start!

I have to stay with P#M140.0 method due to time and space, but what i'm going to try and reference the DBX sensor address based on the pointer.

Hope this makes sense? If i'm way off base here please tell me

Thanks

Gambrinus · Jul 26, 2006

STL??? said:
This is why i think the code wont work in my timeout block. It is a datablock bit address mapped from an ASI PID area.

A M [AR1,P#0.0];

As the Memory bit value is never written i'm assuming its zero, the actual decision based logic is via A "ST1".SENSOR._8S6

So i think the timer can never start!

I have to stay with P#M140.0 method due to time and space, but what i'm going to try and reference the DBX sensor address based on the pointer.

Hope this makes sense? If i'm way off base here please tell me

Thanks

If ASI is mapped to DB it have sense. But I nevertheless suggest you - don't use AR1 directly for address passing to FC, FB. And when you access UDT (is "SENSOR" an UDT?) elements save AR1 before it.

EDIT: I'm stupid sclerotic.

A "ST1".SENSOR._8S6 doesn't change AR1 because address is given directly. AR1 changes when access to structure defined as FC, FB formal parameter occurs.

SimonGoldsworthy · Jul 26, 2006

I'm not clear on the overall picture here. The sensors are addressed in the range M140.0 to M159.7 so that makes 160 sensors. Are there only a certain number of sensors that can be active at any given time ? (otherwise you need a different timer for each sensor)

Gambrinus · Jul 26, 2006

Timers must be addressed indirectly too, it's a second question. But it requires 160 unused timers. May be STL?? haven't so much.

STL??? · Jul 26, 2006

Hi Guys,
Apologies for the confusion, Originally the ASI mapping was to M memory, But it was decided to change this because the machine has 2 stations and two ASI Masters.

Why change it?

The original programmer mapped ASI Master 1 sensors to M140.0 to M159.7 then called the sequence program for the first station.
He then Mapped ASI Master 2 sensors to the same M addresses and then called the (copied) sequence program as station 2.

With the way it was done and monitoring issues, it was decided to give each masters sensors/valves unique addresses.

The changes made have reduced the scan time from 90mS down to 50mS on the older 315DP that is fitted.

Timer wise Gambrinus is right, i have not got many left at all.

Heres what i would like to do

M 140.0 (000) = DB5.DBX22.0
M 140.1 (001) = DB5.DBX22.1
M 140.2 (002) = DB5.DBX22.2
M 140.3 (003) = DB5.DBX22.3

Create a FC, so the contents of AR1 (P#M140.0 - P#M159.7) are utilised to create/point to the DBX address of the Sensor.
[the (000) = Textlist offset]

This then is checked for its status High/Low. If the criteria isnt met a timer counts down a predetermined time and sets a timeout Message unique to that sensor.

Just trying to figure out a way of doing this!!

STL??? · Jul 26, 2006

Simon,
The way the code works is that if a sensor is evaluated and the condition not met, It conditionally jumps to a label where a FC generates a protool textlist value.

So there is only one pointer value worked on each call of the sequencer.
The problem is there is a sensor number on the display for diagnostics, But no actual error message to Log which sensor was defective.

The operators currently are having a field day with this logging downtime as sensor faults!!

We need traceability.

STL Help again!!

STL???

Member

STL???

Member

SimonGoldsworthy

Member

Gambrinus

Member

Gambrinus

Member

STL???

Member

Gambrinus

Member

Gambrinus

Member

STL???

Member

STL???

Member

Gambrinus

Member

SimonGoldsworthy

Member

Gambrinus

Member

STL???

Member

STL???

Member

Similar Topics