Why is Apex Cycling my CO2 on/off So Often

[obrien.david.j]

@SuncrestReef and others, I've got a puzzle for you.

My apex is setup to control the CO2 solenoid for my calcium reactor.   It's fed by a dedicated pH probe in the reactor.

Question, Why does it look like the C02 is being cycled on/off.   (every few minutes)    And in this graph specifically, why did it get turned off at all.

Programming:   if pH > 6.65, turn ON, if <6.45 turn OFF.  Min Time is there to insure solenoid doesn't get cycled faster than 1minute intervals.

Graph shows getting turned off, even though it's < 6.65.

 

[SuncrestReef]

When controlling an output between two values, it's best to not have an initial Set ON or Set OFF, since the conditions cover both states.  I've heard that can somehow cause issues.

[albertareef]

7 minutes ago, SuncrestReef said:

When controlling an output between two values, it's best to not have an initial Set ON or Set OFF, since the conditions cover both states.  I've heard that can somehow cause issues.

Interesting.  So you can just eliminate that line and it might work?

[SuncrestReef]

Just now, albertareef said:

Interesting.  So you can just eliminate that line and it might work?

It's worth a try.  Logically it doesn't seem that it would matter, but I've read enough anecdotal reports of it causing problems.

Fallback OFF
If pHCO2 > 6.65 Then ON
If pHCO2 < 6.45 Then OFF
Min Time 001:00 Then OFF

 

[SuncrestReef]

Just now, SuncrestReef said:

It's worth a try.  Logically it doesn't seem that it would matter, but I've read enough anecdotal reports of it causing problems.

Fallback OFF
If pHCO2 > 6.65 Then ON
If pHCO2 < 6.45 Then OFF
Min Time 001:00 Then OFF

 

Actually, now that I think more about it, there is a good reason to remove the Set OFF.  If it goes On when pH > 6.65, as soon as pH drops to 6.65 or lower it will turn Off right away rather than waiting until it gets down to 6.45.

[SuncrestReef]

@obrien.david.j Another thing to remember when overlaying graphs is that the pH data is being sampled by default every 10 minutes for graphing, but the output logging happens in real time.  The actual programming decisions are based on current pH readings, but the graphs are only 10 minute intervals.  So while it may look like your pH is too low and the output should not turn on, the graph is just showing an average between two sample times.  You can change the sample rate, but it's not recommended to make it too frequent. 

[NateDawg]

1 hour ago, SuncrestReef said:

@obrien.david.j Another thing to remember when overlaying graphs is that the pH data is being sampled by default every 10 minutes for graphing, but the output logging happens in real time.  The actual programming decisions are based on current pH readings, but the graphs are only 10 minute intervals.  So while it may look like your pH is too low and the output should not turn on, the graph is just showing an average between two sample times.  You can change the sample rate, but it's not recommended to make it too frequent. 

Excellent work as usual! Just curious, why is it not advised to increase the frequency of data logging?

[SuncrestReef]

3 minutes ago, NateDawg said:

Excellent work as usual! Just curious, why is it not advised to increase the frequency of data logging?

It's due to the memory capacity on the Apex, and it will also shorten the historic data you will be able to view.

[obrien.david.j]

Here's the results of experiment #1...    Right after posting my question today, I increased the MIN time to 5 minutes.   and I learned something.

New Learning...   Min Time xyx Then OFF...   sets the minimum of time the outlet is OFF.   For some reason I thought it was the min time ON.   

John, Big thanks for flagging the Set OFF command.   It's now Removed.   Also thanks for the datalog interval reminder on the pH.  I'm thinking of doing a limited runtime experiment of logging every minute...  but I know it will chew up datalog memory, and overflow faster.   But I could run for a 30minute window and then restore it.  

 

Until Then, Experiment #2 programming - running now.  (removed "set", anc changed Min Time to be "ON" for 5mins)

• Fallback OFF
• If pHCO2 > 6.65 Then ON
• If pHCO2 < 6.45 Then OFF
• Min Time 005:00 Then ON

 

[SuncrestReef]

Personally, I think it would be best to eliminate the Min Time all together and rely on the pH readings.  This will give you a true sense of how long a normal cycle takes.  By forcing it to remain On or remain Off beyond the pH probe threshold you are making the swings larger.

[albertareef]

9 minutes ago, obrien.david.j said:

Here's the results of experiment #1...    Right after posting my question today, I increased the MIN time to 5 minutes.   and I learned something.

New Learning...   Min Time xyx Then OFF...   sets the minimum of time the outlet is OFF.   For some reason I thought it was the min time ON.   

John, Big thanks for flagging the Set OFF command.   It's now Removed.   Also thanks for the datalog interval reminder on the pH.  I'm thinking of doing a limited runtime experiment of logging every minute...  but I know it will chew up datalog memory, and overflow faster.   But I could run for a 30minute window and then restore it.  

