TON timer e.g. behaves not like I used to on other systems

am_rts

Hi, 

I write in this post just not to make many new topics. 

The point is that  TON timer e.g. behaves not like I used to on other systems.

I provide two links with screencasts

This is the idea what i need , and how it works on Delta PLC

http://screencast.com/t/O4BncVhluSc7

This is how it does not work in TIA V13  with 1214C 

http://screencast.com/t/n5qhYHWYMs0Y

the idea is that timer shoud be ON for one scan , then counts till next ON moment for one scan and so on. 

Best regards,

Evgeniy

Hi.

On the Delta you had this..

On the S7-1200 you tried this..

What will work is this...

The trouble with your attempt lies within the "fine print". Every time you access any element of your timer (like "sin_step_timer1".Q) the timer is refreshed and time value updated. It has been designed that way, so we have to work around that.

Helpful?

As an afterthought... why do you have this piece of code to count time? If you want to count time accurately you can use the TONR (retentive TON).

w

eam_rts

Dear William, thanks for your clear reply. I thought of that something with double calling FB , but just could not figure out in mind, due to habit of thinking in terms of other plc..

So , if we take the network  with 1s timer

tmr.db.q

--|/|-----------------tmr.fb--

1. Q=0 on previous scan, but time in timer already 1 s
2.  NC contact calls tmr.fb to read data of Q, in this moment Q rises in datablock.
3.  tmr.fb which is after NC again invokes code of tmr FB ,but this time because already Q=1, tmr resets and Q=0 so Q never rises for ctu block followed.

Right ?

I wrote this post right before NY2016 in hope that someone replies , because our local Siemens support office had holidays. But 4-6 of Jan they worked and answered me too, their advise was as in attached file.

In this example CTU catches the moment when Q rises, before pair of NC ----- TMR FB clears Q. 

In your example I would need one more M bit to store Q, but yet  had just one line of network in favor of code simplicity, thanks anyway. 

The idea of my example was - to have 360/720 ..so on microsteps of the period for sin() function, on each microstep invoked by tmr sinus function would work. and also random noise generator to add noise to sin() function, then i was going to use this signal to test averaging function. 

BTW could you explain multi-instance term for FB ? I coud not get clear idea, system manual says it means nested FB, when one FB (multi-instance ) could have inside other FB , storing their instances inside own instance, what is the application of this and restiction (for any FB (both user & system like timers ) or only user FB (multi inst) + system FB inside. ) , dependance on plc fw (I have 2.2) and tia portal version (i have V13 SP1)

Evgeniy,

The suggestion you posted will sometimes not work. The idea I have posted is not the only solution but it will work.

I am on location and have limited timed internet.. I will not expand to my usual comprehensive quantity of answering. I hope that others will fill the gaps I am leaving. I encourage you to use the search facility within this forum too as this timer subject has been answered before in detail and FAQs exists. Also find an FAQ relating to code example to generate a pulsing signal.

As to your project goal..
The timer generated pulse is not going to help you.. even a PLC is wrong for the job. You need microsecond timing. A signal processor is meant for this type of task.

If however you HAVE to use a PLC, then read up on cyclic timed interrupt OBs. BUT know that you have to use the setup LAD / SCL instruction to get the microsecond interruptions. Recently there were similar themed questions which I helped out on.. please search for those.

Best of luck to you.
w