So mostly to save money, to adress some grievances people i talked to had with commercial models, but also to have some fun tinkering with stuff, i want to build my own auto alkalinity tester and hopefully log my progress here.
The main inspiration was the KH Keeper which was recently released, but i also looked at the alkatronic, kh guardian, kh director and talked to people using them.
Basic components:
I am still in the planning phase, but here are some of my thoughts:
Main difference between the different commercial models, besides their software, is in the pumps they use. Most have a high precision pump for the reagent, and less accurate pumps for intake and disposal. Exceptions are the kh director which uses GHLs standard quad dosing pump, and the kh guardian, which uses cheap pumps and optically counts reagent drops instead.
For intake and disposal i am currently planning on using Kamoer NKP pumps, like the kh guardian, but instead with a 1mm hose diameter offering lower flow but higher precision. I currently also use those in combination with l298n h-bridges and pca9685 pwm generators for my normal dosing of reef chemicals and i got them to be consistent to approximately 0.25ml which corresponds to 3 drops from the 2mm aperture of the hose holder, I stopped bothering to optimize after achieving that. With a bit of tweaking, that should be accurate enough, but maybe will take a few minutes to fill and empty the reaction vessel. In general, the disposal pump doesn't need to be accurate at all, if it runs too long, all that happens is wasting time, so overshooting by a generous amount is fine. If the 1mm hose is too susceptable to getting clogged from stuff from the tank, or is just too slow, i'll have to look into more expensive pumps like the KCS or KFS line with stepper motors.
For reagent dosing, i am planning to mirror the kh keeper and use the quite pricy Kamoer KAS pump, also with 1mm tubing. One big complaint, that all the people i talked to using this had, was that the thing sounds like a UFO taking off in your living room. I am planning to improve on that by using a high quality stepper motor controller, currently looking at the tcm2209 and similar chips by trinamic that are quite impressive regarding noise reduction in 3d printers, but not quite as cheap as their alternatives. But luckily i only need one (hopefully lol) of them, not five.
The second big difference is in the ph measurement. For the circuit, i am planning to use the reef-pi ph board by @robsworld78, for the probe i am going to experiment with a cheap one of aliexpress, mainly because i am pretty sure that the kh keeper uses the exact same one i ordered for 10€, which saves quite a bit compared to the atlas one or similarly priced ones used in the other commercial testers. If that one dies in an unacceptably short timeframe, i'll probably look at a refillable glass one. But these have yet to arrive, so i can't say anything about their performance yet.
For overall control, i am still a bit torn whether to directly use my pi running reefpi, or even a future version of reefpi itself, or to use a dedicated arduino or pi for this that controls the testing process and have the pi only query for tests/retrieve and log results. Currently tending towards the latter to not get too crowded with the number of available pins, but i might change my mind about that later.
Another thought of mine to improve on the existing design was to add 3 additional low precision dosing pumps to automate regular ph probe calibration, another complaint i heard often from people, probably connected to the cheap probe the kh keeper is using. The NKP pumps are cheap and should be plenty accurate for this purpose, the only question is whether the ph buffers can be held stable enough in dosing containers for a long enough period of time to make it worth it, but if they can, that would provide valuable validation with more expensive probes too, increasing trust in the results.
I'll leave out my thoughts about plumbing and stuff, probably not very interesting or important. The next step will be evaluating the ph board and probe once they get here and looking at the different ways to do the titration manually at first. I hope not everyone died of boredom reading this by now.
The main inspiration was the KH Keeper which was recently released, but i also looked at the alkatronic, kh guardian, kh director and talked to people using them.
Basic components:
- Main processor controlling everything and doing calculations
- Reaction vessel
- ph probe
- ph measurement circuit
- magnetic stirrer
- sample intake/return pump
- sample disposal pump
- reagent dosing pump
- motor control circuits
- reagent container
- waste container
- plumbing
- ph 4 Buffer dosing pump
- ph 7 Buffer dosing pump
- RO/DI flush dosing pump
- sealed buffer containers, potentially with co2-scrubbed pressure equalisation
I am still in the planning phase, but here are some of my thoughts:
Main difference between the different commercial models, besides their software, is in the pumps they use. Most have a high precision pump for the reagent, and less accurate pumps for intake and disposal. Exceptions are the kh director which uses GHLs standard quad dosing pump, and the kh guardian, which uses cheap pumps and optically counts reagent drops instead.
For intake and disposal i am currently planning on using Kamoer NKP pumps, like the kh guardian, but instead with a 1mm hose diameter offering lower flow but higher precision. I currently also use those in combination with l298n h-bridges and pca9685 pwm generators for my normal dosing of reef chemicals and i got them to be consistent to approximately 0.25ml which corresponds to 3 drops from the 2mm aperture of the hose holder, I stopped bothering to optimize after achieving that. With a bit of tweaking, that should be accurate enough, but maybe will take a few minutes to fill and empty the reaction vessel. In general, the disposal pump doesn't need to be accurate at all, if it runs too long, all that happens is wasting time, so overshooting by a generous amount is fine. If the 1mm hose is too susceptable to getting clogged from stuff from the tank, or is just too slow, i'll have to look into more expensive pumps like the KCS or KFS line with stepper motors.
For reagent dosing, i am planning to mirror the kh keeper and use the quite pricy Kamoer KAS pump, also with 1mm tubing. One big complaint, that all the people i talked to using this had, was that the thing sounds like a UFO taking off in your living room. I am planning to improve on that by using a high quality stepper motor controller, currently looking at the tcm2209 and similar chips by trinamic that are quite impressive regarding noise reduction in 3d printers, but not quite as cheap as their alternatives. But luckily i only need one (hopefully lol) of them, not five.
The second big difference is in the ph measurement. For the circuit, i am planning to use the reef-pi ph board by @robsworld78, for the probe i am going to experiment with a cheap one of aliexpress, mainly because i am pretty sure that the kh keeper uses the exact same one i ordered for 10€, which saves quite a bit compared to the atlas one or similarly priced ones used in the other commercial testers. If that one dies in an unacceptably short timeframe, i'll probably look at a refillable glass one. But these have yet to arrive, so i can't say anything about their performance yet.
For overall control, i am still a bit torn whether to directly use my pi running reefpi, or even a future version of reefpi itself, or to use a dedicated arduino or pi for this that controls the testing process and have the pi only query for tests/retrieve and log results. Currently tending towards the latter to not get too crowded with the number of available pins, but i might change my mind about that later.
Another thought of mine to improve on the existing design was to add 3 additional low precision dosing pumps to automate regular ph probe calibration, another complaint i heard often from people, probably connected to the cheap probe the kh keeper is using. The NKP pumps are cheap and should be plenty accurate for this purpose, the only question is whether the ph buffers can be held stable enough in dosing containers for a long enough period of time to make it worth it, but if they can, that would provide valuable validation with more expensive probes too, increasing trust in the results.
I'll leave out my thoughts about plumbing and stuff, probably not very interesting or important. The next step will be evaluating the ph board and probe once they get here and looking at the different ways to do the titration manually at first. I hope not everyone died of boredom reading this by now.
Last edited:
