Hacker News new | past | comments | ask | show | jobs | submit login
Arduino programmable air, pneumatics kit (arduino.cc)
139 points by itronitron on Dec 30, 2019 | hide | past | favorite | 26 comments



I've done a fair amount of MCU based pneumatics control. It's not difficult. For small scale mechatronic applications the main components you want to use are (1) valves (2) vacuum pumps (3) vacuum generators (4) filters (5) pressure regulators

Apparently this kit only includes 1+2, but here's a few quick notes on each.

(1) Valves come in many types and are typically delimited by their orifice type (size, connection/thread type) and pressure rating. There are subcategories of multi-way valve, timed actuation, one way (passive) valves, etc.

(2) Vacuum pumps replace the need for a shop air supply or small pneumatic pressure generator with a pressure tank. However, they are drastically limited in power and need to be mounted carefully to avoid noise and vibration problems.

(3) These turn positive pressure from your supply side in to negative pressure for suction applications.

(4) Filters are often used not only to remove particulate matter from air supply side or dehumidify the supplied air, but also to introduce small amount of lubricant oil to the circuit. Larger filter systems on shop air supplies will use refrigeration to dehumidify more completely.

(5) Pressure regulators can be very expensive if digitally actuated. Manual ones are quite cheap.

For larger applications you usually assume a shop air supply or otherwise externally managed, stable pressure source.

The go to 'reference' vendor for catalogs is Airtak from Taiwan: http://www.airtak.com/

The world of pneumatics is actually a shadow of its former self in complexity terms... before digital control dominated, there was broadly industrially deployed world of nontrivial pneumatic computers which had memory, logic gates, stacks and buffers implemented without any digital control. Apparently the term to search is 'pneumatic logic': https://en.wikipedia.org/wiki/Pneumatics#Pneumatic_logic


If I wanted to learn by playing around with a home air compressor, an arduino, and some some other devices what would you recommend starting with?


Pneuduino and Soft Robotics Toolkit are both compatible with a home air compressor, albeit running at a low pressure(<50PSI).


Interesting. No air cylinders, though. Precision pneumatic control is possible if you have a proportional valve close to the cylinder and position feedback from the cylinder. If you have a double-ended cylinder set up that way, you have a spring with adjustable spring constant, which is very like a muscle. That technology works fine but costs thousands of dollars in industrial grade components. The demand is small and every application seems to need a custom item. Like linear motors.

A gripper with little air cylinders, linear encoders, and proportional valves, in desktop-robot size, would be fun to play with.


This is a bit of a stretch. Proportional pneumatic actuators are highly nonlinear. Their behavior varies dramatically with temperature, pressure, and a host of other factors. Just having feedback and a linear control law is generally not sufficient for, say, robotics applications. People are working on nonlinear adaptive control laws but they are still very much in the research domain.

Source: I do research in alternative robotics actuation and controls.


You have to get the valve very close to the cylinder for this to work at all well, so you don't have the air in the hoses acting as an air spring. The proportional valves have to be fast, which is usually expensive. That gets the lag down to the point that the thing becomes stably controllable. [1]

There was an effort at CWRU to build legged robotics this way around 2008. Site seems to be down.[2]

[1] https://www.hydraulicspneumatics.com/technologies/other-tech...

[2] http://biorobots.case.edu/projects/onrprojects/robot3/


Pneumatics is fascinating.

Years ago I had a several-hour lesson in pneumatic circuits. What kinds of valves there are, how you connect them in order to get some effect, how to modify the pneumatic circuit, so you can replace a broken valve without the whole apparatus moving some actor (this was a safety training, after all).

I don't remember much, I should probably look for the training material in my basement and read it again. Not because I have any practical use, but because it is a cool, specialized world that I haven't had any further contact with.


In the video at the bottom of that link, what is inside of that rubber balloon that causes it to become firm once air is (presumably) removed?


It reminds me of the Stretch Armstrong's nemesis toy, Vac Man: (https://www.youtube.com/watch?v=xkj9ptYsjTg)

It's probably small bits of foam or maybe even sand. When there's air in the balloon, there's room for the sand to move freely, and it's "soft". When the air is sucked out, the sand has to move past other sand, which is a lot of friction to work against, so it's firm.


Google search term: "granular jamming"


Traditionally coffee grounds but I imagine a lot of granular substances would work reasonably well.


It's a cool package for sure, perfect for one of the projects I have envisioned, but $175 is a bit steep for my use, sadly.


Hmmm. For a long time I've been thinking of some sort of actuator that would let me "feel" one bit (zero/one), without looking at it or hearing it, but purely tactile. Maybe this is it - something pressing or not pressing against, say, my arm. Any other ideas?


How about small cheap 5v relays? They can be had for a few dimes each and can be cut into to reveal the moving parts. Even tiny ones have a surprising amount of force. One might be able to connect to the innards to make an array of binary actuators that could rest against skin. It would probably be hard to beat the power efficiency.


Great idea, that might work. Thanks!


Look into braille peripherals, this is roughly how they work. A relay-sized solenoid pushes a selected pin up so the user can read the letter by feel.

Obnoxiously overpriced, just like screen reader software, but there are Cots products that do what you are looking for.



That's exactly the kind of application I had in mind :D


A solenoid.


Only a few days ago I was playing with the little pump I took out of a broken automatic blood pressure cuff. This is very neat, but really I think you could buy a few such devices and scrap them for the parts for less money.


Arduino is expensive because they fund an Italian company and sales fund software and website development and hardware design. Popular kit gets half price junk-quality clones from China soonn after.

Edit: this is is not a Arduino company project, but similar issues apply.


Are the model numbers for the various pneumatic parts available? I can't seem to find them.



Here's a description of all the components- https://www.crowdsupply.com/img/1238/pa-parts_png_project-bo...

https://github.com/tinkrmind/programmable-air/blob/master/RE...

^ also has some other pneumatic components that didn't make it into the kit.


So this is an ad for a crowd funded project, not yet shipping, which costs $199.

I'll probably just wait for the aliexpress knockoffs that run $7.


Hey, droithomme. I'm the creator of Programmable-Air. Happy to tell you that the crowdfunding is fulfilled and extra kits are available for purchase.

The price is high, I agree. But that's mainly because the BOM cost is about 60$, the kit is large and expensive to ship, and it was fully assembled in New York. I'm looking forward for someone to make a cheap knockoff as much as you are!




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: