Skip to Content

Register Login

Industry Online Support

Technical Forum

Navigation

?> How to clear bits of an array of a user defined type?

Created by: William_B at: 12/12/2012 10:03 AM (6 Replies)

Rating (2)

Thanks 2

7 Entries

12/12/2012 10:03 AM	Rate (0)
William_B Posts: 5225 Rating: (1160)	Hi-ya, A purely academical question.. What is the best way (quickest programmed and quickest processing) to clear bits inside an array with a user definable type? My method is seen in the JPG file where all the bits inside c[0] is empty. MOVE_BLK and FILL_BLK works only for arrays with standard types (int / bool ++). For S7-300 / 400 you could say.. L 0 T DIW4 T DIW6 ... For the S7-1200 (under optimised data usage) this is not a possibility. Furthermore - - the idea is to utilise the concept of object orientated programming. Should one create a custom SCL code? ..using type ANY perhaps? What do you guys think? Attachment clear bits in an array of an UDT.jpg (1016 Downloads)
William_B Posts: 5225 Rating: (1160)
Suggestion To thank Quote Answer

12/12/2012 2:35 PM	Rate (1)
0_Einstein Diamond Expert Joined: 7/7/2010 Last visit: 1/27/2023 Posts: 13080 Rating: (2123)	If you create your array of bits instead as a byte or word, you can use a standard FB to set/get using the AT keyword to access the bits, or just add .x0-.x15 to access the bits within.
	science guy
Suggestion To thank Quote Answer
This contribution was helpful to 1 thankful Users William_B

12/12/2012 4:19 PM	Rate (1)
William_B Posts: 5225 Rating: (1160)	huggy_d1 If you create your array of bits instead as a byte or word, you can use a standard FB to set/get using the AT keyword to access the bits, or just add .x0-.x15 to access the bits within. Hi - thank you for your note. It most pushed me in the right direction. It took me a while to figure out what you meant. For those who wish to follow the conversation, I suggest you read in S7-1200 system manual: 4.4.10 Accessing a "slice" of a tagged data type 4.4.11 Accessing a tag with an AT overlay Using the SLICE method, you will loose the capability to give descriptive tags. Compare usage: c[1].x0 against c[1].empty In the second method it is more clear that conveyor 1 is empty. BUT the SLICE method can be employed when the design phase is done. ("Find & Replace" is your friend.) The catch with the AT command is that the FB should NOT be optimised. (At first I could not get this method to work.. maybe someone should write a tutorial on this?) (My mistake was that my 1st trial FB was optomised.) The FB should follow the S7-300 / 400 way of internal memory spacing. When using many IEC_TIMERS this method will kill your memory consumption. So it seems as though it is a give-en-take situation. But thanks.. I have learnt something today. w
William_B Posts: 5225 Rating: (1160)
Suggestion To thank Quote Answer

12/12/2012 4:33 PM	Rate (0)
0_Einstein Diamond Expert Joined: 7/7/2010 Last visit: 1/27/2023 Posts: 13080 Rating: (2123)	There is probably another way to do it if your UDT and all access to it is using the S7-300/400 compatible (vs optimized) configuration. I "like" that I can set up things using the optimized configuration. In practice, I find I intentionally do not use optimized for most everything that matters. To communicate to drives, you often need multiple access methods (direct to bit level and at byte/word level). To communicate to HMIs, you can run into problems with some configurations if you intend on using WinCC Flex or something other than basic panels, or create generic screens that need to work in multiple HMI and PLC types. The list goes on as to why optimized is great in theory, but in practice, typically causes more problems than it fixes.
	science guy
Suggestion To thank Quote Answer
This contribution was helpful to 1 thankful Users William_B

12/13/2012 8:28 AM	Rate (0)
William_B Posts: 5225 Rating: (1160)	huggy_d1 The list goes on as to why optimized is great in theory, but in practice, typically causes more problems than it fixes. I will have to look into that. To my mind optimised method should be a good thing. I will have to check the one vs the other. In my case I utilise many timers. The state bits inside IEC_TIMER type can cause havoc on memory consumption using the non-optimized data structuring. Granted symbolic referencing is slower than absolute referencing. BUT structural changes (moving parameters inside FB/DB) inside a program is handelled with more ease. To further our discussion I have done some memory trials. I have created x4 FBs with DBs with various structures. The structures should have the same outcome as the JPG in my 1st post. [1] A 9 element array of a UDT as in the JPG. [2] Use x9 bytes each with x8 AT-construct [3] A 9 element array of a stucture. The structure has 8 boolean variables. [4] 8 bytes - - the 9th byte is not required as clearing the bits can be done by a MOVE function. (The slice method (ex. BYTE[1].X0) should be employed here to read the bits) ..no coding has been done in each of the eamble FBs mentioned above. The memory consumption (combined FB & DB with UDT in method 1) is as follows: METHOD 1 2 3 4 LOAD MEMORY 7894 6529 3931 2358 WORK MEMORY 27 27 19 17 These are very small numbers when comparing it with the available memory. It becomes more significant however when building big structures. From this method I realise that defining a UDT cost too much memory. Using a UDT only makes sense when creating very many FBs using it. ..off topic.. That also makes me wonder if Siemens should not consider abandoning UDT usage in the motion control library FBs. (From a test a saw 15% of 1MB going to UDTs created by the motion control functions.) That is but my opinion. My personal choice from the above sits with method 3. It is still human readable structured. UDT is not your friend when it comes to conserving memory. w
William_B Posts: 5225 Rating: (1160)
Suggestion To thank Quote Answer

12/19/2012 8:42 AM	Rate (0)
Cicero Gold Member Joined: 10/10/2005 Last visit: 11/21/2022 Posts: 1909 Rating: (131)	Hello, you can define a series of components with various data types in a PLC data type in STEP 7 V11. You can assign any name to the PLC data type. The PLC data types can be declared as data types for variables in the block interface in functions and function blocks. Or they can be used as data types in global data blocks. See the informations and example in the following FAQ (link): http://support.automation.siemens.com/WW/view/en/51339155 With best regards Cicero

Suggestion To thank Quote Answer

1/6/2017 11:42 PM	Rate (0)
CS Moderator Moderator Joined: 9/1/2005 Last visit: 1/9/2023 Posts: 4097 Rating: (192)	New question published by AutoGen_76109 is split to a separate thread with the subject FILL MOVE FUNCTION FROM 1 TO 10. Best regards CS Moderator
	My Forum is the place to personalize your profile, track your topics, and maintain content that interests you. Best regards CS Moderator
Suggestion To thank Quote Answer

|

Share this page:

Share this page on...

Follow us on