 

Until Then, Experiment #2 programming - running now.  (removed "set", anc changed Min Time to be "ON" for 5mins)

• Fallback OFF
• If pHCO2 > 6.65 Then ON
• If pHCO2 < 6.45 Then OFF
• Min Time 005:00 Then ON

 

Very interesting and suggestive that the program itself is somehow driving the on/off cycle vs. the actual pH readings - caveat about long sample interval aside.  That said, you can at least imagine that increasing the off period would potentially lead to larger pH swings if limiting... and it did.

[albertareef]

1 minute ago, SuncrestReef said:

Personally, I think it would be best to eliminate the Min Time all together and rely on the pH readings.  This will give you a true sense of how long a normal cycle takes.  By forcing it to remain On or remain Off beyond the pH probe threshold you are making the swings larger.

I think this is definitely worth doing.  I would say that you could look at reintroducing a longer maximum "on" time as a failsafe (probe failure, broken CO2 line) but drive the "normal" range behavior strictly off the pH reading.  That said, I will let John handle the actual coding for that 😀

[obrien.david.j]

I originally added the MIN time command after reading Johns tutorial, and how he added it for ATO programming in case the the float switches bounced.    There are time the water line on my CaReactor Burps out Excess CO2 gass.  And I was thinking it could glitch the Solenoid, which isn't good.

BTW, the "Min Time 005:00 Then ON" code is already looking like better results.    But need to let a trend build up.

[obrien.david.j]

Success!   

But, which command change generated the desired result?   (removing "Set OFF", or Min Time "ON")   I suspect it's Both.  

time for Experiment #4....  Add "Set OFF" back into the programming.

 

[albertareef]

Progress!  Still not sure you need the min. on but it doesn't seem to be driving the pH past your set point so no harm I guess.  I would assume it's removing the Set off that did the real work however.  looking forward to the results for Exp. 4!

[obrien.david.j]

Success, and experiment over.   "Set OFF" is the culprit.    Removing it forever from this outlet.

Final Programming I have made:

• Fallback OFF
• If pHCO2 > 6.65 Then ON
• If pHCO2 < 6.45 Then OFF
• Min Time 005:00 Then ON

Thanks for your help today.   It was great to do real time updates, see real time responses, and empirically study impacts.

 

[albertareef]

Very cool to see the impact of the setting changes - thanks for sharing!  On a side note, you might want to keep an eye on your Alk levels with the new program - looks like the overall effect will be to keep the avg. pH much closer to your desired set point than it was before so could impact your tank parameters.  

[obrien.david.j]

Agreed.   I'll likely raise the overall pH setup as a result of this.   I'd like to let this settle out for a day so I can establish a better trend, and let the next study begin.   

[SuncrestReef]

@obrien.david.j Are you not going to do a test with Min Time removed?

[NateDawg]

5 hours ago, SuncrestReef said:

It's due to the memory capacity on the Apex, and it will also shorten the historic data you will be able to view.

Makes sense, thanks for the info.

[obrien.david.j]

16 hours ago, SuncrestReef said:

@obrien.david.j Are you not going to do a test with Min Time removed?

Overnight, things look stable.   I wider swings of the Calcium Reactor pH, Solenoid does cycle on/off, but at wider intervals, matching the pH control ranges.   I think I'll leave it alone for another day before touching anything.  (so I can watch Alk too)

[Higher Thinking]

Overnight, things look stable.   I wider swings of the Calcium Reactor pH, Solenoid does cycle on/off, but at wider intervals, matching the pH control ranges.   I think I'll leave it alone for another day before touching anything.  (so I can watch Alk too)

Apex aside, I would reduce the CO2 output rate so that it's closer to what you actually need. The fact that it's flipping on and off so much is evidence that your CO2 is set to a bubble rate that's too much. If you reduce your gas rate to something closer to what your reactor actually needs to keep a desired pH, you can naturally keep the pH you want. Then the Apex just serves as a guide from time to time.

Basically, if the desired pH is a ball, make it more like a bowling bowl between bumpers, rather than a pong ball getting batted back and forth. Hope that analogy makes sense [emoji16][emoji16]

Sent from my SM-G975U using Tapatalk

[obrien.david.j]

Good suggestion.  Just turned my needle valve down a bit.  (this has proven itself to be touchy some times)

And for fun, see if you can tell the difference.    I've made a change, my plan is to now let it ride for a day again.   This one will likely be harder to see.

BEFORE:

 

AFTER:

[albertareef]

Good advice Andrew @Higher Thinking  Based on those videos, I would say you could ratchet that down quite a bit more but see how it goes.

[Higher Thinking]

Good advice Andrew @Higher Thinking  Based on those videos, I would say you could ratchet that down quite a bit more but see how it goes.

Good suggestion.  Just turned my needle valve down a bit.  (this has proven itself to be touchy some times)
And for fun, see if you can tell the difference.    I've made a change, my plan is to now let it ride for a day again.   This one will likely be harder to see.
BEFORE:

 
AFTER:

Oh yeah. That thing is ripping! You definitely don't need it that high. Depending on your effluent rate, I'd be surprised if you need anything close to that.

I'm certainly not one to get caught up in bubble rates or anything, but just throwing out that I've never seen a calcium reactor running so high you couldn't even count them.

Sent from my SM-G975U using Tapatalk