Skip to Content

Register Login

Industry Online Support

Technical Forum

Navigation

Strange behavior of int

Created by: Oleh at: 7/26/2011 12:24 PM (11 Replies)

Rating (2)

Thanks 5

Result pages:: |; 1 | 2; |

12 Entries

Entries per page: 10 | all

7/26/2011 12:24 PM	Rate (0)
Oleh Posts: 12 Rating: (0)	Hi guys, searched and don't found some useful solution. I have unusual problem. I have 2 int variables: HourCounter and CounterF7 in Siamatic Manager simulation they seem to be the same, but in VAT and PLCSim simulation, they are different. Actually, CounterF7 changes very strange, i want him to increase 1 by 1, as HourCounter, but it increases every 32 items (value 20 in HEX). so to clarify: HourCounter CounterF7 1 0 2 0 3 0 . . 31 0 32 32 (20 Hex) 33 32 . 63 32 64 64 . 96 96 . 127 128 160 0 and so on... Attached pics(1,3 from simatic manager; 2,4 from vat) Attachment Pics.zip (28 Downloads)
Oleh Posts: 12 Rating: (0)
Suggestion To thank Quote Answer

7/26/2011 2:12 PM	Rate (0)
hdhosseini Diamond Expert Joined: 1/28/2009 Last visit: 6/5/2023 Posts: 6804 Rating: (1345)	Oleh Hi guys, searched and don't found some useful solution. I have unusual problem. I have 2 int variables: HourCounter and CounterF7 in Siamatic Manager simulation they seem to be the same, but in VAT and PLCSim simulation, they are different. Actually, CounterF7 changes very strange, i want him to increase 1 by 1, as HourCounter, but it increases every 32 items (value 20 in HEX). so to clarify: HourCounter CounterF7 1 0 2 0 3 0 . . 31 0 32 32 (20 Hex) 33 32 . 63 32 64 64 . 96 96 . 127 128 160 0 and so on... Attached pics(1,3 from simatic manager; 2,4 from vat) Dear Oleh One possible reason may be the address you mention overwritten in other part of the program.The value in the VAT is the finally calculated in one CPU's scan time!Upload your codes to check the flow of the program. Regards
	Demo Channel on Youtube
Suggestion To thank Quote Answer

7/26/2011 2:42 PM	Rate (0)
Ayman Elotaify Platinum Expert Joined: 5/28/2008 Last visit: 6/4/2023 Posts: 4448 Rating: (827)	Hello, I agree with hdhosseini. Please check -using cross refence- that the mentioned addresses is not used in other places within your program. May be you are overwrite it. And please let us know what is the type of these words? Are both of them set to be integers? Please check and inform us.
	Ayman Elotaify
Suggestion To thank Quote Answer
This contribution was helpful to 1 thankful Users hdhosseini

7/27/2011 10:05 AM	Rate (0)
Oleh Posts: 12 Rating: (0)	Thanks, I manage the problem changing adresse of CounterF7 variable. So its definetely was overwritten in some other part of code. Can you help me with some other problem? U( L "UhrzeitDB".ZeitHex L L#36000000 ==D ) SPBN G001 L 0 T "HourCounter" G001: NOP 0 this is my code. I want to HourCounter (INT) set to 0 in 10 o'clock for example. What am i doing wrong?
Oleh Posts: 12 Rating: (0)
Suggestion To thank Quote Answer

7/27/2011 10:19 AM	Rate (0)
Aret Posts: 2348 Rating: (259)	Everything depends on what's inside UhrzeitDB".ZeitHex If it's TIME then 1 hour = 3600s, 3600 000 ms, you have 3600 000 0 in your code - it's 10 hours. Also code is a bit wrong - what if you will have 36000001 in your time? Your code wont work. It should be (for example) like this A( L UhrzeitDB".ZeitHex L L#36000000 -D L 0 >D ) A( L UhrzeitDB".ZeitHex L L#36000000 -D L 1000 <D )
Aret Posts: 2348 Rating: (259)
Suggestion To thank Quote Answer

7/27/2011 10:23 AM	Rate (0)
MM1234 Posts: 270 Rating: (52)	Hello Oleh, I don't see why you would need the U(...) there, just load the values, compare them and then either jump to NOPor set the counter. Other than that, are you sure you are comparing right values together? Because rest seems fine to me.
MM1234 Posts: 270 Rating: (52)
Suggestion To thank Quote Answer

7/27/2011 10:37 AM	Rate (0)
Oleh Posts: 12 Rating: (0)	Hi Aret, actually i want that code work only when time is like i set. I want HourCounter set to 0 when the time cames, for example in 11:00:00. But it seems like don't compare at all. I also done this in KOP (attached pic 1.jpg), it looks ok, and it's realy simple code, i don't know what is wrong. MM1234, yes, i know that U() is not really nesessary, but i was just trying to make it work also, when i trying 2.jpg it works ok. Attachment AT.zip (30 Downloads)
Oleh Posts: 12 Rating: (0)	Last edited by: Oleh at: 7/27/2011 11:05 AM Last edited by: Oleh at: 7/27/2011 10:39 AM
Suggestion To thank Quote Answer

7/27/2011 2:47 PM	Rate (2)
Aret Posts: 2348 Rating: (259)	Oleh, please read carefully my previous post. The problem is - you get sequence like 3599950 3599965 3599980 3599995 3600005 ... So there is NO 36000000. My algo (modified correctly and implemented in pic2) will work, if time is between 10 hours and 10 hours 0 minutes 0 seconds 999 ms - it will work on sequence mentioned above. It's just like comparing 2 REALs - proper way is comparing difference <0.000001, comparing 2 reals by ==R is wrong PS. Use OB1x for this task
Aret Posts: 2348 Rating: (259)	Last edited by: Aret at: 7/27/2011 2:49 PM Last edited by: Aret at: 7/27/2011 2:48 PM
Suggestion To thank Quote Answer
This contribution was helpful to 2 thankful Users MM1234 Oleh

7/27/2011 3:05 PM	Rate (0)
Oleh Posts: 12 Rating: (0)	Aret, thanks a lot! I've done it like in 2 picture, thats work ok for me. Just to clarify, we cant use comparing in that case becouse of too much precision? Thats why i will never get the same value, becouse its changing too fast?
Oleh Posts: 12 Rating: (0)
Suggestion To thank Quote Answer

7/27/2011 3:46 PM	Rate (0)
MM1234 Posts: 270 Rating: (52)	That's a good one Aret, it's usualy those "small" things that cause "huge" troubles in searching for the error. :-) Also (seeing the cause as a true general after the battle), less code might be to integer-divide the number from clocks by 1000 and then compare it with number 1000times smaller, it's extra division, but less comparison, it would be interesting to look up the instruction book which one would be faster (and it would also be quite boring/annoying thing to actualy do :D). Aret It's just like comparing 2 REALs - proper way is comparing difference <0.000001, comparing 2 reals by ==R is wrong But not in case they have same exponent, right? Otherwise you would need to divide the one with bigger exponent accordingly to match the smaller exponent to get onto same precision.
MM1234 Posts: 270 Rating: (52)	Last edited by: MM1234 at: 7/27/2011 3:48 PM Can''''t quote :-/
Suggestion To thank Quote Answer

|

Share this page:

Share this page on...

Follow us on