I've been thinking though how to create an Auto Water Change device. I have all the parts I'd need, but I can't figure out how to really truly make one that's fault proof. Making one that works in good cases (happy path) is fairly straightforward, but making one that would work well when something goes wrong (fault tolerant) seems very complicated.
The easiest would be a one pump with a double peristaltic head. One pumphead fill into the pump section of a sump, one drain from the drain line area of the sump. There'd be some mixing, but if done quickly should be ok I'd think. However, there's a bunch of things that can go wrong:
The biggest fundamental thing is there'd be no way to calibrate the two pumpheads against each other, so it seems doomed to ever work right.
Making it more complicated would be two separate pumps. That allows calibration, but now adds that one pump could fail and again things will get out of calibration. Effectively anything oriented around two pumps being well calibrated against each other will have all those issues.
A very different approach would be disable return and use float valves. This would turn off your return, then have your drainage line near that water level, a couple gallons worth below. It would drain until there's no more water it could pull, then fill, then turn the return back on. There'd be a float valve to measure when the level is back to the original, and stop. This effectively gets rid of the calibration issues, but requires that stuff never to get moved, and doesn't help with the issues like the fill reservoir draining or pumps breaking and adds the risk of the return not getting turned on and really causing issues.
My current idea is ultrasonic level measurement of the fill and drain reservoir. Use two 5-gallon buckets, one for fill, one for drain. Put a lid on both with an ultrasonic distance measurer. Calibrate by dumping 1 gallon of water in each bucket at a time so it'd know 1 gallon vs 2 vs ... Then ensure the buckets aren't empty, nor overfilled, and for every bit of water you add, remove the same. It could also test itself at the beginning to make sure both pumps work, and simultaneously fill and drain. The risk would be the sensors go out of calibration, which could be helped by doing double sensors in each.
However, all this seems very complicated and still a lot of ways it could go wrong and lead to your tank getting salinity swings. It seems safer to use it to automate the act of a manual water change, but without you having to carry buckets around.
What is people's experience with AWC devices? Any recommendations of one's that seem very safe? I don't want to steal a company's design and rebuild, but I'm struggling to see how any of these could really be safe without getting very very complicated.
The easiest would be a one pump with a double peristaltic head. One pumphead fill into the pump section of a sump, one drain from the drain line area of the sump. There'd be some mixing, but if done quickly should be ok I'd think. However, there's a bunch of things that can go wrong:
- Pumps pull different amounts
- One pumphead breaks, such as the piping inside breaks or clogs
- The fill reservoir of fresh saltwater goes empty
- The reservoir being dumped into fills
The biggest fundamental thing is there'd be no way to calibrate the two pumpheads against each other, so it seems doomed to ever work right.
Making it more complicated would be two separate pumps. That allows calibration, but now adds that one pump could fail and again things will get out of calibration. Effectively anything oriented around two pumps being well calibrated against each other will have all those issues.
A very different approach would be disable return and use float valves. This would turn off your return, then have your drainage line near that water level, a couple gallons worth below. It would drain until there's no more water it could pull, then fill, then turn the return back on. There'd be a float valve to measure when the level is back to the original, and stop. This effectively gets rid of the calibration issues, but requires that stuff never to get moved, and doesn't help with the issues like the fill reservoir draining or pumps breaking and adds the risk of the return not getting turned on and really causing issues.
My current idea is ultrasonic level measurement of the fill and drain reservoir. Use two 5-gallon buckets, one for fill, one for drain. Put a lid on both with an ultrasonic distance measurer. Calibrate by dumping 1 gallon of water in each bucket at a time so it'd know 1 gallon vs 2 vs ... Then ensure the buckets aren't empty, nor overfilled, and for every bit of water you add, remove the same. It could also test itself at the beginning to make sure both pumps work, and simultaneously fill and drain. The risk would be the sensors go out of calibration, which could be helped by doing double sensors in each.
However, all this seems very complicated and still a lot of ways it could go wrong and lead to your tank getting salinity swings. It seems safer to use it to automate the act of a manual water change, but without you having to carry buckets around.
What is people's experience with AWC devices? Any recommendations of one's that seem very safe? I don't want to steal a company's design and rebuild, but I'm struggling to see how any of these could really be safe without getting very very complicated.
Last edited:
