Need help finding a small, cheap, Arduino-ready solenoid valve or pump assembly

December 12, 2012 11:10 PM

I need some help building a gadget with an Arduino board. Specifically, I need help finding a way to dispense very tiny quantities of fluid (10 μL) once an hour. The valve assembly is the part I'm having trouble with...

I am trying to build a hummingbird feeder that dispenses ~10 μL (+/- 5 μL) of "nectar" (sugar-water) once an hour. I feel like I can figure out the Arduino/motherboard part of the assembly, but I'm having trouble finding the right valve or pump setup. It seems like somebody would have solved this problem to dispense tiny, precise quantities of ink, but I'm having trouble finding similar projects.

I intend to put the nectar in an IV bag (to keep it from being contaminated over 2 weeks in the tropics), which will be inside of a waterproof container along with the electronic components. The nectar will dispense into a feeding nozzle via some sort of electronically-controlled valve or pump.

The feeder should be able to run for at least one, preferably two weeks without maintenance, including changing batteries (hoping and assuming that battery power will be sufficient for this).

The cost of the project needs to be less than $50, and maybe ~$7 of that is already allocated for the non-electronic parts of the feeder.

I've looked at using a solenoid valve, but most of the affordable ones seem to be too big and have really high flow rates for the tiny volume of nectar that I need to dispense. Going with the solenoid valve option also means that I have to find some way to provide pressure inside the container, since all of the small valves I can find require at least 3 PSI of pressure to work.

The volume of liquid I need to dispense is so tiny that the drops won't have enough mass to break the surface tension of the water, so gravity-fed valves probably won't work(I think). I was considering using a tiny pump to regulate the flow, but I haven't had very much luck finding anything small enough that is also affordable (and I might still need to combine this with a solenoid valve anyway, I haven't thought it quite all the way through yet).

Does anyone have any ideas for how I could regulate the nectar flow through my feeder using a simple, entry-level Arduino DIY setup? Good cheap tiny valves? A better way to approach the problem? Cheers for any guidance you can offer.

posted by dialetheia to technology (14 answers total)

Could you buy a used Eppendorf micropipette from eBay or the like? You could then use the Arduino to control a motor that depresses the plunger, to dispense 10 ul.
posted by Blazecock Pileon at 11:26 PM on December 12, 2012 [1 favorite]

You want a Peristaltic Pump. A solenoid could turn a small angle and move forward a small amount of fluid. Pump that through a small enough needle so that you don't run into the drip size problem. You could find a tubing that's say .1ml / 10cm and size the diameter of the rotation and the accuracy of your solenoid to turn X° and press through Ymm of tubing squeezing out Zµl of liquid.
posted by zengargoyle at 11:53 PM on December 12, 2012 [2 favorites]

I am going to suggest a peristaltic pump as well. In lab we have a microplate filler which is used for dispensing small amounts of liquid very precisely into well on a 384 well plate. We normally add 40 uL per well, but I think it goes down to 5 uL per well. You don't need the precision that we requireso I am guessing with some standard tubing and a pump you could easily dispense the required amount.
posted by koolkat at 1:48 AM on December 13, 2012

A peristaltic pump sounds like the perfect thing! Thank you all, these suggestions have all been really helpful. The micropipette might be a little overkill since eventually we'll need a lot of these feeders, but we might be able to do the same sort of thing with combitips if we end up needing to save a bunch of money.

It looks like I will be trying to put together a simple peristaltic tube pump. If anyone can point me to a straightforward DIY how-to, I would be forever grateful!

For the pump, which would be a better mechanism: solenoids or a roller-style? In this case, "best" means cheaper and easier for someone with only basic experience to assemble within the next week or two. If I had to DIY the pump altogether, I would probably need to be able to keep it about as simple as the one at the top of this page, since I'm a pitiful newbie.

It would be even better if I could just buy a preassembled, arduino-ready, consumer-grade peristaltic pump somewhere since we're short on time this year, and we can tweak it to make it cheaper/better for next field season - but I don't see that available anywhere (at least affordably, with the $50 total budget) after a cursory search.

I see some aquarium dosing pump heads (e.g.) that look like they could conceivably work if I was able to hook one up to a stepper motor - would that be an easier way to go about it if I can't find anything ready-made? Thanks again.
posted by dialetheia at 3:09 AM on December 13, 2012

I don't know where we ordered our pumps from, but I was able to find this Company which looks like exactly what you want. They have a stepper motor backed peristaltic pump. Using the arduino you could have ti pump a set amount of rotation every time point. It would take some calibration to adjust to make sure it is approximately the correct volume, but it shouldn't be that hard to fiddle with it and come up with something that works. You don't need microliter accuracy on the volume so it should be OK. Like zengargoyle I suggest you use something with a narrow gauge to eliminate dropping issues related to the surface tension of water. I also woul worry about the nectar clogging the opening with crystalised sugar and increasing the backpressure. Not sure if that would make a big deal or not I suppose it depends on the sugar concentration of the nectar.
posted by koolkat at 5:50 AM on December 13, 2012

Whatever pump you choose, it seems critical that it seal really well since a single drop of leakage in an hour could be several times the amount you want to dispense. The one koolcat linked to offers several different tubing materials, which would probably help you tweak this.

Given that you'll ultimately need many of these feeders, I would not recommend any DIY pump building. Making a bunch of pumps by hand, and having them all operate consistently and reliably... well, I wouldn't bet on that ever happening.
posted by jon1270 at 6:21 AM on December 13, 2012 [1 favorite]

That aquarium dosing pump looks pretty spot on. Quick rough math: 40mm x 40mm = ~20mm radius, 2mm diameter tubing = 1mm radius: 2 * 3.14 * 20 * 3.14 * 1^2 =~ 395 mm^3 per rotation. Since it would be smaller than the 40mm in diameter we can just say that it's 360 mm^3 per rotation. If you can get your stepper motor to step in 10° steps you're right on the money (or you might have to use some gearing). You could probably build something similar out of plastic/wood with material from the hardware store (hole saw, wheels on bearing for hanging sliding doors, or casters and the like) but I doubt it would match the dosing pump or something similar in quality unless you have access to a CNC shop or a 3D printer.
posted by zengargoyle at 6:35 AM on December 13, 2012 [1 favorite]

There are a lot of different types of peristaltic pumps. You might find the types used by IV drip lines to be what you need.

Then the question is, how do you get/build one (for under $50)? I would try some phone calls to companies that make them and play the sympathy angle. You might get lucky.
posted by plinth at 6:35 AM on December 13, 2012

Sorry to ask for details, but why is the amount of liquid important? Are you monitoring the liquid level and making sure there is always a certain level in the reservoir? Are you pumping this amount out when you see activity at the feeder?

If you're just going to pump at a certain rate with no other control, you're just going to waste fluid and waste a lot of engineering time making something that could be done a lot simpler, I'm afraid. Why not design something like a gravity-fed waterer like they make for dogs and cats? The level stays constant and has no moving or electrical parts.
posted by JoeZydeco at 8:00 AM on December 13, 2012

Coffee pots use a check valve and thermal expansion to dispense pulses of water. Your delivery constraints are not critical, so something like this may be better in terms of moving parts and simplicity. Tweaking required. Also, some study.

Peristaltic pumps suffer from diminishing output over time due to tube forming. Not a big deal unless you are making a chemical process ( as i have done ) but at 10 microliters, it is a tiny amount.
posted by FauxScot at 8:54 AM on December 13, 2012

Sorry to ask for details, but why is the amount of liquid important?

