Set coil RESETS whenever logic that sets it goes low.

[Felixs]

I know I am doing something wrong with the logic but I have struggled for days without a solution, I am using a 4-20 ma pressure transducer, filtered analog input.

compare the pressure < ON_setpoint = output turns ON (Set Coil)

Compare pressure > OFF_setpoint=  This should reset the coil

 

BUT

 

instead any time the pressure > ON_setpoint the output turns OFF 

 

 

 

You can watch the screen recording  of a simplified program 

 

 

I also uploaded the actual program that eventually will become the control system of 2 air compressors, lead -lag system which will alternate the order every day. 

 

I am out of ideas! 

I know it is my fault, just don't know where my mistake is being made 

compressors.U90

[kratmel]

Please place one loigic net in to one rung. In your project you place many times two net in one rung. maybe something wrong with compiling this nets.

You can also find on-off control in U90 sample projects (PID control folder)

 

[Felixs]

16 minutes ago, kratmel said:

Please place one loigic net in to one rung. In your project you place many times two net in one rung. maybe something wrong with compiling this nets.

You can also find on-off control in U90 sample projects (PID control folder)

 

I loaded the pump control example and did the same thing, as soon as the comparator a<b goes false output resets 

 

I will clean up the ladder rungs and make sure to use one net per rung !

[Flex727]

I haven't had a chance to review the program itself, but I did watch your (very well done) video. Assuming you don't have any duplicates of O-5 in your program, other than those two rungs, I would suggest the possibility that your analog input has some noise that is causing it to rise above 130 very briefly (too quick to see when online). Keep in mind that you don't get online data from your PLC for every PLC scan.

[Felixs]

14 minutes ago, Flex727 said:

I haven't had a chance to review the program itself, but I did watch your (very well done) video. Assuming you don't have any duplicates of O-5 in your program, other than those two rungs, I would suggest the possibility that your analog input has some noise that is causing it to rise above 130 very briefly (too quick to see when online). Keep in mind that you don't get online data from your PLC for every PLC scan.

I thought about that possibility, but have tried with a higher set point and same result. I am using a high filter for the input too.

I tried using the flow control example, single pump tank level and it behaves the same way.  

[Flex727]

Replace your O-5 Reset Coil with an Increment function for some unused MI, then repeat the process. Does O-5 remain Set? Does the MI increment? If so, then you can rest assured that your analog input is briefly flickering above 130.

If you can't solve the electrical problem, you can workaround by adding a delay timer that requires the input to stay above 130 for a full second or so before triggering the Reset.

[kratmel]

On this pic some net provided by Joe Tauser, you can try to use it for your project. You can change MI1 and MI3 by your setpoints

from this topic

 

 

[Felixs]

Adding the timer as a 1 second delay did the trick! I scratched my head for two days and was starting to loose some sleep over it. Thanks to all of you. 

[Ausman]

Depending on your application needs,  I would also be adding in short cycle protection, as well as variations of mechanical failure recognition.  eg   What will be the maximum run time during normal use?  Give a good safety margin on time and make an interlock that turns off the compressor if it is running too long due to a hose failure etc.

Also, one thing I ALWAYS try to do is not just rely on what I've got the PLC telling  a motor etc to do.  I check for the desired result.  Any discrepancy (including like my hose failure bit above) throws an error, that goes into an automatic checking system that will do an auto reset, try again a few times and if it occurs repeatedly over X time, then it trips for good.  This method gets around odd hiccouphs that aren't real errors, but enough for the control system to raise suspicions.  I also have counters to keep records of such actions and initial faults.

cheers, Aus

[Felixs]

13 hours ago, Ausman said:

Depending on your application needs,  I would also be adding in short cycle protection, as well as variations of mechanical failure recognition.  eg   What will be the maximum run time during normal use?  Give a good safety margin on time and make an interlock that turns off the compressor if it is running too long due to a hose failure etc.

Also, one thing I ALWAYS try to do is not just rely on what I've got the PLC telling  a motor etc to do.  I check for the desired result.  Any discrepancy (including like my hose failure bit above) throws an error, that goes into an automatic checking system that will do an auto reset, try again a few times and if it occurs repeatedly over X time, then it trips for good.  This method gets around odd hiccouphs that aren't real errors, but enough for the control system to raise suspicions.  I also have counters to keep records of such actions and initial faults.

cheers, Aus

Totally agree with you, so far I added a max run time to trigger one alarm, jump to a display then  pick from a list and assign this to a MAX Run Time cause. 

Next thing is to have a LOW PRESSURE ALARM but I am struggling with the noise again, thinking about getting a different power supply. The comparator fails to trigger an alarm with 1 second timer

Run time accumulator for each compressor is also in the works. 

Thanks!

[Felixs]

If someone cares to take a look, I implemented a few ideas from your comments  on this video

 

thanks