It's addressing again...(step 7)

Hi folks,

ok here's one... can I use indirect addressing with a function block call? And can I do it within a loop. I don't think the loop is a problem but not sure about the FB. Basically I was thinking of including the call in an existing loop to save typing, it may have made performance a bit better too, but I therefore need to indirectly address the parameters of the function. Here's the gist of it.


Thanks in advance

Kenny

ps: all the parameters after the call statement are coming up invalid

Kenny
I think you can employ a method recently discussed in JesperMP's thread about UDTs. One of S7Guy's ideas.

If I understand, you have 17 Vac pumps?

Can you create a DB with an array of 17 UDTs. Each UDT being the setpoint, gain, actual etc.

Open this as an instance DB (OPN DI 7) then use UC rather than CALL to call the FB. This doesn't prompt you for the input/output parameters.

After calling the block, increment AR2 by the length of the UDT and call it again, testing for reaching 17.

I think there's an implicit link between an instance data block and AR2.

In your example, Is DB8 only there because Step7 needed one when using CALL FB??.

pp

ps. Step forward S7Guy!

Edit: Not sure how to tie each instance up with the real-world I/O yet.

It can be done almost as you described it. The only thig is that you cant use adressregister values as paramters you have to load and store them in the Temp stack (local data). Your FB call could probably look somthing like this

OPN "Vacuum values" DB7

//Use TAR2 if multi instans capability is used for the FB
//Ar2 is used to keep track off the offset in the multi instans DB
//Saves Ar2 in tempstack so it can be restored later with LAR2
TAR2 #TempAr2 //Temp Dword

L 0
LAR1
LAR2

L 17
strt: T #COUNT

L PIW [AR1,P#512.0]
T DBW [AR2,P#2.0]

L DBW [AR2,P#0.0]
T #Setpoint //Temp Word

L DBW [AR2,P#2.0]
T #Actual //Temp Word

L DBD [AR2,P#6.0]
T #Gain //Temp Dword

L DBD [AR2,P#10.0]
T #Intregral //Temp Dword

L DBD [AR2,P#14.0]
T #Diff //Temp Dword

L DBD [AR2,P#4.0]
T #SignalOut //Temp Dword


CALL "Vacuum controller" , "Vacuum parameters" FB1 / DB8
setpoint := #Setpoint
actual := #Actual
gain := #Gain
integral := #Intregral
diff := #Diff
signal_out:=#SignalOut

L DBW [AR2,P#4.0]
T PQW [AR1,P#514.0]

+AR1 P#4.0
+AR2 P#20.0

L #COUNT
LOOP strt

//Restores Ar2
LAR2 #TempAr2

Thanks guys,

the UDT method is a bit of a mystery but I'll get to it eventually. I saw Jesper's thread but it gave me a sore head. I think the main thing is I hadn't considered the interaction with the instance DB. Bratt's method makes sense and is still less trouble than calling the FB 18 times and filling in 18 sets of addresses. Unfortunately by the time I'd seen this that's exactly what I've done!! Doh! I'll revisit it tho as it streamlines the thing somewhat - the original structure was an FC with the loop which transferred the PIWs to DB and DB to PQW, then the next FC called the FB for each vac regulator. I just thought it would be nifty to incorporate the whole thing in a loop.

Cheers

Kenny