2.12 gripper design ideas [2.12 Design Notebook]
So apparently they made me Gripper Lead-which means I get the honor and pleasure of designing the robotic gripper for our robot arm. The first thing we went to in group discussion was this:
This gripper can shoot hoops! The concept is ideal for what we are trying to do because of the fact that it will help us make the farther baskets and such - however we will have to orient the gripper in 3-Space which is fine. Now I've made one of these jam gripper before and it didn't work so well. This is likely because of how I made it - so ruling the prior nonsense out - the plan is to start by making a test version of one of these grippers - and if it works - design a new one that we can actually use.
Here are my concerns:
- Pressure regulation (we want to be able to suck hard enough to get a good seal)
- Pressure releasing (we want to be able to blow a puff of air into the system to be able to throw the thing like above maybe, at the very least release the sack pressure)
- Granularity/volume ratio (how many particles do we put in the thing and how do we test this?)
- Sealing (especially with the balloon on the end)
- Tearing (we should make the balloon out of some rip-stop yet stretchy material)
concern one: pressure system
So in the interest of doing some background research - I looked into how we can possibly regulate the pressure going into the balloon while also being able to blow a puff of air out the back. I found these things:
So programmable air is this Arduino-based thing someone came up with which allows you to do pressure regulation and fluid control for soft robotics. It even has pressure sensor feedback and such.
This is a pretty good diagram of how it works. We can probably do PID-based control using the pressure sensor with a setup like this. Nothing fancy is required.
- Can control pressure in the lines via Feedback Control
- Is only 175 dollars and completely open source (all design files are online)
- It does have a safety release valve at the bottom in-case the pressure gets too high
- Can do both forward-pressure based control and vacuum pressure based control
- Skeptical that it will generate enough pressure for a true "suck gripper" and at how long it will take to generate this pressure
- It will likely not be able to work for the application of "throwing with air" or releasing high amounts of pressure into the balloon
- Pressure doesn't go very high
- Unsure about the ROS integration ability
There are OTHER designs like this online - the soft robotics inc has some and so does the soft robotics toolkit but they're all very similar, expensive, not powerful enough, not integrated with ROS or some such other issue.
So let's make out OWN!!!! Features we will have:
- Feedback-based pressure regulation of the balloon when gripping
- Feedback-based pressure regulation for a forced-air throw from the gripper
- ROS Communication Node
- Open Source Design files and etc
concept one: adapted from the above system - the pressure regulator
So I've drawn out some first concepts for the pressure regulator. Here they are!
So it's basically the same system as above, but we add what's called a high-pressure-eject-line. That's the separate line that goes into the gripper and is charged with a TANK basically and what it does is it just shoots air into the system at the regulated pressure of the tank. While it's shooting air into the other valves are all closed until the pressure sensor in the first feedback-system senses the increase in pressure and triggers the release valve. The throwing action, however, comes from forcing the air into the gripper from the new line.
So that's design concept one. Here are some pros and cons.
- Independently regulated "throwing pressure" and "gripping pressure"
- Two isolated systems for throwing and gripping so if one doesn't work, it doesn't affect the other
- Charging the "throwing system" can happen while the gripper is still in operation, limited system delays
- Can be used for more than just what we're using it for
- We will have a ROS Node!
- Safety valves are present
- Throwing system must be charged, the time it takes to charge will limit the "cycle time" of the system
- Complex - required three individual pumps for precise control of all aspects
- May be slow because of the PID regulation, we'll need to have large compressor to make the system fast
- Might be difficult to draw high pressures
- Safety valves are required and at the moment are controlled by pressure sensors - we might need physical safety release valves as well that don't depend on the transducers
So we obviously aren't going to go ahead and straight built this. We'll need to do some calculations figure out what pumps we have access to and which ones we want to use. Some parts we will need to look into are:
- Accurate pressure transducers: https://www.amazon.com/Yosoo-Pressure-Transducer-Sensor-Diesel/dp/B01IR2MANC?ref_=fsclp_pl_dp_4
- Pumps and such: https://www.robotshop.com/en/sfe-air-pump.html (8 PSI/16in Hg)
- BOTH Vacuum and Compressor pumps!
- Some power electronics and such
(1 ATM is like 14 PSI)
We will have to calculate things like how much PSI we will need-or we will BUILD AND TEST!!!!
Check out the valve and the other pump online from the actual programmable air thing. We will probably re-design it for our applications but it's good to look at the parts they used.
concept two: gripper design itself
So this is the kind of gripper style we are going for right. This balloon-style granular-jamming gripper system. So we want to design something like this where we have pressure regulation coming into the gripper and that allows it to pick up things with odd shapes and sizes.
Another good thing to check-out is the hackaday project: https://hackaday.io/project/158551-modular-universal-jamming-gripper it seems like they've done all the relevant testing to make this thing work really well. What we are making is something very similar to this but not exactly. It'll be mostly this except there will be an extra line for the high-pressure.
As you can see, mostly the same design. The only differences are the introduced high-pressure line for pneumatic "throwing," and the filter is there so the particles don't clog the rest of the pneumatic system. We will likely change the material of the balloon but also look into granules that aren't coffee and won't degrade over time if time permits.
So the first thing we are going to do is make MAKE's version of the universal robot gripper and test it with a shop vac. If that doesn't work, it might not be worth our time to go through all the things above. Then we will move onto something more like the hackaday version (a more refined one) and then our own design. It's still unclear if we will use the programmable air thing or if we will just make out own, I'm leaning towards the second one.
But we will start here later in the afternoon/evening: https://makezine.com/projects/universal-robot-gripper/
We still haven't decided if we are going to use the programmable air thing or we are going to make our own - but either way!