What are the main causes of Tank Crashes

[TaylorPilot]

Texting while reefing.

 

[Scorpius]

Human error. All tank crashes can be attributed to this one. Think about every issue you've ever had with your tank. Its all caused by YOU!!!

 

[ZaneTer]

Man 3 months that would be hard to keep tank faultless!!

It’s not easy. I engineer control systems for a living so after I had a crash caused by GHL I decided to stop being lazy and use one of my industrial controllers instead. This meant I had to write everything from scratch and produce a graphical interface for my wife in case she needed to do something. It has worked out much better and control has been significantly increased. I can control every aspect remotely and receive email alerts immediately if something starts to go wrong.

 

[TaylorPilot]

@ZaneTer How much does an industrial controller cost?

 

[ZaneTer]

I use a siemens 315-2pn/dp controller. It’s a very modular design. If you need more digital inputs or outputs then you just buy another card. Same with analogue signals.

Overall it works out to be much cheaper than apex or ghl but it does require quite a large amount of knowledge to set up properly. It’s accuracy and capabilities can’t be competed against.

I’ll try find a post I wrote a while ago to show the cost benefits.

 

[ZaneTer]

A 32 digital output card will cost you around $250 based upon a siemens system. A single peristaltic pump is around $5-6. Yes these are very cheap.

This means you can build up an entire rig of 32 dosing heads for around $410 to $442. Each of these heads is capable of dosing with accuracy of less than 1/100th of a millilitre. That is more than 10x more accuracy than the the vendors can offer.

Equivalent from GHL is 8x4 way slave pumps at a total cost of £365x8=£2920 or $3800. Initial cost for a system built upon siemens hardware would be a higher initial cost for a CPU but significantly lower upgrade costs.

Each person will have to calculate for themselves based upon their own needs.

With a custom system it would be very easy to setup automated calibration routines for pH probes using two dosing heads. Almost all of the work can very easily be done by the controller for a relatively small cost. You are really limited by imagination.

For an HMI you could make use of your current home PC instead of spending more money on a tablet or small laptop.

A new analogue card of 8 analogue inputs at 12bit resolution ***see below*** can be picked up for about $400 and can be used for measuring temperature or salinity or ph or ORP. The relevant probes are required as may be a transmitter.

*** analogue card measure in resolution. A 12 bit card can read a value split into 4096 pieces. This means that for a temperature reading (0-100)it goes up or down in increments of (100/4096)=0.0244 degrees. The higher the resolution of your card the higher the price tag. For our application there is no point exceeding 12 bit resolution. Even for pH control on a reactor it would measure in steps of (14pH RANGE/4096)=0.0034. To be honest your probe probably won’t be accurate enough to reliably show this level of resolution.

As mentioned earlier you can set up pretty much any routine you want.

I hope this helps.

You can pick up a controller for around $250 second hand. It all uses open communication methods so you don’t have to worry about any proprietary stuff.

 

[vetteguy53081]

lately it seems main causes are :

Impulse buying
Overfeeding
Chasing parameters
Lack of water changes

 

[maroun.c]

Problem is that by us adding resundancy we also double the risk of error. Like by adding 2 controllera on a heater we benefit from.one side that if one controller gets stuck ON the sencondary on will.kick in to shut the heater when desired temp is reached which isnall godd. But then we have 2 controllers and doyble thw risk of one failing (at the off position ) to freeze ur tank. Same for ATOs with multiple float switches where the more switches the more safety against ATO getting stuck on filling up the tank more than it should and either floodong ur room or potentially killing corals . But then all it takes is one of the switches getting stuck at the Off position to run ur return pump dry and potentially killing ur tank.
Redundant controls like controllers to notify us when things go wrong go a long way when the error is not caused by them and lucky dont fail a lot bur when they do they are alote more grief than other equipment.

 

[ZaneTer]

Problem is that by us adding resundancy we also double the risk of error. Like by adding 2 controllera on a heater we benefit from.one side that if one controller gets stuck ON the sencondary on will.kick in to shut the heater when desired temp is reached which isnall godd. But then we have 2 controllers and doyble thw risk of one failing (at the off position ) to freeze ur tank. Same for ATOs with multiple float switches where the more switches the more safety against ATO getting stuck on filling up the tank more than it should and either floodong ur room or potentially killing corals . But then all it takes is one of the switches getting stuck at the Off position to run ur return pump dry and potentially killing ur tank.
Redundant controls like controllers to notify us when things go wrong go a long way when the error is not caused by them and lucky dont fail a lot bur when they do they are alote more grief than other equipment.

Hi maroun

I understand your thinking but it is a little flawed. I’ll start with I am an automation engineer so this is what I do every day. Your first point about heaters:
The correct way is to run multiple independant heaters. I run 3 off three separate plugs. I also have a chiller. My chiller is stronger than any single heater but not stronger than 2. This means if a heater failed then I am perfectly safe albeit I am wasting electricity. If my chiller failed then I only need 2 heaters to still keep my tank warm, again I am still perfectly safe. I also don’t rely on a single temperature probe. I have 3. With three probes you can tell which probe is wrong by the deviation from the average.

Redundancy isn’t just about throwing more heaters or chillers on a tank. It’s about carefully planning some logic to make the best use of multiple heaters and chillers.

Redundancy is actually about making sure that no single failure can cause catastrophic damage.

If you have any more questions I would be happy to answer them.

Zane

 

[Kdoor]

Leaving RODI going in a large sump on accident / trying to correct salinity too fast was a crash for me -> dead crustaceans -> cyano bloom/nitrates -> Q.Q. I set timers on my phone or google home now with limits for RODI filling up so that I won't forget it's on while doing something.