Because it's for a scientific study, and we need to make sure that the hummingbirds won't be able to drink from the feeder more than once an hour. This will help us minimize the effect the feeder has on their behavior, since that's part of what we want to study, and we hope it will help to prevent them from becoming territorial about the feeder; if it dispensed unlimited nectar, it would become the best food source in town. Long story short, the design of the feeder is critical for answering the question we're asking about hummingbirds and this is what my PI wants, so that's what I am going to make.

I asked Welco for a quote and a sample pump because those look absolutely perfect (and some aquarium enthusiasts said they were able to get a sample pump shipped to them), but I think I saw that those cost $80+ so it would make our feeders quite a bit more expensive unless they had an academic discount or something.

So far, I think my best bet will be the aquarium dosing pump with a stepper motor - zengargoyle, thank you so much for your calculations! That is fantastic and extremely helpful.

Thanks again everyone! Your comments are all very useful and I'd love to hear if anyone has any recommendations for stepper motors that would be appropriate, ways to make/cheaply buy an IV peristaltic pump, or anything else that might be helpful.
posted by dialetheia at 9:39 AM on December 13, 2012

I suggest that while you're working on this, you also mull over alternative plans. I am not a research scientist, but I have worked in product prototyping. My gut reaction is that what you're attempting -- these very precise, robust, electronically controlled dispensers at less that $50 each -- sounds improbable.
posted by jon1270 at 10:23 AM on December 13, 2012

Well, it's not so much that it has to be precise - as I said, +/- 50% is totally fine - it just needs to have very fine resolution, which should be doable if I can find a small enough peristaltic pump (like the one I linked above) & a stepper motor with small enough increments. I think I've found components that would work for both of these things, though now I am starting to worry about whether I'll be able to supply enough power with batteries to run a tiny motor for 2 weeks. Any advice about how to power & wire this thing would be great - especially with respect to questions like whether I'll need a motor driver or shield or anything like that.
posted by dialetheia at 12:34 PM on December 13, 2012

In shopping around, keep in mind that you could add a belt-pulley system or gearing to make things easier. You could arrange for the shaft of a motor to have to spin 100 times to turn the pump through one rotation. Like in record players or tape decks, the big wheel spins slow but the little motor driving it is spinning like crazy. You might be able to just turn on a low torque little DC motor (like I know from tape players and RC cars and such) for a second or two and let it whirl away and only get 10° of rotation on the pump. Or less extreme, you don't necessarily need a stepper motor with such fine control, if you can get cheaper in cost or power with one that turns 1/2 turn at a time you can gear it down with gears or belts. I'd probably try belts, you can find some round ones at many places (vacuum cleaner replacement part) and pulleys can just be disks with a groove around the edge to seat the belt into. This lets you choose from a wide variety of motors and then adjust by size of pulleys depending on what torque or control you have with the motor. Of course you should probably test it for a couple of weeks beforehand. :)

You might want to check a syringe possibility as well, the same using gears to turn lots of motion into a little might work. A threaded rod that pushes the plunger, a nut with a gear attached that is held stationary so that turning it moves the plunder in, and a worm gear on a motor that turns the nut. With fine enough threads on the rod that pushes the plunger it could take like 20 turns of the nut to move the plunger 1", with a large gear on the nut and a worm gear on the motor it could take 100's of revolutions of the motor shaft to rotate the nut once. My 3ml syringe has a travel of 5cm. It's feasible that you could arrange for 336 pulses (14 * 24) of the motor over the 2 weeks would cause the plunger to travel 5cm. I'm not sure how accurate this would be for 10µl/hr but would average out to 3ml/2wk.

14 * 24 * 10 micro liter = 3.36 ml

You might even be able to translate the 336 rotations of the minute shaft of a battery powered clock into the 5cm lateral motion it would take to dispense ~3ml from a standard-ish syringe.

Sigh I can come up with dozens of ways that would probably work taken over the 3ml/2wk but may be hit or miss for getting ~10µl on the hour.
posted by zengargoyle at 8:34 PM on December 13, 2012