Multi-head dosing with one pump and many solenoids?

[Kerinin]

I've been researching dosing pumps. I'm planning to eventually run a Triton tank, but I expect to be doing water changes in the beginning stages (at least). I like the idea of automating the water changes, as well as providing an ATO solution. So if I used peristaltic pumps for everything, I would need 7 dosing heads - 4 for Core7, 1 for ATO, and two for water changes. The dosing pumps to do this reliably and quietly would likely cost a minimum of $800, which seems like a lot of money to me.

My idea is to buy an Arduino, a single fairly high quality dosing pump (for example the one Atlas Scientific makes), 9 normally-closed solenoids and a bunch of push-connect fittings and use the pump to pulls from one source at a time:

                            (manifold)
Core7 A ----- Solenoid A --
Core7 B ----- Solenoid B --|
Core7 C ----- Solenoid C --|           -- Solenoid H -- Tank
Core7 D ----- Solenoid D --|-- Pump --|
Fresh SW ---- Solenoid E --|           -- Solenoid I -- Waste
Fresh RODI -- Solenoid F --|               
Tank -------- Solenoid G --

So for example, to dose Core7 A the Arduino would first activate Solenoids G & H, flush the system with tank water, then activate Solenoids A & H, dose the appropriate amount, then flush again.

ATO would activate Solenoids E & H and pump until the water level reached normal.

Water changes would alternate between activating Solenoids E & H to add SW to the tank and Solenoids G & I to dump old water.

Using the SS Solenoids from STC Valve which run $26/each, the $85 peristaltic pump, $24 Arduino uno and say $20 of assorted fittings and electrical components brings the cost of this system to $352. Using cheaper off-brand solenoids brings that down to $217.

Is this a terrible idea? I feel like there must be a reason that everyone in the industry uses dedicated pumps for each line. What am I missing?

 

[laverda]

Since your one pump would bee getting a lot more use make sure it can handle it. I doubt an $85 dosing pump will last long with 8 times the use of single ones would. I would also thing you would be better off flushing with RO/DI water for most things.
Why not gust set everything to gravity feed and use solenoids to control them. Of course you may not be able to do that for water changes as easily.

 

[Lasse]

It will probably work for a while but -IMO- you need to use solenoids there all parts in contact with your fluids are made of plastic or titan. Core7 1 and 2 must not come into contact with core7 3 a + b - not even a single drop. If - there will be calcium or magnesium carbonate build up.

Your calculation of needed dosing pumps can be changed to 6 - you can mic Core7 3a with Core7 3b and dose the double of the mix.

I would go for the separate heads. Further on the line - you will have need of more heads for dosing other compound - be sure - I know. For the moment - I have 4 GHL Doser 2.x And I would go for a dosing system that can handle many heads from a main unit

Sincerely Lasse

 

[Kerinin]

I would also thing you would be better off flushing with RO/DI water for most things.

I was thinking that tank water would run a lower chance of affecting the tank's salinity. I agree it would probably be better for the life of the valves.

It will probably work for a while but -IMO- you need to use solenoids there all parts in contact with your fluids are made of plastic or titan. Core7 1 and 2 must not come into contact with core7 3 a + b - not even a single drop. If - there will be calcium or magnesium carbonate build up.

Good points - I bet this is why you don't see this in commercial solutions.

 

[peterj66]

You can get a good quality Kamoer stepper pump
from 35 to 50 dollars. If you go with a Kamoer DC pump its even cheaper.

 

[Kerinin]

You can get a good quality Kamoer stepper pump
from 35 to 50 dollars. If you go with a Kamoer DC pump its even cheaper.

Yeah I was looking at those too - that seems like a better option. Do you have experience working with those? I'm curious about their noise level and how hard it is to drive them from a 10v PWM signal.