 

[maroun.c]

Hi maroun

I understand your thinking but it is a little flawed. I’ll start with I am an automation engineer so this is what I do every day. Your first point about heaters:
The correct way is to run multiple independant heaters. I run 3 off three separate plugs. I also have a chiller. My chiller is stronger than any single heater but not stronger than 2. This means if a heater failed then I am perfectly safe albeit I am wasting electricity. If my chiller failed then I only need 2 heaters to still keep my tank warm, again I am still perfectly safe. I also don’t rely on a single temperature probe. I have 3. With three probes you can tell which probe is wrong by the deviation from the average.

Redundancy isn’t just about throwing more heaters or chillers on a tank. It’s about carefully planning some logic to make the best use of multiple heaters and chillers.

Redundancy is actually about making sure that no single failure can cause catastrophic damage.

If you have any more questions I would be happy to answer them.

Zane

Agree with that but what happened in my case was the following, i had 2 schelgo temp controllers with 2 heaters hooked to each and couple other heaters with built in control and apex. Each ran on a separate power line and had apex on top to act as redundant control.
2 of the issues i had in last 3-4 years:
-one schlego controller exploded cutting off power instead of keeping it on (and thats my main concern in redundant control) the remaining heaters on the extra lines were still working but my temp still severely dropped and apex notified me of the issue (thats where notification comes in handy)
Similar scenario with 2 chillers also on separate lines with apex set to shut them off at 25.2 degrees celsius and theyre set to shut down on their own at 25. Apex relay switch toasted shutting down one of the chillers and second chiller was only able to hold temp at 26.5 when halides kicked in. (notification helped and killling the halides also helped)
-apex failed and somehow got corruped sw (so failsafes went all off ) so all plugs were off and whole system was down for 45 min till fusion sent a heartbeat disconnect notification.
Ur example of a chiller overpowering heaters stuck on is valid but heaters stuck off is an issue that it wont solve . I run around 3000 w plus of heaters so having small ones is not an option and having many more channels might trip my backup power in case of a power failure . A bit of a specific scenario i realize.

 

[ZaneTer]

I will be very happy to run through this

Issue 1) Your heaters were undersized. Each heater should barely be capable of maintaining the temperature. You don't want to much larger or smaller. Your chiller should be stronger than a single heater but not stronger than both.

Issue 2) You must always run each item (chiller/heater/pump/whatever) off its own interposing relay. You introduced a single point of failure by using a single relay for two chillers. Good idea for linking the temperature to the halides. You may want to consider stronger chillers in the future.

Issue 3) I wont touch apex or ghl. They are really only hobby grade despite how much they cost. They completely pale in comparison to a modern PLC. If you ran multiple individual heaters where any two heaters are capable of overpowering your chiller then you should have no problem.
**** By the way I really hope you meant 300w and not 3000w unless you are running a public aquarium

My setup is:
3 x 300w heaters
1 x 500w chiller
180g system + 40g sump
3 temperature probes (PT100 -20'c to 100'c)

Each heater comes on at a different delta:
Heater 1 @ 0.2
Heater 2 @ 0.3
Heater 3 @ 0.4
(I dont like to run all my heaters at once if i dont have to)

For bonus points you can also setup detection for stuck on or off outlets by measuring the gradient of the temperature change. You could also use the fact that perhaps your chiller should never be on during winter to help generate an alarm.

 

[Dlealrious]

mmmmm now im thinking a controller for the alerts may be a good idea, i just cant bring myself to spend that much money on an apex. Might wait and see how the new eKoral controller goes.

 

[maroun.c]

Issue 1) Your heaters were undersized. Each heater should barely be capable of maintaining the temperature. You don't want to much larger or smaller. Your chiller should be stronger than a single heater but not stronger than both.

Max i could find is 600 w so i basically run a 600 and a 300 on each channel to not exceed max load of the schlego controllers. Will be switching to heat exchangers as alternative.

Issue 2) You must always run each item (chiller/heater/pump/whatever) off its own interposing relay. You introduced a single point of failure by using a single relay for two chillers. Good idea for linking the temperature to the halides. You may want to consider stronger chillers in the future./QUOTE]
Agree. On list of things to change

Issue 3) I wont touch apex or ghl. They are really only hobby grade despite how much they cost. They completely pale in comparison to a modern PLC. If you ran multiple individual heaters where any two heaters are capable of overpowering your chiller then you should have no problem.
**** By the way I really hope you meant 300w and not 3000w unless you are running a public aquarium

Unfortunately not coding savy so not an option... 3000 was correct its 5 tanks 4 of which in an unheated garage.

My setup is:
3 x 300w heaters
1 x 500w chiller
180g system + 40g sump
3 temperature probes (PT100 -20'c to 100'c)

Each heater comes on at a different delta:
Heater 1 @ 0.2
Heater 2 @ 0.3
Heater 3 @ 0.4
(I dont like to run all my heaters at once if i dont have to)

I like heaters and chillers starting at different deltas but my issue is with a sey of heaters starting and not beig able to heat up the watet but still keeping it stable and second set of heaters doesnt kick in. This has caused some of my heaters to work non stop for extebded time which i believe toasted the controller. I know set heaters higher and have apex set lower so they all start and stop on apex command.

For bonus points you can also setup detection for stuck on or off outlets by measuring the gradient of the temperature change. You could also use the fact that perhaps your chiller should never be on during winter to help generate an alarm.

Something to consider definitely as part of optimizing the setup. Not sure ill be able to achieve this with Apex though.

Thanks for all the replies definitely food for thought.

 

[ZaneTer]

Max i could find is 600 w so i basically run a 600 and a 300 on each channel to not exceed max load of the schlego controllers. Will be switching to heat exchangers as alternative.

Thats a great idea. Heat exchangers are fantastic.