Arduino based auto alkalinity tester DIY

[karimwassef]

Very interesting approach! Wish I had thought of that.

I think it would require a more complex set of staged containers and valves, but not impossible to conceptualize.

I would still need the actuators for the titration chemicals, but I can see using a dosing pump head, multi-stage vials with purge valves and fill valves to duplicate my saltwater and expel actuators without losing accuracy.

I don't have any solenoids in mind, but this prompts a different vector to consider.

[ubasu]

Originally Posted by karimwassef

Very interesting approach! Wish I had thought of that.

I think it would require a more complex set of staged containers and valves, but not impossible to conceptualize.

I would still need the actuators for the titration chemicals, but I can see using a dosing pump head, multi-stage vials with purge valves and fill valves to duplicate my saltwater and expel actuators without losing accuracy.

I don't have any solenoids in mind, but this prompts a different vector to consider.

Yeah I think it would lead to less errors.

Amazing project btw. Holy crap this is cool. The amount of automation that could be based off of this information is incredible! The options are endless!!!!!

[karimwassef]

I don't know if it's less errors than an actuator, but it is simple and inexpensive and should be comparable if calibrated properly.

The risk is that if it's not perfectly level, it'll be a few drops light.

If I had a 3D printer, I could make a very simple cup in cup with line connections, but to make a DIY, I'd use two syringes - one large and one small. The small one cut at the level that would release 5ml (including the line volume before the solenoid):



The purple line is exported waste to drain and the blue line is 5ml into the test vial.

Trapped air bubbles in the line or first chamber may also add error.

[zachts]

That would be an interesting alternative, you'd need to calculate and account for the amount of water in the lines also and make the inside cup just shy of 5ml. It would probably need a pump to transfer as I don't think gravity would necessarily do the trick with such small lines, capillary action and all, along with that last few drops that would stick in the bottom of the inside cup unless it came down to a very tight cone at the bottom(just think a normal syringe if you fill it and let it drain by gravity, a little bit never gets out of the bottom), it might end up being more complex indeed, and possibly not any more accurate...........now you'll have to just make a 3D printer and try both,

[karimwassef]

The syringes naturally have a cone bottom, but it still means the solenoid needs to become a second doser pump head.

Still viable but not so cheap and easy.

Neither are my syringes... But I have them in hand, so I'll keep at it.

[karimwassef]

So... 2000 views and only a handful participating?

[droog]

Originally Posted by karimwassef

Why would you need Alk and phosphates but no Ca?

Alk changes more rapidly than Ca changes more rapidly than Mg. Futher, if you use a balanced Alk additive like Kalk or 2-part Ca is going to track nicely along with Alk. I dose 2-part, don't bother to test for Ca or Mg really.

-Droog

[karimwassef]

My data shows different take up rates for the three. That's probably also a function of water changes that introduce different contributions of each.

[perkint]

Originally Posted by karimwassef

So... 2000 views and only a handful participating?

I think what you're doing is way beyond what most would be willing to do!

One of the things that makes it so interesting!

Tim

[bobbravo2]

Keep up the good work. Have you tested the syringe driven pump action yet?

[karimwassef]

No. It's our end of quarter - so work is on overload these days.

This weekend, I'll try to get the assembly parts for the syringe - actuator bond. Then I can start testing.

[droog]

Originally Posted by karimwassef

My data shows different take up rates for the three. That's probably also a function of water changes that introduce different contributions of each.

Sure. Water changes can affect things, so can abiotic precipitation if if happens. Uptake rates are probably inside the noise limits of hobby test kits too unless taken over a long time period. Its my belief/understanding that the uptake rates should be balanced (all other things being equal).

I could also be wrong, its happened before... :-)

-droog

[droog]

@OP - I don't want to hijack this thread but I've been looking into DIY dosing of liquid foods to keep NPS corals for some time now. There are some DIY solutions using a fridge and dosing pump (sometimes venturi).

It seems that some elements of your design in this thread could be used to make an automatic food dosing system. Agitate a bottle, dose particulate matter via syringe, flush. The device could sit inside a mini-fridge running off a timer.

Not what this thread is about so feel free to ignore this (or I could spin a new thread to discuss). A continuous live food injector would be easier than the Alk measurement device and is likely to be better and easier than the designs currently out there. Could be a prototype / learning platform for Alk monitor. Just a thought.

-d

[karimwassef]

I think dosing pumps work great for live food. I've looked at those threads.

The only reason I'm using actuated syringes is the need for accuracy in titration.

The precession may not be necessary for live food.

You could use this, but I think it's overkill.

I'll post my design and code though, so feel free to adapt it to your needs.

[mm.reefs]

Originally Posted by droog

...

1] Gravity feed rather than syringe?

I wonder if you could fill a small reservoir using an actuator to open/close a value on a low-pressure drain from the DT.

Then another actuator to open close a valve that has a drain that drips a known quantity over time. 1 drop per second, measure and hard code the time needed to make 10ml and 1ml or whatever is called for.

...

-droog

Karim, very cool thread, I’m following it from the beginning!

I find droog's idea very interesting since drops may be counted with optical sensors, so the system may be accurate to the drop!

Keep the good ideas flowing!!!

[karimwassef]

I agree. I am depending on the accuracy of the encoders to make sure my drops align with specific increments of the actuator count position but if that's not tight enough, I might need to add an optical detector.

[droog]

Originally Posted by karimwassef

I think dosing pumps work great for live food. I've looked at those threads.

The only reason I'm using actuated syringes is the need for accuracy in titration.

The precession may not be necessary for live food.

You could use this, but I think it's overkill.

I'll post my design and code though, so feel free to adapt it to your needs.

It seems that most food gets clogged in the line using a dosing pump. I was interested in adapting this for the ability to dose "lumpy stuff" more than for precision. Agree precision is not a high requirement for feeding application.

-droog

[karimwassef]

It depends on whether you're feeding krill and mysis or cyclopese, I guess.

I think the concept of actuated syringes is viable but the scale would need to change. My current syringe and tube openings will likely get clogged with larger food parts, and the volume for each full pull is limited by a 12mm stroke ~ 10ml.

The same goes for the one way valves that control flow path.

I would consider a large diameter solenoid and use gravity with a water post-flush (second solenoid) to export the solid/liquid mix to the tank.

If you think about it, it's not different from the way a toilet bowl functions - it's practical robust engineering. - even if the imagery is not pretty.

[zachts]

Originally Posted by droog

It seems that most food gets clogged in the line using a dosing pump. I was interested in adapting this for the ability to dose "lumpy stuff" more than for precision. Agree precision is not a high requirement for feeding application.

-droog

not to side track too much more but for bulky lumpy foods it just takes a bigger dosing pump with larger bore tubing and more power, like the masterflex heads and an appropriate motor.

[jimbob321]

i worked at a marina, when we pumped out "sewer" we used the same Principal of the pump design, a peristaltic pump. i know they will pump solids and you don't want to be the one changing the tube. lol
OK guy's lets get back on track here..

[karimwassef]

Nice...

I'll post some pics and videos as soon as I get the attachments. The machine screws on these actuators are very small.

I also got all the ribbon cables and connectors in.

If I get around to cutting the acrylic, I'll start testing the precession. We'll see how far I get.

[dartier]

Originally Posted by karimwassef

Peristaltic pumps can't control flow to the level of 0.065ml ... One drop.

If you know of one that does for <$25, I'll give it a try.

Originally Posted by zachts

not perfectly off the shelf but there are pleanty of cheap pump heads you choud juist swap a ~10rpm (or whatever gets you one drop) motor for the stock motor on any number of pumps.......easy, and cheap enough.....after a little calibration with the coding......

Originally Posted by karimwassef

If this doesn't work, I'll try it. There is a difference between precision and accuracy. I don't know if I can calibrate and code well enough to maintain a single drop tolerance and consistency over time from a peristaltic pump.

An actuator with an encoder has a tiny servo that counts turns exactly and that is directly related to position.

Originally Posted by zachts

Stepper type motors would pretty easily do what you need and let you measure revolutions of the motor and subsequently output if time based operation of a low rpm gear motor is not accurate enough.

Based on the one's I've used for dosing the infusion pumps use stepper motors typically, and the units are very, very, accurate when using the tubing sets designed for them. They have to be for medical use, and very reliable.

To do this using peristaltic pumps, you can get fairly inexpensive stepper based pumps on eBay from China ( Stepper pumps). I am using 3 of them (with 3 more on order) coupled with 400 step/rev motors and drop by drop dosing is no problem at all using these. For controlling them, I suggest GRBL, so you can simply communicate with pumps over serial with 1 GRBL controller per 3 pumps heads.

If you are unfamiliar with GRBL, it is a CNC motion control platform that runs on an Arduino. I use it with my CNC router and love it. It is the best way of controlling steppers for exact, repeatable, movement that I have found.

The pump heads appear as the X,Y & Z axis in GRBL, and I send Gcode commands to move an axis n mm. I have each head calibrated so that 1 mm = 1 ml. So to dose a single drop it would be move n axis 0.05mm.

I use simple A4988 stepper drivers (<$2 each on eBay) with a CNC shield, but I have been wanting to make a board to hold 3 drivers along with an Arduino mini but could not think of anything that justified the trouble. Having lots of pumps for doing titrations sounds like a good use

Those mini linear stages you are using are wild. Where did you find them for $20?

Anyway cool project you have underway. I could have used this on a recent vacation where my Alk drifted too low and I lost some frags.

Dennis

[karimwassef]

Great ideas Dennis! Give it a go.
If you can get it done with peristaltic pumps repeatedly, that would make it even easier.

I don't know if cheap pumps can be calibrated well enough to consistently run a single drop at a time without a feedback loop. I think it would work to start off, but 20 titrations later? 50?

Even these actuators worry me in terms of their variability over extended use.

This will only be a success if it can run at least 6 months without needing to be recalibrated or replaced.

[dartier]

I am not sure if this has been mentioned in this thread yet, but there is some good info on developing a opensource colorimeter here https://sites.google.com/site/colorimeterwiki/home. Especially useful is the info on which off the shelf tests work well and what LED colours to use for illumination purposes. They appear to be using a sensor similar to the one on the board that you are using.

They also offer it in kit format. I might pick one up just so that I could not have to use another one of those insufferable colour charts ever again.

Dennis

[karimwassef]

Thanks. I'm hoping that the titration will have a large enough change in color reading that I can use the difference in yellow content (Alk) or blue content (Ca) to capture the event, rather than the actual color.

I'm still waiting for my sensor.